<!DOCTYPE HTML>
<html lang="en" class="sidebar-visible no-js light">
<head>
<!-- Book generated using mdBook -->
<meta charset="UTF-8">
<title>Zsh User's Guide</title>
<meta name="robots" content="noindex" />
<!-- Custom HTML head -->
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
<meta name="description" content="">
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta name="theme-color" content="#ffffff" />
<link rel="icon" href="favicon.svg">
<link rel="shortcut icon" href="favicon.png">
<link rel="stylesheet" href="css/variables.css">
<link rel="stylesheet" href="css/general.css">
<link rel="stylesheet" href="css/chrome.css">
<link rel="stylesheet" href="css/print.css" media="print">
<!-- Fonts -->
<link rel="stylesheet" href="FontAwesome/css/font-awesome.css">
<link rel="stylesheet" href="fonts/fonts.css">
<!-- Highlight.js Stylesheets -->
<link rel="stylesheet" href="highlight.css">
<link rel="stylesheet" href="tomorrow-night.css">
<link rel="stylesheet" href="ayu-highlight.css">
<!-- Custom theme stylesheets -->
</head>
<body>
<!-- Provide site root to javascript -->
<script type="text/javascript">
var path_to_root = "";
var default_theme = window.matchMedia("(prefers-color-scheme: dark)").matches ? "navy" : "light";
</script>
<!-- Work around some values being stored in localStorage wrapped in quotes -->
<script type="text/javascript">
try {
var theme = localStorage.getItem('mdbook-theme');
var sidebar = localStorage.getItem('mdbook-sidebar');
if (theme.startsWith('"') && theme.endsWith('"')) {
localStorage.setItem('mdbook-theme', theme.slice(1, theme.length - 1));
}
if (sidebar.startsWith('"') && sidebar.endsWith('"')) {
localStorage.setItem('mdbook-sidebar', sidebar.slice(1, sidebar.length - 1));
}
} catch (e) { }
</script>
<!-- Set the theme before any content is loaded, prevents flash -->
<script type="text/javascript">
var theme;
try { theme = localStorage.getItem('mdbook-theme'); } catch(e) { }
if (theme === null || theme === undefined) { theme = default_theme; }
var html = document.querySelector('html');
html.classList.remove('no-js')
html.classList.remove('light')
html.classList.add(theme);
html.classList.add('js');
</script>
<!-- Hide / unhide sidebar before it is displayed -->
<script type="text/javascript">
var html = document.querySelector('html');
var sidebar = 'hidden';
if (document.body.clientWidth >= 1080) {
try { sidebar = localStorage.getItem('mdbook-sidebar'); } catch(e) { }
sidebar = sidebar || 'visible';
}
html.classList.remove('sidebar-visible');
html.classList.add("sidebar-" + sidebar);
</script>
<nav id="sidebar" class="sidebar" aria-label="Table of contents">
<div class="sidebar-scrollbox">
<ol class="chapter"><li class="chapter-item expanded affix "><a href="zshguide.html">A User's Guide to the Z-Shell</a></li><li class="chapter-item expanded "><a href="zshguide01.html"><strong aria-hidden="true">1.</strong> A short introduction</a></li><li class="chapter-item expanded "><a href="zshguide02.html"><strong aria-hidden="true">2.</strong> What to put in your startup files</a></li><li class="chapter-item expanded "><a href="zshguide03.html"><strong aria-hidden="true">3.</strong> Dealing with basic shell syntax</a></li><li class="chapter-item expanded "><a href="zshguide04.html"><strong aria-hidden="true">4.</strong> The Z-Shell Line Editor</a></li><li class="chapter-item expanded "><a href="zshguide05.html"><strong aria-hidden="true">5.</strong> Substitutions</a></li><li class="chapter-item expanded "><a href="zshguide06.html"><strong aria-hidden="true">6.</strong> Completion, old and new</a></li><li class="chapter-item expanded "><a href="zshguide07.html"><strong aria-hidden="true">7.</strong> Modules and other bits and pieces Not written</a></li></ol>
</div>
<div id="sidebar-resize-handle" class="sidebar-resize-handle"></div>
</nav>
<div id="page-wrapper" class="page-wrapper">
<div class="page">
<div id="menu-bar-hover-placeholder"></div>
<div id="menu-bar" class="menu-bar sticky bordered">
<div class="left-buttons">
<button id="sidebar-toggle" class="icon-button" type="button" title="Toggle Table of Contents" aria-label="Toggle Table of Contents" aria-controls="sidebar">
<i class="fa fa-bars"></i>
</button>
<button id="theme-toggle" class="icon-button" type="button" title="Change theme" aria-label="Change theme" aria-haspopup="true" aria-expanded="false" aria-controls="theme-list">
<i class="fa fa-paint-brush"></i>
</button>
<ul id="theme-list" class="theme-popup" aria-label="Themes" role="menu">
<li role="none"><button role="menuitem" class="theme" id="light">Light (default)</button></li>
<li role="none"><button role="menuitem" class="theme" id="rust">Rust</button></li>
<li role="none"><button role="menuitem" class="theme" id="coal">Coal</button></li>
<li role="none"><button role="menuitem" class="theme" id="navy">Navy</button></li>
<li role="none"><button role="menuitem" class="theme" id="ayu">Ayu</button></li>
</ul>
<button id="search-toggle" class="icon-button" type="button" title="Search. (Shortkey: s)" aria-label="Toggle Searchbar" aria-expanded="false" aria-keyshortcuts="S" aria-controls="searchbar">
<i class="fa fa-search"></i>
</button>
</div>
<h1 class="menu-title">Zsh User's Guide</h1>
<div class="right-buttons">
<a href="print.html" title="Print this book" aria-label="Print this book">
<i id="print-button" class="fa fa-print"></i>
</a>
</div>
</div>
<div id="search-wrapper" class="hidden">
<form id="searchbar-outer" class="searchbar-outer">
<input type="search" id="searchbar" name="searchbar" placeholder="Search this book ..." aria-controls="searchresults-outer" aria-describedby="searchresults-header">
</form>
<div id="searchresults-outer" class="searchresults-outer hidden">
<div id="searchresults-header" class="searchresults-header"></div>
<ul id="searchresults">
</ul>
</div>
</div>
<!-- Apply ARIA attributes after the sidebar and the sidebar toggle button are added to the DOM -->
<script type="text/javascript">
document.getElementById('sidebar-toggle').setAttribute('aria-expanded', sidebar === 'visible');
document.getElementById('sidebar').setAttribute('aria-hidden', sidebar !== 'visible');
Array.from(document.querySelectorAll('#sidebar a')).forEach(function(link) {
link.setAttribute('tabIndex', sidebar === 'visible' ? 0 : -1);
});
</script>
<div id="content" class="content">
<main>
<hr />
<h1 id="a-users-guide-to-the-z-shell"><a class="header" href="#a-users-guide-to-the-z-shell">A User's Guide to the Z-Shell</a></h1>
<h2 id="peter-stephenson"><a class="header" href="#peter-stephenson">Peter Stephenson</a></h2>
<h2 id="20030323"><a class="header" href="#20030323">2003/03/23</a></h2>
<h1 id="table-of-contents"><a class="header" href="#table-of-contents">Table of Contents</a></h1>
<h2 id="a-hrefzshguide01htmll1chapter-1-a-short-introductiona"><a class="header" href="#a-hrefzshguide01htmll1chapter-1-a-short-introductiona"><a href="zshguide01.html#l1">Chapter 1: A short introduction</a></a></h2>
<h3 id="a-hrefzshguide01htmll211-other-shells-and-other-guidesa"><a class="header" href="#a-hrefzshguide01htmll211-other-shells-and-other-guidesa"><a href="zshguide01.html#l2">1.1: Other shells and other guides</a></a></h3>
<h3 id="a-hrefzshguide01htmll312-versions-of-zsha"><a class="header" href="#a-hrefzshguide01htmll312-versions-of-zsha"><a href="zshguide01.html#l3">1.2: Versions of zsh</a></a></h3>
<h3 id="a-hrefzshguide01htmll413-conventionsa"><a class="header" href="#a-hrefzshguide01htmll413-conventionsa"><a href="zshguide01.html#l4">1.3: Conventions</a></a></h3>
<h3 id="a-hrefzshguide01htmll514-acknowledgmentsa"><a class="header" href="#a-hrefzshguide01htmll514-acknowledgmentsa"><a href="zshguide01.html#l5">1.4: Acknowledgments</a></a></h3>
<h2 id="a-hrefzshguide02htmll6chapter-2-what-to-put-in-your-startup-filesa"><a class="header" href="#a-hrefzshguide02htmll6chapter-2-what-to-put-in-your-startup-filesa"><a href="zshguide02.html#l6">Chapter 2: What to put in your startup files</a></a></h2>
<h3 id="a-hrefzshguide02htmll721-types-of-shell-interactive-and-login-shellsa"><a class="header" href="#a-hrefzshguide02htmll721-types-of-shell-interactive-and-login-shellsa"><a href="zshguide02.html#l7">2.1: Types of shell: interactive and login shells</a></a></h3>
<p><a href="zshguide02.html#l8">2.1.1: What is a login shell? Simple tests</a></p>
<h3 id="a-hrefzshguide02htmll922-all-the-startup-filesa"><a class="header" href="#a-hrefzshguide02htmll922-all-the-startup-filesa"><a href="zshguide02.html#l9">2.2: All the startup files</a></a></h3>
<h3 id="a-hrefzshguide02htmll1023-optionsa"><a class="header" href="#a-hrefzshguide02htmll1023-optionsa"><a href="zshguide02.html#l10">2.3: Options</a></a></h3>
<h3 id="a-hrefzshguide02htmll1124-parametersa"><a class="header" href="#a-hrefzshguide02htmll1124-parametersa"><a href="zshguide02.html#l11">2.4: Parameters</a></a></h3>
<p><a href="zshguide02.html#l12">2.4.1: Arrays</a></p>
<h3 id="a-hrefzshguide02htmll1325-what-to-put-in-your-startup-filesa"><a class="header" href="#a-hrefzshguide02htmll1325-what-to-put-in-your-startup-filesa"><a href="zshguide02.html#l13">2.5: What to put in your startup files</a></a></h3>
<p><a href="zshguide02.html#l14">2.5.1: Compatibility options: <code>SH_WORD_SPLIT</code> and
others</a></p>
<p><a href="zshguide02.html#l15">2.5.2: Options for csh junkies</a></p>
<p><a href="zshguide02.html#l16">2.5.3: The history mechanism: types of history</a></p>
<p><a href="zshguide02.html#l17">2.5.4: Setting up history</a></p>
<p><a href="zshguide02.html#l18">2.5.5: History options</a></p>
<p><a href="zshguide02.html#l19">2.5.6: Prompts</a></p>
<p><a href="zshguide02.html#l20">2.5.7: Named directories</a></p>
<p><a href="zshguide02.html#l21">2.5.8: `Go faster' options for power users</a></p>
<p><a href="zshguide02.html#l22">2.5.9: aliases</a></p>
<p><a href="zshguide02.html#l23">2.5.10: Environment variables</a></p>
<p><a href="zshguide02.html#l24">2.5.11: Path</a></p>
<p><a href="zshguide02.html#l25">2.5.12: Mail</a></p>
<p><a href="zshguide02.html#l26">2.5.13: Other path-like things</a></p>
<p><a href="zshguide02.html#l27">2.5.14: Version-specific things</a></p>
<p><a href="zshguide02.html#l28">2.5.15: Everything else</a></p>
<h2 id="a-hrefzshguide03htmll29chapter-3-dealing-with-basic-shell-syntaxa"><a class="header" href="#a-hrefzshguide03htmll29chapter-3-dealing-with-basic-shell-syntaxa"><a href="zshguide03.html#l29">Chapter 3: Dealing with basic shell syntax</a></a></h2>
<h3 id="a-hrefzshguide03htmll3031-external-commandsa"><a class="header" href="#a-hrefzshguide03htmll3031-external-commandsa"><a href="zshguide03.html#l30">3.1: External commands</a></a></h3>
<h3 id="a-hrefzshguide03htmll3132-builtin-commandsa"><a class="header" href="#a-hrefzshguide03htmll3132-builtin-commandsa"><a href="zshguide03.html#l31">3.2: Builtin commands</a></a></h3>
<p><a href="zshguide03.html#l32">3.2.1: Builtins for printing</a></p>
<p><a href="zshguide03.html#l33">3.2.2: Other builtins just for speed</a></p>
<p><a href="zshguide03.html#l34">3.2.3: Builtins which change the shell's state</a></p>
<p><a href="zshguide03.html#l35">3.2.4: cd and friends</a></p>
<p><a href="zshguide03.html#l36">3.2.5: Command control and information commands</a></p>
<p><a href="zshguide03.html#l37">3.2.6: Parameter control</a></p>
<p><a href="zshguide03.html#l38">3.2.7: History control commands</a></p>
<p><a href="zshguide03.html#l39">3.2.8: Job control and process control</a></p>
<p><a href="zshguide03.html#l40">3.2.9: Terminals, users, etc.</a></p>
<p><a href="zshguide03.html#l41">3.2.10: Syntactic oddments</a></p>
<p><a href="zshguide03.html#l42">3.2.11: More precommand modifiers: <code>exec</code>,
<code>noglob</code></a></p>
<p><a href="zshguide03.html#l43">3.2.12: Testing things</a></p>
<p><a href="zshguide03.html#l44">3.2.13: Handling options to functions and scripts</a></p>
<p><a href="zshguide03.html#l45">3.2.14: Random file control things</a></p>
<p><a href="zshguide03.html#l46">3.2.15: Don't watch this space, watch some other</a></p>
<p><a href="zshguide03.html#l47">3.2.16: And also</a></p>
<h3 id="a-hrefzshguide03htmll4833-functionsa"><a class="header" href="#a-hrefzshguide03htmll4833-functionsa"><a href="zshguide03.html#l48">3.3: Functions</a></a></h3>
<p><a href="zshguide03.html#l49">3.3.1: Loading functions</a></p>
<p><a href="zshguide03.html#l50">3.3.2: Function parameters</a></p>
<p><a href="zshguide03.html#l51">3.3.3: Compiling functions</a></p>
<h3 id="a-hrefzshguide03htmll5234-aliasesa"><a class="header" href="#a-hrefzshguide03htmll5234-aliasesa"><a href="zshguide03.html#l52">3.4: Aliases</a></a></h3>
<h3 id="a-hrefzshguide03htmll5335-command-summarya"><a class="header" href="#a-hrefzshguide03htmll5335-command-summarya"><a href="zshguide03.html#l53">3.5: Command summary</a></a></h3>
<h3 id="a-hrefzshguide03htmll5436-expansions-and-quotesa"><a class="header" href="#a-hrefzshguide03htmll5436-expansions-and-quotesa"><a href="zshguide03.html#l54">3.6: Expansions and quotes</a></a></h3>
<p><a href="zshguide03.html#l55">3.6.1: History expansion</a></p>
<p><a href="zshguide03.html#l56">3.6.2: Alias expansion</a></p>
<p><a href="zshguide03.html#l57">3.6.3: Process, parameter, command, arithmetic and brace
expansion</a></p>
<p><a href="zshguide03.html#l58">3.6.4: Filename Expansion</a></p>
<p><a href="zshguide03.html#l59">3.6.5: Filename Generation</a></p>
<h3 id="a-hrefzshguide03htmll6037-redirection-greater-thans-and-less-thansa"><a class="header" href="#a-hrefzshguide03htmll6037-redirection-greater-thans-and-less-thansa"><a href="zshguide03.html#l60">3.7: Redirection: greater-thans and less-thans</a></a></h3>
<p><a href="zshguide03.html#l61">3.7.1: Clobber</a></p>
<p><a href="zshguide03.html#l62">3.7.2: File descriptors</a></p>
<p><a href="zshguide03.html#l63">3.7.3: Appending, here documents, here strings, read
write</a></p>
<p><a href="zshguide03.html#l64">3.7.4: Clever tricks: exec and other file
descriptors</a></p>
<p><a href="zshguide03.html#l65">3.7.5: Multios</a></p>
<h3 id="a-hrefzshguide03htmll6638-shell-syntax-loops-subshells-and-so-ona"><a class="header" href="#a-hrefzshguide03htmll6638-shell-syntax-loops-subshells-and-so-ona"><a href="zshguide03.html#l66">3.8: Shell syntax: loops, (sub)shells and so on</a></a></h3>
<p><a href="zshguide03.html#l67">3.8.1: Logical command connectors</a></p>
<p><a href="zshguide03.html#l68">3.8.2: Structures</a></p>
<p><a href="zshguide03.html#l69">3.8.3: Subshells and current shell constructs</a></p>
<p><a href="zshguide03.html#l70">3.8.4: Subshells and current shells</a></p>
<h3 id="a-hrefzshguide03htmll7139-emulation-and-portabilitya"><a class="header" href="#a-hrefzshguide03htmll7139-emulation-and-portabilitya"><a href="zshguide03.html#l71">3.9: Emulation and portability</a></a></h3>
<p><a href="zshguide03.html#l72">3.9.1: Differences in detail</a></p>
<p><a href="zshguide03.html#l73">3.9.2: Making your own scripts and functions
portable</a></p>
<h3 id="a-hrefzshguide03htmll74310-running-scriptsa"><a class="header" href="#a-hrefzshguide03htmll74310-running-scriptsa"><a href="zshguide03.html#l74">3.10: Running scripts</a></a></h3>
<h2 id="a-hrefzshguide04htmll75chapter-4-the-z-shell-line-editora"><a class="header" href="#a-hrefzshguide04htmll75chapter-4-the-z-shell-line-editora"><a href="zshguide04.html#l75">Chapter 4: The Z-Shell Line Editor</a></a></h2>
<h3 id="a-hrefzshguide04htmll7641-introducing-zlea"><a class="header" href="#a-hrefzshguide04htmll7641-introducing-zlea"><a href="zshguide04.html#l76">4.1: Introducing zle</a></a></h3>
<p><a href="zshguide04.html#l77">4.1.1: The simple facts</a></p>
<p><a href="zshguide04.html#l78">4.1.2: Vi mode</a></p>
<h3 id="a-hrefzshguide04htmll7942-basic-editinga"><a class="header" href="#a-hrefzshguide04htmll7942-basic-editinga"><a href="zshguide04.html#l79">4.2: Basic editing</a></a></h3>
<p><a href="zshguide04.html#l80">4.2.1: Moving</a></p>
<p><a href="zshguide04.html#l81">4.2.2: Deleting</a></p>
<p><a href="zshguide04.html#l82">4.2.3: More deletion</a></p>
<h3 id="a-hrefzshguide04htmll8343-fancier-editinga"><a class="header" href="#a-hrefzshguide04htmll8343-fancier-editinga"><a href="zshguide04.html#l83">4.3: Fancier editing</a></a></h3>
<p><a href="zshguide04.html#l84">4.3.1: Options controlling zle</a></p>
<p><a href="zshguide04.html#l85">4.3.2: The minibuffer and extended commands</a></p>
<p><a href="zshguide04.html#l86">4.3.3: Prefix (digit) arguments</a></p>
<p><a href="zshguide04.html#l87">4.3.4: Words, regions and marks</a></p>
<p><a href="zshguide04.html#l88">4.3.5: Regions and marks</a></p>
<h3 id="a-hrefzshguide04htmll8944-history-and-searchinga"><a class="header" href="#a-hrefzshguide04htmll8944-history-and-searchinga"><a href="zshguide04.html#l89">4.4: History and searching</a></a></h3>
<p><a href="zshguide04.html#l90">4.4.1: Moving through the history</a></p>
<p><a href="zshguide04.html#l91">4.4.2: Searching through the history</a></p>
<p><a href="zshguide04.html#l92">4.4.3: Extracting words from the history</a></p>
<h3 id="a-hrefzshguide04htmll9345-binding-keys-and-handling-keymapsa"><a class="header" href="#a-hrefzshguide04htmll9345-binding-keys-and-handling-keymapsa"><a href="zshguide04.html#l93">4.5: Binding keys and handling keymaps</a></a></h3>
<p><a href="zshguide04.html#l94">4.5.1: Simple key bindings</a></p>
<p><a href="zshguide04.html#l95">4.5.2: Removing key bindings</a></p>
<p><a href="zshguide04.html#l96">4.5.3: Function keys and so on</a></p>
<p><a href="zshguide04.html#l97">4.5.4: Binding strings instead of commands</a></p>
<p><a href="zshguide04.html#l98">4.5.5: Keymaps</a></p>
<h3 id="a-hrefzshguide04htmll9946-advanced-editinga"><a class="header" href="#a-hrefzshguide04htmll9946-advanced-editinga"><a href="zshguide04.html#l99">4.6: Advanced editing</a></a></h3>
<p><a href="zshguide04.html#l100">4.6.1: Multi-line editing</a></p>
<p><a href="zshguide04.html#l101">4.6.2: The builtin vared and the function zed</a></p>
<p><a href="zshguide04.html#l102">4.6.3: The buffer stack</a></p>
<h3 id="a-hrefzshguide04htmll10347-extending-zlea"><a class="header" href="#a-hrefzshguide04htmll10347-extending-zlea"><a href="zshguide04.html#l103">4.7: Extending zle</a></a></h3>
<p><a href="zshguide04.html#l104">4.7.1: Widgets</a></p>
<p><a href="zshguide04.html#l105">4.7.2: Executing other widgets</a></p>
<p><a href="zshguide04.html#l106">4.7.3: Some special builtin widgets and their
uses</a></p>
<p><a href="zshguide04.html#l107">4.7.4: Special parameters: normal text</a></p>
<p><a href="zshguide04.html#l108">4.7.5: Other special parameters</a></p>
<p><a href="zshguide04.html#l109">4.7.6: Reading keys and using the minibuffer</a></p>
<p><a href="zshguide04.html#l110">4.7.7: Examples</a></p>
<h2 id="a-hrefzshguide05htmll111chapter-5-substitutionsa"><a class="header" href="#a-hrefzshguide05htmll111chapter-5-substitutionsa"><a href="zshguide05.html#l111">Chapter 5: Substitutions</a></a></h2>
<h3 id="a-hrefzshguide05htmll11251-quotinga"><a class="header" href="#a-hrefzshguide05htmll11251-quotinga"><a href="zshguide05.html#l112">5.1: Quoting</a></a></h3>
<p><a href="zshguide05.html#l113">5.1.1: Backslashes</a></p>
<p><a href="zshguide05.html#l114">5.1.2: Single quotes</a></p>
<p><a href="zshguide05.html#l115">5.1.3: POSIX quotes</a></p>
<p><a href="zshguide05.html#l116">5.1.4: Double quotes</a></p>
<p><a href="zshguide05.html#l117">5.1.5: Backquotes</a></p>
<h3 id="a-hrefzshguide05htmll11852-modifiers-and-what-they-modifya"><a class="header" href="#a-hrefzshguide05htmll11852-modifiers-and-what-they-modifya"><a href="zshguide05.html#l118">5.2: Modifiers and what they modify</a></a></h3>
<h3 id="a-hrefzshguide05htmll11953-process-substitutiona"><a class="header" href="#a-hrefzshguide05htmll11953-process-substitutiona"><a href="zshguide05.html#l119">5.3: Process Substitution</a></a></h3>
<h3 id="a-hrefzshguide05htmll12054-parameter-substitutiona"><a class="header" href="#a-hrefzshguide05htmll12054-parameter-substitutiona"><a href="zshguide05.html#l120">5.4: Parameter substitution</a></a></h3>
<p><a href="zshguide05.html#l121">5.4.1: Using arrays</a></p>
<p><a href="zshguide05.html#l122">5.4.2: Using associative arrays</a></p>
<p><a href="zshguide05.html#l123">5.4.3: Substituted substitutions, top- and tailing,
etc.</a></p>
<p><a href="zshguide05.html#l124">5.4.4: Flags for options: splitting and joining</a></p>
<p><a href="zshguide05.html#l125">5.4.5: Flags for options: <code>GLOB_SUBST</code> and
<code>RC_EXPAND_PARAM</code></a></p>
<p><a href="zshguide05.html#l126">5.4.6: Yet more parameter flags</a></p>
<p><a href="zshguide05.html#l127">5.4.7: A couple of parameter substitution tricks</a></p>
<p><a href="zshguide05.html#l128">5.4.8: Nested parameter substitutions</a></p>
<h3 id="a-hrefzshguide05htmll12955-that-substitution-againa"><a class="header" href="#a-hrefzshguide05htmll12955-that-substitution-againa"><a href="zshguide05.html#l129">5.5: That substitution again</a></a></h3>
<h3 id="a-hrefzshguide05htmll13056-arithmetic-expansiona"><a class="header" href="#a-hrefzshguide05htmll13056-arithmetic-expansiona"><a href="zshguide05.html#l130">5.6: Arithmetic Expansion</a></a></h3>
<p><a href="zshguide05.html#l131">5.6.1: Entering and outputting bases</a></p>
<p><a href="zshguide05.html#l132">5.6.2: Parameter typing</a></p>
<h3 id="a-hrefzshguide05htmll13357-brace-expansion-and-arraysa"><a class="header" href="#a-hrefzshguide05htmll13357-brace-expansion-and-arraysa"><a href="zshguide05.html#l133">5.7: Brace Expansion and Arrays</a></a></h3>
<h3 id="a-hrefzshguide05htmll13458-filename-expansiona"><a class="header" href="#a-hrefzshguide05htmll13458-filename-expansiona"><a href="zshguide05.html#l134">5.8: Filename Expansion</a></a></h3>
<h3 id="a-hrefzshguide05htmll13559-filename-generation-and-pattern-matchinga"><a class="header" href="#a-hrefzshguide05htmll13559-filename-generation-and-pattern-matchinga"><a href="zshguide05.html#l135">5.9: Filename Generation and Pattern Matching</a></a></h3>
<p><a href="zshguide05.html#l136">5.9.1: Comparing patterns and regular
expressions</a></p>
<p><a href="zshguide05.html#l137">5.9.2: Standard features</a></p>
<p><a href="zshguide05.html#l138">5.9.3: Extensions usually available</a></p>
<p><a href="zshguide05.html#l139">5.9.4: Extensions requiring <code>EXTENDED_GLOB</code></a></p>
<p><a href="zshguide05.html#l140">5.9.5: Recursive globbing</a></p>
<p><a href="zshguide05.html#l141">5.9.6: Glob qualifiers</a></p>
<p><a href="zshguide05.html#l142">5.9.7: Globbing flags: alter the behaviour of
matches</a></p>
<p><a href="zshguide05.html#l143">5.9.8: The function <code>zmv</code></a></p>
<h2 id="a-hrefzshguide06htmll144chapter-6-completion-old-and-newa"><a class="header" href="#a-hrefzshguide06htmll144chapter-6-completion-old-and-newa"><a href="zshguide06.html#l144">Chapter 6: Completion, old and new</a></a></h2>
<h3 id="a-hrefzshguide06htmll14561-completion-and-expansiona"><a class="header" href="#a-hrefzshguide06htmll14561-completion-and-expansiona"><a href="zshguide06.html#l145">6.1: Completion and expansion</a></a></h3>
<h3 id="a-hrefzshguide06htmll14662-configuring-completion-using-shell-optionsa"><a class="header" href="#a-hrefzshguide06htmll14662-configuring-completion-using-shell-optionsa"><a href="zshguide06.html#l146">6.2: Configuring completion using shell options</a></a></h3>
<p><a href="zshguide06.html#l147">6.2.1: Ambiguous completions</a></p>
<p><a href="zshguide06.html#l148">6.2.2: <code>ALWAYS_LAST_PROMPT</code></a></p>
<p><a href="zshguide06.html#l149">6.2.3: Menu completion and menu selection</a></p>
<p><a href="zshguide06.html#l150">6.2.4: Other ways of changing completion
behaviour</a></p>
<p><a href="zshguide06.html#l151">6.2.5: Changing the way completions are
displayed</a></p>
<h3 id="a-hrefzshguide06htmll15263-getting-started-with-new-completiona"><a class="header" href="#a-hrefzshguide06htmll15263-getting-started-with-new-completiona"><a href="zshguide06.html#l152">6.3: Getting started with new completion</a></a></h3>
<h3 id="a-hrefzshguide06htmll15364-how-the-shell-finds-the-right-completionsa"><a class="header" href="#a-hrefzshguide06htmll15364-how-the-shell-finds-the-right-completionsa"><a href="zshguide06.html#l153">6.4: How the shell finds the right completions</a></a></h3>
<p><a href="zshguide06.html#l154">6.4.1: Contexts</a></p>
<p><a href="zshguide06.html#l155">6.4.2: Tags</a></p>
<h3 id="a-hrefzshguide06htmll15665-configuring-completion-using-stylesa"><a class="header" href="#a-hrefzshguide06htmll15665-configuring-completion-using-stylesa"><a href="zshguide06.html#l156">6.5: Configuring completion using styles</a></a></h3>
<p><a href="zshguide06.html#l157">6.5.1: Specifying completers and their options</a></p>
<p><a href="zshguide06.html#l158">6.5.2: Changing the format of listings: groups
etc.</a></p>
<p><a href="zshguide06.html#l159">6.5.3: Styles affecting particular completions</a></p>
<h3 id="a-hrefzshguide06htmll16066-command-widgetsa"><a class="header" href="#a-hrefzshguide06htmll16066-command-widgetsa"><a href="zshguide06.html#l160">6.6: Command widgets</a></a></h3>
<p><a href="zshguide06.html#l161">6.6.1: <code>_complete_help</code></a></p>
<p><a href="zshguide06.html#l162">6.6.2: <code>_correct_word</code>, <code>_correct_filename</code>,
<code>_expand_word</code></a></p>
<p><a href="zshguide06.html#l163">6.6.3: <code>_history_complete_word</code></a></p>
<p><a href="zshguide06.html#l164">6.6.4: <code>_most_recent_file</code></a></p>
<p><a href="zshguide06.html#l165">6.6.5: <code>_next_tags</code></a></p>
<p><a href="zshguide06.html#l166">6.6.6: <code>_bash_completions</code></a></p>
<p><a href="zshguide06.html#l167">6.6.7: <code>_read_comp</code></a></p>
<p><a href="zshguide06.html#l168">6.6.8: <code>_generic</code></a></p>
<p><a href="zshguide06.html#l169">6.6.9: <code>predict-on</code>, <code>incremental-complete-word</code></a></p>
<h3 id="a-hrefzshguide06htmll17067-matching-control-and-controlling-where-things-are-inserteda"><a class="header" href="#a-hrefzshguide06htmll17067-matching-control-and-controlling-where-things-are-inserteda"><a href="zshguide06.html#l170">6.7: Matching control and controlling where things are inserted</a></a></h3>
<p><a href="zshguide06.html#l171">6.7.1: Case-insensitive matching</a></p>
<p><a href="zshguide06.html#l172">6.7.2: Matching option names</a></p>
<p><a href="zshguide06.html#l173">6.7.3: Partial word completion</a></p>
<p><a href="zshguide06.html#l174">6.7.4: Substring completion</a></p>
<p><a href="zshguide06.html#l175">6.7.5: Partial words with capitals</a></p>
<p><a href="zshguide06.html#l176">6.7.6: Final notes</a></p>
<h3 id="a-hrefzshguide06htmll17768-tutoriala"><a class="header" href="#a-hrefzshguide06htmll17768-tutoriala"><a href="zshguide06.html#l177">6.8: Tutorial</a></a></h3>
<p><a href="zshguide06.html#l178">6.8.1: The dispatcher</a></p>
<p><a href="zshguide06.html#l179">6.8.2: Subcommand completion: <code>_arguments</code></a></p>
<p><a href="zshguide06.html#l180">6.8.3: Completing particular argument types</a></p>
<p><a href="zshguide06.html#l181">6.8.4: The rest</a></p>
<h3 id="a-hrefzshguide06htmll18269-writing-new-completion-functions-and-widgetsa"><a class="header" href="#a-hrefzshguide06htmll18269-writing-new-completion-functions-and-widgetsa"><a href="zshguide06.html#l182">6.9: Writing new completion functions and widgets</a></a></h3>
<p><a href="zshguide06.html#l183">6.9.1: Loading completion functions: <code>compdef</code></a></p>
<p><a href="zshguide06.html#l184">6.9.2: Adding a set of completions: <code>compadd</code></a></p>
<p><a href="zshguide06.html#l185">6.9.3: Functions for generating filenames, etc.</a></p>
<p><a href="zshguide06.html#l186">6.9.4: The <code>zsh/parameter</code> module</a></p>
<p><a href="zshguide06.html#l187">6.9.5: Special completion parameters and
<code>compset</code></a></p>
<p><a href="zshguide06.html#l188">6.9.6: Fancier completion: using the tags and styles
mechanism</a></p>
<p><a href="zshguide06.html#l189">6.9.7: Getting the work done for you: handling arguments
etc.</a></p>
<p><a href="zshguide06.html#l190">6.9.8: More completion utility functions</a></p>
<h3 id="a-hrefzshguide06htmll191610-finallya"><a class="header" href="#a-hrefzshguide06htmll191610-finallya"><a href="zshguide06.html#l191">6.10: Finally</a></a></h3>
<h2 id="a-hrefzshguide07htmll192chapter-7-modules-and-other-bits-and-pieces-not-writtena"><a class="header" href="#a-hrefzshguide07htmll192chapter-7-modules-and-other-bits-and-pieces-not-writtena"><a href="zshguide07.html#l192">Chapter 7: Modules and other bits and pieces <em>Not written</em></a></a></h2>
<h3 id="a-hrefzshguide07htmll19371-control-over-modules-zmodloada"><a class="header" href="#a-hrefzshguide07htmll19371-control-over-modules-zmodloada"><a href="zshguide07.html#l193">7.1: Control over modules: <code>zmodload</code></a></a></h3>
<p><a href="zshguide07.html#l194">7.1.1: Modules defining parameters</a></p>
<p><a href="zshguide07.html#l195">7.1.2: Low-level system interaction</a></p>
<p><a href="zshguide07.html#l196">7.1.3: ZFTP</a></p>
<h3 id="a-hrefzshguide07htmll19772-contributed-bitsa"><a class="header" href="#a-hrefzshguide07htmll19772-contributed-bitsa"><a href="zshguide07.html#l197">7.2: Contributed bits</a></a></h3>
<p><a href="zshguide07.html#l198">7.2.1: Prompt themes</a></p>
<h3 id="a-hrefzshguide07htmll19973-whats-new-in-41a"><a class="header" href="#a-hrefzshguide07htmll19973-whats-new-in-41a"><a href="zshguide07.html#l199">7.3: What's new in 4.1</a></a></h3>
<h2 id="a-hrefzshguide08htmll200appendix-1-obtaining-zsh-and-getting-more-information-not-writtena"><a class="header" href="#a-hrefzshguide08htmll200appendix-1-obtaining-zsh-and-getting-more-information-not-writtena"><a href="zshguide08.html#l200">Appendix 1: Obtaining zsh and getting more information <em>Not written</em></a></a></h2>
<hr />
<div id="chapter_begin" style="break-before: page; page-break-before: always;"></div><!-- START doctoc generated TOC please keep comment here to allow auto update -->
<!-- DON'T EDIT THIS SECTION, INSTEAD RE-RUN doctoc TO UPDATE -->
<p><strong>Table of Contents</strong> <em>generated with <a href="https://github.com/thlorenz/doctoc">DocToc</a></em></p>
<ul>
<li><a href="zshguide01.html#chapter-1-a-short-introduction">Chapter 1: A short introduction</a>
<ul>
<li><a href="zshguide01.html#11-other-shells-and-other-guides">1.1: Other shells and other guides</a></li>
<li><a href="zshguide01.html#12-versions-of-zsh">1.2: Versions of zsh</a></li>
<li><a href="zshguide01.html#13-conventions">1.3: Conventions</a></li>
<li><a href="zshguide01.html#14-acknowledgments">1.4: Acknowledgments</a></li>
</ul>
</li>
</ul>
<!-- END doctoc generated TOC please keep comment here to allow auto update -->
<p><span id="intro"></span><span id="l1"></span></p>
<h1 id="chapter-1-a-short-introduction"><a class="header" href="#chapter-1-a-short-introduction">Chapter 1: A short introduction</a></h1>
<p>The Z-Shell, `zsh' for short, is a command interpreter for UNIX
systems, or in UNIX jargon, a `shell', because it wraps around the
commands you use. More than that, however, zsh is a particularly
powerful shell --- and it's free, and under regular maintenance --- with
lots of interactive features allowing you to do the maximum work with
the minimum fuss. Of course, for that you need to know what the shell
can do and how, and that's what this guide is for.</p>
<p>The most basic basics: I shall assume you have access to a UNIX system,
otherwise the rest of this is not going to be much use. You can also use
zsh under Windows by installing Cygwin, which provides a UNIX-like
environment for programmes --- given the weakness of the standard
Windows command interpreter, this is a good thing to do. There are ports
of older versions of zsh to Windows which run natively, i.e. without a
UNIX environment, although these have a slightly different behaviour in
some respects and I won't talk about them further.</p>
<p>I'll also assume some basic knowledge of UNIX; you should know how the
filesystem works, i.e. what <code>/home/users/pws/.zshrc</code> and <code>../file</code> mean,
and some basic commands, for example <code>ls</code>, and you should have
experience with using <code>rm</code> to delete completely the wrong file by
accident, and that sort of thing. In something like `<code>rm file</code>', I will
often refer to the `command' (<code>rm</code>, of course) and the `argument(s)'
(anything else coming after the command which is used by it), and to the
complete thing you typed in one go as the `command line'.</p>
<p>You're also going to need zsh itself; if you're reading this, you may
well already have it, but if you don't, you or your system administrator
should read <a href="zshguide08.html#appa">Appendix A</a>. For now, we'll suppose
you're sitting in front of a terminal with zsh already running.</p>
<p>Now to the shell. After you log in, you probably see some prompt (a
series of symbols on the screen indicating that you can input a
command), such as `<code>$</code>' or `<code>%</code>', possibly with some other text in
front --- later, we'll see how you can change that text in interesting
ways. That prompt comes from the shell. Type `<code>print hello</code>', then
backspace over `<code>hello</code>' and type `<code>goodbye</code>'. Now hit the `Return'
key (or `Enter' key, I'll just say <code><RET></code> from now on, likewise
<code><TAB></code> for the tab key, <code><SPC></code> for the space key); unless you have a
serious practical-joker problem on your system, you will see
`<code>goodbye</code>', and the shell will come back with another prompt. All of
the time up to when you hit <code><RET></code>, you were interacting with the shell
and its editor, called `Z-Shell Line Editor' or `zle' for short; only
then did the shell go away and tell the <code>print</code> command to print out a
message. So you can see that the shell is important.</p>
<p>However, if all you're doing is typing simple commands like that, why do
you need anything complicated? In that case, you don't; but real life's
not that simple. In the rest of this guide, I describe how, with zsh's
help, you can:</p>
<ul>
<li>customise the environment in which you work, by using startup files,</li>
<li>write your own commands to shorten tasks and store things in shell
variables (`parameters') so you don't have to remember them,</li>
<li>use zle to minimise the amount of typing you have to do --- in zsh,
you can even edit small files that way,</li>
<li>pick the files you want to use for a particular command such as <code>mv</code>
or <code>ls</code> using zsh's very sophisticated filename generation (known
colloquially as `globbing') system,</li>
<li>tell the editor what sort of arguments you use with particular
commands, so that you only need to type part of the name and it will
complete the rest, using zsh's unrivalled programmable completion
system,</li>
<li>use the extra add-ons (`modules') supplied with the latest version
of zsh to do other things you usually can't do in a shell at all.</li>
</ul>
<p>That's only a tiny sample. Since there's so much to say, this guide will
concentrate on the things zsh does best, and in particular the things it
has which other shells don't. The next chapter gives a few of the
basics, by trying to explain how to set the shell up the way you want
it. Like the rest of the guide, it's not intended to be exhaustive, for
which you should look at the shell manual.</p>
<p>Some other things you should probably know straight away. First, the
shell is always running, even when the command you typed is running,
too; the shell simply hangs around waiting for it to finish: you may
know from other shells about putting commands in the <strong>background</strong> by
putting an `<code>&</code>' after the command, which means that the shell doesn't
wait for them to finish. The shell is there even if the command's in the
foreground, but in this case doing nothing.</p>
<p>Second, it doesn't just run other people's commands, it has some of its
own, called <strong>builtin commands</strong> or just <strong>builtins</strong>, and you can even
add your own commands as lists of instructions to the shell called
<strong>functions</strong>; builtins and functions always run in the shell itself.
That's important to know, because things which don't run in the shell
itself can't affect it, and hence can't alter parameters, functions,
aliases, and all the other things I shall talk about.</p>
<p><span id="l2"></span></p>
<h2 id="11-other-shells-and-other-guides"><a class="header" href="#11-other-shells-and-other-guides">1.1: Other shells and other guides</a></h2>
<p>If you want a basic grounding in how shells work, what their syntax is
(i.e. how to write commands), and how to write scripts and functions,
you should read one of the many books on the subject. In particular, you
will get most out of a book that describes the Korn shell (ksh), as zsh
is very similar to this --- so similar that it will be worth my while
pointing out differences as we go along, since they can confuse ksh
users. Recent versions of zsh can emulate ksh (strictly, the 1988
version of ksh, although there are increasingly features from the 1993
version) quite closely, although it's not perfect, and less perfect the
more closely you look. However, it's important to realise that if you
just start up any old zsh there is no guarantee that it will be set up
to work like ksh; unless you or your system adminstrator have changed
some settings, it certainly won't be. You might not see that straight
away, but it affects the shell in subtle ways. I will talk about
emulation a bit more later on.</p>
<p>A few other shells are worth mentioning. The grandfather of all UNIX
shells is sh, now known as the Bourne shell but originally just referred
to as `the shell'. The story is similar to ksh: zsh can emulate sh
quite closely (much more closely than ksh, since sh is considerably
simpler), but in general you need to make sure it's set up to do that
before you can be sure it will emulate sh.</p>
<p>You may also come across the `Bourne-Again Shell', bash. This is a
freely-available enhancement of sh written by the GNU project --- but it
is not always enhanced along the lines of ksh, and hence in many ways it
is very different from zsh. On some free UNIX-like systems such as
Linux/GNU (which is what people usually mean by Linux), the command sh
is really bash, so there you should be extra careful when trying to
ensure that something which runs under the so-called `sh' will also run
under zsh. Some Linux systems also have another simpler Bourne shell
clone, ash; as it's simpler, it's more like the original Bourne shell.</p>
<p>Some more modern operating systems talk about `the POSIX shell'. This
is an attempt to standardize UNIX shells; it's most like the Korn shell,
although, a bit confusingly, it's often just called sh, because the
standard says that it should be. Usually, this just means you get a bit
extra free with your sh and it still does what you expect. Zsh has made
some attempts to fit the standard, but you have to tell it to --- again,
simply starting up `zsh' will not have the right settings for that.</p>
<p>There is another common family of shells with, unfortunately,
incompatible syntax. The source of this family is the C-Shell, csh, so
called because its syntax looks more like the C programming language.
This became widespread when the only other shell available was sh
because csh had better interactive features, such as job control. It was
then enhanced to make tcsh, which has many of the interactive features
you will also find in zsh, and so became very popular. Despite these
common features, the syntax of zsh is very different, so you should not
try and use csh/tcsh commands beyond the very simplest in zsh; but if
you are a tcsh user, you will find virtually every capability you are
used to in zsh somewhere, plus a lot more.</p>
<p><span id="l3"></span></p>
<h2 id="12-versions-of-zsh"><a class="header" href="#12-versions-of-zsh">1.2: Versions of zsh</a></h2>
<p>At the time of writing, the most recent version of zsh available for
widespread use was 4.0.6. You will commonly find two sets of older zsh's
around. The 3.0 series, of which the last release was 3.0.9, was a
stable release, with only bug fixes since the first release of zsh 3.
The 3.1 series were beta versions, with lots of new features; the last
of these, 3.1.9, was not so different from 4.0.1; the main change is
that the shell has now been declared stable, so that as with zsh 3 there
will be a set of bug fixes, labelled 4.0, and a set with new functions
in, labelled 4.1. As 4.0 replaces all zsh 3 versions, I will try to keep
things simple and talk about that; but every now and then it will be
helpful to point out where older versions were different.</p>
<p>One notable feature of zsh is the completion of command line arguments.
The system changed in 3.1.6 and 3.1.7 to make it a lot more
configurable, and (provided you keep your wits about you) a little less
obscure. I therefore won't describe the old completion system, which
used the `compctl' command, in any detail; a very brief introduction is
given in the zsh FAQ. The old system remains available, however we
strongly recommend new users to start with the new one. See <a href="zshguide06.html#comp">chapter
6</a> `Completion, old and new' for the lowdown on
new-style completion.</p>
<p>There won't be a big difference between 4.0 and 4.1, just bug fixes and
a few evolutionary changes, plus some extra modules. There will be some
notes in <a href="zshguide07.html#ragbag">chapter 7</a> about new features in 4.1,
but nothing you write for 4.0 is likely to become obsolete in the
foreseeable future.</p>
<p><span id="l4"></span></p>
<h2 id="13-conventions"><a class="header" href="#13-conventions">1.3: Conventions</a></h2>
<p>Most of what I say will be reasonably self-contained (which means I use
phrases like `as I said before' and `as I'll discuss later on' more
than a real stylist would like, and the number times I refer to other
chapters is excessive), but there are some points I should perhaps draw
your attention to before you leap in.</p>
<p>I will often write chunks of code as you would put them in a file for
execution (a `script' or a `function', the differences to be discussed
<em>passim</em>):</p>
<pre><code> if [[ $ZSH_VERSION = 3.* ]]; then
print This is a release of the third version of zsh.
else
print This is either very new or very old.
fi
</code></pre>
<p>but sometimes I will show both what you type into a shell interactively,
and what the shell throws back at you:</p>
<pre><code> % print $ZSH_VERSION
3.1.9
% print $CPUTYPE
i586
</code></pre>
<p>Here, `<code>%</code>' shows the prompt the shell puts up to tell you it is
expecting input (and the space immediately after is part of it).
Actually, you probably see something before the percent sign like the
name of the machine or your user name, or maybe something fancier. I've
pruned it to the minimum to avoid confusion, and kept it as reminder
that this is the line you type.</p>
<p>If you're reading an electronic version of this guide, and want to copy
lines with the `<code>%</code>' in front into a terminal to be executed, there's a
neat way of doing this where you don't even have to edit the line first:</p>
<pre><code> alias %=' '
</code></pre>
<p>Then <code>%</code> at the start of a line is turned into nothing whatsoever; the
space just indicates that any following aliases should be expanded. So
the line `<code>% print $CPUTYPE</code>' will ignore the `<code>%</code>' and execute the
rest of the line. (I hope it's obvious, but your <em>own</em> prompt is always
ignored; this is just if you copy the prompts from the guide into the
shell.)</p>
<p>There are lots of different types of object in zsh, but one of the most
common is parameters, which I will always show with a `<code>$</code>' sign in
front, like `<code>$ZSH_VERSION</code>', to remind you they are parameters. You
need to remember that when you're setting or fiddling with the parameter
itself, rather than its value, you omit the `<code>$</code>'. When you do and
don't need it should become clearer as we go along.</p>
<p>The other objects I'll show specially are shell options --- choices
about how the shell is to work --- which I write like this:
`<code>SH_WORD_SPLIT</code>', `<code>NO_NOMATCH</code>', `<code>ZLE</code>'. Again, that's not the
whole story since whenever the shell expects options you can write them
in upper or lower case with as many or as few underscores as you like;
and often in code chunks I'll use the simplest form instead:
`<code>shwordsplit</code>', `<code>nonomatch</code>', `<code>zle</code>'. If you're philosophical you
can think of it as expressing the category difference between talking
about programming and actual programming, but really it's just me being
inconsistent.</p>
<p>You may find it odd that I use three hyphens to signify a dash. That's
actually a convention used in the printed version of this guide, which
is made with LaTeX. One day, I will turn this into a macro and it will
appear properly in other versions; but then, one day the universe will
come to an end.</p>
<p><span id="l5"></span></p>
<h2 id="14-acknowledgments"><a class="header" href="#14-acknowledgments">1.4: Acknowledgments</a></h2>
<p>I am grateful for comments from various zsh users. In particular, I have
had detailed comments and corrections from Bart Schaefer, Sven `Mr
Completion' Wischnowsky and Oliver Kiddle. It's usual to add that any
remaining errors are my own, but that's so stark staringly obvious as to
be ridiculous. I mean, who wrote this? Never mind.</p>
<p>Most of this written on one or another release of Linux Mandrake (a
derivative of Red Hat), with the usual GNU and XFree86 tools. Since all
of this was free, it only seems fair to say `thank you' for the gift.
It also works a lot better than the operating system that came with this
particular PC.</p>
<div id="chapter_begin" style="break-before: page; page-break-before: always;"></div><!-- START doctoc generated TOC please keep comment here to allow auto update -->
<!-- DON'T EDIT THIS SECTION, INSTEAD RE-RUN doctoc TO UPDATE -->
<p><strong>Table of Contents</strong> <em>generated with <a href="https://github.com/thlorenz/doctoc">DocToc</a></em></p>
<ul>
<li><a href="zshguide02.html#chapter-2-what-to-put-in-your-startup-files">Chapter 2: What to put in your startup files</a>
<ul>
<li><a href="zshguide02.html#21-types-of-shell-interactive-and-login-shells">2.1: Types of shell: interactive and login shells</a>
<ul>
<li><a href="zshguide02.html#211-what-is-a-login-shell-simple-tests">2.1.1: What is a login shell? Simple tests</a></li>
</ul>
</li>
<li><a href="zshguide02.html#22-all-the-startup-files">2.2: All the startup files</a></li>
<li><a href="zshguide02.html#23-options">2.3: Options</a></li>
<li><a href="zshguide02.html#24-parameters">2.4: Parameters</a>
<ul>
<li><a href="zshguide02.html#241-arrays">2.4.1: Arrays</a></li>
</ul>
</li>
<li><a href="zshguide02.html#25-what-to-put-in-your-startup-files">2.5: What to put in your startup files</a>
<ul>
<li><a href="zshguide02.html#251-compatibility-options-sh_word_split-and-others">2.5.1: Compatibility options: <code>SH_WORD_SPLIT</code> and others</a></li>
<li><a href="zshguide02.html#252-options-for-csh-junkies">2.5.2: Options for csh junkies</a></li>
<li><a href="zshguide02.html#253-the-history-mechanism-types-of-history">2.5.3: The history mechanism: types of history</a></li>
<li><a href="zshguide02.html#254-setting-up-history">2.5.4: Setting up history</a></li>
<li><a href="zshguide02.html#255-history-options">2.5.5: History options</a></li>
<li><a href="zshguide02.html#256-prompts">2.5.6: Prompts</a></li>
<li><a href="zshguide02.html#257-named-directories">2.5.7: Named directories</a></li>
<li><a href="zshguide02.html#258-%5Cgo-faster-options-for-power-users">2.5.8: `Go faster' options for power users</a></li>
<li><a href="zshguide02.html#259-aliases">2.5.9: aliases</a></li>
<li><a href="zshguide02.html#2510-environment-variables">2.5.10: Environment variables</a></li>
<li><a href="zshguide02.html#2511-path">2.5.11: Path</a></li>
<li><a href="zshguide02.html#2512-mail">2.5.12: Mail</a></li>
<li><a href="zshguide02.html#2513-other-path-like-things">2.5.13: Other path-like things</a></li>
<li><a href="zshguide02.html#2514-version-specific-things">2.5.14: Version-specific things</a></li>
<li><a href="zshguide02.html#2515-everything-else">2.5.15: Everything else</a></li>
</ul>
</li>
</ul>
</li>
</ul>
<!-- END doctoc generated TOC please keep comment here to allow auto update -->
<p><span id="init"></span><span id="l6"></span></p>
<h1 id="chapter-2-what-to-put-in-your-startup-files"><a class="header" href="#chapter-2-what-to-put-in-your-startup-files">Chapter 2: What to put in your startup files</a></h1>
<p>There are probably various changes you want to make to the shell's
behaviour. All shells have `startup' files, containing commands which
are executed as soon as the shell starts. Like many others, zsh allows
each user to have their own startup files. In this chapter, I discuss
the sorts of things you might want to put there. This will serve as an
introduction to what the shell does; by the end, you should have an
inkling of many of the things which will be discussed in more detail
later on and why they are interesting. Sometimes you will find out more
than you want to know, such as how zsh differs from other shells you're
not going to use. Explaining the differences here saves me having to lie
about how the shell works and correcting it later on: most people will
simply want to know how the shell normally works, and note that there
are other ways of doing it.</p>
<p><span id="l7"></span></p>
<h2 id="21-types-of-shell-interactive-and-login-shells"><a class="header" href="#21-types-of-shell-interactive-and-login-shells">2.1: Types of shell: interactive and login shells</a></h2>
<p>First, you need to know what is meant by an <strong>interactive</strong> and a
<strong>login</strong> shell. Basically, the shell is just there to take a list of
commands and run them; it doesn't really care whether the commands are
in a file, or typed in at the terminal. In the second case, when you are
typing at a prompt and waiting for each command to run, the shell is
<strong>interactive</strong>; in the other case, when the shell is reading commands
from a file, it is, consequently, <strong>non-interactive</strong>. A list of
commands used in this second way --- typically by typing something like
<code>zsh filename</code>, although there are shortcuts --- is called a <strong>script</strong>,
as if the shell was acting in a play when it read from it (and shells
can be real hams when it comes to playacting). When you start up a
script from the keyboard, there are actually two zsh's around: the
interactive one you're typing at, which is waiting for another,
non-interactive one to finish running the script. Almost nothing that
happens in the second one affects the first; they are different copies
of zsh.</p>
<p>Remember that when I give examples for you to type, I often show them as
they would appear in a script, without prompts in front. What you
actually see on the screen if you type them in will have a lot more in
front.</p>
<p>When you first log into the computer, the shell you are presented with
is interactive, but it is also a login shell. If you type `<code>zsh</code>', it
starts up a new interactive shell: because you didn't give it the name
of a file with commands in, it assumes you are going to type them
interactively. Now you've got two interactive shells at once, one
waiting for the other: it doesn't sound all that useful, but there are
times when you are going to make some radical changes to the shell's
settings temporarily, and the easiest thing to do is to start another
shell, do what you want to do, and exit back to the original, unaltered,
shell --- so it's not as stupid as it sounds.</p>
<p>However, that second shell will not be a login shell. How does zsh know
the difference? Well, the programme that logs you in after you type your
password (called, predictably, <strong>login</strong>), actually sticks a `<code>-</code>' in
front of the name of the shell, which zsh recognises. The other way of
making a shell a login shell is to run it yourself with the option <code>-l</code>;
typing `<code>zsh -l</code>' will start a zsh that also thinks it's a login shell,
and later I'll explain how to turn on options within the shell, which
you can do with the login option too. Otherwise, any zsh you start
yourself will not be a login shell. If you are using X-Windows, and have
a terminal emulator such as xterm running a shell, that is probably not
a login shell. However, it's actually possible to get xterm to start a
login shell by giving it the option <code>-ls</code>, so if you type `<code>xterm -ls &</code>', you will get a window running a login shell (the <code>&</code> means the
shell in the first window doesn't wait for it to finish).</p>
<p>The first main difference between a login shell and any other
interactive shell is the one to do with startup files, described below.
The other one is what you do when you're finished. With a login shell
you can type `<code>logout</code>' to exit the shell; with another you type
`<code>exit</code>'. However, `<code>exit</code>' works for all shells, interactive,
non-interactive, login, whatever, so a lot of people just use that. In
fact, the only difference is that `<code>logout</code>' will tell you `<code>not login shell</code>' if you use it anywhere else and fail to exit. The command
`<code>bye</code>' is identical to `<code>exit</code>', only shorter and less standard. So
my advice is just to use `<code>exit</code>'.</p>
<p>As somebody pointed out to me recently, login shells don't have to be
interactive. You can always start a shell in the two ways that make it a
login shell; the ways that make it an interactive shell or not are
independent. In fact, some start-up scripts for windowing systems run a
non-interactive login shell to incorporate definitions from the
appropriate login scripts before executing the commands to start the
windowing session.</p>
<p><span id="l8"></span></p>
<h3 id="211-what-is-a-login-shell-simple-tests"><a class="header" href="#211-what-is-a-login-shell-simple-tests">2.1.1: What is a login shell? Simple tests</a></h3>
<p>Telling if the shell you are looking at is interactive is usually easy:
if there's a prompt, it's interactive. As you may have gathered, telling
if it's a login shell is more involved because you don't always know how
the shell was started or if the option got changed. If you want to know,
you can type the following (one line at a time if you like, see below),</p>
<pre><code> if [[ -o login ]]; then
print yes
else
print no
fi
</code></pre>
<p>which will print `yes' or `no' according to whether it's a login shell
or not; the syntax will be explained as we go along. There are shorter
ways of doing it, but this illustrates the commonest shell syntax for
testing things, something you probably often want to do in a startup
file. What you're testing goes inside the `<code>[[ ... ]]</code>'; in this case,
the <code>-o</code> tells the shell to test an option, here <code>login</code>. The next line
says what to do if the test succeeded; the line after the `<code>else</code>' what
to do if the test failed. This syntax is virtually identical to ksh; in
this guide, I will not give exhaustive details on the tests you can
perform, since there are many of them, but just show some of the most
useful. As always, see the manual --- in this case, `Conditional
Expressions' in the <code>zshmisc</code> manual pages.</p>
<p>Although you usually know when a shell is interactive, in fact you can
test that in exactly the same way, too: just use `<code>[[ -o interactive ]]</code>'. This is one option you can't change within the shell; if you turn
off reading from the keyboard, where is the shell supposed to read from?
But you can at least test it.</p>
<p>Aside for beginners in shell programming: maybe the semicolon looks a
bit funny; that's because the `<code>then</code>' is really a separate command.
The semicolon is just instead of putting it on a new line; the two are
interchangeable. In fact, I could have written,</p>
<pre><code> if [[ -o login ]]; then; print yes; else; print no; fi
</code></pre>
<p>which does exactly the same thing. I could even have missed out the
semicolons after `<code>then</code>' and `<code>else</code>', because the shell knows that a
command must come after each of those --- though the semicolon or
newline <em>before</em> the <code>then</code> is often important, because the shell does
<code>not</code> know a command has to come next, and might mix up the <code>then</code> with
the arguments of the command after the `<code>if</code>': it may look odd, but the
`<code>[[</code> <em>...</em> <code>]]</code>' is actually a command. So you will see various ways
of dividing up the lines in shell programmes. You might also like to
know that <code>print</code> is one of the builtin commands referred to before; in
other words, the whole of that chunk of programme is executed by the
shell itself. If you're using a newish version of the shell, you will
notice that zsh tells you what it's waiting for, i.e. a `<code>then</code>' or an
`<code>else</code>' clause --- see the explanation of <code>$PS2</code> below for more on
this. Finally, the spaces I put before the `<code>print</code>' commands were
simply to make it look prettier; any number of spaces can appear before,
after, or between commands and arguments, as long as there's at least
one between ordinary words (the semicolon is recognised as special, so
you don't need one before that, though it's harmless if you do put one
in).</p>
<p>Second aside for users of sh: you may remember that tests in sh used a
single pair of brackets, `<code>if [ ... ]; then ...</code>', or equivalently as a
command called <strong>test</strong>, `<code>if test ...; then ...</code>'. The Korn shell was
deliberately made to be different, and zsh follows that. The reason is
that `<code>[[</code>' is treated specially, which allows the shell to do some
extra checks and allows more natural syntax. For example, you may know
that in sh it's dangerous to test a parameter which may be empty: `[
$var = foo ]' will fail if <code>$var</code> is empty, because in that case the
word is missed out and the shell never knows it was supposed to be
there; with `<code>[[</code> <em>...</em> <code>]]</code>', this is quite safe because the shell is
aware there's a word before the `<code>=</code>', even if it's empty. Also, you
can use `<code>&&</code>' and `<code>||</code>' to mean logical `and' and `or', which
agrees with the usual UNIX/C convention; in sh, they would have been
taken as starting a new command, not as part of the test, and you have
to use the less clear `<code>-a</code>' and `<code>-o</code>'. Actually, zsh provides the
old form of test for backward compatibility, but things will work a lot
more smoothly if you don't use it.</p>
<p><span id="l9"></span></p>
<h2 id="22-all-the-startup-files"><a class="header" href="#22-all-the-startup-files">2.2: All the startup files</a></h2>
<p>Now here's a list of the startup files and when they're run. You'll see
they fall into two classes: those in the <code>/etc</code> directory, which are put
there by the system administrator and are run for all users, and those
in your home directory, which zsh, like many shells, allows you to
abbreviate to a `<code>~</code>'. It's possible that the latter files are
somewhere else; type `<code>print $ZDOTDIR</code>' and if you get something other
than a blank line, or an error message telling you the parameter isn't
set, it's telling you a directory other than `<code>~</code>' where your startup
files live. If <code>$ZDOTDIR</code> (another parameter) is not already set, you
won't want to set it without a good reason.</p>
<ul>
<li><strong><code>/etc/zshenv</code></strong><br />
Always run for every zsh.</li>
<li><strong><code>~/.zshenv</code></strong><br />
Usually run for every zsh (see below).</li>
<li><strong><code>/etc/zprofile</code></strong><br />
Run for login shells.</li>
<li><strong><code>~/.zprofile</code></strong><br />
Run for login shells.</li>
<li><strong><code>/etc/zshrc</code></strong><br />
Run for interactive shells.</li>
<li><strong><code>~/.zshrc</code></strong><br />
Run for interactive shells.</li>
<li><strong><code>/etc/zlogin</code></strong><br />
Run for login shells.</li>
<li><strong><code>~/.zlogin</code></strong><br />
Run for login shells.</li>
</ul>
<p>Now you know what login and interactive shells are, this should be
straightforward. You may wonder why there are both <code>~/.zprofile</code> and
<code>~/.zlogin</code>, when they are both for login shells: the answer is the
obvious one, that one is run before, one after <code>~/.zshrc</code>. This is
historical; Bourne-type shells run <code>/etc/profile</code>, and csh-type shells
run <code>~/.login</code>, and zsh tries to cover the bases with its own startup
files.</p>
<p>The complication is hinted at by the `see below'. The file
<code>/etc/zshenv</code>, as it says, is always run at the start of any zsh.
However, if the option <code>NO_RCS</code> is set (or, equivalently, the <code>RCS</code>
option is unset: I'll talk about options shortly, since they are
important in startup files), none of the others are run. The most common
way of setting this option is with a flag on the command line: if you
start the shell as `<code>zsh -f</code>', the option becomes set, so only
<code>/etc/zshenv</code> is run and the others are skipped. Often, scripts do this
as a way of trying to get a basic shell with no frills, as I'll describe
below; but if something is set in <code>/etc/zshenv</code>, there's no way to avoid
it. This leads to the First Law of Zsh Administration: put as little as
possible in the file <code>/etc/zshenv</code>, as every single zsh which starts up
has to read it. In particular, if the script assumes that only the basic
options are set and <code>/etc/zshenv</code> has altered them, it might well not
work. So, at the absolute least, you should probably surround any option
settings in <code>/etc/zshenv</code> with</p>
<pre><code> if [[ ! -o norcs ]]; then
... <commands to run if NO_RCS is not set,
such as setting options> ...
fi
</code></pre>
<p>and your users will be eternally grateful. Settings for interactive
shells, such as prompts, have no business in <code>/etc/zshenv</code> unless you
<em>really</em> insist that all users have them as defaults for every single
shell. Script writers who want to get round problems with options being
changed in <code>/etc/zshenv</code> should put `<code>emulate zsh</code>' at the top of the
script.</p>
<p>There are two files run at the end: <code>~/.zlogout</code> and <code>/etc/zlogout</code>, in
that order. As their names suggest, they are counterparts of the
<code>zlogin</code> files, and therefore are only run for login shells --- though
you can trick the shell by setting the <code>login</code> option. Note that whether
you use <code>exit</code>, <code>bye</code> or <code>logout</code> to leave the shell does not affect
whether these files are run: I wasn't lying (this time) when I said that
the error message was the only difference between <code>exit</code> and <code>logout</code>.
If you want to run a file at the end of any other type of shell, you can
do it another way:</p>
<pre><code> TRAPEXIT() {
# commands to run here, e.g. if you
# always want to run .zlogout:
if [[ ! -o login ]]; then
# don't do this in a login shell
# because it happens anyway
. ~/.zlogout
fi
}
</code></pre>
<p>If you put that in <code>.zshrc</code>, it will force <code>.zlogout</code> to be run at the
end of all interactive shells. Traps will be mentioned later, but this
is rather a one-off; it's really just a hack to get commands run at the
end of the shell. I won't talk about logout files, however, since
there's little that's standard to put in them; some people make them
clear the screen to remove sensitive information with the `<code>clear</code>'
command. Other than that, you might need to tidy a few files up when you
exit.</p>
<p><span id="l10"></span></p>
<h2 id="23-options"><a class="header" href="#23-options">2.3: Options</a></h2>
<p>It's time to talk about options, since I've mentioned them several
times. Each option describes one particular shell behaviour; they are
all Boolean, i.e. can either be on or off, with no other state. They
have short names and in the documentation and this guide they are
written in uppercase with underscores separating the bits (except in
actual code, where I'll write them in the short form). However, neither
of those is necessary. In fact, <code>NO_RCS</code> and <code>norcs</code> and <code>__N_o_R_c_S__</code>
mean the same thing and are all accepted by the shell.</p>
<p>The second thing is that an option with `<code>no</code>' in front just means the
opposite of the option without. I could also have written the test `<code>[[ ! -o norcs ]]</code>' as `<code>[[ -o rcs ]]</code>'; the `<code>!</code>' means `not', as in C.
You can only have one `<code>no</code>'; `<code>nonorcs</code>' is meaningless.
Unfortunately, there is an option `<code>NOMATCH</code>' which has `<code>no</code>' as part
of its basic name, so in this case the opposite really is
`<code>NO_NOMATCH</code>'; <code>NOTIFY</code>, of course, is also a full name in its own
right.</p>
<p>The usual way to set and unset options is with the commands <strong>setopt</strong>
and <strong>unsetopt</strong> which take a string of option names. Some options also
have flags, like the `<code>-f</code>' for <code>NO_RCS</code>, which these commands also
accept, but it's much clearer to use the full name and the extra time
and space is negligible. The command `<code>set -o</code>' is equivalent to
<code>setopt</code>; this comes from ksh. Note that <code>set</code> with no `<code>-o</code>' does
something else --- that sets the positional parameters, which is zsh's
way of passing arguments to scripts and functions.</p>
<p>Almost everybody sets some options in their startup files. Since you
want them in every interactive shell, at the least, the choice is
between putting them in <code>~/.zshrc</code> or <code>~/.zshenv</code>. The choice really
depends on how you use non-interactive shells. They can be started up in
unexpected places. For example, if you use Emacs and run commands from
inside it, such as <strong>grep</strong>, that will start a non-interactive shell,
and may require some options. My rule of thumb is to put as many options
as possible into <code>~/.zshrc</code>, and transfer them to <code>~/.zshenv</code> if I find
I need them there. Some purists object to setting options in <code>~/.zshenv</code>
at all, since it affects scripts; but, as I've already hinted, you have
to work a bit harder to make sure scripts are unaffected by that sort of
thing anyway. In the following, I just assume they are going to be in
<code>~/.zshrc</code>.</p>
<p><span id="l11"></span></p>
<h2 id="24-parameters"><a class="header" href="#24-parameters">2.4: Parameters</a></h2>
<p>One more thing you'll need to know about in order to write startup files
is parameters, also known as variables. These are mostly like variables
in other programming languages. Simple parameters can be stored like
this (an <strong>assignment</strong>):</p>
<pre><code> foo='This is a parameter.'
</code></pre>
<p>Note two things: first, there are no spaces around the `<code>=</code>'. If there
was a space before, zsh would think `<code>foo</code>' was the name of a command
to execute; if there was a space after it, it would assign an empty
string to the parameter <code>foo</code>. Second, note the use of quotes to stop
the spaces inside the string having the same effect. Single quotes, as
here, are the nuclear option of quotes: everything up to another single
quote is treated as a simple string --- newlines, equal signs,
unprintable characters, the lot, in this example all would be assigned
to the variable; for example,</p>
<pre><code> foo='This is a parameter.
This is still the same parameter.'
</code></pre>
<p>So they're the best thing to use until you know what you're doing with
double quotes, which have extra effects. Sometimes you don't need them,
for example,</p>
<pre><code> foo=oneword
</code></pre>
<p>because there's nothing in `<code>oneword</code>' to confuse the shell; but you
could still put quotes there anyway.</p>
<p>Users of csh should note that you don't use `<code>set</code>' to set parameters.
This is important because there is a <code>set</code> command, but it works
differently --- if you try `<code>set var="this wont't work"</code>', you won't
get an error but you won't set the parameter, either. Type `<code>print $1</code>'
to see what you did set instead.</p>
<p>To get back what was stored in a parameter, you use the name somewhere
on the command line with a `<code>$</code>' tacked on the front --- this is called
an <strong>expansion</strong>, or to be more precise, since there are other types of
expansion, a <strong>parameter expansion</strong>. For example, after the first
assignment above.</p>
<pre><code> print -- '$foo is "'$foo'"'
</code></pre>
<p>gives</p>
<pre><code> $foo is "This is a parameter."
</code></pre>
<p>so you can see what I meant about the effect of single quotes. Note the
asymmetry --- there is no `<code>$</code>' when assigning the parameter, but there
is a `<code>$</code>' in front to have it expanded it into the command line. You
may find the word `substitution' used instead of `expansion'
sometimes; I'll try and stick with the terminology in the manual.</p>
<p>Two more things while we're at it. First, why did I put `<code>-``-</code>' after
the <code>print</code>? That's because <strong>print</strong>, like many UNIX commands, can take
options after it which begin with a `<code>-</code>'. `<code>-``-</code>' says that there
are no more options; so if what you're trying to print begins with a
`<code>-</code>', it will still print out. Actually, in this case you can see it
doesn't, so you're safe; but it's a good habit to get into, and I wish I
had. As always in zsh, there are exceptions; for example, if you use the
<code>-R</code> option to print before the `<code>-``-</code>', it only recognizes BSD-style
options, which means it doesn't understand `<code>-``-</code>'. Indeed, zsh
programmers can be quite lax about standards and often use the old, but
now non-standard, single `<code>-</code>' to show there are no more options.
Currently, this works even after <code>-R</code>.</p>
<p>The next point is that I didn't put spaces between the single quotes and
the <code>$foo</code> and it was still expanded --- expansion happens anywhere the
parameter is not quoted; it doesn't have to be on its own, just
separated from anything which might make it look like a different
parameter. This is one of those things that can help make shell scripts
look so barbaric.</p>
<p>As well as defining your own parameters, there are also a number which
the shell sets itself, and some others which have a special effect when
you set them. All the above still applies, though. For the rest of this
guide, I will indicate parameters with the `<code>$</code>' stuck in front, to
remind you what they are, but you should remember that the `<code>$</code>' is
missing when you set them, or, indeed, any time when you're referring to
the name of the parameter instead of its value.</p>
<p><span id="l12"></span></p>
<h3 id="241-arrays"><a class="header" href="#241-arrays">2.4.1: Arrays</a></h3>
<p>There is a special type of parameter called an <strong>array</strong> which zsh
inherited from both ksh and csh. This is a slightly shaky marriage,
since some of the things those two shells do with them are not
compatible, and zsh has elements of both, so you need to be careful if
you've used arrays in either. The option <code>KSH_ARRAYS</code> is something you
can set to make them behave more like they do in ksh, but a lot of zsh
users write functions and scripts assuming it isn't set, so it can be
dangerous.</p>
<p>Unlike normal parameters (known as <strong>scalars</strong>), arrays have more than
one word in them. In the examples above, we made the parameter <code>$foo</code>
get a string with spaces in, but the spaces weren't significant. If we'd
done</p>
<pre><code> foo=(This is a parameter.)
</code></pre>
<p>(note the absence of quotes), it would have created an array. Again,
there must be no space between the `<code>=</code>' and the `(', though inside
the parentheses spaces separate words just like they do on a command
line. The difference isn't obvious if you try and print it --- it looks
just the same --- but now try this:</p>
<pre><code> print -- ${foo[4]}
</code></pre>
<p>and you get `<code>parameter.</code>'. The array stores the words separately, and
you can retrieve them separately by putting the number of the element of
the array in square brackets. Note also the braces `<code>{...}</code>' --- zsh
doesn't always require them, but they make things much clearer when
things get complicated, and it's never wrong to put them in: you could
have said `<code>${foo}</code>' when you wanted to print out the complete
parameter, and it would be treated identically to `<code>$foo</code>'. The braces
simply screen off the expansion from whatever else might be lying around
to confuse the shell. It's useful too in expressions like `<code>${foo}s</code>'
to keep the `<code>s</code>' from being part of the parameter name; and, finally,
with <code>KSH_ARRAYS</code> set, the braces are compulsory, though unfortunately
arrays are indexed from 0 in that case.</p>
<p>You can use quotes when defining arrays; as before, this protects
against the shell thinking the spaces are between different elements of
the array. Try:</p>
<pre><code> foo=('first element' 'second element')
print -- ${foo[2]}
</code></pre>
<p>Arrays are useful when the shell needs to keep a whole series of
different things together, so we'll meet some you may want to put in a
startup file. Users of ksh will have noticed that things are a bit
different in zsh, but for now I'll just assume you're using the normal
zsh way of doing things.</p>
<p><span id="l13"></span></p>
<h2 id="25-what-to-put-in-your-startup-files"><a class="header" href="#25-what-to-put-in-your-startup-files">2.5: What to put in your startup files</a></h2>
<p>At the last count there were over 130 options and several dozen
parameters which are special to the shell, and many of them deal with
things I won't talk about till much later. But as a guide to get you
started, and an indication of what's to come, here are some options and
parameters you might want to think about setting in <code>~/.zshrc</code>.</p>
<p><span id="l14"></span></p>
<h3 id="251-compatibility-options-sh_word_split-and-others"><a class="header" href="#251-compatibility-options-sh_word_split-and-others">2.5.1: Compatibility options: <code>SH_WORD_SPLIT</code> and others</a></h3>
<p>I've already mentioned that zsh works differently from ksh, its nearest
standard relative, and that some of these differences can be confusing
to new users, for example the use of arrays. Some options like
<code>KSH_ARRAYS</code> exist to allow you to have things work the ksh way. Most of
these are fairly finnicky, but one catches out a lot of people. Above, I
said that after</p>
<pre><code> foo='This is a parameter.'
</code></pre>
<p>then <code>$foo</code> would be treated as one word. In traditional Bourne-like
shells including sh, ksh and bash, however, the shell will split <code>$foo</code>
on any spaces it finds. So if you run a command</p>
<pre><code> command $foo
</code></pre>
<p>then in zsh the command gets a single argument `<code>This is a parameter.</code>', but in the other shells it gets the first argument
`<code>This</code>', the second argument `<code>is</code>', and so on. If you like this, or
are so used to it it would be confusing to change, you should set the
option <code>SH_WORD_SPLIT</code> in your <code>~/.zshrc</code>. Most experienced zsh users
use arrays when they want word splitting, since as I explained you have
control over what is split and what is not; that's why <code>SH_WORD_SPLIT</code>
is not set by default. Users of other shells just get used to putting
things in double quotes,</p>
<pre><code> command "$foo"
</code></pre>
<p>which, unlike single quotes, allow the `<code>$</code>' to remain special, and
have the side effect that whatever is in quotes will remain a single
word (though there's an exception to that, too: the parameter <code>$@</code>).</p>
<p>There are a lot of other options doing similar things to keep users of
standard shells happy. Many of them simply turn features off, because
the other shell doesn't have them and hence unexpected things might
happen, or simply tweak a feature which is a little different or doesn't
usually matter. Currently such options include <code>NO_BANG_HIST</code>,
<code>BSD_ECHO</code> (sh only), <code>IGNORE_BRACES</code>, <code>INTERACTIVE_COMMENTS</code>,
<code>KSH_OPTION_PRINT</code>, <code>NO_MULTIOS</code>, <code>POSIX_BUILTINS</code>, <code>PROMPT_BANG</code>,
<code>SINGLE_LINE_ZLE</code> (I've written them how they would appear as an
argument to <code>setopt</code> to put the option the way the other shell expects,
so some have `<code>NO_</code>' in front). Most people probably won't change those
unless they notice something isn't working how they expect.</p>
<p>Some others have more noticeable effects. Here are a few of the ones
most likely to make you scratch your head if you're changing from
another Bourne-like shell.</p>
<p><strong><code>BARE_GLOB_QUAL</code>, <code>GLOB_SUBST</code>, <code>SH_FILE_EXPANSION</code>, <code>SH_GLOB</code>,
<code>KSH_GLOB</code></strong></p>
<p>These are all to do with how pattern matching works. You probably
already know that the pattern `<code>*.c</code>' will be expanded into all the
files in the current directory ending in `<code>.c</code>'. Simple uses like this
are the same in all shells, and the way filenames are expanded is often
referred to as `globbing' for historical reasons (apparently it stood
for `global replacement'), hence the name of some of these options.</p>
<p>However, zsh and ksh differ over more complicated patterns. For example,
to match either file <code>foo.c</code> or file <code>bar.c</code>, in ksh you would say
<code>@(foo|bar).c</code>. The usual zsh way of doing things is <code>(foo|bar).c</code>. To
turn on the ksh way of doing things, set the option <code>KSH_GLOB</code>; to turn
off the zsh way, set the options <code>SH_GLOB</code> and <code>NO_BARE_GLOB_QUAL</code>. The
last of those turns off <strong>qualifiers</strong>, a very powerful way of selecting
files by type (for example, directories or executable files) instead of
by name which I'll talk about in <a href="zshguide05.html#subst">chapter 5</a>.</p>
<p>The other two need a bit more explanation. Try this:</p>
<pre><code> foo='*'
print $foo
</code></pre>
<p>In zsh, you usually get a `<code>*</code>' printed, while in ksh the `<code>*</code>' is
expanded to all the files in the directory, just as if you had typed
`<code>print *</code>'. This is a little like <code>SH_WORD_SPLIT</code>, in that ksh is
pretending that the value of <code>$foo</code> appears on the command line just as
if you typed it, while zsh is using what you assigned to <code>foo</code> without
allowing it to be changed any more. To allow the word to be expanded in
zsh, too, you can set the option <code>GLOB_SUBST</code>. As with <code>SH_WORD_SPLIT</code>,
the way around the ksh behaviour if you don't want the value changed is
to use double quotes: <code>"$foo"</code>.</p>
<p>You are less likely to have to worry about <code>SH_FILE_EXPANSION</code>. It
determines when the shell expands things like <code>~/.zshrc</code> to the full
path, e.g. <code>/home/user2/pws/.zshrc</code>. In the case of zsh, this is usually
done quite late, after most other forms of expansion such as parameter
expansion. That means if you set <code>GLOB_SUBST</code> and do</p>
<pre><code> foo='~/.zshrc'
print $foo
</code></pre>
<p>you would normally see the full path, starting with a `<code>/</code>'. If you
<em>also</em> set <code>SH_FILE_EXPANSION</code>, however, the `<code>~</code>' is tested much
earlier, before <code>$foo</code> is replaced when there isn't one yet, so that
`<code>~/.zshrc</code>' would be printed. This (with both options) is the way ksh
works. It also means I lied when I said ksh treats <code>$foo</code> exactly as if
its value had been typed, because if you type <code>print ~/.zshrc</code> the
`<code>~</code>' does get expanded. So you see how convenient lying is.</p>
<p><strong><code>NOMATCH</code>, <code>BAD_PATTERN</code></strong></p>
<p>These also relate to patterns which produce file names, but in this case
they determine what happens when the pattern doesn't match a file for
some reason. There are two possible reasons: either no file happened to
match, or you didn't use a proper pattern. In both cases, zsh, unlike
ksh, prints an error message. For example,</p>
<pre><code> % print nosuchfile*
zsh: no matches found: nosuchfile*
% print [-
zsh: bad pattern: [-
</code></pre>
<p>(Remember the `<code>%</code>' lines are what you type, with a prompt in front
which comes from the shell.) You can see there are two different error
messages: you can stop the first by setting <code>NO_NOMATCH</code>, and the second
by setting <code>NO_BAD_PATTERN</code>. In both cases, that makes the shell print
out what you originally type without any expansion when there are no
matching files.</p>
<p><strong><code>BG_NICE</code>, <code>NOTIFY</code></strong></p>
<p>All UNIX shells allow you to start a <em>background</em> job by putting `<code>&</code>'
at the end of the line; then the shell doesn't wait for the job to
finish, so you can type something else. In zsh, such jobs are usually
run at a lower priority (a `higher nice value' in UNIX-speak), so that
they don't use so much of the processor's time as foreground jobs (all
the others, without the `<code>&</code>') do. This is so that jobs like editing or
using the shell don't get slowed down, which can be highly annoying. You
can turn this feature off by setting <code>NO_BG_NICE</code>.</p>
<p>When a background job finishes, zsh usually tells you immediately by
printing a message, which interrupts whatever you're doing. You can stop
this by setting <code>NO_NOTIFY</code>. Actually, this is an option in most
versions of ksh, too, but it's a little less annoying in zsh because if
it happens while you're typing something else to the shell, the shell
will reprint the line you were on as far as you've got. For example:</p>
<pre><code> % sleep 3 &
[1] 40366
% print The quick brown
[1] + 40366 done sleep 3
% print The quick brown
</code></pre>
<p>The command sleep simply does nothing for however many seconds you tell
it, but here it did it in the background (zsh printed a message to tell
you). After you typed for three seconds, the job exited, and with
<code>NOTIFY</code> set it printed out another message: the `<code>done</code>' is the key
thing, as it tells you the job has finished. But zsh was smart enough to
know the display was messed up, so it reprinted the line you were
editing, and you can continue. If you were already running another
programme in the foreground, however, that wouldn't know that zsh had
printed the message, so the display would still be messed up.</p>
<p><strong><code>HUP</code></strong></p>
<p>Signals are the way of persuading a job to do something it doesn't want
to, such as die; when you type <code>^C</code>, it sends a signal (called <code>SIGINT</code>
in this case) to the job. In zsh, if you have a background job running
when the shell exits, the shell will assume you want that to be killed;
in this case it is sent a particular signal called `<code>SIGHUP</code>' which
stands for `hangup' (as in telephone, not as in Woody Allen) and is the
UNIX equivalent of `time to go home'. If you often start jobs that
should go on even when the shell has exited, then you can set the option
<code>NO_HUP</code>, and background jobs will be left alone.</p>
<p><strong><code>KSH_ARRAYS</code></strong></p>
<p>I've already mentioned this, but here are the details. Suppose you have
defined an array <code>arr</code>, for example with</p>
<pre><code> arr=(foo bar)
</code></pre>
<p>although the syntax in ksh, which zsh also allows, is</p>
<pre><code> set -A arr foo bar
</code></pre>
<p>In zsh, <code>$arr</code> gives out the whole array; in ksh it just produces the
first element. In zsh, <code>${arr[1]}</code> refers to the first element of the
array, i.e. <code>foo</code>, while in ksh the first element is referred to as
<code>${arr[0]}</code> so that <code>${arr[1]}</code> gives you <code>bar</code>. Finally, in zsh you can
get away with <code>$arr[1]</code> to refer to an element, while ksh insists on the
braces. By setting <code>KSH_ARRAYS</code>, zsh will switch to the ksh way of doing
things. This is one option you need to be particularly careful about
when writing functions and scripts.</p>
<p><strong><code>FUNCTION_ARG_ZERO</code></strong></p>
<p>Shell functions are a useful way of specifying a set of commands to be
run by the shell. Here's a simple example:</p>
<pre><code> % fn() { print My name is $0; }
% fn
My name is fn
</code></pre>
<p>Note the special syntax: the `<code>()</code>' appears after a function name to
say you are defining one, then a set of commands appears between the
`<code>{ ... }</code>'. When you type the name of the function, those commands are
executed. If you know the programming language C, the syntax will be
pretty familiar, although note that the `<code>()</code>' is a bit of a delusion:
you might think you would put arguments to the function in there, but
you can't, it must always appear simply as `<code>()</code>'. If you don't know C,
it doesn't matter; nothing from C really applies in detail, it's just a
superficial resemblance.</p>
<p>In this case, zsh printed the special parameter `<code>$0</code>' (`argument
zero') and, as you see, that turned into the name of the function. Now
<code>$0</code> outside a function means the name of the shell, or the name of the
script for a non-interactive shell, so if you type `<code>print $0</code>' it will
probably say `<code>zsh</code>'. In most versions of ksh, this is <code>$0</code>'s only use;
it doesn't change in functions, and `fn' would print `ksh'. To get
this behaviour, you can set <code>NO_FUNCTION_ARG_ZERO</code>. There's probably no
reason why you would want to, but zsh functions quite often test their
own name, so this is one reason why they might not work.</p>
<p>There's another difference when defining functions, irrespective of
<code>FUNCTION_ARG_ZERO</code>: in zsh, you can get away without the final `<code>;</code>'
before the end of the definition of <code>fn</code>, because it knows the `<code>}</code>'
must finish the last command as well as the function; but ksh is not so
forgiving here. Lots of syntactic know-alls will probably be able to
tell you why that's a good thing, but fortunately I can't.</p>
<p><strong><code>KSH_AUTOLOAD</code></strong></p>
<p>There's an easy way of loading functions built into both ksh and zsh.
Instead of putting them all together in a big startup file, you can put
a single line in that,</p>
<pre><code> autoload fn
</code></pre>
<p>and the function `<code>fn</code>' will only be loaded when you run it by typing
its name as a command. The shell needs to know where the function is
stored. This is done by a special parameter called <code>$fpath</code>, an array
which is a list of directories; it will search all the directories for a
file called <code>fn</code>, and use that as the function definition. If you want
to try this you can type `<code>autoload fn; fpath=(. $fpath)</code>' and write a
file called <code>fn</code> in the current directory.</p>
<p>Unfortunately ksh and zsh disagree a bit about what should be in that
file. The normal zsh way of doing things is just putting the body of the
function there. So if the file <code>fn</code> is autoloadable and contains,</p>
<pre><code> # this is a simple function
print My name is $0
</code></pre>
<p>then typing `<code>fn</code>' will have exactly the same effect as the function
<code>fn</code> above, printing `<code>My name is fn</code>'. Zsh users tend to like this
because the function is written the same way as a script; if instead you
had typed <code>zsh fn</code>, to call the file as a script with a new copy of zsh
of its own, it would have worked the same way. The first line is a
comment; it's ignored, and in zsh not even autoloaded when the function
is run, so it's not only much clearer to add explanatory contents, it
also doesn't use any more memory either. It uses more disk space, of
course, but nowadays even home PCs come with the sort of disk size which
allows you a little indulgence with legibility.</p>
<p>However, ksh does things differently, and here the file <code>fn</code> needs to
contain</p>
<pre><code> fn() {
# this is a simple function
print My name is $0
}
</code></pre>
<p>in other words, exactly what you would type to define the function. The
advantage of this form is that you can put other things in the file,
which will then be run straight away and forgotten about, such as
defining things that <code>fn</code> may need to use but which don't need to be
redefined every single time you run <code>fn</code>. The option to force zsh to
work the ksh way here is called <code>KSH_AUTOLOAD</code>. (If you wanted to try
the second example, you would need to type `<code>unfunction fn; autoload fn</code>' to remove the function from memory and mark it for autoloading
again.)</p>
<p>Actually, zsh is a little bit cleverer. If the option <code>KSH_AUTOLOAD</code> is
not set, but the file contains just a function definition in the ksh
form and nothing else (like the last one above, in fact), then zsh
assumes that it needs to run the function just loaded straight away. The
other possibility would be that you wanted to define a function which
did nothing other than define a function of the same name, which is
assumed to be unlikely --- and if you really want to do that, you will
need to trick zsh by putting a do-nothing command in the same file, such
as a `<code>:</code>' on the last line.</p>
<p>A final complication --- sorry, but this one actually happens --- is
that sometimes in zsh you want to define not just the function to be
called, but some others to help it along. Then you need to do this:</p>
<pre><code> fn() {
# this is the function after which the file is named
}
helper() {
# goodness knows what this does
}
fn "$@"
# this actually calls the function the first time,
# with any arguments passed (see the subsection
# `Function Parameters' in the section `Functions'
# of the next chapter for the "$@").
</code></pre>
<p>That last non-comment line is unnecessary with <code>KSH_AUTOLOAD</code>. The
functions supplied with zsh assume that <code>KSH_AUTOLOAD</code> is not set,
however, so you shouldn't turn it on unless you need to. You could just
make <code>fn</code> into the whole body, as usual, and define <code>helper</code> inside
that; the problem is that <code>helper</code> would be redefined each time you
executed <code>fn</code>, which is inefficient. A better way of avoiding the
problem would be to define helper as a completely separate function,
itself autoloaded: in both zsh and ksh, it makes no difference whether a
function is defined inside another function or outside it, unlike (say)
Pascal or Scheme.</p>
<p><strong><code>LOCAL_OPTIONS</code>, <code>LOCAL_TRAPS</code></strong></p>
<p>These two options also refer to functions, and here the ksh way of doing
things is usually preferable, so many people set at least
<code>LOCAL_OPTIONS</code> in a lot of their functions. The first versions of zsh
didn't have these, which is why you need to turn them on by hand.</p>
<p>If <code>LOCAL_OPTIONS</code> is set in a function (or was already set before the
function, and not unset inside it), then any options which are changed
inside the function will be put back the way they were when the function
finishes. So</p>
<pre><code> fn() {
setopt localoptions kshglob
...
}
</code></pre>
<p>allows you to use a function with the ksh globbing syntax, but will make
sure that the option <code>KSH_GLOB</code> is restored to whatever it was before
when the function exits. This works even if the function was interrupted
by typing <code>^C</code>. Note that <code>LOCAL_OPTIONS</code> will itself be restored to the
way it was.</p>
<p>The option <code>LOCAL_TRAPS</code>, which first appeared in version 3.1.6, is for
a similar reason but refers to (guess what) <strong>traps</strong>, which are a way
of stopping signals sent to the shell, for example by typing <code>^C</code> to
cancel something (<code>SIGINT</code>, short for `signal interrupt'), or <code>^Z</code> to
suspend it temporarily (<code>SIGTSTP</code>, `signal terminal stop'), or <code>SIGHUP</code>
which we've already met, and so on. To do something of your own when the
shell gets a <code>^C</code>, you can do</p>
<pre><code> trap 'print I caught a SIGINT' INT
</code></pre>
<p>and the set of commands in quotes will be run when the <code>^C</code> arrives (you
can even try it without running anything). If the string is empty (just
<code>'``'</code> with nothing inside), the signal will be ignored; typing <code>^C</code> has
no effect. To put it back to normal, the command is `<code>trap - INT</code>'.</p>
<p>Traps are most useful in functions, where you may temporarily (say) not
want things to stop when you hit <code>^C</code>, or you may want to clear up
something before returning from the function. So now you can guess what
<code>LOCAL_TRAPS</code> does; with</p>
<pre><code> fn() {
setopt localoptions localtraps
trap '' INT
...
}
</code></pre>
<p>the shell will ignore <code>^C</code>'s to the end of the function, but then put
back the trap that was there before, or remove it completely if there
was none. Traps are described in more detail in <a href="zshguide03.html#syntax">chapter
3</a>.</p>
<p>There is a very convenient shorthand for making options and traps local,
as well as for setting the others to their standard values: put
`<code>emulate -L zsh</code>' at the start of a function. This sets the option
values back to the ones set when zsh starts, but with <code>LOCAL_OPTIONS</code>
and <code>LOCAL_TRAPS</code> set too, so you now know exactly how things are going
to work for the rest of the function, whatever options are set in the
outside world. In fact, this only changes the options which affect
normal programming; you can set every option which it makes sense to set
to its standard value with `<code>emulate -RL zsh</code>' (it doesn't, for
example, make sense to change options like <code>login</code> at this point).
Furthermore, you can make the shell behave as much like ksh as it knows
how to by doing `<code>emulate -L ksh</code>', with or without the <code>-R</code>.</p>
<p>The <code>-L</code> option to <code>emulate</code> actually only appears in versions from
3.0.6 and 3.1.6. Before that you needed</p>
<pre><code> emulate zsh
setopt localoptions
</code></pre>
<p>since <code>localtraps</code> didn't exist, and indeed doesn't exist in 3.0.6
either.</p>
<p><strong><code>PROMPT_PERCENT</code>, <code>PROMPT_SUBST</code></strong></p>
<p>As promised, setting prompts will be discussed later, but for now there
are two ways of getting information into prompts, such as the parameter
<code>$PS1</code> which determines the usual prompt at the start of a new command
line. One is by using <em>percent escapes</em>, which means a `<code>%</code>' followed
by another character, maybe with a number between the two. For example,
the default zsh prompt is `<code>%m%# </code>'. The first percent escape turns
into the name of the host computer, the second usually turns into a
`<code>%</code>', but a `<code>#</code>' for the superuser. However, ksh doesn't have these,
so you can turn them off by setting <code>NO_PROMPT_PERCENT</code>.</p>
<p>The usual ksh way of doing things, on the other hand, is by putting
parameters in the prompt to be substituted. To get zsh to do this, you
have to set <code>PROMPT_SUBST</code>. Then assigning</p>
<pre><code> PS1='${PWD}% '
</code></pre>
<p>is another way of putting the name of the current directory (`<code>$PWD</code>'
is presumably named after the command `pwd' to `print working
directory') into the prompt. Note the single quotes, so that this
happens when the prompt is shown, not when it is assigned. If they
weren't there, or were double quotes, then the <code>$PWD</code> would be expanded
to the directory when the assignment took place, probably your home
directory, and wouldn't change to reflect the directory you were
actually in. Of course, you need the quotes for the space, too, else it
just gets swallowed up when the assignment is executed.</p>
<p>As there is potentially much more information available in parameters
than the fixed number of predefined percent escapes, you may wish to set
<code>PROMPT_SUBST</code> anyway. Furthermore, you can get the output of commands
into prompts since other forms of expansion are done on them, not just
that of parameters; in fact, prompts with <code>PROMPT_SUBST</code> are expanded
pretty much the same as a string inside double quotes every time the
prompt is displayed.</p>
<p><strong><code>RM_STAR_SILENT</code></strong></p>
<p>Everybody at some time or another deletes more files than they mean to
(and <em>that's</em> a gross understatement); my favourite is:</p>
<pre><code> rm *>o
</code></pre>
<p>That `<code>></code>' should be a `.', but I still had the shift key pressed.
This removes all files, echoing the output (there isn't any) into a file
`o'. Delightfully, the empty file `o' is not removed. (Don't try this
at home.)</p>
<p>There is a protection mechanism built into zsh to stop you deleting all
the files in a directory by accident. If zsh finds that the command is
`<code>rm</code>', and there is a `<code>*</code>' on the command line (there may be other
stuff as well), then it will ask you if you really want to delete all
those files. You can turn this off by setting <code>RM_STAR_SILENT</code>.
Overreliance on this option is a bad idea; it's only a last line of
defence.</p>
<p><strong><code>SH_OPTION_LETTERS</code></strong></p>
<p>Many options also have single letters to stand for them; you can set an
option in this way by, for example, `<code>set -f</code>', which sets <code>NO_RCS</code>.
However, even where sh, ksh and zsh share options, not all have the same
letters. This option allows the single letter options to be more like
those in sh and ksh. Look them up in the manual if you want to know, but
I already recommended that you use the full names for options anyway.</p>
<p><strong><code>SH_WORD_SPLIT</code></strong></p>
<p>I've already talked about this, see above, but it's mentioned here so
you don't forget it, since it's an important difference.</p>
<p><strong>Starting zsh as ksh</strong></p>
<p>Finally on the subject of compatibility, you might like to know that as
well as `<code>emulate</code>' there is another way of forcing zsh to behave as
much like sh or ksh as possible. This is by actually calling zsh under
the name ksh. You don't need to rename zsh, you can make a link from the
name zsh to the name ksh, which will be enough to convince it.</p>
<p>There is an easier way when you are doing this from within zsh itself.
The parameter <code>$ARGV0</code> is special; it is the value which will be passed
as the first argument of a command which is run by the shell. Normally
this is the name of the command, but it doesn't have to be since the
command only finds out what it is after it has already been run. You can
use it to trick a programme into thinking its name is different. So</p>
<pre><code> ARGV0=ksh zsh
</code></pre>
<p>will start a copy of zsh that tries to make itself like ksh. Note this
doesn't work unless you're already in zsh, as the <code>$ARGV0</code> won't be
special.</p>
<p>I haven't mentioned putting a parameter assignment before a command
name, but that simply assigns the parameter (strictly an environment
variable in this case) for the duration of the command; the value
<code>$ARGV0</code> won't be set after that command (the ksh-like zsh) finishes, as
you can easily test with <code>print</code>. While I'm here, I should mention a few
of its other features. First, the parameter is automatically exported to
the environment, meaning it's available for other programmes started by
zsh (including, in this case, the new zsh) --- see the section on
environment variables below. Second, this doesn't do what you might
expect:</p>
<pre><code> FOO=bar print $FOO
</code></pre>
<p>because of the order of expansion: the command line and its parameters
are expanded before execution, giving whatever value <code>$FOO</code> had before,
probably none, then FOO=bar is put into the environment, and then the
command is executed but doesn't use the new value of <code>$FOO</code>.</p>
<p><span id="l15"></span></p>
<h3 id="252-options-for-csh-junkies"><a class="header" href="#252-options-for-csh-junkies">2.5.2: Options for csh junkies</a></h3>
<p>As well as old ksh users, there are some options available to make old
csh and tcsh users feel more at home. As you will already have noticed,
the syntax is very different, so you are never going to feel completely
at home and it might be best just to remember the fact. But here is a
brief list. The last, <code>CSH_NULL_GLOB</code>, is actually quite useful.</p>
<p><strong><code>CSH_JUNKIE_HISTORY</code></strong></p>
<p>Zsh has the old csh mechanism for referring to words on a previous
command line using a `<code>!</code>'; it's less used, now the editor is more
powerful, but is still a convenient shorthand for extracting short bits
from the previous line. This mechanism is sometimes called
<strong>bang-history</strong>, since busy people sometimes like to say `<code>!</code>' as
`bang'. This option affects how a single `<code>!</code>' works. For example,</p>
<pre><code> % print foo bar
% print open closed
% print !-2:1 !:2
</code></pre>
<p>In the last line, `<code>!-2</code>' means two entries ago, i.e. the line `<code>print foo bar</code>'. The `<code>:1</code>' chooses the first word after the command, i.e.
`<code>foo</code>'. In the second expression, no number is given after the `<code>!</code>'.
Usually zsh interprets that to mean that the same item just selected, in
this case -2, should be used. With <code>CSH_JUNKIE_HISTORY</code> set, it refers
instead to the last command. Note that if you hadn't given that -2, it
would refer to the last command in any case, although the explicit way
of referring to the last command is `<code>!!</code>' --- you have to use that if
there are no `<code>:</code>' bits following. In summary, zsh usually gives you
`<code>print foo bar</code>'; with <code>CSH_JUNKIE_HISTORY</code> you get `<code>print foo closed</code>'.</p>
<p>There's another option controlling this, <code>BANG_HIST</code>. If you unset that,
the mechanism won't work at all. There's also a parameter, <code>$histchars</code>.
The first character is the main history expansion character, normally
`<code>!</code>' of course; the second is for rapid substitutions (normally `<code>^</code>'
--- use of this is described below); the third is the character
introducing comments, normally `<code>#</code>'. Changing the third character is
definitely not recommended. There's little real reason to change any.</p>
<p><strong><code>CSH_JUNKIE_LOOPS</code></strong></p>
<p>Normal zsh loops look something like this,</p>
<pre><code> while true; do
print Never-ending story
done
</code></pre>
<p>which just prints the message over and over (type it line-by-line at the
prompt, if you like, then <code>^C</code> to stop it). With <code>CSH_JUNKIE_LOOPS</code> set,
you can instead do</p>
<pre><code> while true
print Never-ending story
end
</code></pre>
<p>which will, of course, make your zsh code unlike most other people's, so
for most users it's best to learn the proper syntax.</p>
<p><strong><code>CSH_NULL_GLOB</code></strong></p>
<p>This is another of the family of options like <code>NO_NOMATCH</code>, already
mentioned. In this case, if you have a command line consisting of a set
of patterns, at least one of them must match at least one file, or an
error is caused; any that don't match are removed from the command line.
The default is that all of them have to match. There is one final member
of this set of options, <code>NULL_GLOB</code>: all non-matching patterns are
removed from the command line, no error is caused. As a summary, suppose
you enter the command `<code>print file1* file2*</code>' and the directory
contains just the file <code>file1.c</code>.</p>
<ol>
<li>By default, there must be files matching both patterns, so an error
is reported.</li>
<li>With <code>NO_NOMATCH</code> set, any patterns which don't match are left
alone, so `<code>file1.c file2*</code>' is printed.</li>
<li>With <code>CSH_NULL_GLOB</code> set, <code>file1*</code> matched, so <code>file2*</code> is silently
removed; `<code>file1.c</code>' is reported. If that had not been there, an
error would have been reported.</li>
<li>With <code>NULL_GLOB</code> set, any patterns which don't match are removed, so
again `<code>file1.c</code>' is printed, but in this case if that had not been
there a blank line would have been printed, with no error.</li>
</ol>
<p><code>CSH_NULL_GLOB</code> is good thing to have set since it can keep you on the
straight and narrow without too many unwanted error messages, so this
time it's not just for csh junkies.</p>
<p><strong><code>CSH_JUNKIE_QUOTES</code></strong></p>
<p>Here just for completeness. Csh and friends don't allow multiline
quotes, as zsh does; if you don't finish a pair of quotes before a new
line, csh will complain. This option makes zsh do the same. But
multi-line quotes are very useful and very common in zsh scripts and
functions; this is only for people whose minds have been really screwed
up by using csh.</p>
<p><span id="l16"></span></p>
<h3 id="253-the-history-mechanism-types-of-history"><a class="header" href="#253-the-history-mechanism-types-of-history">2.5.3: The history mechanism: types of history</a></h3>
<p>The name `history mechanism' refers to the fact that zsh keeps a
`history' of the commands you have typed. There are three ways of
getting these back; all these use the same set of command lines, but the
mechanisms for getting at them are rather different. For some reason,
items in the history list (a complete line of input typed and executed
at once) have become known as `events'.</p>
<p><strong>Editing the history directly</strong></p>
<p>First, you can use the editor; usually hitting up-arrow will take you to
the previous line, and down-arrow takes you back. This is usually the
easiest way, since you can see exactly what you're doing. I will say a
great deal more about the editor in <a href="zshguide04.html#zle">chapter 4</a>;
the first thing to know is that its basic commands work either like
emacs, or like vi, so if you know one of those, you can start editing
lines straight away. The shell tries to guess whether to use emacs or vi
from the environment variables <code>$VISUAL</code> or <code>$EDITOR</code>, in that order;
these traditionally hold the name of your preferred editor for
programmes which need you to edit text. In the old days, <code>$VISUAL</code> was a
full-screen editor and <code>$EDITOR</code> a line editor, like <code>ed</code> of blessed
memory, but the distinction is now very blurred. If either contains the
string <code>vi</code>, the line editor will start in vi mode, else it will start
in emacs mode. If you're in the wrong mode, `<code>bindkey -e</code>' in
<code>~/.zshrc</code> takes you to emacs mode and `<code>bindkey -v</code>' to vi mode. For
vi users, the thing to remember is that you start in insert mode, so
type `<code>ESC</code>' to be able to enter vi commands.</p>
<p><strong>`Bang'-history</strong></p>
<p>Second, you can use the csh-style `bang-history' mechanism (unless you
have set the option <code>NO_BANG_HIST</code>); the `bang' is the exclamation
mark, `!', also known as `pling' or `shriek' (or factorial, but
that's another story). Thus `<code>!!</code>' retrieves the last command line and
executes it; `<code>!-2</code>' retrieves the second last. You can select words:
`<code>!!:1</code>' picks the first word after the command of the last command (if
you were paying attention above, you will note you just need one `<code>!</code>'
in that case); <code>0</code> after colon would pick the command word itself;
`<code>*</code>' picks all arguments after the command; `<code>$</code>' picks the last
word. You can even have ranges: `<code>!!:1-3</code>' picks those three words, and
things like `<code>!!:3-$</code>' work too.</p>
<p>After the word selector, you can have a second set of colons and then
some special commands called <strong>modifiers</strong> --- these can be very useful
to remember, since they can be applied to parameters and file patterns
to, so here's some more details. The `<code>:t</code>' (tail) modifier picks the
last part of a filename, everything after the last slash; conversely,
`<code>:h</code>' (head) picks everything before that. So with a history entry,</p>
<pre><code> % print /usr/bin/cat
/usr/bin/cat
% print !!:t
print cat
cat
</code></pre>
<p>Note two things: first, the bang-history mechanism always prints what
it's about to execute. Secondly, you don't need the word selector; the
shell can tell that the `<code>:t</code>' is a modifier, and assumes you want it
applied to the entire previous command. (Be careful here, since actually
the <code>:t</code> will reduce the expression to everything after the last slash
in <em>any</em> word, which is a little unexpected.)</p>
<p>With parameters:</p>
<pre><code> % foo=/usr/bin/cat
% print ${foo:h}
/usr/bin
</code></pre>
<p>(you can usually omit the `<code>{</code>' and `<code>}</code>', but it's clearer and safer
with them). And finally with files --- this won't work if you set
<code>NO_BARE_GLOB_QUAL</code> for sh-like behaviour:</p>
<pre><code> % print /usr/bin/cat(:t)
cat
</code></pre>
<p>where you need the parentheses to tell the shell the `<code>:t</code>' isn't just
part of the file name.</p>
<p>For a complete list, see the <code>zshexpn</code> manual, or the section
<code>Modifiers</code> in the printed or Info versions of the manual, but here are
a few more of the most useful. `<code>:r</code>' removes the suffix of a file,
turning <code>file.c</code> into <code>file</code>; `<code>:l</code>' and `<code>:u</code>' make the word(s) all
lowercase or all uppercase; `<code>:s/foo/bar/</code>' substitutes the first
occurrence of <code>foo</code> with <code>bar</code> in the word(s); `<code>:gs/foo/bar</code>'
substitutes all occurrences (the `<code>g</code>' stands for global); `<code>:&</code>'
repeats the last such substitution, even if you did it on a previous
line; `<code>:g&</code>' also works. So</p>
<pre><code> % print this is this line
this is this line
% !!:s/this/that/
print that is this line
that is this line
% print this is no longer this line
this is no longer this line
% !!:g&
print that is no longer that line
that is no longer that line
</code></pre>
<p>Finally, there is a shortcut: <code>^old^new^</code> is exactly equivalent to
<code>!!:s/old/new/</code>; you can even put another modifier after it. The `<code>^</code>'
is actually the second character of <code>$histchars</code> mentioned above. You
can miss out the last `<code>^</code>' if there's nothing else to follow it. By
the way, you can put modifiers together, but each one needs the colon
with it: <code>:t:r</code> applied to `<code>dir/file.c</code>' produces `<code>file</code>', and
repeated applications of <code>:h</code> get you shorter and shorter paths.</p>
<p>Before we leave bang-history, note the option <code>HIST_VERIFY</code>. If that's
set, then after a substitution the line appears again with the changes,
instead of being immediately printed and executed. As you just have to
type <code><RET></code> to execute it, this is a useful trick to save you executing
the wrong thing, which can easily happen with complicated bang-history
lines; I have this set myself.</p>
<p>And one last tip: the shell's expansion and completion, which I will
enthuse about at length later on, allows you to expand bang-history
references straight away by hitting <code>TAB</code> immediately after you've typed
the complete reference, and you can usually type control together with
slash (on some keyboards, you are restricted to <code>^Xu</code>) to put it back
the way it was if you don't like the result --- this is part of the
editor's `undo' feature.</p>
<p><strong>Ksh-style history commands</strong></p>
<p>The third form of history uses the <code>fc</code> builtin. It's the most
cumbersome: you have to tell the command which complete lines to
execute, and may be given a chance to edit them first (but using an
external editor, not in the shell). You probably won't use it that way,
but there are three things which are actually controlled by <code>fc</code> which
you might use: first, the `<code>r</code>' command repeats the last command
(ignoring <code>r</code>'s), which is a bit like `<code>!!</code>'. Secondly, the command
called `<code>history</code>' is also really <code>fc</code> in disguise. It gives you a list
of recent commands. They have numbers next to them; you can use these
with bang-history instead of using negative numbers to count backward in
the way I originally explained, the advantage being they don't change as
you enter more commands. You can give ranges of numbers to <code>history</code>,
the first number for where to start listing, and the second where to
stop: a particular example is `<code>history 1</code>', which lists all commands
(even if the first command it still remembers is higher than 1; it just
silently omits all those). The third use of <code>fc</code> is for reading and
writing your history so you can keep it between sessions.</p>
<p><span id="l17"></span></p>
<h3 id="254-setting-up-history"><a class="header" href="#254-setting-up-history">2.5.4: Setting up history</a></h3>
<p>In fact, the shell is able to read and write history without being told.
You need to tell it where to save the history, however, and for that you
have to set the parameter <code>$HISTFILE</code> to the name of the file you want
to use (a common choice is `<code>~/.history</code>'). Next, you need to set the
parameter <code>$SAVEHIST</code> to the number of lines of your history you want
saved. When these two are set, the shell will read <code>$HISTSIZE</code> lines
from <code>$HISTFILE</code> at the start of an interactive session, and save the
last <code>$SAVEHIST</code> lines you executed at the end of the session. For it to
read or write in the middle, you will either need to set one of the
options described below (<code>INC_APPEND_HISTORY</code> and <code>SHARE_HISTORY</code>), or
use the <code>fc</code> command: <code>fc -R</code> and <code>fc -W</code> read and write the history
respectively, while <code>fc -A</code> appends it to the the file (although pruning
it if it's longer than <code>$SAVEHIST</code>); <code>fc -WI</code> and <code>fc -AI</code> are similar,
but the <code>I</code> means only write out events since the last time history was
written.</p>
<p>There is a third parameter <code>$HISTSIZE</code>, which determines the number of
lines the shell will keep within one session; except for special reasons
which I won't talk about, you should set <code>$SAVEHIST</code> to be no more than
<code>$HISTSIZE</code>, though it can be less. The default value for <code>$HISTSIZE</code> is
30, which is a bit stingy for the memory and disk space of today's
computers; zsh users often use anything up to 1000. So a simple set of
parameters to set in <code>.zshrc</code> is</p>
<pre><code> HISTSIZE=1000
SAVEHIST=1000
HISTFILE=~/.history
</code></pre>
<p>and that is enough to get things working. Note that you <em>must</em> set
<code>$SAVEHIST</code> and <code>$HISTFILE</code> for automatic reading and writing of history
lines to work.</p>
<p><span id="l18"></span></p>
<h3 id="255-history-options"><a class="header" href="#255-history-options">2.5.5: History options</a></h3>
<p>There are also many options affecting history; these increased
substantially with version 3.1.6, which provided for the first time
<code>INC_APPEND_HISTORY</code>, <code>SHARE_HISTORY</code>, <code>HIST_EXPIRE_DUPS_FIRST</code>,
<code>HIST_IGNORE_ALL_DUPS</code>, <code>HIST_SAVE_NO_DUPS</code> and <code>HIST_NO_FUNCTIONS</code>. I
have already described <code>BANG_HIST</code>, <code>CSH_JUNKIE_HISTORY</code> and
<code>HIST_VERIFY</code> and I won't talk about them again.</p>
<p><strong><code>APPEND_HISTORY</code>, <code>INC_APPEND_HISTORY</code>, <code>SHARE_HISTORY</code></strong></p>
<p>Normally, when it writes a history file, zsh just overwrites everything
that's there. <code>APPEND_HISTORY</code> allows it to append the new history to
the old. The shell will make an effort not to write out lines which
should be there already; this can get complicated if you have lots of
zshs running in different windows at once. This option is a good one for
most people to use. <code>INC_APPEND_HISTORY</code> means that instead of doing
this when the shell exits, each line is added to the history in this way
as it is executed; this means, for example, that if you start up a zsh
inside the main shell its history will look like that of the main shell,
which is quite useful. It also means the ordering of commands from
different shells running at the same time is much more logical ---
basically just the order they were executed --- so for 3.1.6 and higher
this option is recommended.</p>
<p><code>SHARE_HISTORY</code> takes this one stage further: as each line is added, the
history file is checked to see if anything was written out by another
shell, and if so it is included in the history of the current shell too.
This means that zsh's running in different windows but on the same host
(or more generally with the same home directory) share the same history.
Note that zsh tries not to confuse you by having unexpected history
entries pop up: if you use <code>!</code>-style history, the commands from other
session don't appear in the history list until you explicitly type the
<code>history</code> command to display them, so that you can be sure what command
you are actually reexecuting. The Korn shell always behaves as if
<code>SHARE_HISTORY</code> is set, presumably because it doesn't store history
internally.</p>
<p><strong><code>EXTENDED_HISTORY</code></strong></p>
<p>This makes the format of the history entry more complicated: in addition
to just the command, it saves the time when the command was started and
how long it ran for. The <code>history</code> command takes three options which use
this: <code>history -d</code> prints the start time of the command; <code>history -f</code>
prints that as well as the date; <code>history -D</code> (which you can combine
with <code>-f</code> or <code>-d</code>) prints the command's elapsed time. The date format
can be changed with <code>-E</code> for European (<em>day</em>.<em>month</em>.<em>year</em>) and <code>-i</code>
for international (<em>year</em>-<em>month</em>-<em>day</em>) formats. The main reasons why
you <em>wouldn't</em> want to set this would be shortage of disk space, or
because you wanted your history file to be read by another shell.</p>
<p><strong><code>HIST_IGNORE_DUPS</code>, <code>HIST_IGNORE_ALL_DUPS</code>, <code>HIST_EXPIRE_DUPS_FIRST</code>,
<code>HIST_SAVE_NO_DUPS</code>, <code>HIST_FIND_NO_DUPS</code></strong></p>
<p>These options give ways of dealing with the duplicate lines that often
appear in the history. The simplest is <code>HIST_IGNORE_DUPS</code>, which tells
the shell not to store a history line if it's the same as the previous
one, thus collapsing a lot of repeated commands down to one; this is a
very good option to have set. It does nothing when duplicate lines are
not adjacent, so for example alternating pairs of commands will always
be stored. The next two options can help here: <code>HIST_IGNORE_ALL_DUPS</code>
simply removes copies of lines still in the history list, keeping the
newly added one, while <code>HIST_EXPIRE_DUPS_FIRST</code> is more subtle: it
preferentially removes duplicates when the history fills up, but does
nothing until then. <code>HIST_SAVE_NO_DUPS</code> means that whatever options are
set for the current session, the shell is not to save duplicated lines
more than once; and <code>HIST_FIND_NO_DUPS</code> means that even if duplicate
lines have been saved, searches backwards with editor commands don't
show them more than once.</p>
<p><strong><code>HIST_ALLOW_CLOBBER</code>, <code>HIST_REDUCE_BLANKS</code></strong></p>
<p>These allow the history mechanism to make changes to lines as they are
entered. The first affects output redirections, where you use the symbol
<code>></code> to redirect the output of a command or set of commands to a named
file, or use <code>>``></code> to append the output to that file. If you have the
<code>NO_CLOBBER</code> option set, then</p>
<pre><code> touch newfile
echo hello >newfile
</code></pre>
<p>fails, because the `<code>touch</code>' command has created <code>newfile</code> and
<code>NO_CLOBBER</code> won't let you overwrite (clobber) it in the next line. With
<code>HIST_ALLOW_CLOBBER</code>, the second line appears in the history as</p>
<pre><code> echo hello >|newfile
</code></pre>
<p>where the <code>>|</code> overrides <code>NO_CLOBBER</code>. So to get round the <code>NO_CLOBBER</code>
you can just go back to the previous line and execute it without editing
it.</p>
<p>The second option, <code>HIST_REDUCE_BLANKS</code>, will tidy up the line when it
is entered into the history by removing any excess blanks that mean
nothing to the shell. This can also mean that the line becomes a
duplicate of a previous one even if it would not have been in its
untidied form. It is smart enough not to remove blanks which are
important, i.e. are quoted.</p>
<p><strong><code>HIST_IGNORE_SPACE</code>, <code>HIST_NO_STORE</code>, <code>HIST_NO_FUNCTIONS</code></strong></p>
<p>These three options allow you to say that certain lines shouldn't go
into the history at all. <code>HIST_IGNORE_SPACE</code> means that lines which
begin with a space don't go into the history; the idea is that you
deliberately type a space, which is not otherwise significant to the
shell, before entering any line you want to be forgotten immediately
afterwards. In zsh 4.0.1 this is implemented so that you can always
recall the immediately preceding line for editing, even if it had a
space; but when the next line is executed and entered into the history,
the line beginning with the space is forgotten.</p>
<p><code>HIST_NO_STORE</code> tells the shell not to store <code>history</code> or <code>fc</code> commands.
while <code>HIST_NO_FUNCTIONS</code> tells it not to store function definitions as
these, though usually infrequent, can be tiresomely long. A function
definition is anything beginning `<code>function funcname {...</code>' or
`<code>funcname () { ...</code>'.</p>
<p><strong><code>NO_HIST_BEEP</code></strong></p>
<p>Finally, <code>HIST_BEEP</code> is used in the editor: if you try to scroll up or
down beyond the end of the history list, the shell will beep. It is on
by default, so use <code>NO_HIST_BEEP</code> to turn it off.</p>
<p><span id="l19"></span></p>
<h3 id="256-prompts"><a class="header" href="#256-prompts">2.5.6: Prompts</a></h3>
<p>Most people have some definitions in <code>.zshrc</code> for altering the prompt
you see at the start of each line. I've already mentioned
<code>PROMPT_PERCENT</code> (set by default) and <code>PROMPT_SUBST</code> (unset by default);
I'll assume here you haven't changed these settings, and point out some
of the possibilities with <strong>prompt escapes</strong>, sequences that start with
a `<code>%</code>'. If you get really sophisticated, you might need to turn on
<code>PROMPT_SUBST</code>.</p>
<p>The main prompt is in a parameter called either <code>$PS1</code> or <code>$PROMPT</code> or
<code>$prompt</code>; the reason for having all these names is historical --- they
come from different shells --- so I'll just stick with the shortest.
There is also <code>$RPS1</code>, which prints a prompt at the right of the screen.
The point of this is that it automatically disappears if you type so far
along the line that you run into it, so it can help make the best use of
space for showing long things like directories.</p>
<p><code>$PS2</code> is shown when the shell is waiting for some more input, i.e. it
knows that what you have typed so far isn't a complete line: it may
contain the start of a quoted expression, but not the end, or the start
of some syntactic structure which is not yet finished. Usually you will
keep it different from <code>$PS1</code>, but all the same escapes are understood
in all five prompts.</p>
<p><code>$PS3</code> is shown within a loop started by the shell's <code>select</code> mechanism,
when the shell wants you to input a choice: see the <code>zshmisc</code> manual
page as I won't say much about that.</p>
<p><code>$PS4</code> is useful in debugging: there is an option <code>XTRACE</code> which causes
the shell to print out lines about to be executed, preceded by <code>$PS4</code>.
Only from version 3.1.6 has it started to be substituted in the same way
as the other prompts, though this turns out to be very useful --- see
`Location in script or function' in the following list.</p>
<p>Here are some of the things you might want to include in your prompts.
Note that you can try this out before you alter the prompt by using
`<code>print -P</code>': this expands strings just are as they are in prompts. You
will probably need to put the string in single quotes.</p>
<p><strong>The time</strong></p>
<p>Zsh allows you lots of different ways of putting the time into your
prompt with percent escapes. The simplest are <code>%t</code> and <code>%T</code>, the time in
12 and 24 hour formats, and <code>%*</code>, the same as <code>%T</code> but with seconds; you
can also have the date as (e.g.) `<code>Wed 22</code>' using <code>%w</code>, as `<code>9/22/99</code>'
(US format) using %W, or as `<code>99-09-22</code>' (International format) using
%D. However, there is another way of using %D to get many more
possibilities: a following string in braces, `<code>%D{...}</code>' can contain a
completely different set of percent escapes all of which refer to
elements of the time and date. On most systems, the documentation for
the <code>strftime</code> function will tell you what these are. zsh has a few of
its own, given in the <code>zshmisc</code> manual page in the <code>PROMPT EXPANSION</code>
section. For example, I use <code>%D{%L:%M}</code> which gives the time in hours
and minutes, with the hours as a single digit for 1 to 9; it looks more
homely to my unsophisticated eyes.</p>
<p>You can have more fun by using the `<code>%</code>(<em>numX</em>.<em>true</em>.<em>false</em>)' syntax,
where <em>X</em> is one of <code>t</code> or <code>T</code>. For <code>t</code>, if the time in minutes is the
same as <em>num</em> (default zero), then <em>true</em> is used as the text for this
section of the prompt, while <em>false</em> is used otherwise. <code>T</code> does the
same for hours. Hence</p>
<pre><code> PS1='%(t.Ding!.%D{%L:%M})%# '
</code></pre>
<p>prints the message `<code>Ding!</code>' at zero minutes past the hour, and a more
conventional time otherwise. The `<code>%#</code>' is the standard sequence which
prints a `<code>#</code>' if you are the superuser (root), or a `<code>%</code>' for
everyone else, which occurs in a lot of people's prompts. Likewise, you
could use `<code>%</code>(<code>30t.Dong!.</code>...' for a message at half past the hour.</p>
<p><strong>The current directory</strong></p>
<p>The sequence `<code>%~</code>' prints out the directory, with any home or named
directories (see below) shortened to the form starting with <code>~</code>; the
sequence `<code>%/</code>' doesn't do that shortening, so usually `<code>%~</code>' is
better. Directories can be long, and there are various ways to deal with
it. First, if you are using a windowing system you can put the directory
in the title bar, rather than anywhere inside the window. Second, you
can use <code>$RPS1</code> which disappears when you type near it. Third, you can
pick segments out of `<code>%~</code>' or `<code>%/</code>' by giving them a number after
the `<code>%</code>': for example, `<code>%1~</code>' just picks out the last segment of the
path to the current directory.</p>
<p>The fourth way gives you the most control. Prompts or parts of prompts,
not just bits showing the directory, can be truncated to any length you
choose. To truncate a path on the left, use something like
`<code>%10<</code><em>...</em><code><%~</code>'. That works like this: the `<code>%<``<</code>' is the basic
form for truncation. The 10 after the `<code>%</code>' says that anything
following is limited to 10 characters, and the characters `<em>...</em>' are
to be displayed whenever the prompt would otherwise be longer than that
(you can leave this empty). This applies to anything following, so now
the <code>%~</code> can't be longer than 10 characters, otherwise it will be
truncated (to 7 characters, once the `<code>...</code>' has been printed). You can
turn off truncation with `<code>%<``<</code>', i.e. no number after the `<code>%</code>';
truncation then applies to the entire region between where it was turned
on and where it was turned off (this has changed from older versions of
zsh, where it just applied to individual `<code>%</code>' constructs).</p>
<p><strong>What are you waiting for?</strong></p>
<p>The prompt <code>$PS2</code> appears when the shell is waiting for you to finish
entering something, and it's useful to know what the shell is waiting
for. The sequence `<code>%_</code>' shows this. It's part of the default <code>$PS2</code>,
which is `<code>%_> </code>'. Hence, if you type `<code>if true; then</code>' and <code><RET></code>,
the prompt will say `<code>then> </code>'. You can also use it in the trace
prompt, <code>$PS4</code>, to show the same information about what is being
executed in a script or function, though as there is usually enough
information there (as described next) it's not part of the default. In
this case, a number after the `<code>%</code>' will limit the depth shown, so with
`<code>%1_</code>' only the most recent thing will be mentioned.</p>
<p><strong>Location in script or function</strong></p>
<p>The default <code>$PS4</code> contains `<code>%N</code>' and `<code>%i</code>', which tell you the name
of the most recently started function, script, or sourced file, and the
line number being executed inside it; they are not very useful in other
prompts. However, `<code>%i</code>' in <code>$PS1</code> will tell you the current
interactive line number, which zsh keeps track of, though doesn't
usually show you; the parameter <code>$LINENO</code> contains the same information.</p>
<p>Another point to bear about `<code>%i</code>' in mind is that the line number
shown applies to the version of a function first read in, not how it
appears with the `<code>functions</code>' command, which is tidied up. If you use
autoloaded functions, however, the file containing the function will
usually be what you want to alter, so this shouldn't be a problem when
debugging.</p>
<p>Remember, the <code>$PS4</code> display only happens when the <code>XTRACE</code> option is
set; as options may be local to functions, and always are to scripts,
you will often need to put an explicit `<code>setopt xtrace</code>' at the top of
whatever you are debugging. Alternatively, you can use `<code>typeset -ft</code>
<em>funcname</em>' to turn on tracing for that function (something I only just
discovered); use `<code>typeset +ft</code> <em>funcname</em>' to turn it off again.</p>
<p><strong>Other bits and pieces</strong></p>
<p>There are many other percent escapes described in the <code>zshmisc</code> manual
page, mostly straightforward. For example, `<code>%h</code>' shows you the history
entry number, useful if you are using bang-history; `<code>%m</code>' shows you
the current host name up to any dot; `<code>%n</code>' shows the username.</p>
<p>There are two other features I happen to use myself. First, it's
sometimes convenient to know when the last command failed. Every command
returns a status, which is a number: zero for success, some other number
for some type of failure. You can get this from the parameter `<code>$?</code>' or
`<code>$status</code>' (again, they refer to the same thing). It's also available
in the prompt as `<code>%?</code>', and there's also one of the so-called
`ternary' expressions with parentheses I described for time, which pick
different strings depending on a test. Here the test is, reasonably
enough, `<code>%</code>(<code>?...</code>'. Putting these two together, you can get a message
which is only displayed when the exit status is non-zero; I've put an
extra set of parentheses around the number just to make it clearer,
where the `)' needs to be turned into `<code>%</code>)' to stop it marking the
end of the group:</p>
<pre><code> PS1='%(?..(%?%))%# '
</code></pre>
<p>It's also sometimes convenient to know if you're in a subshell, that is
if you've started another shell within the main one by typing `<code>zsh</code>'.
You can do this by using another ternary expression:</p>
<pre><code> PS1='%(2L.+.)%# '
</code></pre>
<p>This checks the parameter <code>SHLVL</code>, which is incremented every time a new
zsh starts, so if there was already one running (which would have set
<code>SHLVL</code> to 1), it will now be 2; and if <code>SHLVL</code> is at least 2, an extra
`<code>+</code>' is printed in front of the prompt, otherwise nothing. If you're
using a windowing system, you may need to turn the 2 into 3 as there may
be a zsh already running when you first log in, so that the shells in
the windows have <code>SHLVL</code> set to 2 already. This depends a good deal on
how your windowing system is set up; finding out more is left as an
exercise for the reader.</p>
<p><strong>Colours</strong></p>
<p>Many terminals can now display colours, and it is quite useful to be
able to put these into prompts to distinguish those from the surrounding
text. I often find a programme has just dumped a whole load of output on
my terminal and it's not obvious where it starts. Being able to find the
prompt just before helps a lot.</p>
<p>Colors, like bold or underlined text, use escape sequences which don't
move the cursor. The golden rule for inserting any such escape sequences
into prompts is to surround them with `<code>%{</code>' at the start and `<code>%}</code>'
at the end. Otherwise, the shell will be confused about the length of
the line. This affects what happens when the line editor needs to redraw
the line, and also changes the position of the right prompt <code>$RPS1</code>, if
you use that. You don't need that with the special sequences <code>%B</code> and
<code>%b</code>, which start and stop bold text, because the shell already knows
what to do with those; it's only random characters which you happen to
know don't move the cursor, though the shell doesn't, that cause the
problem.</p>
<p>In the case of colours, there is a shell function <code>colors</code> supplied with
the standard distribution to help you. When loaded and run, it defines
associative array parameters <code>$fg</code> and <code>$bg</code> which you use to extract
the escape sequences for given colours, for example
<code>${fg[red]}${bg[yellow]}</code> produces the sequences for red text on a
yellow background. So for example,</p>
<pre><code> PS1="%{${bg[white]}${fg[red]}%}%(?..(%?%))\
%{${fg[yellow]}${bg[black]}%}%# "
</code></pre>
<p>produces a red-on-white `<code>(1)</code>' if the previous programme exited with
status 1, but nothing if it exited with status 0, followed by a
yellow-on-black `<code>%</code>' or `<code>#</code>' if you are the superuser. Note the use
of the double quotes here to force the parameters to be expanded
straight away --- the escape sequences are fixed, so they don't need to
be re-extracted from the parameters every time the prompt is shown.</p>
<p>Even if your terminal does support colour, there's no guarantee all the
possibilities work, although the basic ANSI colour scheme is fairly
standard. The colours understood are: cyan, white, yellow, magenta,
black, blue, red, grey, green. You can also used `default', which puts
the terminal back how it was to begin with. In addition, you can use the
basic colours with the parameters <code>$bg_bold</code> and <code>$fg_bold</code> for bold
varieties of the colours and <code>$bg_no_bold</code> and <code>$fg_no_bold</code> to switch
explicitly back to non-bold.</p>
<p><strong>Themes</strong></p>
<p>There are also a set of themes provided as functions to set up your
prompt to various predefined possibilities. These make use of the
colours set up as described above. See the <code>zshcontrib</code> manual page for
how to do this (search for `prompt themes').</p>
<p><span id="l20"></span></p>
<h3 id="257-named-directories"><a class="header" href="#257-named-directories">2.5.7: Named directories</a></h3>
<p>As already mentioned, `<code>~/</code>' at the start of a filename expands to your
home directory. More generally, `<code>~</code><em>user</em><code>/</code>' allows you to refer to
the home directory of any other user. Furthermore, zsh lets you define
your own named directories which use this syntax. The basic idea is
simple, since any parameter can be a named directory:</p>
<pre><code> dir=/tmp/mydir
print ~dir
</code></pre>
<p>prints `/tmp/mydir'. So far, this isn't any different from using the
parameter as <code>$dir</code>. The difference comes if you use the `<code>%~</code>'
construct, described above, in your prompt. Then when you change into
that directory, instead of seeing the message `<code>/tmp/mydir</code>', you will
see the abbreviation `<code>~dir</code>'.</p>
<p>The shell will not register the name of the directory until you force it
to by using `<code>~dir</code>' yourself at least once. You can do the following
in your <code>.zshrc</code>:</p>
<pre><code> dir=/tmp/mydir
bin=~/myprogs/bin
: ~dir ~bin
</code></pre>
<p>where `<code>:</code>' is a command that does nothing --- but its arguments are
checked for parameters and so on in the usual way, so that the shell can
put <code>dir</code> and <code>bin</code> into its list of named directories. A more simple
way of doing this is to set the option <code>AUTO_NAME_DIRS</code>; then any
parameter created which refers to a directory will automatically be
turned into a name. The directory must have an absolute path, i.e. its
expanded value, after turning any `<code>~</code>'s at the start into full paths,
must begin with a `<code>/</code>'. The parameter <code>$PWD</code>, which shows the current
directory, is protected from being turned into <code>~PWD</code>, since that would
tell you nothing.</p>
<p><span id="l21"></span></p>
<h3 id="258-go-faster-options-for-power-users"><a class="header" href="#258-go-faster-options-for-power-users">2.5.8: `Go faster' options for power users</a></h3>
<p>Here are a few more random options you might want to set in your
<code>.zshrc</code>.</p>
<p><strong><code>NO_BEEP</code></strong></p>
<p>Normally zsh will beep if it doesn't like something. This can get
extremely annoying; `<code>setopt nobeep</code>' will turn it off. I refer to this
informally as the <code>OPEN_PLAN_OFFICE_NO_VIGILANTE_ATTACKS</code> option.</p>
<p><strong><code>AUTO_CD</code></strong></p>
<p>If this option is set, and you type something with no arguments which
isn't a command, zsh will check to see if it's actually a directory. If
it is, the shell will change to that directory. So `<code>./bin</code>' on its own
is equivalent to `<code>cd ./bin</code>', as long as the directory `<code>./bin</code>'
really exists. This is particularly useful in the form `<code>..</code>', which
changes to the parent directory.</p>
<p><strong><code>CD_ABLE_VARS</code></strong></p>
<p>This is another way of saving typing when changing directory, though
only one character. If a directory doesn't exist when you try to change
to it, zsh will try and find a parameter of that name and use that
instead. You can also have a `<code>/</code>' and other bits after the parameter.
So `cd <code>foo/dir</code>', if there is no directory `<code>foo</code>' but there is a
parameter <code>$foo</code>, becomes equivalent to `cd <code>$foo/dir</code>'.</p>
<p><strong><code>EXTENDED_GLOB</code></strong></p>
<p>Patterns, to match the name of files and other things, can be very
sophisticated in zsh, but to get the most out of them you need to use
this option, as otherwise certain features are not enabled, so that
people used to simpler patterns (maybe just `<code>*</code>', `<code>?</code>' and
`<code>[...]</code>') are not confused by strange happenings. I'll say much more
about zsh's pattern features, but this is to remind you that you need
this option if you're doing anything clever with `<code>~</code>', `<code>#</code>', `<code>^</code>'
or globbing flags --- and also to remind you that those characters can
have strange effects if you have the option set.</p>
<p><strong><code>MULTIOS</code></strong></p>
<p>I mentioned above that to get zsh to behave like ksh you needed to set
<code>NO_MULTIOS</code>, but I didn't say what the <code>MULTIOS</code> option did. It has two
different effects for output and input.</p>
<p>First, for output. Here it's an alternative to the <code>tee</code> programme. I've
mentioned once, but haven't described in detail, that you could use
<code>>filename</code> to tell the shell to send output into a file with a given
name instead of to the terminal. With <code>MULTIOS</code> set, you can have more
than one of those redirections on the command line:</p>
<pre><code> echo foo >file1 >file2
</code></pre>
<p>Here, `<code>foo</code>' will be written to <strong>both</strong> the named files; zsh copies
the output. The pipe mechanism, which I'll describe better in <a href="zshguide03.html#syntax">chapter
3</a>, is a sort of redirection into another
programme instead of into a file: <code>MULTIOS</code> affects this as well:</p>
<pre><code> echo foo >file1 | sed 's/foo/bar/'
</code></pre>
<p>Here, `<code>foo</code>' is again written to <code>file1</code>, but is also sent into the
pipe to the programme <code>sed</code> (`stream editor') which substitutes
`<code>foo</code>' into `<code>bar</code>' and (since there is no output redirection in this
part) prints it to the terminal.</p>
<p>Note that the second example above has several times been reported as a
bug, often in a form like:</p>
<pre><code> some_command 2>&1 >/dev/null | sed 's/foo/bar/'
</code></pre>
<p>The intention here is presumably to send standard error to standard
output (the `<code>2>&1</code>', a very commonly used shell hieroglyphic), and not
send standard output anywhere (the `<code>>/dev/null</code>'). (If you haven't met
the concept of `standard error', it's just another output channel which
goes to the same place as normal output unless you redirect it; it's
used, for example to send error messages to the terminal even if your
output is going somewhere else.) In this example, too, the <code>MULTIOS</code>
feature forces the original standard output to go to the pipe. You can
see this happening if we put in a version of `<code>some_command</code>':</p>
<pre><code> { echo foo error >&2; echo foo not error; } 2>&1 >/dev/null |
sed 's/foo/bar/'
</code></pre>
<p>where you can consider the stuff inside the `<code>{...}</code>' as a black box
that sends the message `foo error' to standard error, and `foo not
error' to standard output. With <code>MULTIOS</code>, however, the result is</p>
<pre><code> error bar
not error bar
</code></pre>
<p>because both have been sent into the pipe. Without <code>MULTIOS</code> you get the
expected result,</p>
<pre><code> error bar
</code></pre>
<p>as any other Bourne-style shell would produce. There</p>
<p>On input, <code>MULTIOS</code> arranges for a series of files to be read in order.
This time it's a bit like using the programme <code>cat</code>, which combines all
the files listed after it. In other words,</p>
<pre><code> cat file1 file2 | myprog
</code></pre>
<p>(where <code>myprog</code> is some programme that reads all the files sent to it as
input) can be replaced by</p>
<pre><code> myprog <file1 <file2
</code></pre>
<p>which does the same thing. Once again, a pipe counts as a redirection,
and the pipe is read from first, before any files listed after a `<code><</code>':</p>
<pre><code> echo then this >testfile
echo this first | cat <testfile
</code></pre>
<p><strong><code>CORRECT</code>, <code>CORRECT_ALL</code></strong></p>
<p>If you have <code>CORRECT</code> set, the shell will check all the commands you
type and if they don't exist, but there is one with a similar name, it
will ask you if you meant that one instead. You can type `<code>n</code>' for no,
don't correct, just go ahead; `<code>y</code>' for yes, correct it then go ahead;
`<code>a</code>' for abort, don't do anything; `<code>e</code>' for edit, return to the
editor to edit the same line again. Users of the new completion system
should note this is not the same correction you get there: it's just
simple correction of commands.</p>
<p><code>CORRECT_ALL</code> applies to all the words on the line. It's a little less
useful, because currently the shell has to assume that they are supposed
to be filenames, and will try to correct them if they don't exist as
such, but of course many of the arguments to a command are not
filenames. If particular commands generate too many attempts to correct
their arguments, you can turn this off by putting `<code>nocorrect</code>' in
front of the command name. An alias is a very good way of doing this, as
described next.</p>
<p><span id="l22"></span></p>
<h3 id="259-aliases"><a class="header" href="#259-aliases">2.5.9: aliases</a></h3>
<p>An alias is used like a command, but it expands into some other text
which is itself used as a command. For example,</p>
<pre><code> alias foo='print I said foo'
foo
</code></pre>
<p>prints (guess what) `<code>I said foo</code>'. Note the syntax for definition ---
you need the `<code>=</code>', and you need to make sure the whole alias is
treated by the shell as one word; you can give a whole list of aliases
to the same `<code>alias</code>' command. You may be able to think of some aliases
you want to define in your startup files; <code>.zshrc</code> is probably the right
place. If you have <code>CORRECT_ALL</code> set, the way to avoid the `<code>mkdir</code>'
command spell-checking its arguments --- which is useless, because they
<em>have</em> to be non-existent for the command to work --- is to define:</p>
<pre><code> alias mkdir='nocorrect mkdir'
</code></pre>
<p>This shows one useful feature about aliases: the alias can contain
something of the same name as itself. When it is encountered in the
expansion text (the right hand side), the shell knows it is not to
expand the alias again, but this time to treat it as a real command.
Note that functions do <em>not</em> have this property: functions are more
powerful than aliases and in some cases it is useful for them to call
themselves, It's a common mistake to have functions call themselves over
and over again until the shell complains. I'll describe ways round this
in <a href="zshguide03.html#syntax">chapter 3</a>.</p>
<p>One other way functions are more powerful than aliases is that functions
can take arguments while aliases can't --- in other words, there is no
way of referring inside the alias to what follows it on the command
line, unlike a function, and also unlike aliases in csh (because that
has no functions, that's why). It is just blindly expanded, and the
remainder of the command line stuck on the end. Hence aliases in zsh are
usually kept for quite simple things, and functions are written for
anything more complicated. You couldn't do that trick with
`<code>nocorrect</code>' using a function, though, since the function is called
too late: aliases are expanded straight away, so the <code>nocorrect</code> is
found in time to be useful. You can almost think of them as just plain
typing abbreviations.</p>
<p>Normal aliases only work when in command position, i.e. at the start of
the command line (more strictly, when zsh is expecting a command). There
are other things called `global aliases', which you define by the
`<code>-g</code>' option to <code>alias</code>, which will be expanded at any position on the
command line. You should think seriously before defining these, as they
can have a drastic effect. Note, however, that quoting a word, or even a
single character, will stop an alias being expanded for it.</p>
<p>I only tend to use aliases in interactive shells, so I define them from
<code>.zshrc</code>, but you may want to use <code>.zshenv</code> if you use aliases more
widely. In fact, to keep my <code>.zshrc</code> neat I save all the aliases in a
separate file called <code>.aliasrc</code> and in <code>.zshrc</code> I have:</p>
<pre><code> if [[ -r ~/.aliasrc ]]; then
. ~/.aliasrc
fi
</code></pre>
<p>which checks if there is a readable file <code>~/.aliasrc</code>, and if there is,
it runs it in exactly the same way the normal startup files are run. You
can use `<code>source</code>' instead of `<code>.</code>' if it means more to you; `<code>.</code>' is
the traditional Bourne and Korn shell name, however.</p>
<p><span id="l23"></span></p>
<h3 id="2510-environment-variables"><a class="header" href="#2510-environment-variables">2.5.10: Environment variables</a></h3>
<p>Often, the manual for a programme will tell you to define certain
environment variables, usually a collection of uppercase letters with
maybe numbers and the odd underscore. These can pass information to the
programme without you needing to use extra arguments. In zsh,
environment variables appear as ordinary shell parameters, although they
have to be defined slightly differently: strictly, the environment is a
special region outside the shell, and zsh has to be told to put a copy
there as well as keeping one of its own. The usual syntax is</p>
<pre><code> export VARNAME='value'
</code></pre>
<p>in other words, like an ordinary assignment, but with `<code>export</code>' in
front. Note there is no `<code>$</code>' before the name of the environment
variable; all `<code>export</code>' and similar statements work the same way. The
easiest place to put these is in <code>.zshenv</code> --- hence it's name.
Environment variables will be passed to any programmes run from a shell,
so it may be enough to define them in <code>.zlogin</code> or <code>.zprofile</code>: however,
any shell started for you non-interactively won't run those, and there
are other possible problems if you use a windowing system which is
started by a shell other than zsh or which doesn't run a shell start-up
file at all --- I had to tweak mine to make it do so. So <code>.zshenv</code> is
the safest place; it doesn't take long to define environment variables.
Other people will no doubt give you completely contradictory views, but
that's people for you.</p>
<p>Note that you can't export arrays. If you export a parameter, then
assign an array to it, nothing will appear in the environment; you can
use the external command `<code>printenv VARNAME</code>' (again no `<code>$</code>' because
the command needs to know the name, not the value) to check. There's a
more subtle problem with arrays, too. The <code>export</code> builtin is just a
special case of the builtin <strong>typeset</strong>, which defines a variable
without marking it for export to the environment. You might think you
could do</p>
<pre><code> typeset array=(this doesn\'t work)
</code></pre>
<p>but you can't --- the special array syntax is only understood when the
assignment does not follow a command, not in normal arguments like the
case here, so you have to put the array assignment on the next line.
This is a very easy mistake to make. More uses of <code>typeset</code> will be
described in <a href="zshguide03.html#syntax">chapter 3</a>; they include creating
local parameters in functions, and defining special attributes (of which
the `export' attribute is just one) for parameters.</p>
<p><span id="l24"></span></p>
<h3 id="2511-path"><a class="header" href="#2511-path">2.5.11: Path</a></h3>
<p>It helps to be able to find external programmes, i.e. anything not part
of the shell, any command other than a builtin, function or alias. The
<code>$path</code> array is used for this. Actually, what the system needs is the
environment variable <code>$PATH</code>, which contains a list of directories in
which to search for programmes, separated from each other by a colon.
These directories are the individual components of the array <code>$path</code>. So
if <code>$path</code> contains</p>
<pre><code> path=(/bin /usr/bin /usr/local/bin .)
</code></pre>
<p>then <code>$PATH</code> will automatically contain the effect of</p>
<pre><code> PATH=/bin:/usr/bin:/usr/local/bin:.
</code></pre>
<p>without you having to set that. The idea is simply that, while the
system needs <code>$PATH</code> because it doesn't understand arrays, it's much
more flexible to be able to use arrays within the shell and hence pretty
much forget about the <code>$PATH</code> form.</p>
<p>Changes to the path are similar to changes to environment variables
described above, so all that applies. There's a slight difficulty in
setting <code>$path</code> in <code>.zshenv</code> however, even though the reasons given
above for doing so still apply. Usually, the path will be set for you,
either by the system, or by the system administrator in one of the
global start up files, and if you change path you will simply want to
add to it. But if your <code>.zshenv</code> contains</p>
<pre><code> path=(~/bin ~/progs/bin $path)
</code></pre>
<p>--- which is the right way of adding something to the front of <code>$path</code>
--- then every time <code>.zshenv</code> is called, <code>~/bin</code> and <code>~/progs/bin</code> are
stuck in front, so if you start another zsh you will have two sets
there.</p>
<p>You can add tests to see if something's already there, of course. Zsh
conveniently allows you to test for the existence of elements in an
array. By preceding an array index by <code>(r)</code> (for reverse), it will try
to find a matching element and return that, else an empty string. Here's
a way of doing that (but don't add this yet, see the next paragraph):</p>
<pre><code> for dir in ~/bin ~/progs/bin; do
if [[ -z ${path[(r)$dir]} ]]; then
path=($dir $path)
fi
done
</code></pre>
<p>That <code>for</code>... <code>do</code> ... <code>done</code> is another special shell construct. It
takes each thing after `<code>in</code>' and assigns it in turn to the parameter
named before the `<code>in</code>' --- <code>$dir</code>, but because this is a form of
assignment, the `<code>$</code>' is left off --- so the first time round it has
the effect of <code>dir=~/bin</code>, and the next time <code>dir=~/progs/bin</code>. Then it
executes what's in the loop. The test <code>-z</code> checks that what follows is
empty: in this case it will be if the directory <code>$dir</code> is not yet in
<code>$path</code>, so it goes ahead and adds it in front. Note that the
directories get added in the reverse of the order they appear.</p>
<p>Actually, however, zsh takes all that trouble away from you. The
incantation `<code>typeset -U path</code>', where the <code>-U</code> stands for unique,
tells the shell that it should not add anything to <code>$path</code> if it's there
already. To be precise, it keeps only the left-most occurrence, so if
you added something at the end it will disappear and if you added
something at the beginning, the old one will disappear. Thus the
following works nicely in <code>.zshenv</code>:</p>
<pre><code> typeset -U path
path=(~/bin ~/progs/bin $path)
</code></pre>
<p>and you can put down that `<code>for</code>' stuff as a lesson in shell
programming. You can list all the variables which have uniqueness turned
on by typing `<code>typeset +U</code>', with `<code>+</code>' instead of `<code>-</code>', because in
the latter case the shell would show the values of the parameters as
well, which isn't what you need here. The <code>-U</code> flag will also work with
colon-separated arrays, like <code>$PATH</code>.</p>
<p><span id="l25"></span></p>
<h3 id="2512-mail"><a class="header" href="#2512-mail">2.5.12: Mail</a></h3>
<p>Zsh will check for new mail for you. If all you need is to be reminded
of something arriving in your normal folder every now and then, you just
need to set the parameter <code>$MAIL</code> to wherever that is: it's typically
one of <code>/usr/spool/mail</code>, <code>/var/spool/mail</code>, or <code>/var/mail</code>.</p>
<p>The array <code>$mailpath</code> allows more possibilities. Like <code>$path</code>, it has a
colleague in uppercase, <code>$MAILPATH</code>, which is a colon-separated array.
The system doesn't need that, this time, so it's mainly there so that
you can export it to another version of zsh; exporting arrays won't
work. As may by now be painfully clear, if you set in <code>.zshenv</code> or
<code>.zshrc</code>, you don't need to export it, because it's set in each instance
of the shell. The elements of <code>$mailpath</code> work like <code>$MAIL</code>, so you can
specify different places where mail arrives. That's most useful if you
have a programme like <code>filter</code> or <code>procmail</code> running to redistribute
arriving mail to different folders. You can specify a different message
for each folder by putting `<code>?</code><em>message</em>' at the end. For example, mine
looks like this.</p>
<pre><code> mailpref=/temp/pws/Mail
mailpath=($mailpref/newmail
$mailpref/zsh-new'?New zsh mail'
$mailpref/list-new'?New list mail'
$mailpref/urth-new'?New Urth mail')
</code></pre>
<p>Note that zsh knows the array isn't finished until the `)', even though
the elements are on different lines; this is one very good reason for
setting <code>$mailpath</code> rather than <code>$MAILPATH</code>, which needs one long chunk.</p>
<p>The other parameter of interest is <code>$MAILCHECK</code>, which gives the
frequency in seconds when zsh should check for new mail. The default is</p>
<ol start="60">
<li>Actually, zsh only checks just after a command has finished running
and it is about to print a prompt. Since checking files doesn't take
long, you can usually set this to its minimum value, which is
<code>MAILCHECK=1</code>; zero doesn't work because it switches off checking. One
reason why you wouldn't want to do that might be because <code>$MAIL</code> and
<code>$mailpath</code> can contain directories instead of ordinary files; these
will be checked recursively for any files with something new in them, so
this can be slow.</li>
</ol>
<p>Finally, there is one associated option, <code>MAIL_WARNING</code> (though
<code>MAIL_WARN</code> is also accepted for the same thing for reasons of
compatibility with less grammatical shells). The shell remembers when it
found the mail file was checked; next time it checks, it compares the
date. If there is no new mail, but the date of the file changed anyway,
it will print a warning message. This will happen if you read the mail
with your mail reader and put the messages somewhere else. Presumably
you <em>know</em> you did that, so the warning may not be all that useful.</p>
<p><span id="l26"></span></p>
<h3 id="2513-other-path-like-things"><a class="header" href="#2513-other-path-like-things">2.5.13: Other path-like things</a></h3>
<p>There are other pairs like <code>$path</code> and <code>$PATH</code>. I will keep back talk of
<code>$cdpath</code> until I say more about the way zsh handles directories. When I
mentioned <code>$fpath</code>, I didn't say there was also <code>$FPATH</code>, but there is.
Then there is <code>$manpath</code> and <code>$MANPATH</code>; these aren't used by the shell
at all, but <code>$MANPATH</code>, if exported, is used by the <strong>man</strong> external
command, and <code>$manpath</code> gives an easier way to set it.</p>
<p>From 3.1.6 there is a mechanism to define your own such combinations; if
this had been available before, there would have been no need to build
in <code>$manpath</code> and <code>$MANPATH</code>. In <code>.zshenv</code> you would put,</p>
<pre><code> export -TU TEXINPUTS texinputs
</code></pre>
<p>to define such a pair. The <code>-T</code> (for tie) is the key to that; I've used
`<code>export</code>' even though the basic variable declaration command is
`<code>typeset</code>' because you nearly always want to get the colon-separated
version (<code>$TEXINPUTS</code> here) visible to the environment, and I've set
<code>-U</code> as described above for <code>$path</code> because it's a neat feature anyway.
Now you can assign to the array <code>$texinputs</code> and let the programme (TeX
or its derivatives) see <code>$TEXINPUTS</code>. Another useful variable to do this
with is <code>$LD_LIBRARY_PATH</code>, which on most modern versions of UNIX (and
Linux) tells the system where to find the libraries which provide extra
functions when it runs a programme.</p>
<p><span id="l27"></span></p>
<h3 id="2514-version-specific-things"><a class="header" href="#2514-version-specific-things">2.5.14: Version-specific things</a></h3>
<p>Since zsh changes faster than almost any other command interpreter known
to humankind, you will often find you need to find out what version you
are using. This can get a bit verbose; indeed, the parameter you need to
check, which is now <code>$ZSH_VERSION</code>, used simply to be called <code>$VERSION</code>
before version <code>3.0</code>. If you are not using legacy software of that kind,
you can probably get away with tests like this:</p>
<pre><code> if [[ $ZSH_VERSION == 3.1.<5->* ||
$ZSH_VERSION == 3.<2->* ||
$ZSH_VERSION == <4->* ]]; then
# set feature which appeared first in 3.1.5
fi
</code></pre>
<p>It's like that to be futureproof: it says that if this is a 3.1 release,
it has to be at least 3.1.5, but any 3.2 release (there weren't any), or
any release 4 or later, will also be OK. The `<code><5-></code>' etc. are advanced
pattern matching tests: pattern matching uses the same symbols as
globbing, but to test other things, here what's on the left of the
`<code>==</code>'. This one matches any number which is at least 5, for example 6
or 10 or 252, but not 1 or 4. There are also development releases;
nowadays the version numbers look like <em>X</em><code>.</code><em>Y</em><code>.</code><em>Z</em>-<em>tag</em>-<em>N</em> (<em>tag</em>
is some short word, the others are numbers) but unless you're keeping up
with development you won't need to look for those, since they aren't
released officially. That `<code>==</code>' in the test could also be just `<code>=</code>',
but the manual says the former is preferred, so I've used them here,
even though people usually don't bother.</p>
<p>Version 4 of zsh provides a function <code>is-at-least</code> to do this for you:
it looks only at the numbers <em>X</em>, <em>Y</em> and <em>Z</em> (and <em>N</em> if it exists),
ignoring all letters and punctuation. You give it the minimum version of
the shell you need and it returns true if the current shell is recent
enough. For example, `<code>is-at-least 3.1.6-pws-9</code>' will return true if
the current version of zsh is 3.1.6-dev-20 (or 3.1.9, or 4.0.1, and so
on), which is the correct behaviour. As with any other shell function,
you have to arrange for <code>is-at-least</code> to be <code>autoload</code>ed if you want to
use it.</p>
<p><span id="l28"></span></p>
<h3 id="2515-everything-else"><a class="header" href="#2515-everything-else">2.5.15: Everything else</a></h3>
<p>There are many other possibilities for things to go in startup files; in
particular, I haven't touched on defining things for the line editor and
setting up completion. There's quite a lot to explain for those, so I'll
come back to those in the appropriate chapters. You just need to
remember that all that stuff should go in <code>.zshrc</code>, since you need it
for all interactive shells, and for no others.</p>
<div id="chapter_begin" style="break-before: page; page-break-before: always;"></div><!-- START doctoc generated TOC please keep comment here to allow auto update -->
<!-- DON'T EDIT THIS SECTION, INSTEAD RE-RUN doctoc TO UPDATE -->
<p><strong>Table of Contents</strong> <em>generated with <a href="https://github.com/thlorenz/doctoc">DocToc</a></em></p>
<ul>
<li><a href="zshguide03.html#chapter-3-dealing-with-basic-shell-syntax">Chapter 3: Dealing with basic shell syntax</a>
<ul>
<li><a href="zshguide03.html#31-external-commands">3.1: External commands</a></li>
<li><a href="zshguide03.html#32-builtin-commands">3.2: Builtin commands</a>
<ul>
<li><a href="zshguide03.html#321-builtins-for-printing">3.2.1: Builtins for printing</a></li>
<li><a href="zshguide03.html#322-other-builtins-just-for-speed">3.2.2: Other builtins just for speed</a></li>
<li><a href="zshguide03.html#323-builtins-which-change-the-shells-state">3.2.3: Builtins which change the shell's state</a></li>
<li><a href="zshguide03.html#324-cd-and-friends">3.2.4: cd and friends</a></li>
<li><a href="zshguide03.html#325-command-control-and-information-commands">3.2.5: Command control and information commands</a></li>
<li><a href="zshguide03.html#326-parameter-control">3.2.6: Parameter control</a></li>
<li><a href="zshguide03.html#327-history-control-commands">3.2.7: History control commands</a></li>
<li><a href="zshguide03.html#328-job-control-and-process-control">3.2.8: Job control and process control</a></li>
<li><a href="zshguide03.html#329-terminals-users-etc">3.2.9: Terminals, users, etc.</a></li>
<li><a href="zshguide03.html#3210-syntactic-oddments">3.2.10: Syntactic oddments</a></li>
<li><a href="zshguide03.html#3211-more-precommand-modifiers-exec-noglob">3.2.11: More precommand modifiers: <code>exec</code>, <code>noglob</code></a></li>
<li><a href="zshguide03.html#3212-testing-things">3.2.12: Testing things</a></li>
<li><a href="zshguide03.html#3213-handling-options-to-functions-and-scripts">3.2.13: Handling options to functions and scripts</a></li>
<li><a href="zshguide03.html#3214-random-file-control-things">3.2.14: Random file control things</a></li>
<li><a href="zshguide03.html#3215-dont-watch-this-space-watch-some-other">3.2.15: Don't watch this space, watch some other</a></li>
<li><a href="zshguide03.html#3216-and-also">3.2.16: And also</a></li>
</ul>
</li>
<li><a href="zshguide03.html#33-functions">3.3: Functions</a>
<ul>
<li><a href="zshguide03.html#331-loading-functions">3.3.1: Loading functions</a></li>
<li><a href="zshguide03.html#332-function-parameters">3.3.2: Function parameters</a></li>
<li><a href="zshguide03.html#333-compiling-functions">3.3.3: Compiling functions</a></li>
</ul>
</li>
<li><a href="zshguide03.html#34-aliases">3.4: Aliases</a></li>
<li><a href="zshguide03.html#35-command-summary">3.5: Command summary</a></li>
<li><a href="zshguide03.html#36-expansions-and-quotes">3.6: Expansions and quotes</a>
<ul>
<li><a href="zshguide03.html#361-history-expansion">3.6.1: History expansion</a></li>
<li><a href="zshguide03.html#362-alias-expansion">3.6.2: Alias expansion</a></li>
<li><a href="zshguide03.html#363-process-parameter-command-arithmetic-and-brace-expansion">3.6.3: Process, parameter, command, arithmetic and brace expansion</a></li>
<li><a href="zshguide03.html#364-filename-expansion">3.6.4: Filename Expansion</a></li>
<li><a href="zshguide03.html#365-filename-generation">3.6.5: Filename Generation</a></li>
</ul>
</li>
<li><a href="zshguide03.html#37-redirection-greater-thans-and-less-thans">3.7: Redirection: greater-thans and less-thans</a>
<ul>
<li><a href="zshguide03.html#371-clobber">3.7.1: Clobber</a></li>
<li><a href="zshguide03.html#372-file-descriptors">3.7.2: File descriptors</a></li>
<li><a href="zshguide03.html#373-appending-here-documents-here-strings-read-write">3.7.3: Appending, here documents, here strings, read write</a></li>
<li><a href="zshguide03.html#374-clever-tricks-exec-and-other-file-descriptors">3.7.4: Clever tricks: exec and other file descriptors</a></li>
<li><a href="zshguide03.html#375-multios">3.7.5: Multios</a></li>
</ul>
</li>
<li><a href="zshguide03.html#38-shell-syntax-loops-subshells-and-so-on">3.8: Shell syntax: loops, (sub)shells and so on</a>
<ul>
<li><a href="zshguide03.html#381-logical-command-connectors">3.8.1: Logical command connectors</a></li>
<li><a href="zshguide03.html#382-structures">3.8.2: Structures</a></li>
<li><a href="zshguide03.html#383-subshells-and-current-shell-constructs">3.8.3: Subshells and current shell constructs</a></li>
<li><a href="zshguide03.html#384-subshells-and-current-shells">3.8.4: Subshells and current shells</a></li>
</ul>
</li>
<li><a href="zshguide03.html#39-emulation-and-portability">3.9: Emulation and portability</a>
<ul>
<li><a href="zshguide03.html#391-differences-in-detail">3.9.1: Differences in detail</a></li>
<li><a href="zshguide03.html#392-making-your-own-scripts-and-functions-portable">3.9.2: Making your own scripts and functions portable</a></li>
</ul>
</li>
<li><a href="zshguide03.html#310-running-scripts">3.10: Running scripts</a></li>
</ul>
</li>
</ul>
<!-- END doctoc generated TOC please keep comment here to allow auto update -->
<p><span id="syntax"></span><span id="l29"></span></p>
<h1 id="chapter-3-dealing-with-basic-shell-syntax"><a class="header" href="#chapter-3-dealing-with-basic-shell-syntax">Chapter 3: Dealing with basic shell syntax</a></h1>
<p>This chapter is a more thorough examination of much of what appeared in
the <a href="zshguide02.html#init">chapter 2</a>; to be more specific, I assume
you're sitting in front of your terminal about to use the features you
just set up in your initialisation files and want to know enough to get
them going. Actually, you will probably spend most of the time editing
command lines and in particular completing commands --- both of these
activities are covered in later chapters. For now I'm going to talk
about commands and the syntax that goes along with using them. This will
let you write shell functions and scripts to do more of your work for
you.</p>
<p>In the following there are often several consecutive paragraphs about
quite minor features. If you find you read this all through the first
time, maybe you need to get out more. Most people will probably find it
better to skim through to find what the subject matter is, then come
back if they later find they want to know more about a particular aspect
of the shell's commands and syntax.</p>
<p>One aspect of the syntax is left to <a href="zshguide05.html#subst">chapter 5</a>:
there's just so much to it, and it can be so useful if you know enough
to get it right, that it can't all be squashed in here. The subject is
expansion, covering a multitude of things such as parameter expansion,
globbing and history expansions. You've already met the basics of these
in <a href="zshguide02.html#init">chapter 2</a>; but if you want to know how to
pick a particular file with a globbing expression with pinpoint
accuracy, or how to make a single parameter expansion reduce a long
expression to the words you need, you should read that chapter; it's
more or less self-contained, so you don't necessarily need to know
everything in this one.</p>
<p>We start with the most basic issue in any command line interpreter,
running commands. As you know, you just type words separated by spaces,
where the first word is a command and the remainder are arguments to it.
It's important to distinguish between the types of command.</p>
<p><span id="l30"></span></p>
<h2 id="31-external-commands"><a class="header" href="#31-external-commands">3.1: External commands</a></h2>
<p>External commands are the easiest, because they have the least
interaction with the shell --- many of the commands provided by the
shell itself, which are described in the next section, are built into
the shell especially to avoid this difficulty.</p>
<p>The only major issue is therefore how to find them. This is done through
the parameters <code>$path</code> and <code>$PATH</code>, which, as I described in <a href="zshguide02.html#init">chapter
2</a>, are tied together because although the first
one is more useful inside the shell --- being an array, its various
parts can be manipulated separately --- the second is the one that is
used by other commands called by the shell; in the jargon, <code>$PATH</code> is
`exported to the environment', which means exactly that other commands
called by the shell can see its value.</p>
<p>So suppose your <code>$path</code> contains</p>
<pre><code> /home/pws/bin /usr/local/bin /bin /usr/bin
</code></pre>
<p>and you try to run `<code>ls</code>'. The shell first looks in <code>/home/pws/bin</code> for
a command called <code>ls</code>, then in <code>/usr/local/bin</code>, then in <code>/bin</code>, where
it finds it, so it executes <code>/bin/ls</code>. Actually, the operating system
itself knows about paths if you execute a command the right way, so the
shell doesn't strictly need to.</p>
<p>There is a subtlety here. The shell tries to remember where the commands
are, so it can find them again the next time. It keeps them in a
so-called `hash table', and you find the word `hash' all over the
place in the documentation: all it means is a fast way of finding some
value, given a particular key. In this case, given the name of a
command, the shell can find the path to it quickly. You can see this
table, in the form `<em>key</em><code>=</code><em>value</em>', by typing `<code>hash</code>'.</p>
<p>In fact the shell only does this when the option <code>HASH_CMDS</code> is set, as
it is by default. As you might expect, it stops searching when it finds
the directory with the command it's looking for. There is an extra
optimisation in the option <code>HASH_ALL</code>, also set by default: when the
shell scans a directory to find a command, it will add all the other
commands in that directory to the hash table. This is sensible because
on most UNIX-like operating systems reading a whole lot of files in the
same directory is quite fast.</p>
<p>The way commands are stored has other consequences. In particular, zsh
won't look for a new command if it already knows where to find one. If I
put a new <code>ls</code> command in <code>/usr/local/bin</code> in the above example, zsh
would continue to use <code>/bin/ls</code> (assuming it had already been found). To
fix this, there is the command <code>rehash</code>, which actually empties the
command hash table, so that finding commands starts again from scratch.
Users of csh may remember having to type <code>rehash</code> quite a lot with new
commands: it's not so bad in zsh, because if no command was already
hashed, or the existing one disappeared, zsh will automatically scan the
path again; furthermore, zsh performs a <code>rehash</code> of its own accord if
<code>$path</code> is altered. So adding a new duplicate command somewhere towards
the head of <code>$path</code> is the main reason for needing <code>rehash</code>.</p>
<p>One thing that can happen if zsh hasn't filled its command hash table
and so doesn't know about all external commands is that the <code>AUTO_CD</code>
option, mentioned in the previous chapter and again below, can think you
are trying to change to a particular directory with the same name as the
command. This is one of the drawbacks of <code>AUTO_CD</code>.</p>
<p>To be a little bit more technical, it's actually not so obvious that
command hashing is needed at all; many modern operating systems can find
commands quickly without it. The clincher in the case of zsh is that the
same hash table is necessary for command completion, a very commonly
used feature. If you type `<code>compr<TAB></code>', the shell completes this to
`<code>compress</code>'. It can only do this if it has a list of commands to
complete, and this is the hash table. (In this case it didn't need to
know where to find the command, just its name, but it's only a little
extra work to store that too.) If you were following the previous
paragraphs, you'll realise zsh doesn't necessarily know <em>all</em> the
possible commands at the time you hit <code>TAB</code>, because it only looks when
it needs to. For this purpose, there is another option, <code>HASH_LIST_ALL</code>,
again set by default, which will make sure the command hash table is
full when you try to complete a command. It only needs to do this once
(unless you alter <code>$path</code>), but it does mean the first command
completion is slow. If <code>HASH_LIST_ALL</code> is not set, command completion is
not available: the shell could be rewritten to search the path
laboriously every single time you try to complete a command name, but it
just doesn't seem worth it.</p>
<p>The fact that <code>$PATH</code> is passed on from the shell to commands called
from it (strictly only if the variable is marked for export, as it
usually is --- this is described in more detail with the <code>typeset</code>
family of builtin commands below) also has consequences. Some commands
call subcommands of their own using <code>$PATH</code>. If you have that set to
something unusual, so that some of the standard commands can't be found,
it could happen that a command which <em>is</em> found nonetheless doesn't run
properly because it's searching for something it can't find in the path
passed down to it. That can lead to some strange and confusing error
messages.</p>
<p>One important thing to remember about external commands is that the
shell continues to exist while they are running; it just hangs around
doing nothing, waiting for the job to finish (though you can tell it not
to, as we'll see). The command is given a completely new environment in
which to run; changes in that don't affect the shell, which simply
starts up where it left off after the command has run. So if you need to
do something which changes the state of the shell, an external command
isn't good enough. This brings us to builtin commands.</p>
<p><span id="l31"></span></p>
<h2 id="32-builtin-commands"><a class="header" href="#32-builtin-commands">3.2: Builtin commands</a></h2>
<p>Builtin commands, or builtins for short, are commands which are part of
the shell itself. Since builtins are necessary for controlling the
shell's own behaviour, introducing them actually serves as an
introduction to quite a lot of what is going on in the shell. So a fair
fraction of what would otherwise appear later in the chapter has
accumulated here, one way or another. This does make things a little
tricksy in places; count how many times I use the word `<code>subtle</code>' and
keep it for your grandchildren to see.</p>
<p>I just described one reason for builtins, but there's a simpler one:
speed. Going through the process of setting up an entirely new
environment for the command at the beginning, swapping between this
command and anything else which is being run on the computer, then
destroying it again at the end is considerable overkill if all you want
to do is, say, print out a message on the screen. So there are builtins
for this sort of thing.</p>
<p><span id="l32"></span></p>
<h3 id="321-builtins-for-printing"><a class="header" href="#321-builtins-for-printing">3.2.1: Builtins for printing</a></h3>
<p>The commands `<code>echo</code>' and `<code>print</code>' are shell builtins; they just show
what you typed, after the shell has removed all the quoting. The
difference between the two is really historical: `<code>echo</code>' came first,
and only handled a few simple options; ksh provided `<code>print</code>', which
had more complex options and so became a different command. The
difference remains between the two commands in zsh; if you want wacky
effects, you should look to <code>print</code>. Note that there is usually also an
external command called <code>echo</code>, which may not be identical to zsh's;
there is no standard external command called <code>print</code>, but if someone has
installed one on your system, the chances are it sends something to the
printer, not the screen.</p>
<p>One special effect is `<code>print -z</code>' puts the arguments onto the editing
buffer stack, a list maintained by the shell of things you are about to
edit. Try:</p>
<pre><code> print -z print -z print This is a line
</code></pre>
<p>(it may look as if something needs quoting, but it doesn't) and hit
return three times. The first time caused everything after the first
`<code>print -z</code>' to appear for you to edit, and so on.</p>
<p>For something more useful, you can write functions that give you a line
to edit:</p>
<pre><code> fn() { print -z print The time now is $(date); }
</code></pre>
<p>Now when you type `<code>fn</code>', the line with the date appears on the command
line for you to edit. The option `<code>-s</code>' is a bit similar; the line
appears in the history list, so you will see it if you use up-arrow, but
it doesn't reappear automatically.</p>
<p>A few other useful options, some of which you've already seen, are</p>
<ul>
<li><strong><code>-r</code></strong><br />
don't interpret special character sequences like `<code>\n</code>'</li>
<li><strong><code>-P</code></strong><br />
use `<code>%</code>' as in prompts</li>
<li><strong><code>-n</code></strong><br />
don't put a newline at the end in case there's more output to follow</li>
<li><strong><code>-c</code></strong><br />
print the output in columns --- this means that `<code>print -c *</code>' has
the effect of a sort of poor person's `<code>ls</code>', only faster</li>
<li><strong><code>-l</code></strong><br />
use one line per argument instead of one column, which is sometimes
useful for sticking lists into files, and for working out what part
of an array parameter is in each element.</li>
</ul>
<p>If you don't use the <code>-r</code> option, there are a whole lot of special
character sequences. Many of these may be familiar to you from C.</p>
<ul>
<li><strong><code>\n</code></strong><br />
newline</li>
<li><strong><code>\t</code></strong><br />
tab</li>
<li><strong><code>\e</code> or <code>\E</code></strong><br />
escape character</li>
<li><strong><code>\a</code></strong><br />
ring the bell (alarm), usually a euphemism for a hideous beep</li>
<li><strong><code>\b</code></strong><br />
move back one character.</li>
<li><strong><code>\c</code></strong><br />
don't print a newline --- like the <code>-n</code> option, but embedded in the
string. This alternative comes from Berkeley UNIX.</li>
<li><strong><code>\f</code></strong><br />
form feed, the phrase for `advance to next page' from the days when
terminals were called teletypes, maybe more familiar to you as <code>^L</code></li>
<li><strong><code>\r</code></strong><br />
carriage return --- when printed, the annoying <code>^M</code>'s you get in DOS
files, but actually rather useful with `<code>print</code>', since it will
erase everything to the start of the line. The combination of the
<code>-n</code> option and a <code>\r</code> at the start of the print string can give the
illusion of a continuously changing status line.</li>
<li><strong><code>\v</code></strong><br />
vertical tab, which I for one have never used (I just tried it now
and it behaved like a newline, only without assuming a carriage
return, but that's up to your terminal).</li>
</ul>
<p>In fact, you can get any of the 255 characters possible, although your
terminal may not like some or all of the ones above 127, by specifying a
number after the backslash. Normally this consists of three octal
characters, but you can use two hexadecimal characters after <code>\x</code>
instead --- so `<code>\n</code>', `<code>\012</code>' and `<code>\x0a</code>' are all newlines. `<code>\</code>'
itself escapes any other character, i.e. they appear as themselves even
if they normally wouldn't.</p>
<p>Two notes: first, don't get confused because `<code>n</code>' is the fourteenth
letter of the alphabet; printing `<code>\016</code>' (fourteen in octal) won't do
you any good. The remedy, after you discover your text is unreadable
(for VT100-like terminals including xterm), is to print `<code>\017</code>'.</p>
<p>Secondly, those backslashes can land you in real quoting difficulties.
Normally a backslash on the command line escapes the next character ---
this is a <em>different</em> form of escaping to <code>print</code>'s --- so</p>
<pre><code> print \n
</code></pre>
<p>doesn't produce a newline, it just prints out an `<code>n</code>'. So you need to
quote that. This means</p>
<pre><code> print \\
</code></pre>
<p>passes a single backslash to quote, and</p>
<pre><code> print \\n
</code></pre>
<p>or</p>
<pre><code> print '\n'
</code></pre>
<p>prints a newline (followed by the extra one that's usually there). To
print a real backslash, you would thus need</p>
<pre><code> print \\\\
</code></pre>
<p>Actually, you can get away with the two if there's nothing else after
--- <code>print</code> just shrugs its shoulders and outputs what it's been given
--- but that's not a good habit to get into. There are other ways of
doing this: since single quotes quote anything, including backslashes
(they are the only way of making backslashes behave like normal
characters), and since the `<code>-r</code>' option makes print treat characters
normally,</p>
<pre><code> print -r '\'
</code></pre>
<p>has the same effect. But you need to remember the two levels of quoting
for backslashes. Quotes aren't special to <code>print</code>, so</p>
<pre><code> print \'
</code></pre>
<p>is good enough for printing a quote.</p>
<p><strong><code>echotc</code></strong></p>
<p>There's an oddity called `<code>echotc</code>', which takes as its argument
`termcap' capabilities. This now lives in its own module,
<code>zsh/termcap</code>.</p>
<p>Termcap is a now rather old-fashioned way of giving the commands
necessary for performing various standard operations on terminals:
moving the cursor, clearing to the end of the line, turning on standout
mode, and so on. It has now been replaced almost everywhere by
`terminfo', a completely different way of specifying capabilities, and
by `curses', a more advanced system for manipulating objects on a
character terminal. This means that the arguments you need to give to
<code>echotc</code> can be rather hard to come by; try the <code>termcap</code> manual page;
if there are two, it's probably the one in section five which gives the
codes, i.e. `<code>man 5 zsh</code>' or `<code>man -s 5 zsh</code>' on Solaris. Otherwise
you'll have to search the web. The reason the <code>zsh</code> manual doesn't give
a list is that the shell only uses a few well-known sequences, and there
are very many others which will work with <code>echotc</code>, because the
sequences are interpreted by a the terminal, not the shell.</p>
<p>This chunk gives you a flavour:</p>
<pre><code> zmodload -i zsh/termcap
echotc md
echo -n bold
echotc mr
echo -n reverse
echotc me
echo
</code></pre>
<p>First we make sure the module is loaded into the shell; on some older
operating systems, this only works if it was compiled in when zsh was
installed. The option <code>-i</code> to <code>zmodload</code> stops the shell from
complaining if the module was already loaded. This is a sensible way of
ensuring you have the right facilities available in a shell function,
since loading a module makes it available until it is explicitly
unloaded.</p>
<p>You should see `<code>bold</code>' in bold characters, and `<code>reverse</code>' in bold
reverse video. The `<code>md</code>' capability turns on bold mode; `<code>mr</code>' turns
on reverse video; `<code>me</code>' turns off both modes. A more typical zsh way
of doing this is:</p>
<pre><code> print -P '%Bbold%Sreverse%b%s'
</code></pre>
<p>which should show the same thing, but using prompt escapes --- prompts
are the most common use of special fonts. The `<code>%S</code>' is because zsh
calls reverse `standout' mode, because it does. (On a colour xterm, you
may find `bold' is interpreted as `blue'.)</p>
<p>There's a lot more you can do with <code>echotc</code> if you really try. The shell
has just acquired a way of printing terminfo sequences, predictably
called <code>echoti</code>, although it's only available on systems where zsh needs
terminfo to compile --- this happens when the termcap code is actually a
part of terminfo. The good news about this is that terminfo tends to be
better documented, so you have a good chance of finding out the
capabilities you want from the <code>terminfo</code> manual page. The <code>echoti</code>
command lives in another predictably named module, <code>zsh/terminfo</code>.</p>
<p><span id="l33"></span></p>
<h3 id="322-other-builtins-just-for-speed"><a class="header" href="#322-other-builtins-just-for-speed">3.2.2: Other builtins just for speed</a></h3>
<p>There are only a few other builtins which are there just to make things
go faster. Strictly, tests could go into this category, but as I
explained in the last chapter it's useful to have tests in the form</p>
<pre><code> if [[ $var1 = $var2 ]]; then
print doing something
fi
</code></pre>
<p>be treated as a special syntax by the shell, in case <code>$var1</code> or <code>$var2</code>
expands to nothing which would otherwise confuse it. This example
consists of two features described below: the test itself, between the
double square brackets, which is true if the two substituted values are
the same string, and the `<code>if</code>' construct which runs the commands in
the middle (here just the <code>print</code>) if that test was true.</p>
<p>The builtins `<code>true</code>' and `<code>false</code>' do nothing at all, except return a
command status zero or one, respectively. They're just used as
placeholders: to run a loop forever --- <code>while</code> will also be explained
in more detail later --- you use</p>
<pre><code> while true; do
print doing something over and over
done
</code></pre>
<p>since the test always succeeds.</p>
<p>A synonym for `<code>true</code>' is `<code>:</code>'; it's often used in this form to give
arguments which have side effects but which shouldn't be used ---
something like</p>
<pre><code> : ${param:=value}
</code></pre>
<p>which is a common idiom in all Bourne shell derivatives. In the
parameter expansion, <code>$param</code> is given the value <code>value</code> if it was empty
before, and left alone otherwise. Since that was the only reason for the
parameter expansion, you use <code>:</code> to ignore the argument. Actually, the
shell blithely builds the command line --- the colon, followed by
whatever the value of <code>$param</code> is, whether or not the assignment
happened --- then executes the command; it just so happens that `<code>:</code>'
takes no notice of the arguments it was given. If you're switching from
ksh, you may expect certain synonyms like this to be aliases, rather
than builtins themselves, but in zsh they are actually builtins; there
are no aliases predefined by the shell. (You can still get rid of them
using `<code>disable</code>', as described below.)</p>
<p><span id="l34"></span></p>
<h3 id="323-builtins-which-change-the-shells-state"><a class="header" href="#323-builtins-which-change-the-shells-state">3.2.3: Builtins which change the shell's state</a></h3>
<p>A more common use for builtins is that they change something inside the
shell, or report information about what's going on in the shell. There
is one vital thing to remember about external commands. It applies, too,
to other cases we'll meet where the shell `forks', literally splitting
itself into two parts, where the forked-off part behaves just like an
external command. In both of these cases, the command is in a different
<em>process</em>, UNIX's basic unit of things that run. (In fact, even Windows
knows about processes nowadays, although they interact a little bit
differently with one another.)</p>
<p>The vital thing is that no change in a separate process started by the
shell affects the shell itself. The most common case of this is the
current directory --- every process has its own current directory. You
can see this by starting a new zsh:</p>
<pre><code> % pwd # show the current directory
~
% zsh # start a new shell, which
# is a separate process
% cd tmp
% pwd # now I'm in a different
# directory...
~/tmp
% exit # leave the new shell...
% pwd # now I'm back where I was...
~
</code></pre>
<p>Hence the <code>cd</code> command must be a shell builtin, or this would happen
every time you ran it.</p>
<p>Here's a more useful example. Putting parentheses around a command asks
the shell to start a different process for it. That's useful when you
specifically <em>don't</em> want the effects propagating back:</p>
<pre><code> (cd some-other-dir; run-some-command)
</code></pre>
<p>runs the command, but doesn't change the directory the `real' shell is
in, only its forked-off `subshell'. Hence,</p>
<pre><code> % pwd
~
% (cd /; pwd)
/
% pwd
~
</code></pre>
<p>There's a more subtle case:</p>
<pre><code> cd some-other-dir | print Hello
</code></pre>
<p>Remember, the `<code>|</code>' (`pipe') connects the output of the first command
to the input of the next --- though actually no information is passed
that way in this example. In zsh, all but the last portion of the
`pipeline' thus created is run in different processes. Hence the <code>cd</code>
doesn't affect the main shell. I'll refer to it as the `parent' shell,
which is the standard UNIX language for processes; when you start
another command or fork off a subshell, you are creating `children'
(without meaning to be morbid, the children usually die first in this
case). Thus, as you would guess,</p>
<pre><code> print Hello | cd some-other-dir
</code></pre>
<p><em>does</em> have the effect of changing the directory. Note that other shells
do this differently; it is always guaranteed to work this way in zsh,
because many people rely on it for setting parameters, but many shells
have the <em>left</em> hand of the pipeline being the bit that runs in the
parent shell. If both sides of the pipe symbol are external commands of
some sort, both will of course run in subprocesses.</p>
<p>There are other ways you change the state of the shell, for example by
declaring parameters of a particular type, or by telling it how to
interpret certain commands, or, of course, by changing options. Here are
the most useful, grouped in a vaguely logical fashion.</p>
<p><span id="l35"></span></p>
<h3 id="324-cd-and-friends"><a class="header" href="#324-cd-and-friends">3.2.4: cd and friends</a></h3>
<p>You will not by now be surprised to learn that the `<code>cd</code>' command
changes directory. There is a synonym, `<code>chdir</code>', which as far as I
know no-one ever uses. (It's the same name as the system call, so if you
had been programming in C or Perl and forgot that you were now using the
shell, you might use `<code>chdir</code>'. But that seems a bit far-fetched.)</p>
<p>There are various extra features built into <code>cd</code> and <code>chdir</code>. First, if
you miss out the directory to which you want to change, you will be
taken to your home directory, although it's not as if `<code>cd ~</code>' is all
that hard to type.</p>
<p>Next, the command `<code>cd -</code>' is special: it takes you to the last
directory you were in. If you do a sequence of <code>cd</code> commands, only the
immediately preceding directory is remembered; they are not stacked up.</p>
<p>Thirdly, there is a shortcut for changing between similarly named
directories. If you type `<code>cd <old> <new></code>', then the shell will look
for the first occurrence of the string `<code><old></code>' in the current
directory, and try to replace it with `<code><new></code>'. For example,</p>
<pre><code> % pwd
~/src/zsh-3.0.8/Src
% cd 0.8 1.9
~/src/zsh-3.1.9/Src
</code></pre>
<p>The <code>cd</code> command actually reported the new directory, as it usually does
if it's not entirely obvious where it's taken you.</p>
<p>Note that only the <em>first</em> match of <code><old></code> is taken. It's an easy
mistake to think you can change from
<code>/home/export1/pws/mydir1/something</code> to
<code>/home/export1/pws/mydir2/something</code> with `<code>cd 1 2</code>', but that first
`<code>1</code>' messes it up. Arguably the shell could be smarter here. Of
course, `<code>cd r1 r2</code>' will work in this case.</p>
<p><code>cd</code>'s friend `<code>pwd</code>' (print working directory) tells you what the
current working directory is; this information is also available in the
shell parameter <code>$PWD</code>, which is special and automatically updated when
the directory changes. Later, when you know all about expansion, you
will find that you can do tricks with this to refer to other
directories. For example, <code>${PWD/old/new}</code> uses the parameter
substitution mechanism to refer to a different directory with <code>old</code>
replaced by <code>new</code> --- and this time <code>old</code> can be a pattern, i.e.
something with wildcard matches in it. So if you are in the
<code>zsh-3.0.8/Src</code> directory as above and want to copy a file from the
<code>zsh-3.1.9/Src</code> directory, you have a shorthand:</p>
<pre><code> cp ${PWD/0.8/1.9}/myfile.c .
</code></pre>
<p><strong>Symbolic links</strong></p>
<p>Zsh tries to track directories across symbolic links. If you're not
familiar with these, you can think of them as a filename which behaves
like a pointer to another file (a little like Windows' shortcuts, though
UNIX has had them for much longer and they work better). You create them
like this (<code>ln</code> is not a builtin command, but its use to make symbolic
links is very standard these days):</p>
<pre><code> ln -s existing-file-name name-of-link
</code></pre>
<p>for example</p>
<pre><code> ln -s /usr/bin/ln ln
</code></pre>
<p>creates a file called <code>ln</code> in the current directory which does nothing
but point to the file <code>/usr/bin/ln</code>. Symbolic links are very good at
behaving as much like the original file as you usually want; for
example, you can run the <code>ln</code> link you've just created as if it were
<code>/usr/bin/ln</code>. They show up differently in a long file listing with
`<code>ls -l</code>', the last column showing the file they point to.</p>
<p>You can make them point to any sort of file at all, including
directories, and that is why they are mentioned here. Suppose you create
a symbolic link from your home directory to the root directory and
change into it:</p>
<pre><code> ln -s / ~/mylink
cd ~/mylink
</code></pre>
<p>If you don't know it's a link, you expect to be able to change to the
parent directory by doing `<code>cd ..</code>'. However, the operating system ---
which just has one set of directories starting from <code>/</code> and going down,
and ignores symbolic links after it has followed them, they really are
just pointers --- thinks you are in the root directory <code>/</code>. This can be
confusing. Hence zsh tries to keep track of where <em>you</em> probably think
you are, rather than where the system does. If you type `<code>pwd</code>', you
will see `<code>/home/you/mylink</code>' (wherever your home directory is), not
`<code>/</code>'; if you type `<code>cd ..</code>', you will find yourself back in your home
directory.</p>
<p>You can turn all this second-guessing off by setting the option
<code>CHASE_LINKS</code>; then `<code>cd ~/mydir; pwd</code>' will show you to be in <code>/</code>,
where changing to the parent directory has no effect; the parent of the
root directory is the root directory, except on certain slightly
psychedelic networked file systems. This does have advantages: for
example, `<code>cd ~/mydir; ls ..</code>' always lists the root directory, not
your home directory, regardless of the option setting, because <code>ls</code>
doesn't know about the links you followed, only zsh does, and it treats
the <code>..</code> as referring to the root directory. Having <code>CHASE_LINKS</code> set
allows `<code>pwd</code>' to warn you about where the system thinks you are.</p>
<p>An aside for non-UNIX-experts (over 99.9% of the population of the world
at the last count): I said `symbolic links' instead of just `links'
because there are others called `hard links'. This is what `<code>ln</code>'
creates if you don't use the <code>-s</code> option. A hard link is not so much a
pointer to a file as an alternative name for a file. If you do</p>
<pre><code> ln myfile othername
ls -l
</code></pre>
<p>where <code>myfile</code> already exists you can't tell which of <code>myfile</code> and
<code>othername</code> is the original --- and in fact the system doesn't care. You
can remove either, and the other will be perfectly happy as the name for
the file. This is pretty much how renaming files works, except that
creating the hard link is done for you in that case. Hard links have
limitations --- you can't link to directories, or to a file on another
disk partition (and if you don't know what a disk partition is, you'll
see what a limitation that can be). Furthermore, you usually want to
know which is the original and which is the link --- so for most users,
creating symbolic links is more useful. The only drawback is that
following the pointers is a tiny bit slower; if you think you can notice
the difference, you definitely ought to slow down a bit.</p>
<p>The target of a symbolic link, unlike a hard link, doesn't actually have
to exist and no checking is performed until you try to use the link. The
best thing to do is to run `<code>ls -lL</code>' when you create the link; the
<code>-L</code> part tells <code>ls</code> to follow links, and if it worked you should see
that your link is shown as having exactly the same characteristics as
the file it points to. If it is still shown as a link, there was no such
file.</p>
<p>While I'm at it, I should point out one slight oddity with symbolic
links: the name of the file linked to (the first name), if it is not an
absolute path (beginning with <code>/</code> after any <code>~</code> expansion), is treated
relative to the directory where the link is created --- not the current
directory when you run <code>ln</code>. Here:</p>
<pre><code> ln -s ../mydir ~/links/otherdir
</code></pre>
<p>the link <code>otherdir</code> will refer to <code>mydir</code> in <em>its own</em> parent directory,
i.e. <code>~/links</code> --- not, as you might think, the parent of the directory
where you were when you ran the command. What makes it worse is that the
second word, if is not an absolute path, <em>is</em> interpreted relative to
the directory where you ran the command.</p>
<p><strong>$cdpath and AUTO_CD</strong></p>
<p>We're nowhere near the end of the magic you can do with directories yet
(and, in fact, I haven't even got to the zsh-specific parts). The next
trick is <code>$cdpath</code> and <code>$CDPATH</code>. They look a lot like <code>$path</code> and
<code>$PATH</code> which you met in the last chapter, and I mentioned them briefly
back in the last chapter in that context: <code>$cdpath</code> is an array of
directories, while <code>$CDPATH</code> is colon-separated list behaving otherwise
like a scalar variable. They give a list of directories whose
subdirectories you may want to change into. If you use a normal cd
command (i.e. in the form `<code>cd </code><em>dirname</em>', and <em>dirname</em> does not
begin with a <code>/</code> or <code>~</code>, the shell will look through the directories in
<code>$cdpath</code> to find one which contains the subdirectory <em>dirname</em>. If
<code>$cdpath</code> isn't set, as you'd guess, it just uses the current directory.</p>
<p>Note that <code>$cdpath</code> is always searched in order, and you can put a <code>.</code>
in it to represent the current directory. If you do, the current
directory will always be searched <em>at that point</em>, not necessarily
first, which may not be what you expect. For example, let's set up some
directories:</p>
<pre><code> mkdir ~/crick ~/crick/dna
mkdir ~/watson ~/watson/dna
cdpath=(~/crick .)
cd ~/watson
cd dna
</code></pre>
<p>So I've moved to the directory <code>~/watson</code>, which contains the
subdirectory <code>dna</code>, and done `<code>cd dna</code>'. But because of <code>$cdpath</code>, the
shell will look first in <code>~/crick</code>, and find the <code>dna</code> there, and take
you to that copy of the self-reproducing directory, not the one in
<code>~/watson</code>. Most people have <code>.</code> at the start of their <code>cdpath</code> for that
reason. However, at least <code>cd</code> warns you --- if you tried it, you will
see that it prints the name of the directory it's picked in cases like
this.</p>
<p>In fact, if you don't have <code>.</code> in your directory at all, the shell will
always look there first; there's no way of making <code>cd</code> never change to a
subdirectory of the current one, short of turning <code>cd</code> into a function.
Some shells don't do this; they use the directories in <code>$cdpath</code>, and
only those.</p>
<p>There's yet another shorthand, this time specific to zsh: the option
<code>AUTO_CD</code> which I mentioned in the last chapter. That way a command
without any arguments which is really a directory will take you to that
directory. Normally that's perfect --- you would just get a `command
not found' message otherwise, and you might as well make use of the
option. Just occasionally, however, the name of a directory clashes with
the name of a command, builtin or external, or a shell function, and
then there can be some confusion: zsh will always pick the command as
long as it knows about it, but there are cases where it doesn't, as I
described above.</p>
<p>What I didn't say in the last chapter is that <code>AUTO_CD</code> respects
<code>$cdpath</code>; in fact, it really is implemented so that `<em>dirname</em>' on its
own behaves as much like `<code>cd</code> <em>dirname</em>' as is possible without tying
the shell's insides into knots.</p>
<p><strong>The directory stack</strong></p>
<p>One very useful facility that zsh inherited from the C-shell family
(traditional Korn shell doesn't have it) is the directory stack. This is
a list of directories you have recently been in. If you use the command
`<code>pushd</code>' instead of `<code>cd</code>', e.g. `<code>pushd</code> <em>dirname</em>', then the
directory you are in is saved in this list, and you are taken to
<em>dirname</em>, using <code>$CDPATH</code> just as <code>cd</code> does. Then when you type
`<code>popd</code>', you are taken back to where you were. The list can be as long
as you like; you can <code>pushd</code> any number of directories, and each <code>popd</code>
will take you back through the list (this is how a `stack', or more
precisely a `last-in-first-out' stack usually operates in computer
jargon, hence the name `directory stack').</p>
<p>You can see the list --- which always starts with the current directory
--- with the <code>dirs</code> command. So, for example:</p>
<pre><code> cd ~
pushd ~/src
pushd ~/zsh
dirs
</code></pre>
<p>displays</p>
<pre><code> ~/zsh ~/src ~
</code></pre>
<p>and the next <code>popd</code> will take you back to <code>~/src</code>. If you do it, you
will see that <code>pushd</code> reports the list given by <code>dirs</code> automatically as
it goes along; you can turn this off with the option <code>PUSHD_SILENT</code>,
when you will have to rely on typing <code>dirs</code> explicitly.</p>
<p>In fact, a lot of the use of this comes not from using simple <code>pushd</code>
and <code>popd</code> combinations, but from two other features. First, `<code>pushd</code>'
on its own swaps the top two directories on the stack. Second, <code>pushd</code>
with a numeric argument preceded by a `<code>+</code>' or `<code>-</code>' can take you to
one of the other directories in the list. The command `<code>dirs -v</code>' tells
you the numbers you need; <code>0</code> is the current directory. So if you get,</p>
<pre><code> 0 ~/zsh
1 ~/src
2 ~
</code></pre>
<p>then `<code>pushd +2</code>' takes you to <code>~</code>. (A little suspension of disbelief
that I didn't just use <code>AUTO_CD</code> and type `<code>..</code>' is required here.) If
you use a <code>-</code>, it counts from the other end of the list; <code>-0</code> (with
apologies to the numerate) is the last item, i.e. the same as <code>~</code> in
this case. Some people are used to having the `<code>-</code>' and `<code>+</code>'
arguments behave the other way around; the option <code>PUSHD_MINUS</code> exists
for this.</p>
<p>Apart from <code>PUSHD_SILENT</code> and <code>PUSHD_MINUS</code>, there are a few other
relevant options. Setting <code>PUSHD_IGNORE_DUPS</code> means that if you <code>pushd</code>
to a directory which is already somewhere in the list, the duplicate
entry will be silently removed. This is useful for most human operations
--- however, if you are using <code>pushd</code> in a function or script to
remember previous directories for a future matching <code>popd</code>, this can be
dangerous and you probably want to turn it off locally inside the
function.</p>
<p><code>AUTO_PUSHD</code> means that any directory-changing command, including an
auto-cd, is treated as a <code>pushd</code> command with the target directory as
argument. Using this can make the directory stack get very long, and
there is a parameter <code>$DIRSTACKSIZE</code> which you can set to specify a
maximum length. The oldest entry (the highest number in the `<code>dirs -v</code>'
listing) is automatically removed when this length is exceeded. There is
no limit unless this is explicitly set.</p>
<p>The final <code>pushd</code> option is <code>PUSHD_TO_HOME</code>. This makes <code>pushd</code> on its
own behave like <code>cd</code> on its own in that it takes you to your home
directory, instead of swapping the top two directories. Normally a
series of `<code>pushd</code>' commands works pretty much like a series of `<code>cd -</code>' commands, always taking you the directory you were in before, with
the obvious difference that `<code>cd -</code>' doesn't consult the directory
stack, it just remembers the previous directory automatically, and hence
it can confuse <code>pushd</code> if you just use `<code>cd -</code>' instead.</p>
<p>There's one remaining subtlety with <code>pushd</code>, and that is what happens to
the rest of the list when you bring a particular directory to the front
with something like `<code>pushd +2</code>'. Normally the list is simply cycled,
so the directories which were +3, and +4 are now right behind the new
head of the list, while the two directories which were ahead of it get
moved to the end. If the list before was:</p>
<pre><code> dir1 dir2 dir3 dir4
</code></pre>
<p>then after <code>pushd +2</code> you get</p>
<pre><code> dir3 dir4 dir1 dir2
</code></pre>
<p>That behaviour changed during the lifetime of zsh, and some of us
preferred the old behaviour, where that one directory was yanked to the
front and the rest just closed the gap:</p>
<pre><code> # Old behaviour
dir3 dir1 dir2 dir4
</code></pre>
<p>so that after a while you get a `greatest hits' group at the front of
the list. If you like this behaviour too (I feel as if I'd need to have
written papers on group theory to like the new behaviour) there is a
function <code>pushd</code> supplied with the source code, although it's short
enough to repeat here --- this is in the form for autoloading in the zsh
fashion:</p>
<pre><code> # pushd function to emulate the old zsh behaviour.
# With this, pushd +/-n lifts the selected element
# to the top of the stack instead of cycling
# the stack.
emulate -R zsh
setopt localoptions
if [[ ARGC -eq 1 && "$1" == [+-]<-> ]] then
setopt pushdignoredups
builtin pushd ~$1
else
builtin pushd "$@"
fi
</code></pre>
<p>The `<code>&&</code>' is a logical `and', requiring both tests to be true. The
tests are that there is exactly one argument to the function, and that
it has the form of a `<code>+</code>' or a `<code>-</code>' followed by any number (`<code><-></code>'
is a special zsh pattern to match any number, an extension of forms like
`<code><1-100></code>' which matches any number in the range 1 to 100 inclusive).</p>
<p><strong>Referring to other directories</strong></p>
<p>Zsh has two ways of allowing you to refer to particular directories.
They have in common that they begin with a <code>~</code> (in very old versions of
zsh, the second form actually used an `<code>=</code>', but the current way is
much more logical).</p>
<p>You will certainly be aware, because I've made a lot of use of it, that
a `<code>~</code>' on its own or followed by a <code>/</code> refers to your own home
directory. An extension of this --- again from the C-shell, although the
Korn shell has it too in this case --- is that <code>~name</code> can refer to the
home directory of any user on the system. So if your user name is <code>pws</code>,
then <code>~</code> and <code>~pws</code> are the same directory.</p>
<p>Zsh has an extension to this; you can actually name your own
directories. This was described in <a href="zshguide02.html#init">chapter 2</a>, Ã
propos of prompts, since that is the major use:</p>
<pre><code> host% PS1='%~? '
~? cd zsh/Src
~/zsh/Src? zsrc=$PWD
~/zsh/Src? echo ~zsrc
/home/pws/zsh/Src
~zsrc?
</code></pre>
<p>Consult that chapter for the ways of forcing a parameter to be
recognised as a named directory.</p>
<p>There's a slightly more sophisticated way of doing this directly:</p>
<pre><code> hash -d zsrc=~/zsh/Src
</code></pre>
<p>makes <code>~zsrc</code> appear in prompts as before, and in this case there is no
parameter <code>$zsrc</code>. This is the purist's way (although very few zsh users
are purists). You can guess what `<code>unhash -d zsrc</code>' does; this works
with directories named via parameters, too, but leaves the parameter
itself alone.</p>
<p>It's possible to have a named directory with the same name as a user. In
that case `<code>~name</code>' refers to the directory you named explicitly, and
there is no easy way of getting <code>name</code>'s home directory without removing
the name you defined.</p>
<p>If you're using named directories with one of the <code>cd</code>-like commands or
<code>AUTO_CD</code>, you can set the option <code>CDABLEVARS</code> which allows you to omit
the leading <code>~</code>; `<code>cd zsrc</code>' with this option would take you to
<code>~zsrc</code>. The name is a historical artifact and now a misnomer; it really
is named directories, not parameters (i.e. variables), which are used.</p>
<p>The second way of referring to directories with <code>~</code>'s is to use numbers
instead of names: the numbers refer to directories in the directory
stack. So if <code>dirs -v</code> gives you</p>
<pre><code> 0 ~zsf
1 ~src
</code></pre>
<p>then <code>~+1</code> and <code>~-0</code> (not very mathematical, but quite logical if you
think about it) refer to <code>~src</code>. In this case, unlike pushd arguments,
you can omit the <code>+</code> and use <code>~1</code>. The option <code>PUSHD_MINUS</code> is
respected. You'll see this was used in the <code>pushd</code> function above: the
trick was that <code>~+3</code>, for example, refers to the same element as <code>pushd +3</code>, hence <code>pushd ~+3</code> pushed that directory onto the front of the list.
However, we set <code>PUSHD_IGNORE_DUPS</code>, so that the value in the old
position was removed as well, giving us the effect we wanted of simply
yanking the directory to the front with no trick cycling.</p>
<p><span id="l36"></span></p>
<h3 id="325-command-control-and-information-commands"><a class="header" href="#325-command-control-and-information-commands">3.2.5: Command control and information commands</a></h3>
<p>Various builtins exist which control how you access commands, and which
show you information about the commands which can be run.</p>
<p>The first two are strictly speaking `precommand modifiers' rather than
commands: that means that they go before a command line and modify its
behaviour, rather than being commands in their own right. If you put
`<code>command</code>' in front of a command line, the command word (the next one
along) will be taken as the name of an external command, however it
would normally be interpreted; likewise, if you put `<code>builtin</code>' in
front, the shell will try to run the command as a builtin command.
Normally, shell functions take precedence over builtins which take
precedence over external commands. So, for example, if your printer
control system has the command `<code>enable</code>' (as many System V versions
do), which clashes with a builtin I am about to talk about, you can run
`<code>command enable lp</code>' to enable a printer; otherwise, the builtin
enable would have been run. Likewise, if you have defined <code>cd</code> to be a
function, but this time want to call the normal builtin <code>cd</code>, you can
say `<code>builtin cd mydir</code>'.</p>
<p>A common use for <code>command</code> is inside a shell function of the same name.
Sometimes you want to enhance an ordinary command by sticking some extra
stuff around it, then calling that command, so you write a shell
function of the same name. To call the command itself inside the shell
function, you use `<code>command</code>'. The following works, although it's
obviously not all that useful as it stands:</p>
<pre><code> ls() {
command ls "$[@]"
}
</code></pre>
<p>so when you run `<code>ls</code>', it calls the function, which calls the real
<code>ls</code> command, passing on the arguments you gave it.</p>
<p>You can gain longer lasting control over the commands which the shell
will run with the `<code>disable</code>' and `<code>enable</code>' commands. The first
normally takes builtin arguments; each such builtin will not be
recognised by the shell until you give an `<code>enable</code>' command for it. So
if you want to be able to run the external <code>enable</code> command and don't
particularly care about the builtin version, `<code>disable enable</code>' (sorry
if that's confusing) will do the trick. Ha, you're thinking, you can't
run `<code>enable enable</code>'. That's correct: some time in the dim and distant
past, <code>builtin enable enable</code>' would have worked, but currently it
doesn't; this may change, if I remember to change it. You can list all
disabled builtins with just `<code>disable</code>' on its own --- most of the
builtins that do this sort of manipulation work like that.</p>
<p>You can manipulate other sets of commands with <code>disable</code> and <code>enable</code> by
giving different options: aliases with the option <code>-a</code>, functions with
<code>-f</code>, and reserved words with <code>-r</code>. The first two you probably know
about, and I'll come to them anyway, but `reserved words' need
describing. They are essentially builtin commands which have some
special syntactic meaning to the shell, including some symbols such as
`<code>{</code>' and `<code>[[</code>'. They take precedence over everything else except
aliases --- in fact, since they're syntactically special, the shell
needs to know very early on that it has found a reserved word, it's no
use just waiting until it tries to execute a command. For example, if
the shell finds `<code>[[</code>' it needs to know that everything until `<code>]]</code>'
must be treated as a test rather than as ordinary command arguments.
Consequently, you wouldn't often want to disable a reserved word, since
the shell wouldn't work properly. The most obvious reason why you might
would be for compatibility with some other shell which didn't have one.
You can get a complete list with:</p>
<pre><code> whence -wm '*' | grep reserved
</code></pre>
<p>which I'll explain below, since I'm coming to `<code>whence</code>'.</p>
<p>Furthermore, I tend to find that if I want to get rid of aliases or
functions I use the commands `<code>unalias</code>' and `<code>unfunction</code>' to get rid
of them permanently, since I always have the original definitions stored
somewhere, so these two options may not be that useful either. Disabling
builtins is definitely the most useful of the four possibilities for
<code>disable</code>.</p>
<p>External commands have to be manipulated differently. The types given
above are handled internally by the shell, so all it needs to do is
remember what code to call. With external commands, the issue instead is
how to find them. I mentioned <code>rehash</code> above, but didn't tell you that
the <code>hash</code> command, which you've already seen with the <code>-d</code> option, can
be used to tell the shell how to find an external command:</p>
<pre><code> hash foo=/path/to/foo
</code></pre>
<p>makes <code>foo</code> execute the command using the path shown (which doesn't even
have to end in `<code>foo</code>'). This is rather like an alias --- most people
would probably do this with an alias, in fact --- although a little
faster, though you're unlikely to notice the difference. You can remove
this with <code>unhash</code>. One gotcha here is that if the path is rehashed,
either by calling <code>rehash</code> or when you alter <code>$path</code>, the entire hash
table is emptied, including anything you put in in this way; so it's not
particularly useful.</p>
<p>In the midst of all this, it's useful to be able to find out what the
shell thinks a particular command name does. The command `<code>whence</code>'
tells you this; it also exists, with slightly different options, under
the names <code>where</code>, <code>which</code> and <code>type</code>, largely to provide compatibility
with other shells. I'll just stick to <code>whence</code>.</p>
<p>Its standard output isn't actually sparklingly interesting. If it's a
command somehow known to the shell internally, it gets echoed back, with
the alias expanded if it was an alias; if it's an external command it's
printed with the full path, showing where it came from; and if it's not
known the command returns status 1 and prints nothing.</p>
<p>You can make it more useful with the <code>-v</code> or <code>-c</code> options, which are
more verbose; the first prints out an information message, while the
second prints out the definitions of any functions it was asked about
(this is also the effect of using `<code>which</code>' instead of `<code>whence</code>). A
very useful option is <code>-m</code>, which takes any arguments as patterns using
the usual zsh pattern format, in other words the same one used for
matching files. Thus</p>
<pre><code> whence -vm "*"
</code></pre>
<p>prints out every command the shell knows about, together with what it
thinks of it.</p>
<p>Note the quotes around the `<code>*</code>' --- you have to remember these
anywhere where the pattern is not to be used to generate filenames on
the command line, but instead needs to be passed to the command to be
interpreted. If this seems a rather subtle distinction, think about what
would happen if you ran</p>
<pre><code> # Oops. Better not try this at home.
# (Even better, don't do it at work either.)
whence -vm *
</code></pre>
<p>in a directory with the files `<code>foo</code>' and (guess what) `<code>bar</code>' in it.
The shell hasn't decided what command it's going to run when it first
looks at the command line; it just sees the `<code>*</code>' and expands the line
to</p>
<pre><code> whence -vm foo bar
</code></pre>
<p>which isn't what you meant.</p>
<p>There are a couple of other tricks worth mentioning: <code>-p</code> makes the
shell search your path for them, even if the name is matched as
something else (say, a shell function). So if you have <code>ls</code> defined as a
function,</p>
<pre><code> which -p ls
</code></pre>
<p>will still tell what `<code>command ls</code>' would find. Also, the option <code>-a</code>
searches for all commands; in the same example, this would show you both
the <code>ls</code> command and the <code>ls</code> function, whereas <code>whence</code> would normally
only show the function because that's the one that would be run. The
<code>-a</code> option also shows if it finds more than one external command in
your path.</p>
<p>Finally, the option <code>-w</code> is useful because it identifies the type of a
command with a single word: <code>alias</code>, <code>builtin</code>, <code>command</code>, <code>function</code>,
<code>hashed</code>, <code>reserved</code> or <code>none</code>. Most of those are obvious, with
<code>command</code> being an ordinary external command; <code>hashed</code> is an external
command which has been explicitly given a path with the <code>hash</code> builtin,
and <code>none</code> means it wasn't recognised as a command at all. Now you know
how we extracted the reserved words above.</p>
<p>A close relative of <code>whence</code> is <code>functions</code>, which applies, of course,
to shell functions; it usually lists the definitions of all functions
given as arguments, but its relatives (of which <code>autoload</code> is one)
perform various other tricks, to be described in the section on shell
functions below. Be careful with <code>function</code>, without the `s', which is
completely different and not like <code>command</code> or <code>builtin</code> --- it is
actually a keyword used to <em>define</em> a function.</p>
<p><span id="l37"></span></p>
<h3 id="326-parameter-control"><a class="header" href="#326-parameter-control">3.2.6: Parameter control</a></h3>
<p>There are various builtins for controlling the shells parameters. You
already know how to set and use parameters, but it's a good deal more
complicated than that when you look at the details.</p>
<p><strong>Local parameters</strong></p>
<p>The principal command for manipulating the behaviour of parameters is
`<code>typeset</code>'. Its easiest usage is to declare a parameter; you just give
it a list of parameter names, which are created as scalar parameters.
You can create parameters just by assigning to them, but the major point
of `<code>typeset</code>' is that if a parameter is created that way inside a
function, the parameter is restored to its original value, or removed if
it didn't previously exist, at the end of the function --- in other
words, it has `local scope' like the variables which you declare in
most ordinary programming languages. In fact, to use the jargon it has
`dynamical' rather than `syntactic' scope, which means that the same
parameter is visible in any function called within the current one; this
is different from, say, C or FORTRAN where any function or subroutine
called wouldn't see any variable declared in the parent function.</p>
<p>The following makes this more concrete.</p>
<pre><code> var='Original value'
subfn() {
print $var
}
fn() {
print $var
typeset var='Value in function'
print $var
subfn
}
fn
print $var
</code></pre>
<p>This chunk of code prints out</p>
<pre><code> Original value
Value in function
Value in function
Original value
</code></pre>
<p>The first three chunks of the code just define the parameter <code>$var</code>, and
two functions, <code>subfn</code> and <code>fn</code>. Then we call <code>fn</code>. The first thing this
does is print out <code>$var</code>, which gives `<code>Original value</code>' since we
haven't changed the original definition. However, the <code>typeset</code> next
does that; as you see, we can assign to the parameter during the
typeset. Thus when we print <code>$var</code> out again, we get `<code>Value in function</code>'. Then <code>subfn</code> is called, which prints out the same value as
in <code>fn</code>, because we haven't changed it --- this is where C or FORTRAN
would differ, and wouldn't recognise the variable because it hadn't been
declared in that function. Finally, <code>fn</code> exits and the original value is
restored, and is printed out by the final `<code>print</code>'.</p>
<p>Note the value changes twice: first at the <code>typeset</code>, then again at the
end of <code>fn</code>. The value of <code>$var</code> at any point will be one of those two
values.</p>
<p>Although you can do assignments in a <code>typeset</code> statement, you can't
assign to arrays (I already said this in the last chapter):</p>
<pre><code> typeset var=(Doesn\'t work\!)
</code></pre>
<p>because the syntax with the parentheses is special; it only works when
the line consists of nothing but assignments. However, the shell doesn't
complain if you try to assign an array to a scalar, or vice versa; it
just silently converts the type:</p>
<pre><code> typeset var='scalar value'
var=(array value)
</code></pre>
<p>I put in the assignment in the typeset statement to rub the point in
that it creates scalars, but actually the usual way of setting up an
array in a function is</p>
<pre><code> typeset var
var=()
</code></pre>
<p>which creates an empty scalar, then converts that to an empty array.
Recent versions of the shell have `<code>typeset -a var</code>' to do that in one
go --- but you <em>still</em> can't assign to it in the same statement.</p>
<p>There are other catches associated with the fact that <code>typeset</code> and its
relatives are just ordinary commands with ordinary sets of arguments.
Consider this:</p>
<pre><code> % typeset var=`echo two words`
% print $var
two
</code></pre>
<p>What has happened to the `<code>words</code>'? The answer is that backquote
substitution, to be discussed below, splits words when not quoted. So
the <code>typeset</code> statement is equivalent to</p>
<pre><code> % typeset var=two words
</code></pre>
<p>There are two ways to get round this; first, use an ordinary assignment:</p>
<pre><code> % typeset var
% var=`echo two words`
</code></pre>
<p>which can tell a scalar assignment, and hence knows not to split words,
or quote the backquotes,</p>
<pre><code> % typeset var="`echo two words`"
</code></pre>
<p>There are three important types we haven't talked about; both of these
can only be created with <code>typeset</code> or one of the similar builtins I'll
list in a moment. They are integer types, floating point types, and
associative array types.</p>
<p><strong>Numeric parameters</strong></p>
<p>Integers are created with `<code>typeset -i</code>', or `<code>integer</code>' which is
another way of saying the same thing. They are used for arithmetic,
which the shell can do as follows:</p>
<pre><code> integer i
(( i = 3 * 2 + 1 ))
</code></pre>
<p>The double parentheses surround a complete arithmetic expression: it
behaves as if it's quoted. The expression inside can be pretty much
anything you might be used to from arithmetic in other programming
languages. One important point to note is that parameters don't need to
have the <code>$</code> in front, even when their value is being taken:</p>
<pre><code> integer i j=12
(( i = 3 * ( j + 4 ) ** 2 ))
</code></pre>
<p>Here, <code>j</code> will be replaced by 12 and <code>$i</code> gets the value 768 (sixteen
squared times three). One thing you might not recognise is the <code>**</code>,
which is the `to the power of' operator which occurs in FORTRAN and
Perl. Note that it's fine to have parentheses inside the double
parentheses --- indeed, you can even do</p>
<pre><code> (( i = (3 * ( j + 4 )) ** 2 ))
</code></pre>
<p>and the shell won't get confused because it knows that any parentheses
inside must be in balanced pairs (until you deliberately confuse it with
your buggy code).</p>
<p>You would normally use `<code>print $i</code>' to see what value had been given to
<code>$i</code>, of course, and as you would expect it gets printed out as a
decimal number. However, <code>typeset</code> allows you to specify another base
for printing out. If you do</p>
<pre><code> typeset -i 16 i
print $i
</code></pre>
<p>after the last calculation, you should see <code>16#900</code>, which means 900 in
base 16 (hexadecimal). That's the only effect the option `<code>-i 16</code>' has
on <code>$i</code> --- you can assign to it and use it in arithmetical expressions
just as normal, but when you print it out it appears in this form. You
can use this base notation for inputting numbers, too:</p>
<pre><code> (( i = 16#ff * 2#10 ))
</code></pre>
<p>which means 255 (<code>ff</code> in hexadecimal) times 2 (<code>10</code> in binary). The
shell understands C notation too, so `<code>16#ff</code>' could have been
expressed `<code>0xff</code>'.</p>
<p>Floating point variables are very similar. You can declare them with
`<code>typeset -F</code>' or `<code>typeset -E</code>'. The only difference between the two
is, again, on output; <code>-F</code> uses a fixed point notation, while <code>-E</code> uses
scientific (mnemonic: exponential) notation. The builtin `<code>float</code>' is
equivalent to `<code>typeset -E</code>' (because Korn shell does it, that's why).
Floating point expressions also work the way you are probably used to:</p>
<pre><code> typeset -E e
typeset -F f
(( e = 32/3, f = 32.0/3.0 ))
print $e $f
</code></pre>
<p>prints</p>
<pre><code> 1.000000000e+01 10.6666666667
</code></pre>
<p>Various points: the `<code>,</code>' can separate different expressions, just like
in C, so the <code>e</code> and <code>f</code> assignments are performed separately. The <code>e</code>
assignment was actually an integer division, because neither 32 nor 3 is
a floating point number, which must contain a dot. That means an integer
division was done, producing 10, which was then converted to a floating
point number only at the end. Again, this is just how grown-up languages
work, so it's no use cursing. The <code>f</code> assignment was a full floating
point performance. Floating point parameters weren't available before
version <code>3.1.7</code>.</p>
<p>Although this is really a matter for a later chapter, there is a library
of floating point functions you can load (actually it's just a way of
linking in the system mathematical library). The usual incantation is
`<code>zmodload zsh/mathfunc</code>'; you may not have `dynamic loading' of
libraries on your system, which may mean that doesn't work. If it does,
you can do things like</p>
<pre><code> (( pi = 4.0 * atan(1.0) ))
</code></pre>
<p>Broadly, all the functions which appear in most system mathematical
libraries (see the manual page for <code>math</code>) are available in zsh.</p>
<p>Like all other parameters created with <code>typeset</code> or one of its cousins,
integer and floating point parameters are local to functions. You may
wonder how to create a global parameter (i.e. one which is valid outside
as well as inside the function) which has an integer or floating point
value. There's a recent addition to the shell (in version 3.1.6) which
allows this: use the flag <code>-g</code> to typeset along with any others. For
example,</p>
<pre><code> fn() {
typeset -Fg f
(( f = 42.75 ))
}
fn
print $f
</code></pre>
<p>If you try it, you will see the value of <code>$f</code> has survived beyond the
function. The <code>g</code> stands for global, obviously, although it's not quite
that simple:</p>
<pre><code> fn() {
typeset -Fg f
}
outerfn() {
typeset f='scalar value'
fn
print $f
}
outerfn
</code></pre>
<p>The function <code>outerfn</code> creates a local scalar value for <code>f</code>; that's what
<code>fn</code> sees. So it was not really operating on a `global' value, it just
didn't create a new one for the scope of <code>fn</code>. The error message comes
because it tried to preserve the value of <code>$f</code> while changing its type,
and the value wasn't a proper floating point expression. The error
message,</p>
<pre><code> fn: bad math expression: operator expected at `value'
</code></pre>
<p>comes about because assigning to numeric parameters always does an
arithmetic evaluation. Operating on `<code>scalar value</code>' it found
`<code>scalar</code>' and assumed this was a parameter, then looked for an
operator like `<code>+</code>' to come next; instead it found `<code>value</code>'. If you
want to experiment, change the string to `<code>scalar + value</code>' and set
`<code>value=42</code>', or whatever, then try again. This is a little confusing
(which is a roundabout way of saying it confused me), but consistent
with how zsh usually treats parameters.</p>
<p>Actually, to a certain extent you don't need to use the integer and
floating point parameters. Any time zsh needs a numeric expression it
will force a scalar to the right value, and any time it produces a
numeric expression and assigns it to a scalar, it will convert the
result to a string. So</p>
<pre><code> typeset num=3 # This is the *string* `3'.
(( num = num + 1 )) # But this works anyway
# ($num is still a string).
</code></pre>
<p>This can be useful if you have a parameter which is sometimes a number,
sometimes a string, since zsh does all the conversion work for you.
However, it can also be confusing if you always want a number, because
zsh can't guess that for you; plus it's a little more efficient not to
have to convert back and forth; plus you lose accuracy when you do,
because if the number is stored as a string rather than in the internal
numeric representation, what you say is what you get (although zsh tends
to give you quite a lot of decimal places when converting implicitly to
strings). Anyway, I'd recommend that if you know a parameter has to be
an integer or floating point value you should declare it as such.</p>
<p>There is a builtin called <code>let</code> to handle mathematical expressions, but
since</p>
<pre><code> let "num = num + 1"
</code></pre>
<p>is equivalent to</p>
<pre><code> (( num = num + 1 ))
</code></pre>
<p>and the second form is easier and more memorable, you probably won't
need to use it. If you do, remember that (unlike BASIC) each
mathematical expression should appear as one argument in quotes.</p>
<p><strong>Associative arrays</strong></p>
<p>The one remaining major type of parameter is the associative array; if
you use Perl, you may call it a `hash', but we tend not to since that's
really a description of how it's implemented rather than what it does.
(All right, what it does is hash things. Now shut up.)</p>
<p>These have to be declared by a typeset statement --- there's no getting
round it. There are some quite eclectic builtins that produce a
filled-in associative array for you, but the only way to tell zsh you
want your very own associative array is</p>
<pre><code> typeset -A assoc
</code></pre>
<p>to create <code>$assoc</code>. As to what it does, that's best shown by example:</p>
<pre><code> typeset -A assoc
assoc=(one eins two zwei three drei)
print ${assoc[two]}
</code></pre>
<p>which prints `<code>zwei</code>'. So it works a bit like an ordinary array, but
the numeric <em>subscript</em> of an ordinary array which would have appeared
inside the square bracket is replaced by the string <em>key</em>, in this case
<code>two</code>. The array assignment was a bit deceptive; the `values' were
actually pairs, with `<code>one</code>' being the key for the value `<code>eins</code>', and
so on. The shell will complain if there are an odd number of elements in
such a list. This may also be familiar from Perl. You can assign values
one at a time:</p>
<pre><code> assoc[four]=vier
</code></pre>
<p>and also unset one key/value pair:</p>
<pre><code> unset 'assoc[one]'
</code></pre>
<p>where the quotes stop the square brackets from being interpreted as a
pattern on the command line.</p>
<p>Expansion has been held over, but you might like to know about the ways
of getting back what you put in. If you do</p>
<pre><code> print $assoc
</code></pre>
<p>you just see the values --- that's exactly the same as with an ordinary
array, where the subscripts 1, 2, 3, etc. aren't shown. Note they are in
random order --- that's the other main difference from ordinary arrays;
associative arrays have no notion of an order unless you explicitly sort
them.</p>
<p>But here the keys may be just as interesting. So there is:</p>
<pre><code> print ${(k)assoc}
print ${(kv)assoc}
</code></pre>
<p>giving (if you've followed through all the commands above):</p>
<pre><code> four two three
four vier two zwei three drei
</code></pre>
<p>which print out the keys instead of the values, and the key and value
pairs much as you entered them. You can see that, although the order of
the pairs isn't obvious, it's the same each time. From this example you
can work out how to copy an associative array into another one:</p>
<pre><code> typeset -A newass
newass=(${(kv)assoc})
</code></pre>
<p>where the `<code>(kv)</code>' is important --- as is the <code>typeset</code> just before the
assignment, otherwise <code>$newass</code> would be a badass ordinary array. You
can also prove that <code>${(v)assoc}</code> does what you would probably expect.
There are lots of other tricks, but they are mostly associated with
clever types of parameter expansion, to be described in <a href="zshguide05.html#subst">chapter
5</a>.</p>
<p><strong>Other typeset and type tricks</strong></p>
<p>There are variants of <code>typeset</code>, some mentioned sporadically above.
There is nothing you can do with any of them that you can't do with
<code>typeset</code> --- that wasn't always the case; we've tried to improve the
orthogonality of the options. They differ in the options which are set
by default, and the additional options which are allowed. Here's a list:
<code>declare</code>, <code>export</code>, <code>float</code>, <code>integer</code>, <code>local</code>, <code>readonly</code>. I won't
confuse you by describing all in detail; see the manual.</p>
<p>If there is an odd one out, it's <code>export</code>, which not only marks a
parameter for export but has the <code>-g</code> flag turned on by default, so that
that parameter is not local to the function; in other words, it's
equivalent to <code>typeset -gx</code>. However, one holdover from the days when
the options weren't quite so logical is that <code>typeset -x</code> behaves like
<code>export</code>, in other words the <code>-g</code> flag is turned on by default. You can
fix this by unsetting the option <code>GLOBAL_EXPORT</code> --- the option only
exists for compatibility; logically it should always be unset. This is
partly because in the old days you couldn't export local parameters, so
<code>typeset -x</code> either had to turn on <code>-g</code> or turn off <code>-x</code>; that was fixed
for the 3.1.9 release, and (for example) `<code>local -x</code>' creates a local
parameter which is exported to the environment; both the parameter
itself, and the value in the environment, will be restored when the
function exits. The builtin <code>local</code> is essentially a form of <code>typeset</code>
which renounces the <code>-g</code> flag and all its works.</p>
<p>Another old restriction which has gone is that you couldn't make special
parameters, in particular <code>$PATH</code>, local to a function; you just
modified the original parameter. Now if you say `<code>typeset PATH</code>',
things happen the way you probably expect, with <code>$PATH</code> having its usual
effect, and being restored to its old value when the function exits.
Since <code>$PATH</code> is still special, though, you should make sure you assign
something to it in the function before calling external commands, else
it will be empty and no commands will be found. It's possible that you
specifically don't want some parameter you make local to have the
special property; 3.1.7 and after allow the typeset flag <code>-h</code> to hide
the specialness for that parameter, so in `<code>typeset -h PATH</code>', <code>PATH</code>
would be an ordinary variable for the duration of the enclosing
function. Internally, the same value as was previously set would
continue to be used for finding commands, but it wouldn't be exported.</p>
<p>The second main use of <code>typeset</code> is to set attributes for the
parameters. In this case it can operate on an existing parameter, as
well as creating a new one. For example,</p>
<pre><code> typeset -r msg='This is an important message.'
</code></pre>
<p>sets the readonly flag (-r) for the parameter <code>msg</code>. If the parameter
didn't exist, it would be created with the usual scoping rules; but if
it did exist at the current level of scoping, it would be made readonly
with the value assigned to it, meaning you can't set that particular
copy of the parameter. For obvious reasons, it's normal to assign a
value to a readonly parameter when you first declare it. Here's a
reality check on how this affects scoping:</p>
<pre><code> msg='This is an ordinary parameter'
fn() {
typeset msg='This is a local ordinary parameter'
print $msg
typeset -r msg='This is a local readonly parameter'
print $msg
msg='Watch me cause an error.'
}
fn
print $msg
msg='This version of the parameter'\
' can still be overwritten'
print $msg
</code></pre>
<p>outputs</p>
<pre><code> This is a local ordinary parameter
This is a local readonly parameter
fn:5: read-only variable: msg
This is an ordinary parameter
This version of the parameter can still be overwritten
</code></pre>
<p>Unfortunately there was a bug with this code until recently --- thirty
seconds ago, actually: the second <code>typeset</code> in <code>fn</code> incorrectly added
the readonly flag to the existing <code>msg</code> <em>before</em> attempting to set the
new value, which was wrong and inconsistent with what happens if you
create a new local parameter. Maybe it's reassuring that the shell can
get confused about local parameters, too. (I don't find it reassuring in
the slightest, since <code>typeset</code> is one of the parts of the code where I
tend to fix the bugs, but maybe you do.)</p>
<p>Anyway, when the bug is fixed, you should get the output shown, because
the first typeset created a local variable which the second typeset made
readonly, so that the final assignment caused an error. Then the <code>$msg</code>
in the function went out of scope, and the ordinary parameter, with no
readonly restriction, was visible again.</p>
<p>I mentioned another special typeset option in the previous chapter:</p>
<pre><code> typeset -T TEXINPUTS texinputs
</code></pre>
<p>to tie together the scalar <code>$TEXINPUTS</code> and the array <code>$texinputs</code> in
the same way that <code>$PATH</code> and <code>$path</code> work. This is a one-off; it's the
only time <code>typeset</code> takes exactly two parameter names on the command
line. All other uses of typeset take a list of parameters to which any
flags given are applied. See the manual for the remaining flags,
although most of the more interesting ones have been discussed.</p>
<p>The other thing you need to know about flags is that you use them with a
`<code>+</code>' sign to turn off the corresponding attribute. So</p>
<pre><code> typeset +r msg
</code></pre>
<p>allows you to set <code>$msg</code> again. From version <code>4.1</code>, you won't be able to
turn off the readonly attribute for a special parameter; that's because
there's too much scope for confusion, including attempting to set
constant strings in the code. For example, `<code>$ZSH_VERSION</code>' always
prints a fixed string; attempting to change that is futile.</p>
<p>The final use of typeset is to list parameters. If you type `<code>typeset</code>'
on its own, you get a complete list of parameters and their values. From
3.1.7, you can turn on the flag <code>-H</code> for a parameter, which means to
hide its value while you're doing this. This can be useful for some of
the more enormous parameters, particularly special parameters which I'll
talk about in the section in <a href="zshguide07.html#ragbag">chapter 7</a> on
modules, which tend to swamp the display <code>typeset</code> produces.</p>
<p>You can also list parameters of a particular type, by listing the flags
you want to know about. For example,</p>
<pre><code> typeset -r
</code></pre>
<p>lists all readonly parameters. You might expect `<code>typeset +r</code>' to list
parameters which <em>don't</em> have that attribute, but actually it lists the
same parameters but without showing their value. `<code>typeset +</code>' lists
all parameters in this way.</p>
<p>Another good way of finding out about parameters is to use the special
expansion `<code>${(t)</code><em>param</em><code>}</code>', for example</p>
<pre><code> print ${(t)PATH}
</code></pre>
<p>prints `<code>scalar-export-special</code>': <code>$PATH</code> is a scalar parameter, with
the <code>-x</code> flag set, and has a special meaning to the shell. Actually,
`<code>special</code>' means something a bit more than that: it means the internal
code to get and set the parameter behaves in a way which has side
effects, either to the parameter itself or elsewhere in the shell. There
are other parameters, like <code>$HISTFILE</code>, which are used by the shell, but
which are get and set in a normal way --- they are only special in that
the value is looked at by the shell; and, after all, any old shell
function can do that, too. Contrast this with <code>$PATH</code> which has all that
paraphernalia to do with hashing commands to take care of when it's set,
as I discussed above, and I hope you'll see the difference.</p>
<p><strong>Reading into parameters</strong></p>
<p>The `<code>read</code>' builtin, as its name suggests, is the opposite to
`<code>print</code>' (there's no `<code>write</code>' command in the shell, though there is
often an external command of that name to send a message to another
user), but reading, unlike printing, requires something in the shell to
change to take the value, so unlike <code>print</code>, <code>read</code> is forced to be a
builtin. Inevitably, the values are read into a parameter. Normally they
are taken from standard input, very often the terminal (even if you're
running a script, unless you redirected the input). So the simplest case
is just</p>
<pre><code> read param
</code></pre>
<p>and if you type a line, and hit return, it will be put into <code>$param</code>,
without the final newline.</p>
<p>The <code>read</code> builtin actually does a bit of processing on the input. It
will usually strip any initial or final whitespace (spaces or tabs) from
the line read in, though any in the middle are kept. You can read a set
of values separated by whitespace just by listing the parameters to
assign them to; the last parameter gets all the remainder of the line
without it being split. Very often it's easiest just to read into an
array:</p>
<pre><code> % read -A array
this is a line typed in now, \
by me, in this space
% print ${array[1]} ${array[12]}
this space
</code></pre>
<p>(I'm assuming you're using the native zsh array format, rather than the
one set with <code>KSH_ARRAYS</code>, and shall continue to assume this.)</p>
<p>It's useful to be able to print a prompt when you want to read
something. You can do this with `<code>print -n</code>', but there's a shorthand:</p>
<pre><code> % read line'?Please enter a line: '
Please enter a line: some words
% print $line
some words
</code></pre>
<p>Note the quotes surround the `<code>?</code>' to prevent it being taken as part of
a pattern on the command line. You can quote the whole expression from
the beginning of `<code>line</code>', if you like; I just write it like that
because I know parameter names don't need quoting, because they can't
have funny characters in. It's almost logical.</p>
<p>Another useful trick with <code>read</code> is to read a single character; the
`<code>-k</code>' option does this, and in fact you can stick a number immediately
after the `<code>k</code>' which specifies a number to read. Even easier, the
`<code>-q</code>' option reads a single character and returns status 0 if it was
<code>y</code> or <code>Y</code>, and status 1 otherwise; thus you can read the answer to
yes/no questions without using a parameter at all. Note, however, that
if you don't supply a parameter, the reply gets assigned in any case to
<code>$REPLY</code> if it's a scalar --- as it is with <code>-q</code> --- or <code>$reply</code> if it's
an array --- i.e. if you specify <code>-A</code>, but no parameter name. These are
more examples of the non-special parameters which the shell uses --- it
sets <code>$REPLY</code> or <code>$reply</code>, but only in the same way you would set them;
there are no side-effects.</p>
<p>Like <code>print</code>, <code>read</code> has a <code>-r</code> flag for raw mode. However, this just
has one effect for <code>read</code>: without it, a <code>\</code> at the end of the line
specifies that the next line is a continuation of the current one (you
can do this when you're typing at the terminal). With it, <code>\</code> is not
treated specially.</p>
<p>Finally, a more sophisticated note about word-splitting. I said that,
when you are reading to many parameters or an array, the word is split
on whitespace. In fact the shell splits words on any of the characters
found in the (genuinely special, because it affects the shell's guts)
parameter <code>$IFS</code>, which stands for `input field separator'. By default
--- and in the vast majority of uses --- it contains space, tab, newline
and a null character (character zero: if you know that these are usually
used to mark the end of strings, you might be surprised the shell
handles these as ordinary characters, but it does, although printing
them out usually doesn't show anything). However, you can set it to any
string: enter</p>
<pre><code> fn() {
local IFS=:
read -A array
print -l $array
}
fn
</code></pre>
<p>and type</p>
<pre><code>one word:two words:three words:four
</code></pre>
<p>The shell will show you what's in the array it's read, one `word' per
line:</p>
<pre><code> one word
two words
three words
four
</code></pre>
<p>You'll see the bananas, er, words (joke for the over-thirties) have been
treated as separated by a colon, not by whitespace. Making <code>$IFS</code> local
didn't work in old versions of zsh, as with other specials; you had to
save it and restore it.</p>
<p>The <code>read</code> command in zsh doesn't let you do line editing, which some
shells do. For that, you should use the <code>vared</code> command, which runs the
line editor to edit a parameter, with the <code>-c</code> option, which allows
<code>vared</code> to create a new parameter. It also takes the option <code>-p</code> to
specify a prompt, so one of the examples above can be rewritten</p>
<pre><code> vared -c -p 'Please enter a line: ' line
</code></pre>
<p>which works rather like read but with full editing support. If you give
the option <code>-h</code> (history), you can even retrieve values from previous
command lines. It doesn't have all the formatting options of read,
however, although when reading an array (use the option <code>-a</code> with <code>-c</code>
if creating a new array) it will perform splitting.</p>
<p><strong>Other builtins to control parameters</strong></p>
<p>The remaining builtins which handle parameters can be dealt with more
swiftly.</p>
<p>The builtin <code>set</code> simply sets the special parameter which is passed as
an argument to functions or scripts, and which you access as <code>$*</code> or
<code>$@</code>, or <code>$<number></code> (Bourne-like format), or via <code>$argv</code> (csh-like
format), known however you set them as the `positional parameters':</p>
<pre><code> % set a whole load of words
% print $1
a
% print $*
a whole load of words
% print $argv[2,-2]
whole load of
</code></pre>
<p>It's exactly as if you were in a function and had called the function
with the arguments `<code>a whole load of words</code>'. Actually, set can also be
used to set shell options, either as flags, e.g. `<code>set -x</code>', or as
words after `<code>-o</code>' , e.g. `<code>set -o xtrace</code>' does the same as the
previous example. It's generally easier to use <code>setopt</code>, and the upshot
is that you need to be careful when setting arguments this way in case
they begin with a `<code>-</code>'. Putting `<code>-``-</code>' before the real arguments
fixes this.</p>
<p>One other use of <code>set</code> is to set any array, via</p>
<pre><code> set -A any_array words to assign to any_array
</code></pre>
<p>which is equivalent to (and the standard Korn shell version of)</p>
<pre><code> any_array=(words to assign to any_array)
</code></pre>
<p>One case where the <code>set</code> version is more useful is if the name of an
array itself comes from a parameter:</p>
<pre><code> arrname=myarray
set -A $arrname words to assign
</code></pre>
<p>has no easy equivalent in the other form; the left hand side of an
ordinary assignment won't expand a parameter:</p>
<pre><code> # Doesn't work; syntax error
$arrname=(words to assign)
</code></pre>
<p>This worked in old versions of zsh, but that was on the non-standard
side. The <code>eval</code> command, described below, gives another way around
this.</p>
<p>Next comes `<code>shift</code>', which simply moves an array up one element,
deleting the original first one. Without an array name, it operates on
the positional parameters. You can also give it a number to shift other
than one, before the array name.</p>
<pre><code> shift array
</code></pre>
<p>is equivalent to</p>
<pre><code> array=(${array[2,-1]})
</code></pre>
<p>(almost --- I'll leave the subtleties here for the chapter on expansion)
which picks the second to last elements of the array and assigns them
back to the original array. Note, yet again, that <code>shift</code> operates using
the <em>name</em>, not the <em>value</em> of the array, so no `<code>$</code>' should appear in
front, otherwise you get something similar to the trick I showed for
`<code>set -A</code>'.</p>
<p>Finally, <code>unset</code> unsets a parameter, and I already showed you could
unset a key/value pair of an associative array. There is one subtlety to
be mentioned here. Normally, <code>unset</code> just makes the parameter named
disappear off the face of the earth. However, if you call <code>unset</code> in a
function, its ghost lives on in the sense that any parameter you create
in the same name will be scoped as the original parameter was. Hence:</p>
<pre><code> var='global value'
fn() {
typeset var='local value'
unset var
var='what about this?'
}
fn
print $var
</code></pre>
<p>The final statement prints `<code>global value</code>': even though the local copy
of <code>$var</code> was unset, the shell remembers that it was local, so the
second <code>$var</code> in the function is also local and its value disappears at
the end of the function.</p>
<p><span id="l38"></span></p>
<h3 id="327-history-control-commands"><a class="header" href="#327-history-control-commands">3.2.7: History control commands</a></h3>
<p>The easiest way to access the shell's command history is by editing it
directly. The second easiest way is to use the `<code>!</code>'-history mechanism.
Other ways of manipulating it are based around the <code>fc</code> builtin, which
probably once stood for something (according to Oliver Kiddle, `fix
command', which is as good as anything). I talked quite a bit about it
in the last chapter, and don't really have anything to add. Just note
that the two other commands based around it are <code>history</code> and <code>r</code>.</p>
<p><span id="l39"></span></p>
<h3 id="328-job-control-and-process-control"><a class="header" href="#328-job-control-and-process-control">3.2.8: Job control and process control</a></h3>
<p>One of the major contributions of the C-shell was job control. You need
to know about foreground and background tasks, and again I introduced
these in the last chapter along with the options that control them. Here
is an introduction to the relevant builtins.</p>
<p>You start a background job in two ways. First, directly, by putting an
`<code>&</code>' after it:</p>
<pre><code> sleep 10 &
</code></pre>
<p>and secondly by starting it in the normal way (i.e. in the foreground),
then typing <code>^Z</code>, and using the <code>bg</code> command to put it in the
background. Between typing <code>^Z</code> and <code>bg</code>, the job is still there, but is
not running; it is `suspended' or `stopped' (systems use different
descriptions for the same thing), waiting for you to decide what to do
with it. In either case, the job then continues without the shell
waiting for it. It will still try and read from or write to the terminal
if that's how you started it; you need to use the shell's redirection
facilities right at the start if you want to change that, there's
nothing you can do after the job has already started.</p>
<p>By the way, `sleep' isn't a builtin. Oddly enough, you can suspend a
builtin command or sequence of commands (such as shell function) with
<code>^Z</code>, although since the shell has to continue executing your commands
as well as being suspended, it does the only thing it can do --- fork,
so that the commands you suspend are put into the background. Probably
you will only rarely do this with builtins. No other shell, so far as I
know, has this feature.</p>
<p>A job will stop if it needs to read from the terminal. You see a message
like:</p>
<pre><code> [1] + 1348 suspended (tty input) jobname and arguments
</code></pre>
<p>which means the job is suspended very much like you had just typed <code>^Z</code>.
You need to bring the job into the forground, as described below, so
that you can type something to it.</p>
<p>By the way, the key to type to suspend a command may not be <code>^Z</code>; it
usually is, but that can be changed. Run `<code>stty -a</code>' and look for what
is listed after `<code>susp =</code>' --- probably, but not necessarily, <code>^Z</code>. So
if you want to use another character --- it must be a single character;
this is handled deep in the terminal interface, not in the shell --- you
can run</p>
<pre><code> stty susp '^]'
</code></pre>
<p>or whatever. You will note from the <code>stty</code> output that various other job
control characters can be changed similarly. The <code>stty</code> command is
external and its format for both output and input can vary quite a bit
from system to system.</p>
<p>Instead of putting the command into the background, you can bring it
back to the foreground again with <code>fg</code>. This is useful for temporarily
stopping what you are doing so you can do something else. These days you
would probably do it in another window; in the old days when people
logged in from simple terminals this was even more useful. A typical
example of this is</p>
<pre><code> more file # look at file
^Z # suspend
[1] + 8592 suspended more file # message printed
... # do something else
fg %1 # resume the `more'
</code></pre>
<p>The `<code>%</code>' is the usual way of referring to jobs. The number after it is
what appeared in square brackets with the suspended message; I don't
know why the shell doesn't use the `<code>%</code>' notation there, too. You also
see that with the `continued' message when you put something into the
background, and again at the end with the `done' message which tells
you a background job is finished. The `<code>%</code>' can take other forms; the
most common is to follow it by the name of a command, such as `<code>%more</code>'
in this case. The forms <code>%+</code> and <code>%-</code> refer to the most recent and
second most recent jobs --- the `<code>+</code>' in the `suspended' message is
telling you that the <code>more</code> job could be referred to like that.</p>
<p>Most of the job control commands will actually assume you are talking
about `<code>%+</code>' if you don't give an argument, so assuming I hadn't
started any other commands in the background, I could just have put
`<code>fg</code>' at the end of the sequence of commands above. This actually cuts
both ways: <code>fg</code> is the default operation on jobs referred to with the
`<code>%</code>' notation, so just typing `<code>%1</code>' with no command name would have
worked, too.</p>
<p>You can jog your memory about what's going on with the `<code>jobs</code>'
command. It looks like a series of messages of the form beginning with
the number in square brackets; usually the jobs will either be
`running' or `suspended'. This will tell you the numbers you need.</p>
<p>One other useful thing you can do with a job is to tell the shell to
forget about it. This is only really useful if it is already running in
the background; then you can run `<code>disown</code>' with the job identifier.
It's useful for jobs you want to continue after you've logged out, as
well as jobs that have their own windows which you can therefore control
directly. With disowned jobs, the shell doesn't warn you that they are
still there when you log out. You can actually disown a background job
when you start it by putting `<code>&|</code>' or `<code>&!</code>' at the end of the line
instead of simply `<code>&</code>'. Note that if the job was suspended when you
disowned it, it will stay disowned; this is pretty pointless, so you
probably should run `<code>bg</code>' on it first.</p>
<p>The next most likely thing you want to do with a job is kill it, or
maybe suspend it when it's already in the background and you can't just
type <code>^Z</code>. This is where the <code>kill</code> builtin comes in. There's more to
this than there is to the builtins mentioned above. First, you can use
<code>kill</code> with other processes that weren't started from the current shell.
In that case, you would use a number to identify it, with no <code>%</code> ---
that's why the <code>%</code>'s were there in the other cases. Of course, you need
to find out the number; the usual way is with the <code>ps</code> command, which is
not a builtin but which appears on all UNIX-like systems. As a stupid
example, here I start a disowned process which does very little, look
for it, then kill it:</p>
<pre><code> % sleep 60 &|
% ps -f
UID PID PPID C STIME TTY TIME CMD
pws 623 614 0 22:12 pts/0 00:00:00 zsh
pws 8613 623 0 23:12 pts/0 00:00:00 sleep 60
pws 8615 623 0 23:12 pts/0 00:00:00 ps -f
% kill 8613
% ps -f
UID PID PPID C STIME TTY TIME CMD
pws 623 614 0 22:12 pts/0 00:00:00 zsh
pws 8616 623 0 23:12 pts/0 00:00:00 ps -f
</code></pre>
<p>The process has disappeared the second time I look. Notice that in the
usual lugubrious UNIX way the shell didn't bother to tell you the
process had been killed; however, it will report an error if it failed
to send it the signal. Sending it the signal is all the shell cares
about; the shell won't warn if you if the process decided it didn't want
to die when told to, so it's still a good idea to check.</p>
<p>Sometimes you want to wait for a process to exit; the <code>wait</code> builtin can
do this, and like <code>kill</code> can take a process number as well as a job
number. However, that's a bit deceptive --- you can't actually wait for
a process which wasn't started directly from the shell. Indeed, the
mechanism for waiting is all bound up with the way UNIX handles
processes; unless its parent waits for it, a process becomes a `zombie'
and hangs around until the system's foster parent, the `init' process
(always process number 1) waits for it instead. It's all a little bit
baroque, but for the shell user, wait just means you can hang on until
something you started has finished. Indeed, that's how foreground
processes work: the shell in effect uses the internal version of <code>wait</code>
to hang around until the job exits. (Well, actually that's a lie; the
system wakes it up from whatever it's doing to tell it a child has
finished, so all it has to do is doze off to wait.)</p>
<p>Furthermore, you can wait for a process even if job control isn't
running. Job control, basically anything involving those <code>%</code>'s, is only
useful when you are sitting at a terminal fiddling with commands; it
doesn't operate when you run scripts, say. Then the shell has much less
freedom in how to control its jobs, but it can still wait for a
background process, and it can still use <code>kill</code> on a process if it knows
its number. For this purpose, the shell stores the ID of the last
process started in the background in the parameter <code>$!</code>; there's
probably a good reason for the `<code>!</code>', but I don't know what it is. This
happens regardless of job control.</p>
<p><strong>Signals</strong></p>
<p>The <code>kill</code> command can do a good deal more than just kill a process.
That is the default action, which is why the command has that name. But
what it's really doing is sending a `signal' to a process. Signals are
the simplest way of communicating to another process; in fact, they are
about the only simple way if you haven't made special arrangements for
the process to read messages from you. Signal names are written like
<code>SIGINT</code>, <code>SIGTSTP</code>, <code>SIGKILL</code>; to send a particular signal to a
process, you remove the <code>SIG</code>, stick a hyphen in front, and use that as
the first argument to <code>kill</code>, e.g.:</p>
<pre><code> kill -KILL 8613
</code></pre>
<p>Some of the things you already know about are actually doing just that.
When you type <code>^C</code> to stop a process, you are actually sending it a
<code>SIGINT</code> for `interrupt', as if you had done</p>
<pre><code> kill -INT 8613
</code></pre>
<p>The usual signal sent by <code>kill</code> is not, as you might have guessed,
<code>SIGKILL</code>, but actually <code>SIGTERM</code> for `terminate'; <code>SIGKILL</code> is
stronger as the process can't block that signal, as it can with many
(we'll see how the shell can do that in a moment). It's familiar to UNIX
hackers as `<code>kill -9</code>', because all the signals also have numbers. You
can see the list of signals in zsh by doing:</p>
<pre><code> % print $signals
EXIT HUP INT QUIT ILL TRAP ABRT BUS FPE KILL USR1
SEGV USR2 PIPE ALRM TERM STKFLT CLD CONT STOP TSTP
TTIN TTOU URG XCPU XFSZ VTALRM PROF WINCH POLL PWR
UNUSED ZERR DEBUG
</code></pre>
<p>Your list will probably be different from mine; this is for Linux, and
the list is very system-specific, even though the first nine are
generally the same, and many of the others are virtually always present.
Actually, <code>SIGEXIT</code> is an invention by the shell for you to allow the
shell to do something when a function exits (see the section on `traps'
below); you can't actually use `<code>kill -EXIT</code>'. Thus <code>SIGHUP</code> is the
first real signal, and indeed that's number one, so you have to shift
the contents of <code>$signals</code> along one to get the right numbers. <code>SIGTERM</code>
and <code>SIGINT</code> usually have the same effect, stopping the process, unless
that has decided to handle the signal some other way.</p>
<p>The last two signals are bogus, too: <code>SIGZERR</code> is to allow the shell to
do something on an error (non-zero exit status), while with <code>SIGDEBUG</code>
you can do it on every command. Again, the `something' to be executed
is a `trap', as I'll discuss in a short while.</p>
<p>Typing <code>^Z</code> to suspend a process actually sends the process a <code>SIGTSTP</code>
(terminal stop, since it usually comes from the terminal), while
<code>SIGSTOP</code> is similar but usually doesn't come from a terminal. Even
restarting a process as with <code>bg</code> sends it a signal, in this case
<code>SIGCONT</code>. It seems a bit odd to signal a process to restart; why can't
the operating system just restart it when you ask? The real answer is
probably that signals provide an easy way for you to talk to the
operating system without grovelling around in the dirt too much.</p>
<p>Before I talk about how you make the shell handle signals it receives,
there is one extra oddment: the <code>suspend</code> builtin effectively sends the
shell a signal to suspend it, as if you'd typed <code>^Z</code>, though as you've
probably found by now that doesn't suspend the shell itself. It's only
useful to do this if the shell is running under some other programme,
else there's no way of restoring it and suspending is effectively the
same as exiting the shell. For this reason, the shell won't let you call
<code>suspend</code> in a login shell, because it assumes that is running as the
top level (though in the previous chapter you learnt there's actually
nothing that special about login shells; you can start one just with
`zsh -l'). If you're logged in remotely via <code>rsh</code> or <code>ssh</code>, it's
usually more convenient to use the keystrokes `<code>~^Z</code>' which those
define, rather than zsh's mechanism; they have to be at the beginning of
a line, so hit return first if necessary. This returns you to your local
terminal; you can resume the remote login with `<code>fg</code>' just like any
other programme.</p>
<p><strong>Traps</strong></p>
<p>The way of making the shell handle signals is called `traps'. There are
actually two mechanisms for this. I'll present the more standard one and
then talk about the advantages and drawbacks of the other one at the
end.</p>
<p>The standard version (shared with other shells) is via the `<code>trap</code>'
builtin. The first argument is a chunk of shell code to execute, which
obviously needs to be quoted when you pass it as an argument, and the
remaining arguments are a list of signals to handle, minus the <code>SIG</code>
prefix. So:</p>
<pre><code> trap "echo I\\'m trapped." INT
</code></pre>
<p>tells the shell what to do on <code>SIGINT</code>, i.e. <code>^C</code>. Note the extra layer
of quoting: the double quotes surround the code, so that when they are
stripped <code>trap</code> sees the chunk</p>
<pre><code> echo I\'m trapped
</code></pre>
<p>Usually the shell would abort what it was doing and return to the main
prompt when you hit <code>^C</code>. Now, however, it will simply print the message
and carry on. You can try this, for example, with</p>
<pre><code> read line
</code></pre>
<p>If you hit <code>^C</code> while it's waiting for input, you'll see the message go
up, but the shell will still wait for you to type a line.</p>
<p>A warning about this: <code>^C</code> is only trapped within the shell itself. If
you start up an external programme, it will have its own mechanism for
handling signals, and if it usually aborts on <code>^C</code> it still will. But
there's a sting in the tail: do</p>
<pre><code> cat
</code></pre>
<p>which waits for input to output again (you need to use <code>^D</code> to exit
normally). If you type <code>^C</code> here, the command will be aborted, as I said
--- but you still get the message `<code>I'm trapped</code>'. That's because the
shell is able to tell that the command got that particular signal, and
calls the trap when the <code>cat</code> exits. Not all shells do this;
furthermore, some commands which handle signals themselves won't give
the shell enough information to know that a signal arrived, and in that
case the trap won't be called. Such commands are usually the more
sophisticated things like editors or screen managers or whatever; you
just have to find out by trial and error.</p>
<p>You can also make the shell ignore the signal completely. To do this,
the first argument should be an empty string:</p>
<pre><code> trap '' INT
</code></pre>
<p>Now <code>^C</code> will have no effect, and <em>this</em> time the effect <em>is</em> passed on
directly to commands called from the shell --- try the <code>cat</code> example and
you won't be able to interrupt it; type <code>^D</code> or use the lesser known but
more powerful <code>^\</code> (control with backslash), which sends <code>SIGQUIT</code>. If
it hasn't been disabled, this will also produce a file <code>core</code>, which
contains debugging information about what the programme was doing when
it exited --- never call your own files <code>core</code>. You can trap <code>SIGQUIT</code>
too, if you want. (The shell itself usually ignores <code>SIGQUIT</code>; it's only
useful for external commands.)</p>
<p>Now the other sort of trap. I could have written for the first example:</p>
<pre><code> TRAPINT() {
print I\'m trapped.
}
</code></pre>
<p>As you can see, this is just a function: functions beginning <code>TRAP</code> are
special. However, it's a real function too; you can call it by hand with
the command `TRAPINT', and it will run perfectly happily with no funny
side effects.</p>
<p>There is a difference between the way the two types work. In the
`<code>trap</code>' sort of trap, the code is just evaluated just as if it
appeared as instructions to the shell at the point where the trap
happened. So if you were in a function, you would see the environment of
that function with its local variables; if you set a local variable with
<code>typeset</code>, it would be visible in the function just as if it were
created there.</p>
<p>However, in the function type of trap, the code is provided with its own
function environment. Now if you use <code>typeset</code> the parameter created is
local only to the trap. In most cases, that's all the difference there
is; it's up to you to decide which is more convenient. As you can see,
the function type of trap doesn't require the extra layer of quoting, so
looks a little smarter. Conveniently, the `<code>trap</code>' command on its own
lists all traps in the form of the shell code you'd need to recreate
them, and you can see which sort is which.</p>
<p>There are two cases where the difference sticks out. One is that the
function type has some extra wiring to allow you both to trap a signal,
and pretend to anyone watching that the shell didn't handle it. An
example will show this:</p>
<pre><code> TRAPINT() {
print "Signal caught, stopping anyway."
return $(( 128 + $1 ))
}
</code></pre>
<p>That second line may look as rococo as the Amalienburg, but it's meaning
is this: <code>$1</code>, the first argument to the function, is set to the number
of the signal. In this case it will be 2 because that's the standard
number for <code>SIGINT</code>. That means the arithmetic substitution <code>$((...))</code>
returns 130, the command `<code>return 130</code>' is executed, and the function
returns with status 130. Returning with non-zero status is special in
function traps: it tells the shell you want to abort the surrounding
command even though the trap was handled, and that you want the status
associated with that to be 130. It so happens that this is how UNIX
handles returns from normal traps. Without setting a trap, do</p>
<pre><code> % cat
^C
% print $?
</code></pre>
<p>and you'll see that this, too, has given the status 130, 128 plus the
value of <code>SIGINT</code>. So if you <em>do</em> have the trap set, you'll see the
message, but the command will abort --- even if it was running inside
the shell.</p>
<p>Try</p>
<pre><code> % read line
^C
</code></pre>
<p>to see that happening. If you look at the status in <code>$?</code> you'll find
it's actually 1, not 130; that's because the <code>read</code> command, when it
aborted, overrode the return value from the trap. But it does that with
an untrapped <code>^C</code>, too, so that's not really an exception to what I've
just said.</p>
<p>If you've been paying attention, you'll realise that traps set with the
<code>trap</code> builtin can't do it in quite this way, because the function they
return from would be whatever function you were in. You can see that:</p>
<pre><code> trap 'echo Returning...; return;' INT
fn() {
print In fn...
read param
print Leaving fn..
}
</code></pre>
<p>If you run <code>fn</code> and hit <code>^C</code>, the signal is trapped and the message
printed, but because of the <code>return</code>, the shell quits <code>fn</code> immediately
and you don't see the final message. If you missed out the `<code>return;</code>'
(try it), the shell would carry on with the rest of <code>fn</code> after you typed
something to <code>read</code>. Of course you can use this mechanism to leave
functions after trapping a signal; it just so happens that in this case
the mechanism with <code>TRAPINT</code> is a little closer to what untrapped
signals do and hence a little neater.</p>
<p>One final flourish of late Baroque splendour: the trap for <code>SIGEXIT</code>,
the one called when a function (or the shell itself, in fact) exits is a
bit special because in the case of exiting a function it will be called
in the environment of the calling function. So if you need to do
something like set a local variable for an enclosing function you can
have</p>
<pre><code> trap 'typeset param_in_enclosing_func=value' EXIT
</code></pre>
<p>do it for you; you couldn't do that with <code>TRAPEXIT</code> because the code
would have its own function, so that even though it would be called
after the first function exited, it wouldn't run directly in the
enclosing one but in a separate <code>TRAPEXIT</code> function. You can even set an
EXIT trap for the enclosing function by defining a nested `<code>trap .. EXIT</code>' inside that trap itself.</p>
<p>I lied, because there is one more special thing about <code>TRAPEXIT</code>: it's
always reset after you exit a function and the trap itself has been
called. Most traps just hang around until you explicitly unset them.
There is an option, <code>LOCAL_TRAPS</code>, which makes traps set inside
functions as well insulated as possible from those outside, or inside
deeper functions. In other words, the old trap is saved and then
restored when you exit the function; the scoping works pretty much like
that for <code>typeset</code>, and in the same way traps for the enclosing scope,
apart from any for <code>EXIT</code>, remain in effect inside a function unless you
explicitly override them; and, again in the same way, if you unset it
inside the function it will still be restored on exit.</p>
<p><code>LOCAL_TRAPS</code> is the fixed behaviour of some other shells. In zsh,
without the option set:</p>
<pre><code> trap 'echo Hi.' INT
fn() {
trap 'echo Bye.' INT
}
</code></pre>
<p>Calling <code>fn</code> simply replaces the trap defined outside the function with
the one defined inside while:</p>
<pre><code> trap 'echo Hi.' INT
fn() {
setopt localtraps
trap 'echo Bye.' INT
}
</code></pre>
<p>puts the original `Hi' trap back after the function exits.</p>
<p>I haven't told you how to unset a trap for good: the answer is</p>
<pre><code> trap - INT
</code></pre>
<p>As you would guess, you can use <code>unfunction</code> with function-type traps;
that will correctly remove the trap as well as deleting the function.
However, `<code>trap -</code>' works with both, so that's the recommended way.</p>
<p><strong>Limits on processes</strong></p>
<p>One other way that jobs started by the shell can be controlled is by
using limits. These are actually limits set by the operating system, but
the shell gives you a way of controlling them: the <code>limit</code> and <code>unlimit</code>
commands. Type `<code>limit</code>' on its own to see a summary. I get:</p>
<pre><code> cputime unlimited
filesize unlimited
datasize unlimited
stacksize 8MB
coredumpsize 0kB
memoryuse unlimited
maxproc 2048
descriptors 1024
memorylocked unlimited
addressspace unlimited
</code></pre>
<p>where the item on the left of each line is what is being limited, and on
the right is the value. The manual page to look at, at least on Linux is
for the function <code>getrusage</code>; that's the function the shell is calling
when you run <code>limit</code> or <code>unlimit</code>.</p>
<p>In this case, the items are:</p>
<ul>
<li><strong><code>cputime</code></strong><br />
the total CPU time used by a process</li>
<li><strong><code>filesize</code></strong><br />
maximum size of a file</li>
<li><strong><code>datasize</code></strong><br />
the maximum size of data in use by a programme</li>
<li><strong><code>stacksize</code></strong><br />
the maximum size of the stack, which is the area of memory used to
store information during function calls</li>
<li><strong><code>coredumpsize</code></strong><br />
the maximum size of a <code>core</code> file, which is an image of memory left
by a programme that crashes, allowing you to debug it with <code>gdb</code>,
<code>dbx</code>, <code>ddd</code> or some other debugger</li>
<li><strong><code>memoryuse</code></strong><br />
the maximum main memory, i.e. programme memory which is in active
use and hasn't been `swapped out' to disk</li>
<li><strong><code>maxproc</code></strong><br />
the maximum number of simultaneous processes</li>
<li><strong><code>descriptors</code></strong><br />
the maximum number of simultaneously open files (`descriptors' are
the internal mechanism for referring to an open file on UNIX-like
systems)</li>
<li><strong><code>memorylocked</code></strong><br />
the maximum amount of memory locked in (I don't know what that is,
either)</li>
<li><strong><code>addressspace</code></strong><br />
the total amount of virtual memory, i.e. any memory whether it is
main memory, or refers to somewhere on a disk, or indeed anything
else.</li>
</ul>
<p>You may well see other names; the shell decides when it is compiled what
limits are supported by the system.</p>
<p>Of those, the one I use most commonly is <code>coredumpsize</code>: sometimes when
I'm debugging a file I want a crashed programme to produce a `<code>core</code>'
files so I can run <code>gdb</code> or <code>dbx</code> on it (`<code>unlimit coredumpsize</code>'),
while other times they are just untidy (`<code>limit coredumpsize 0</code>').
Probably you would only alter any of the others if you knew there was a
problem, for example a number-crunching programme used so much memory
that the rest of the system was badly affected and you wanted to limit
<code>datasize</code> to 64 megabyte or whatever. You could write this as:</p>
<pre><code> limit datasize 64m
</code></pre>
<p>There is a distinction made between `hard' and `soft' limits. Both
have the same effect on programmes, but you can remove or reduce `soft'
limits, while only the superuser (the system administrator's login,
root) can do that to `hard' limits. Usually, therefore, <code>limit</code> and
<code>unlimit</code> manipulate soft limits; to show or set hard limits, give the
option <code>-h</code>. If I do `<code>limit -h</code>', I get the same list of limits as
above, but with <code>stacksize</code> and <code>coredumpsize</code> unlimited --- that means
I can reduce or remove the limits on those if I want, they're just set
for my own convenience.</p>
<p>Why is <code>stacksize</code> set in this way? As I said, it refers to the memory
in which the functions in programmes store variables and any other local
information. If one function calls another, it uses more memory. You can
get into a situation where functions call themselves recursively and
there is no way out until the machine runs out of memory; limiting
<code>stacksize</code> prevents this. You can actually see this with zsh itself
(probably better not to try this if you'd rather the shell you're
running didn't crash):</p>
<pre><code> % fn() { fn; }
% fn
</code></pre>
<p>defines a function which keeps calling itself. To do this, all the
functions <em>inside</em> zsh are calling themselves as well, using more and
more stack memory. Actually, zsh uses other forms of memory inside each
function and my version of zsh crashes due to exhaustion of that memory
instead. However, it depends on the system how this works out.</p>
<p><strong>Times</strong></p>
<p>One way of returning information on process resources is with the
`<code>times</code>' command. It simply shows the total CPU time used by the shell
and by the programmes called for it --- in that order, and without
description, so you need to remember. On each line, the first number is
the time spent in user space and the second is the time spent in system
space. If you're not concerned about the details of programmes the
difference is pretty irrelevant, but if you are, then the difference is
very roughly that between the time spent in the code you actually see
before you compile a programme, and the time spent in `hidden' code
where the system is doing something for you. It's not such an obvious
distinction, because many library routines, such as mathematical
functions, are run in user mode as no privileged access to internal bits
of the system is required. Typically, system time is concerned with the
details of input and output --- though even there it's not so simple,
because the C output routines <code>printf</code>, <code>puts</code>, <code>fread</code> and others have
user mode code which then calls the system routines <code>read</code>, <code>write</code> and
so on.</p>
<p>You can measure the time taken by a particular external command by
putting `<code>time</code>', in the singular this time, in front of it; this is
essentially another precommand modifier, and is a shell reserved word
rather than a builtin. This gives fairly obvious information. You can
specify the information using the <code>$TIMEFMT</code> parameter, which has its
own percent escapes, different from the ones used in prompts. It exists
partly because the shell allowed you to access all sorts of other
information about the process which ran, such as `page faults' ---
occasions when the system had to fetch a part of the programme or data
from disk because it wasn't in the main memory. However, that
disappeared because it was too much work to convert the feature to
configure itself automatically for different operating systems. It may
be time to resurrect it.</p>
<p>You can also force the time to be shown automatically by setting the
parameter <code>$REPORTTIME</code>; if a command runs for more than this many
seconds, the <code>$TIMEFMT</code> output will be shown automatically.</p>
<p><span id="l40"></span></p>
<h3 id="329-terminals-users-etc"><a class="header" href="#329-terminals-users-etc">3.2.9: Terminals, users, etc.</a></h3>
<p><strong>Watching for other users</strong></p>
<p>Although this is more associated with parameters than builtins, the
`<code>log</code>' command will tell you whether any of a group of people you want
to watch out for have logged in or out. To use this, you set the
<code>$watch</code> array parameter to a list of user names, or `<code>all</code>' for
everyone, or `<code>notme</code>' for everyone except yourself. Even if you don't
use <code>log</code>, any changes will be reported just before the shell prints a
prompt. It will be printed using the <code>$WATCHFMT</code> parameter: once again,
this takes its own set of percent escapes, listed in the <code>zshparam</code>
manual.</p>
<p><strong><code>ttyctl</code></strong></p>
<p>There is a command <code>ttyctl</code> which is designed to keep badly behaved
external commands from messing up the terminal settings. Most programmes
are careful to restore any settings they change, but there are
exceptions. After `<code>ttyctl -f</code>', the terminal is frozen; zsh will
restore the settings, no matter what an external programme does with it.
This includes deliberate attempts to change the terminal settings with
the `<code>stty</code>' command, so the default is unfrozen, `<code>ttyctl -u</code>'.</p>
<p><span id="l41"></span></p>
<h3 id="3210-syntactic-oddments"><a class="header" href="#3210-syntactic-oddments">3.2.10: Syntactic oddments</a></h3>
<p>This section collects together a few builtins which, rather than
controlling the behaviour of some feature of the shell, have some other
special effect.</p>
<p><strong>Controlling programme flow</strong></p>
<p>The four functions here are <code>exit</code>, <code>return</code>, <code>break</code>, <code>continue</code> and
<code>source</code> or <code>.</code>: they determine what the shell does next. You've met
<code>exit</code> --- leave the shell altogether --- and <code>return</code> --- leave the
current function. Be very careful not to confuse them. Calling <code>exit</code> in
a shell function is usually bad:</p>
<pre><code> % fn() { exit; }
% fn
</code></pre>
<p>This makes you entire shell session go away, not just the function. If
you write C programmes, you should be very familiar with both, although
there is one difference in this case: <code>return</code> at the top level in an
interactive shell actually does nothing, rather than leaving the shell
as you might expect. However, in a script, return outside a function
<em>does</em> cause the entire script to stop. The reason for this is that zsh
allows you to write autoloaded functions in the same form as scripts, so
that they can be used as either; this wouldn't work if <code>return</code> did
nothing when the file was run as a script. Other shells don't do this:
<code>return</code> does nothing at the top level of a script, as well as
interactively. However, other shells don't have the feature that
function definition files can be run as scripts, either.</p>
<p>The next two commands, <code>break</code> and <code>continue</code>, are to do with constructs
like `<code>if</code>'-blocks and loops, and it will be much easier if I introduce
them when I talk about those below. They will also already be familiar
to C programmers. (If you are a FORTRAN programmer, however, <code>continue</code>
is <em>not</em> the statement you are familiar with; it is instead equivalent
to <code>CYCLE</code> in FORTRAN90.)</p>
<p>The final pair of commands are <code>.</code> and <code>source</code>. They are similar to one
another and cause another file to be read as a stream of commands in the
current shell --- not as a script, for which a new shell would be
started which would finish at the end of the script. The two are
intended for running a series of commands which have some effect on the
current shell, exactly like the startup files. Indeed, it's a very
common use to have a call to one or other in a startup file; I have in
my <code>~/.zshrc</code></p>
<pre><code> [[ -f ~/.aliasrc ]] && . ~/.aliasrc
</code></pre>
<p>which tests if the file <code>~/.aliasrc</code> exists, and if so runs the commands
in it; they are treated exactly as if they had appeared directly at that
point in <code>.zshrc</code>.</p>
<p>Note that your <code>$path</code> is used to find the file to read from; this is a
little surprising if you think of this as like a script being run, since
zsh doesn't search for a script, it uses the name exactly as you gave
it. In particular, if you don't have `<code>.</code>' in your <code>$path</code> and you use
the form `<code>.</code>' rather than `<code>source</code>' you will need to say explicitly
when you want to source a file in the current directory:</p>
<pre><code> . ./file
</code></pre>
<p>otherwise it won't be found.</p>
<p>It's a little bit like running a function, with the file as the function
body. Indeed, the shell will set the positional parameters <code>$*</code> in just
the same way. However, there's a crucial difference: there is no local
parameter scope. Any variables in a sourced file, as in one of the
startup files, are in the same scope as the point from which it was
started. You can, therefore, source a file from inside a function and
have the parameters in the sourced file local, but normally the only way
of having parameters only for use in a sourced file is to unset them
when you are finished.</p>
<p>The fact that both <code>.</code> and <code>source</code> exist is historical: the former
comes from the Bourne shell, and the latter from the C shell, which
seems deliberately to have done everything differently. The point noted
above, that source always searches the current directory (and searches
it first), is the only difference.</p>
<p><strong>Re-evaluating an expression</strong></p>
<p>Sometimes it's very useful to take a string and run it as if it were a
set of shell commands. This is what <code>eval</code> does. More precisely, it
sticks the arguments together with spaces and calls them. In the case of
something like</p>
<pre><code> eval print Hello.
</code></pre>
<p>this isn't very useful; that's no different from a simple</p>
<pre><code> print Hello.
</code></pre>
<p>The difference comes when what's on the command line has something to be
expanded, like a parameter:</p>
<pre><code> param='print Hello.'
eval $param
</code></pre>
<p>Here, the <code>$param</code> is expanded just as it would be for a normal command.
Then <code>eval</code> gets the string `<code>print Hello.</code>' and executes it as shell
command line. Everything --- really everything --- that the shell would
normally do to execute a command line is done again; in effect, it's run
as a little function, except that no local context for parameters is
created. If this sounds familiar, that's because it's exactly the way
traps defined in the form</p>
<pre><code> trap 'print Hello.' EXIT
</code></pre>
<p>are called. This is one simple way out of the hole you can sometimes get
yourself into when you have a parameter which contains the name of
another parameter, instead of some data, and you want to get your hands
on the data:</p>
<pre><code> # somewhere above...
origdata='I am data.'
# but all you know about is
paramname=origdata
# so to extract the data you can do...
eval data=\$$paramname
</code></pre>
<p>Now <code>$data</code> contains the value you want. Make sure you understand the
series of expansions going on: this sort of thing can get very
confusing. First the command line is expanded just as normal. This turns
the argument to <code>eval</code> into `<code>data=$origdata</code>'. The `<code>$</code>' that's still
there was quoted by a backslash; the backslash was stripped and the
`<code>$</code>' left; the <code>$paramname</code> was evaluated completely separately ---
quoted characters like the <code>\$</code> don't have any effect on expansions ---
to give <code>origdata</code>. Eval calls the new line `<code>data=$origdata</code>' as a
command in its own right, with the now obvious effect. If you're even
slightly confused, the best thing to do is simply to quote everything
you don't want to be immediately expanded:</p>
<pre><code> eval 'data=$'$paramname
</code></pre>
<p>or even</p>
<pre><code> eval 'data=${'$paramname'}'
</code></pre>
<p>may perhaps make your intentions more obvious.</p>
<p>It's possible when you're starting out to confuse `<code>eval</code>' with the
<code>`...`</code> and <code>$(...)</code> commands, which also take the command in the
middle `<code>...</code>' and evaluate it as a command line. However, these two
(they're identical except for the syntax) then insert the output of that
command back into the command line, while <code>eval</code> does no such thing; it
has no effect at all on where input and output go. Conversely, the two
forms of command substitution don't do an extra level of expansion.
Compare:</p>
<pre><code> % foo='print bar'
% eval $foo
bar
</code></pre>
<p>with</p>
<pre><code> % foo='print bar'
% echo $($foo)
zsh: command not found: print bar
</code></pre>
<p>The <code>$</code>(<em>...</em>) substitution took <code>$foo</code> as the command line. As you are
now painfully aware, zsh doesn't split scalar parameters, so this was
turned into the single word `<code>print bar</code>', which isn't a command. The
blank line is `<code>echo</code>' printing the empty result of the failed
substitution.</p>
<p><span id="l42"></span></p>
<h3 id="3211-more-precommand-modifiers-exec-noglob"><a class="header" href="#3211-more-precommand-modifiers-exec-noglob">3.2.11: More precommand modifiers: <code>exec</code>, <code>noglob</code></a></h3>
<p>Sometimes you want to run a command <em>instead</em> of the shell. This
sometimes happens when you write a shell script to process the arguments
to an external command, or set parameters for it, then call that
command. For example:</p>
<pre><code> export MOZILLA_HOME=/usr/local/netscape
netscape "$@"
</code></pre>
<p>Run as a script, this sets an environment variable, then starts
<code>netscape</code>. However, as always the shell waits for the command to
finish. That's rather wasteful here, since there's nothing more for the
shell to do; you'd rather it simply magically turned into the <code>netscape</code>
command. You can actually do this:</p>
<pre><code> export MOZILLA_HOME=/usr/local/netscape
exec netscape "$@"
</code></pre>
<p>`<code>exec</code>' tells the shell that it doesn't need to wait; it can just make
the command to run replace the shell. So this only uses a single
process.</p>
<p>Normally, you should be careful not to use <code>exec</code> interactively, since
normally you don't want the shell to go away. One legitimate use is to
replace the current zsh with a brand new one if (say) you've set a whole
load of options you don't like and want to restore the ones you usually
have on startup:</p>
<pre><code> exec zsh
</code></pre>
<p>Or you may have the bad taste to start a completely different shell
altogether. Conversely, a good piece of news about <code>exec</code> is that it is
common to all shells, so you can use it from another shell to start zsh
in the way I've just shown.</p>
<p>Like `<code>command</code>' and `<code>builtin</code>', `<code>exec</code>' is a `precommand
modifier' in that it alters the way a command line is interpreted.
Here's one more:</p>
<pre><code> noglob print *
</code></pre>
<p>If you've remembered what `glob' means, this is all fairly obvious. It
instructs the shell not to turn the `<code>*</code>' into a list of all the files
in the directory, but instead to let well alone. You can do this by
quoting the `<code>*</code>', of course; often <code>noglob</code> is used as part of an
alias to set up commands where you never need filename generation and
don't want to have to bother quoting everything. However, note that
<code>noglob</code> has no effect on any other type of expansion: parameter
expansion and backquote (<code>`....`</code>) expansion, for example, happen as
normal; the only thing that doesn't is turning patterns into a list of
matching files. So it doesn't take away the necessity of knowing the
rules of shell expansion. If you need that, the best thing to do is to
use <code>read</code> or <code>vared</code> (see below) to read a line into a parameter, which
you pass to your function:</p>
<pre><code> read -r param
print $param
</code></pre>
<p>The <code>-r</code> makes sure <code>$param</code> is the unadulterated input.</p>
<p><span id="l43"></span></p>
<h3 id="3212-testing-things"><a class="header" href="#3212-testing-things">3.2.12: Testing things</a></h3>
<p>I told you in the last chapter that the right way to write tests in zsh
was using the `<code>[[ ... ]]</code>' form, and why. So you can ignore the two
builtins `<code>test</code>' and `<code>[</code>', even though they're the ones that
resemble the Bourne shell. You can safely write</p>
<pre><code> if [[ $foo = '' ]]; then
print The parameter foo is empty. O, misery me.
fi
</code></pre>
<p>or</p>
<pre><code> if [[ -z $foo ]]; then
print Alack and alas, foo still has nothing in it.
fi
</code></pre>
<p>instead of monstrosities like</p>
<pre><code> if test x$foo != x; then
echo The emptiness of foo. Yet are we not all empty\?
fi
</code></pre>
<p>because even if <code>$foo</code> does expand to an empty string, which is what is
implied if the tests are true, `<code>[[ ... ]]</code>' remembers there was
something there and gets the syntax right. Rather than a builtin, this
is actually a reserved word --- in fact it has to be, to be
syntactically special --- but you probably aren't too bothered about the
difference.</p>
<p>There are two sorts of tests, both shown above: those with three
arguments, and those with two. The three-argument forms all have some
comparison in the middle; in addition to `<code>=</code>' (or `<code>==</code>', which means
the same here, and which according to the manual page we should be
using, though none of us does), there are `<code>!=</code>' (not equal), `<code><</code>',
`<code>></code>', `<code><=</code>' and `<code>>=</code>'. All these do <em>string</em> comparisons, i.e.
they compare the sort order of the strings.</p>
<p>Since there are better ways of sorting things in zsh, the `<code>=</code>' and
`<code>!=</code>' forms are by far the most common. Actually, they do something a
bit more than string comparison: the expression on the right can be a
pattern. The patterns understood are just the same as for matching
filenames, except that `<code>/</code>' isn't special, so it can be matched by a
`<code>*</code>'. Note that, because `<code>=</code>' and `<code>!=</code>' are treated specially by
the shell, you shouldn't quote the patterns: you might think that unless
you do, they'll be turned into file names, but they won't. So</p>
<pre><code> if [[ biryani = b* ]]; then
print Word begins with a b.
fi
</code></pre>
<p>works. If you'd written <code>'b*'</code>, including the quotes, it wouldn't have
been treated as a pattern; it would have tested for a string which was
exactly the two letters `<code>b*</code>' and nothing else. Pattern matching like
this can be very powerful. If you've done any Bourne shell programming,
you may remember the only way to use patterns there was via the
`<code>case</code>' construction: that's still in zsh (see below), and uses the
same sort of patterns, but the test form shown above is often more
useful.</p>
<p>Then there are other three-argument tests which do numeric comparison.
Rather oddly, these use letters rather than mathematical symbols:
`<code>-eq</code>', `<code>-lt</code>' and `<code>-le</code>' compare if two numbers are equal, less
than, or less than or equal, to one another. You can guess what `<code>-gt</code>'
and `<code>-ge</code>' do. Note this is the other way round to Perl, which much
more logically uses `<code>==</code>' to test for equality of numbers (not `<code>=</code>',
since that's always an assignment operator in Perl) and `<code>eq</code>' (minus
the minus) to test for equality of strings. Unfortunately we're now
stuck with it this way round. If you are only comparing numbers, it's
better to use the `<code>(( ... ))</code>' expression, because that has a proper
understanding of arithmetic. However,</p>
<pre><code> if [[ $number -gt 3 ]]; then
print Wow, that\'s big
fi
</code></pre>
<p>and</p>
<pre><code> if (( $number > 3 )); then
print Wow, that\'s STILL big
fi
</code></pre>
<p>are essentially equivalent. In the second case, the status is zero
(true) if the number in the expression was non-zero (sorry if I'm
confusing you again) and vice versa. This means that</p>
<pre><code> if (( 3 )); then
print It seems that 3 is non-zero, Watson.
fi
</code></pre>
<p>is a perfectly valid test. As in C, the test operators in arithmetic
return 1 for true and 0 for false, i.e. `<code>$number > 3</code>' is 1 if
<code>$number</code> is greater than 3 and 0 otherwise; the inversion to shell
logic, zero for true, only occurs at the final step when the expression
has been completely evaluated and the `<code>(( ... ))</code>' command returns. At
least with `<code>[[ ... ]]</code>' you don't need to worry about the extra
negation; you can simply think in logical terms (although that's hard
enough for a lot of people).</p>
<p>Finally, there are a few other odd comparisons in the three-argument
form:</p>
<pre><code> if [[ file1 -nt file2 ]]; then
print file1 is newer than file2
fi
</code></pre>
<p>does the test implied by the example; there is also `<code>-ot</code>' to test for
an older file, and there is also the little-used `<code>-ef</code>' which tests
for an `equivalent file', meaning that they refer to the same file ---
in other words, are linked; this can be a hard or a symbolic link, and
in the second case it doesn't matter which of the two is the symbolic
link. (If you were paying attention above, you'll know it can't possibly
matter in the first case.)</p>
<p>In addition to these tests, which are pretty recognisable from most
programming languages --- although you'll just have to remember that the
`<code>=</code>' family compares strings and not numbers --- there are another set
which are largely peculiar to UNIXy scripting languages. These are all
in the form of a hyphen followed by a letter as the test, which always
takes a single argument. I showed one: `-z $var' tests whether
`<code>$var</code>' has zero length. It's opposite is `-n $var' which tests for
non-zero length. Perhaps this is as good a time as any to point out that
the arguments to these commands can be any single word expression, not
just variables or filenames. You are quite at liberty to test</p>
<pre><code> if [[ -z "$var is sqrt(`print bibble`)" ]]; then
print Flying pig detected.
fi
</code></pre>
<p>if you like. In fact, the tests are so eager to make sure that they only
have a one word argument that they will treat things like arrays, which
usually return a whole set of words, as if they were in double quotes,
joining the bits with spaces:</p>
<pre><code> array=(two words)
if [[ $array = 'two words' ]]; then
print "The array \$array is OK. O, joy."
fi
</code></pre>
<p>Apart from `<code>-z</code>' and `<code>-n</code>', most of the two-argument tests are to do
with files: `<code>-e</code>' tests that the file named next exists, whatever type
of file it is (it might be a directory or something weirder); `<code>-f</code>'
tests if it exists and is a regular file (so it isn't a directory or
anything weird this time); `<code>-x</code>' tests whether you can execute it.
There are all sorts of others which are listed in the manual page for
various properties of files. Then there are a couple of others: ``-o</p>
<option>`' you've met and tests whether the option is set, and \``-t
<fd>`' tests whether the file descriptor is attached to a terminal. A
file descriptor is a number which for the shell must be between 0 and 9
inclusive (others may exist, but you can't access them directly); 0 is
the standard input, 1 the standard output, and 2 the channel on which
errors are usually printed. Hence \``[[ -t 0 ]]`' tests whether the
input is coming from a terminal.
<p>There are only four other things making up tests. `<code>&&</code>' and `<code>||</code>'
mean logical `and' and `or', `<code>!</code>' negates the effect of a test, and
parentheses `<code>( ... )</code>' can be used to surround a set of tests which
are to be treated as one. These are all essentially the same as in C. So</p>
<pre><code> if [[ 3 -gt 2 && ( me > you || ! -z bah ) ]]; then
print will I, won\'t I...
fi
</code></pre>
<p>will, because 3 is numerically greater than 2; the expression in
parentheses is evaluated and though `me' actually comes before `you'
in the alphabet, so the first test fails, `<code>-z bah</code>' is false because
you gave it a non-empty string, and hence `<code>! -z bah</code>' is true. So both
sides of the `<code>&&</code>' are true and the test succeeds.</p>
<p><span id="l44"></span></p>
<h3 id="3213-handling-options-to-functions-and-scripts"><a class="header" href="#3213-handling-options-to-functions-and-scripts">3.2.13: Handling options to functions and scripts</a></h3>
<p>It's often convenient to have your own functions and scripts process
single-letter options in the way a lot of builtin commands (as well as a
great many other UNIX-style commands) do. The shell provides a builtin
for this called `<code>getopts</code>'. This should always be called in some kind
of loop, usually a `<code>while</code>' loop. It's easiest to explain by example.</p>
<pre><code> testopts() {
# $opt will hold the current option
local opt
while getopts ab: opt; do
# loop continues till options finished
# see which pattern $opt matches...
case $opt in
(a)
print Option a set
;;
(b)
print Option b set to $OPTARG
;;
# matches a question mark
# (and nothing else, see text)
(\?)
print Bad option, aborting.
return 1
;;
esac
done
(( OPTIND > 1 )) && shift $(( OPTIND - 1 ))
print Remaining arguments are: $*
}
</code></pre>
<p>There's quite a lot here if you're new to shell programming. You might
want to read the stuff on structures like <code>while</code> and <code>case</code> below and
then come back and look at this. First let's see what it does.</p>
<pre><code> % testopts -b foo -a -- args
Option b set to foo
Option a set
Remaining arguments are: args
</code></pre>
<p>Here's what's happening. `<code>getopts ab: opt</code>' is the argument to the
`<code>while</code>'. That means that the <code>getopts</code> gets run; if it succeeds
(returns status zero), then the loop after the `<code>do</code>' is run. When
that's finished, the <code>getopts</code> command is run again, and so on until it
fails (returns a non-zero status). It will do that when there are no
more options left on the command line. So the loop processes the options
one by one. Each time through, the number of the next argument to look
at is left in the parameter <code>$OPTIND</code>, so this gradually increases;
that's how <code>getopts</code> knows how far it has got.</p>
<p>The first argument to the <code>getopts</code> is `<code>ab:</code>'. That means `<code>a</code>' is an
option which doesn't take an argument, while `<code>b</code>' is an argument which
takes a single argument, signified by the colon after it. You can have
any number of single-letter (or even digit) arguments, which are
case-sensitive; for example `<code>ab:c:ABC:</code>' are six different options,
three with arguments. If the option found has an argument, that is
stored in the parameter <code>$OPTARG</code>; <code>getopts</code> then increments <code>$OPTIND</code>
by however much is necessary, which may be 2 or just 1 since `<code>-b foo</code>'
and `<code>-bfoo</code>' are both valid ways of giving the argument.</p>
<p>If an option is found, we use the `<code>case</code>' mechanism to find out what
it was. The idea of this is simple, even if the syntax has the look of
an 18th-century French chateau: the argument `<code>$opt</code>' is tested against
all of the patterns in the `<code>pattern</code>)' lines until one matches, and
the commands are executed until the next `<code>;;</code>'. It's the shell's
equivalent of C's `<code>switch</code>'. In this example, we just print out what
the <code>getopts</code> brought in. Note the last line, which is called if <code>$opt</code>
is a question mark --- it's quoted because `<code>?</code>' on its own can stand
for any single character. This is how <code>getopts</code> signals an unknown
option. If you try it, you'll see that <code>getopts</code> prints its own error
message, so ours was unnecessary: you can turn the former off by putting
a colon right at the start of the list of options, making it `<code>:ab:</code>'
here.</p>
<p>Actually, having this last pattern as an <em>un</em>quoted `<code>?</code>' isn't such a
bad idea. Suppose you add a letter to the list that <code>getopts</code> should
handle and forget to add a corresponding item in the <code>case</code> list for it.
If the last item matches any character, you will get the behaviour for
an unhandled option, which is probably the best thing to do. Otherwise
nothing in the <code>case</code> list will match, the shell will sail blithely on
to the next call to <code>getopts</code>, and when you try to use the function with
the new option you will be left wondering quite what happened to it.</p>
<p>The last piece of the <code>getopts</code> jigsaw is the next line, which tests if
<code>$OPTIND</code> is larger than 1, i.e. an option was found and <code>$OPTIND</code> was
advanced --- it is automatically set to 1 at the start of every function
or script. If it was, the `<code>shift</code>' builtin with a numeric argument,
but no array name, moves the positional parameters, i.e. the function's
arguments, to shift away the options that have been processed. The
<code>print</code> in the next line shows you that only the remaining non-option
arguments are left. You don't need to do that --- you can just start
using the remaining arguments from <code>$argv[$OPTIND]</code> on --- but it's a
pretty good way of doing it.</p>
<p>In the call, I showed a line with `<code>-``-</code>' in it: that's the standard
way of telling <code>getopts</code> that the options are finished; even if later
words start with a <code>-</code>, they are not treated as options. However,
<code>getopts</code> stops anyway when it reaches a word not beginning with `<code>-</code>',
so that wasn't necessary here. But it works anyway.</p>
<p>You can do all of what <code>getopts</code> does without <em>that</em> much difficulty
with a few extra lines of shell programming, of course. The best
argument for using <code>getopts</code> is probably that it allows you to group
single-letter options, e.g. `<code>-abc</code>' is equivalent to `<code>-a -b -c</code>' if
none of them was defined to have an argument. In this case, <code>getopts</code>
has to remember the position <em>inside</em> the word on the command line for
you to call it next, since the `<code>a</code>' `<code>b</code>' and `<code>c</code>' still appear on
different calls. Rather unsatisfactorily, this is hidden inside the
shell (as it is in other shells --- we haven't fixed <em>all</em> of everybody
else's bad design decisions); you can't get at it or reset it without
altering <code>$OPTIND</code>. But if you read the small print at the top of the
guide, you'll find I carefully avoided saying everything was
satisfactory.</p>
<p>While we're at it, why do blocks starting with `<code>if</code>' and `<code>then</code>' end
with `<code>fi</code>', and blocks starting with `<code>case</code>' end with `<code>esac</code>',
while those starting with `<code>while</code>' and `<code>do</code>' end with `<code>done</code>', not
`<code>elihw</code>' (perfectly pronounceable in Welsh, after all) or `<code>od</code>'?
Don't ask me.</p>
<p><span id="l45"></span></p>
<h3 id="3214-random-file-control-things"><a class="header" href="#3214-random-file-control-things">3.2.14: Random file control things</a></h3>
<p>We're now down into the odds and ends. If you know UNIX at all, you will
already be familiar with the <code>umask</code> command and its effect on file
creation, but as it is a builtin I will describe it here. Create a file
and look at it:</p>
<pre><code> % touch tmpfile
% ls -l tmpfile
-rw-r--r-- 1 pws pws 0 Jul 19 21:19 tmpfile
</code></pre>
<p>(I've shortened the output line for the TeX version of this document.)
You'll see that the permissions are read for everyone, write-only for
the owner. How did the command (<code>touch</code>, not a builtin, creates an empty
file if there was none, or simply updates the modification time of an
existing file) know what permissions to set?</p>
<pre><code> % umask
022
% umask 077
% rm tmpfile; touch tmpfile
% ls -l tmpfile
-rw------- 1 pws pws 0 Jul 19 21:22 tmpfile
</code></pre>
<p><code>umask</code> was how it knew. It gives an octal number corresponding to the
permissions which should <em>not</em> be given to a newly created file (only
newly created files are affected; operations on existing files don't
involve <code>umask</code>). Each digit is made up of a 4 for read, 2 for write, 1
for executed, in the same order that <code>ls</code> shows for permissions: user,
then group, then everyone else. (On this Linux/GNU-based system, like
many others, users have their own groups, so both are called `<code>pws</code>'.)
So my original `022' specified that everything should be allowed except
writing for group and other, while `077' disallowed any operation by
group and other. These are the two most common settings, although here
`002' or `007' would be useful because of the way groups are specific
to users, making it easier to grant permission to specific other users
to write in my directories. (Except there aren't any other users.)</p>
<p>You can also use <code>chmod</code>-like permission changing expressions in
<code>umask</code>. So</p>
<pre><code> % umask go+rx
</code></pre>
<p>would restore group and other permissions for reading and executing,
hence returning the mask to 022. Note that because it is <em>adding</em>
permissions, just like <code>chmod</code> does, it is <em>removing</em> numbers from the
umask.</p>
<p>You might have wondered about execute permissions, since `<code>touch</code>'
didn't give any, even where it was allowed by <code>umask</code>. That's because
only operations which create executable programmes, such as running a
compiler and linker, set that bit; the normal way of opening a new file
--- internally, the UNIX <code>open</code> function, with the <code>O_CREAT</code> flag set
--- doesn't touch it. For the same reason, if you create shell scripts
which you want to be able to execute by typing the name, you have to
make them executable yourself:</p>
<pre><code> % chmod +x myscript
</code></pre>
<p>and, indeed, you can think of <code>chmod</code> as <code>umask</code>'s companion for files
which already exist. It doesn't need to be a builtin, because the files
you are operating on are external to <code>zsh</code>; <code>umask</code>, on the other hand,
operates when you create a file from within <code>zsh</code> or any child process,
so needs to be a builtin. The fact that it's inherited means you can set
<code>umask</code> before you start an editor, and files created by that editor
will reflect the permissions.</p>
<p>Note that the value set by <code>umask</code> is also inherited and used by
<code>chmod</code>. In the example of <code>chmod</code> I gave, I didn't see <em>which</em> type of
execute permission to add; <code>chmod</code> looks at my <code>umask</code> and decides based
on that --- in other words, with 022, everybody would be allowed to
execute <code>myscript</code>, while with 077, only I would, because of the 1's in
the number: (0+0+0)+(4+2+1)+(4+2+1). Of course, you can be explicit with
chmod and say `<code>chmod u+x myscript</code>' and so on.</p>
<p>Something else that may or may not be obvious: if you run a script by
passing it as an argument to the shell,</p>
<pre><code> % zsh myscript
</code></pre>
<p>what matters is <em>read</em> permission. That's what the shell's doing to the
script to find the commands, after all. Execute permission applies when
the system (or, in some cases, including zsh, the parent shell where you
typed `<code>myscript</code>') has to decide whether to find a shell to run the
script by itself.</p>
<p><span id="l46"></span></p>
<h3 id="3215-dont-watch-this-space-watch-some-other"><a class="header" href="#3215-dont-watch-this-space-watch-some-other">3.2.15: Don't watch this space, watch some other</a></h3>
<p>Finally for builtins, some things which really belong elsewhere. There
are three commands you use to control the shell's editor. These will be
described in <a href="zshguide04.html#zle">chapter 4</a>, where I talk all about
the editor.</p>
<p>The <code>bindkey</code> command allows you to attach a set of keystrokes to a
command. It understands an abbreviated notation for the keystrokes.</p>
<pre><code> % bindkey '^Xc' copy-prev-word
</code></pre>
<p>This binds the keystrokes consisting of <code>Ctrl</code> held down with <code>x</code>, then
<code>c</code>, to the command which copies the previous word on the line to the
current position. The commands are listed in the <code>zshzle</code> manual page.
<code>bindkey</code> can also do things with keymaps, which are a complete set of
mappings between keys and commands like the one I showed.</p>
<p>The <code>vared</code> command is an extremely useful builtin for editing a shell
variable. Usually much the easiest way to change <code>$path</code> (or <code>$PS1</code>, or
whatever) is to run `<code>vared path</code>': note the lack of a `<code>$</code>', since
otherwise you would be editing whatever <code>$path</code> was expanded to. This is
because very often you want to leave most of what's there and just
change the odd character or word. Otherwise, you would end up doing this
with ordinary parameter substitutions, which are a lot more complicated
and error prone. Editing a parameter is exactly like editing a command
line, except without the prompt at the start.</p>
<p>Finally, there is the <code>zle</code> command. This is the most mysterious, as it
offers a fairly low-level interface to the line editor; you use it to
define your own editing commands. So I'll leave this alone for now.</p>
<p><span id="l47"></span></p>
<h3 id="3216-and-also"><a class="header" href="#3216-and-also">3.2.16: And also</a></h3>
<p>There is one more standard builtin that I haven't covered: <code>zmodload</code>,
which allows you to manipulate add-on packages for zsh. Many extras are
supplied with the shell which aren't normally loaded to keep down the
use of memory and to avoid having too many rarely used builtins, etc.,
getting in the way. In the last chapter I will talk about some of these.
To be more honest, a lot of the stuff in between actually uses these
addons, generically referred to as modules --- the line editor, zle, is
itself a separate module, though heavily dependent on the main shell ---
and you've probably forgotten I mentioned above using `<code>zmodload zsh/mathfunc</code>' to load mathematical functions.</p>
<p><span id="l48"></span></p>
<h2 id="33-functions"><a class="header" href="#33-functions">3.3: Functions</a></h2>
<p>Now it's time to look at functions in more detail. The various issues to
be discussed are: loading functions, handling parameters, compiling
functions, and repairing bike tyres when the rubber solution won't stick
to the surface. Unfortunately I've already used so much space that I'll
have to skip the last issue, however topical it might be for me at the
moment.</p>
<p><span id="l49"></span></p>
<h3 id="331-loading-functions"><a class="header" href="#331-loading-functions">3.3.1: Loading functions</a></h3>
<p>Well, you know what happens now. You can define functions on the command
line:</p>
<pre><code> fn() {
print I am a function
}
</code></pre>
<p>which you call under the name `<code>fn</code>'. As you type, the shell knows that
it's in the middle of a function definition, and prompts you until you
get to the closing brace.</p>
<p>Alternatively, and much more normally, you put a file called <code>fn</code>
somewhere in a directory listed in the <code>$fpath</code> array. At this point,
you need to be familiar with the <code>KSH_AUTOLOAD</code> option described in the
last chapter. From now on, I'm just going to assume your autoloadable
function files contain just the body of the function, i.e.
<code>KSH_AUTOLOAD</code> is not set. Then the file <code>fn</code> would contain:</p>
<pre><code> print I am a function
</code></pre>
<p>and nothing else.</p>
<p>Recent versions of zsh, since <code>3.1.6</code>, set up <code>$fpath</code> for you. It
contains two parts, although the second may have multiple directories.
The first is, typically, <code>/usr/local/share/zsh/site-functions</code>, although
the prefix may vary. This is empty unless your system administrator has
put something in it, which is what it's there for.</p>
<p>The remaining part may be either a single directory such as
<code>/usr/local/share/zsh/3.1.9/functions</code>, or a whole set of directories
starting with that path. It simply depends whether the person installing
zsh wanted to keep all the functions in the same directory, or have them
sorted according to what they do. These directories are full of
functions. However, none of the functions is autoloaded automatically,
so unless you specifically put `<code>autoload ...</code>' in a startup file, the
shell won't actually take any notice of them. As you'll see, part of the
path is the shell version. This makes it very easy to keep multiple
versions of zsh with functions which use features that may be different
between the two versions. By the way, if these directories don't exist,
you should check <code>$fpath</code> to see if they are in some other location, and
if you can't find any correspondence between what's in <code>$fpath</code> and
what's on the disk even when you start the shell with <code>zsh -f</code> to
suppress loading of startup files, complain to the system administrator:
he or she has either not installed them properly, or has made
<code>/etc/zshenv</code> stomp on <code>$fpath</code>, both of which are thoroughly evil
things to do. (<code>/etc/zshrc</code>, <code>/etc/zprofile</code> and <code>/etc/zlogin</code> shouldn't
stomp on <code>$fpath</code> either, of course. In fact, they shouldn't do very
much; that's up to the user.)</p>
<p>One point about <code>autoload</code> is the `<code>-U</code>' option. This turns off the use
of any aliases you have defined when the function is actually loaded ---
the flag is remembered with the name of the function for future
reference, rather than being interpreted immediately by the <code>autoload</code>
command. Since aliases can pretty much redefine any command into any
other, and are usually interpreted while a function is being defined or
loaded, you can see that without this flag there is fair scope for
complete havoc.</p>
<pre><code> alias ls='echo Your ls command has been requisitioned.'
lsroot() {
ls -l /
}
lsroot
</code></pre>
<p>That's not what the function writer intended. (Yes, I know it actually
<em>is</em>, because I wrote it to show the problem, but that's not what I
<em>meant</em>.) So <code>-U</code> is recommended for all standard functions, where you
have no easy way of telling quite what could be run inside.</p>
<p>Recently, the functions for the new completion system (described in
<a href="zshguide06.html#comp">chapter 6</a>) have been changing the fastest. They
either begin with <code>comp</code> or an underscore, `<code>_</code>'. If the <code>functions</code>
directory is subdivided, most of the subdirectories refer to this. There
are various other classes of functions distributed with the shell:</p>
<ul>
<li>
<p>Functions beginning <code>zf</code> are associated with zftp, a builtin system
for FTP transfers. Traditional FTP clients, ones which don't use a
graphical interface, tend to be based around a set of commands on a
command line --- exactly what zsh is good at. This also makes it
very easy to write macros for FTP jobs --- they're just shell
functions. This is described in the final chapter along with other
modules. It's based around a single builtin, <code>zftp</code>, which is loaded
from the module <code>zsh/zftp</code>.</p>
</li>
<li>
<p>Functions beginning <code>prompt</code>, which may be in the <code>Prompts</code>
subdirectory, are part of a `prompt themes' system which makes it
easy for you to switch between preexisting prompts. You load it with
`<code>autoload -U promptinit; promptinit</code>'. Then `<code>prompt -h</code>' will
tell you what to do next. If you have new completion loaded (with
`<code>autoload -U compinit; compinit</code>', what else) the arguments to
`<code>prompt</code>' can be listed with <code>^D</code> and completed with a TAB; they
are various sorts of prompt which you may or may not like.</p>
</li>
<li>
<p>Functions with long names and hyphens, like <code>predict-on</code> and
<code>incremental-complete-word</code>. These are editor functions; you use
them with</p>
<pre><code> zle -N predict-on
bindkey <keystroke> predict-on
</code></pre>
<p>Here, the <code>predict-on</code> function automatically looks back in the
history list for matching lines as you type. You should also bind
<code>predict-off</code>, which is loaded when <code>predict-on</code> is first called.
<code>incremental-complete-word</code> is a fairly simple attempt at showing
possible completions for the current word as you type; it could do
with improving.</p>
</li>
<li>
<p>Everything else; these may be in the <code>Misc</code> subdirectory. These are
a very mixed bag which you should read to see if you like any. One
of the most useful is <code>zed</code>, which allows you to edit a small file
(it's really just a simple front-end to <code>vared</code>). The <code>run-help</code>
file shows you the sort of thing you might want to define for use
with the <code>\eh</code> (<code>run-help</code>) keystroke. <code>is-at-least</code> is a function
for finding out if the version of the shell running is recent
enough, assuming you know what version you need for a given feature.
Several of the other functions refer to the old completion system
--- which you won't need, since you will be reading <a href="zshguide06.html#comp">chapter
6</a> and using the new completion system, of
course.</p>
</li>
</ul>
<p>If you have your own functions --- and if you use zsh a lot, you almost
certainly will eventually --- it's a good idea to add your own personal
directory to the front of <code>$fpath</code>, so that everything there takes
precedence over the standard functions. That allows you to override a
completion function very easily, just by copying it and editing it. I
tend to do something like this in my <code>.zshenv</code>:</p>
<pre><code> [[ $fpath = *pws* ]] || fpath=(~pws/bin/fns $fpath)
</code></pre>
<p>to protect against the possibility that the directory I want to add is
already there, in case I source that startup file again, and there are
other similar ways. (You may well find your own account isn't called
<code>pws</code>, however.)</p>
<p>Chances are you will always want your own functions to be autoloaded.
There is an easy way of doing this: put it just after the line I showed
above:</p>
<pre><code> autoload ${fpath[1]}/*(:t)
</code></pre>
<p>The <code>${fpath[1]}/*</code> expands to all the files in the directory at the
head of the <code>$fpath</code> array. The <code>(:t)</code> is a `glob modifier': applied to
a filename generation pattern, it takes the tail (basename) of all the
files in the list. These are exactly the names of the functions you want
to autoload. It's up to you whether you want the <code>-U</code> argument here.</p>
<p><span id="l50"></span></p>
<h3 id="332-function-parameters"><a class="header" href="#332-function-parameters">3.3.2: Function parameters</a></h3>
<p>I covered local parameters in some detail when I talked about <code>typeset</code>,
so I won't talk about that here. I didn't mention the other parameters
which are effectively local to a function, the ones that pass down the
arguments to the function, so here is more detail. They work pretty much
identically in scripts.</p>
<p>There are basically two forms. There is the form inherited from Bourne
shell via Korn shell, with the typically uninformative names: <code>$#</code>,
<code>$*</code>, <code>$@</code> and the numerical parameters <code>$1</code> etc. --- as high a number
as the shell will parse is allowed, not just single digits. Then there
is the form inherited from the C shell: <code>$ARGC</code> and <code>$argv</code>. I'll mainly
use the Bourne shell versions, which are far more commonly used, and
come back to some oddities of the C shell variety at the end.</p>
<p><code>$#</code> tells you how many arguments were passed to the function, while
<code>$*</code> gives those arguments as an array. This was the only array
available in the Bourne shell, otherwise there would probably have been
a more obvious way of doing it. To get the size and the number of
elements of the array you don't use <code>${#*}</code> and <code>${*[1]}</code> etc. (well,
you usually don't --- zsh is typically permissive here), you use <code>$1</code>,
<code>$2</code>. Despite the syntax, these are rather like ordinary array elements;
if you refer to one off the end, you will get an empty string, but no
error, unless you have the option <code>NO_UNSET</code> set. It is this not-quite
array which gets shifted if you use the <code>shift</code> builtin without an
argument: the old <code>$1</code> disappears, the old <code>$2</code> becomes <code>$1</code>, and so on,
while <code>$#</code> is reduced by one. If there were no arguments (<code>$#</code> was
zero), nothing happens.</p>
<p>The form <code>$@</code> is very similar to <code>$*</code>, and you can use it in place of
that in most contexts. There is one place where they differ. Let's
define a function which prints the number of its arguments, then the
arguments.</p>
<pre><code> args() {
print $# $*
}
</code></pre>
<p>Now some arguments. We'll do this for the current shell --- it's a
slightly odd idea, that you can set the arguments for what's already
running, or that an interactive shell has arguments at all, but
nonetheless it's true:</p>
<pre><code> set arguments to the shell
print $*
</code></pre>
<p>sets <code>$*</code> and hence prints the message `<code>arguments to the shell</code>'. We
now pass <em>these</em> arguments on to the function in two different ways:</p>
<pre><code> args $*
args $@
</code></pre>
<p>This outputs</p>
<pre><code> 4 arguments to the shell
4 arguments to the shell
</code></pre>
<p>-- no surprises so far. Remember, too, that zsh doesn't split words on
spaces unless you ask it too. So:</p>
<pre><code> % set 'one word'
% args $*
1 one word
% args $@
1 one word
</code></pre>
<p>Now here's the difference:</p>
<pre><code> % set two words
% args "$*"
1 two words
% args "$@"
2 two words
</code></pre>
<p>In quotes, <code>"$*"</code> behaves as a normal array, joining the words with
spaces. However, <code>"$@"</code> doesn't --- it still behaves as if it was
unquoted. You can't see from the arguments themselves in this case, but
you can from the digit giving the number of arguments the function has.</p>
<p>This probably seems pretty silly. Why quote something to have it behave
like an unquoted array? The original answer lies back in Bourne shell
syntax, and relates to the vexed question of word splitting. Suppose we
turn on Bourne shell behaviour, and try the example of a word with
spaces again:</p>
<pre><code> % setopt shwordsplit
% set 'one word'
% args $*
2 one word
% args $@
2 one word
% args "$*"
1 one word
% args "$@"
1 one word
</code></pre>
<p>Aha! <em>This</em> time <code>"$@"</code> kept the single word with the space intact. In
other words, <code>"$@"</code> was a slightly primitive mechanism for suppressing
splitting of words, while allowing the splitting of arrays into
elements. In zsh, you would probably quite often use <code>$*</code>, not <code>"$@"</code>,
safe in the knowledge that nothing was split until you asked it to be;
and if you wanted it split, you would use the special form of
substitution <code>${=*}</code> which does that:</p>
<pre><code> % unsetopt shwordsplit
% args $*
1 one word
% args ${=*}
2 one word
</code></pre>
<p>(I can't tell you why the `<code>=</code>' was chosen for this purpose, except
that it consists of two split lines, or in an assignment it splits two
things, or something.) This works with any parameter, whether scalar or
array, quoted or unquoted.</p>
<p>However, that's actually not quite the whole story. There are times when
the shell removes arguments, because there's nothing there:</p>
<pre><code> % set hello '' there
% args $*
2 hello there
</code></pre>
<p>The second element of the array was empty, as if you'd typed</p>
<pre><code> 2=
</code></pre>
<p>--- yes, you can assign to the individual positional parameters
directly, instead of using <code>set</code>. When the array was expanded on the
command line, the empty element was simply missed out altogether. The
same happens with all empty variables, including scalars:</p>
<pre><code> % empty=
% args $empty
0
</code></pre>
<p>But there are times when you don't want that, any more than you want
word splitting --- you want <em>all</em> arguments passed just as you gave
them. This is another side effect of the <code>"$@"</code> form.</p>
<pre><code> % args "$@"
3 hello there
</code></pre>
<p>Here, the empty element was passed in as well. That's why you often find
<code>"$@"</code> being used in zsh when wordsplitting is already turned off.</p>
<p>Another note: why does the following not work like the example with
<code>$*</code>?</p>
<pre><code> % args hello '' there
3 hello there
</code></pre>
<p>The quotes were kept here. Why? The reason is that the shell doesn't
elide an argument if there were quotes, even if the result was empty:
instead, it provides an empty string. So this empty string was passed as
the second argument. Look back at:</p>
<pre><code> set hello '' there
</code></pre>
<p>Although you probably didn't think about it at the time, the same thing
was happening here. Only with the <code>'``'</code> did we get an empty string
assigned to <code>$2</code>; later, this was missed out when passing <code>$*</code> to the
function. The same difference occurs with scalars:</p>
<pre><code> % args $empty
0
% args "$empty"
1
</code></pre>
<p>The <code>$empty</code> expanded to an empty string in each case. In the first case
it was unquoted and was removed; this is like passing an empty part of
<code>$*</code> to a command. In the second case, the quotes stopped that from
being removed completely; this is similar to setting part of <code>$*</code> to an
empty string using <code>''</code>.</p>
<p>That's all thoroughly confusing the first time round. Here's a table to
try and make it a bit clearer.</p>
<pre><code> | Number of arguments
| if $* contains...
| (two words)
Expression Word | 'one word'
on line splitting? | empty string
--------------------------------------------------
$* n | 2 1 0
$@ n | 2 1 0
"$*" n | 1 1 1
"$@" n | 2 1 1
|
$* y | 2 2 0
$@ y | 2 2 0
"$*" y | 1 1 1
"$@" y | 2 1 1
|
${=*} n | 2 2 0
${=@} n | 2 2 0
"${=*}" n | 2 2 1
"${=@}" n | 2 2 1
</code></pre>
<p>On the left is shown the expression to be passed to the function, and in
the three right hand columns the number of arguments the function will
get if the positional parameters are set to an array of two words, a
single word with a space in the middle, or a single word which is an
empty string (the effect of `<code>set -- '``'</code>') respectively. The second
column shows whether word splitting is in effect, i.e. whether the
<code>SH_WORD_SPLIT</code> option is set. The first four lines show the normal zsh
behaviour; the second four show the normal sh/ksh behaviour, with word
splitting turned on --- only the case where a word has a space in it
changes, and then only when no quotes are supplied. The final four show
what happens when you use the `<code>${=..}</code>' method to turn on word
splitting, for convenience: that's particularly simple, since it always
forces words to be split, even inside quotation marks.</p>
<p>I would recommend that anyone not wedded to the Bourne shell behaviour
use the top set as standard: in particular, `<code>$*</code>' for normal array
behaviour with removal of empty items, `<code>"$@"</code>' for normal array
behaviour with empty items left as empty items, and `<code>"$*"</code>' for
turning arrays into single strings. If you need word-splitting, you
should use `<code>${=*}</code>' or `<code>"${=@}"</code>' for splitting with/without removal
of empty items (obviously there's no counterpart to the quoted-array
behaviour here). Then keep <code>SH_WORD_SPLIT</code> turned off. If you are wedded
to the Bourne shell behaviour, you're on your own.</p>
<p><strong>It's a bug</strong></p>
<p>There's a bug in handling of the the form <code>${1+"$@"}</code>. This looks rather
an arcane and unlikely combination, but actually it is commonly used to
get around a bug in some versions of the Bourne shell (which is not in
zsh): that <code>"$@"</code> generates a single empty argument if there are no
arguments. The form shown tests whether there is a first argument, and
if so substitutes <code>"$@"</code>, else it doesn't substitute anything, avoiding
the bug.</p>
<p>Unfortunately, in zsh, when <code>shwordsplit</code> is set --- which is the time
you usually run across attempts like this to standardise the way
different shells work --- this will actually cause too much
word-splitting. The way the shell is written at the moment, the embedded
<code>"$@"</code> will force extra splitting on spaces inside the arguments. So if
the first argument is `<code>one word</code>', and <code>shwordsplit</code> is set,
<code>${1+"$@"}</code> produces <em>two</em> words `<code>one</code>' and `<code>word</code>'.</p>
<p>Oliver Kiddle spotted a way of getting round this which has been adapted
for use in the GNU autoconf package: in your initialisation code, have</p>
<pre><code> [ x$ZSH_VERSION != x ] && alias -g '${1+"$@"}'='"$@"'
</code></pre>
<p>This uses a global alias to turn <code>${1+"$@"}</code> wherever it occurs as a
single word into <code>"$@"</code> which doesn't have the problem. Aliasing occurs
so early in processing that the fact that most of the characters have a
special meaning to the shell is irrelevant; the shell behaves as if it
read in <code>"$@"</code>. The only catch is that for this to work the script or
function must use <em>exactly</em> the character string <code>${1+"$@"}</code>, with no
leading or trailing word characters (whitespace, obviously, or
characters which terminate parsing such as `<code>;</code>' are all right). Some
day, we may fix the underlying bug, but it's not very easy with the way
the parameter substitution code is written at the moment.</p>
<p><strong>Parameters inherited from csh</strong></p>
<p>The final matter is the C-shell syntax. There are two extra variables
but, luckily, there is not much extra in the way of complexity. <code>$ARGC</code>
is essentially identical to <code>$#</code>, and <code>$argv</code> corresponds to <code>$*</code>, but
is a real array this time, so instead of <code>$1</code> you have <code>${argv[1]}</code> and
so on. They use the convention that scalars used by the shell are all
uppercase, while arrays are all lowercase. This feature is probably the
only reason anyone would need these variants. For example,
<code>${argv[2,-1]}</code> means all arguments from the second to the last,
inclusive: negative indices count from the end, and a comma indicates a
slice of an array, so that <code>${argv[1,-1]}</code> is always the same as the
full array. Otherwise, my advice would be to stick with the Bourne shell
variants, however cryptic they may look at first sight, for the usual
reason that zsh isn't really like the C shell and if you pretend it is,
you will come a cropper sooner or later.</p>
<p>It looks like you're missing <code>"$@"</code>, but actually you can do that with
<code>"${argv[@]}"</code>. This, like negative indices and slices, works with all
arrays.</p>
<p>There's one slight oddity with <code>$ARGC</code> and <code>$argv</code>, which isn't really a
deliberate feature of the shell at all, but just in case you run into
it: although the values in them are of course local to functions, the
variables <code>$ARGC</code> and <code>$argv</code> <em>themselves</em> are actually treated like
global variables. That means if you apply a <code>typeset -g</code> command to
them, it will affect the behaviour of <code>$ARGC</code> and <code>$argv</code> in all
functions, even though they have different values. It's probably not a
good idea to rely on this behaviour.</p>
<p>nusubsect(Arguments to all commands work the same)</p>
<p>I've been a little tricky here, because I've been talking about two
levels of functions at once: <code>$*</code> and friends as set in the current
function, or even at the top level, as well as how they are passed down
to commands such as my <code>args</code> function. Of course, in the second case
the same behaviour applies to all commands, not just functions. What I
mean is, in</p>
<pre><code> fn() {
cat $*
cat "$*"
}
</code></pre>
<p>the `<code>cat</code>' command will see the differences in behaviour between the
two calls just as <code>args</code> would. That should be obvious.</p>
<p><strong>It's not a bug</strong></p>
<p>Let me finally mention again a feature I noted in passing:</p>
<pre><code> 1='first argument'
</code></pre>
<p>sets the first command argument for the current shell or function,
independently of any others. People sometimes complain that</p>
<pre><code> 1000000='millionth argument'
</code></pre>
<p>suddenly makes the shell use a lot more memory. That's not a bug at all:
you've asked the shell to set the millionth element of an array, but not
any others, so the shell creates an array a million elements long with
the first 999,999 empty, except for any arguments which were already
set. It's not surprising this takes up a lot of memory.</p>
<p><span id="l51"></span></p>
<h3 id="333-compiling-functions"><a class="header" href="#333-compiling-functions">3.3.3: Compiling functions</a></h3>
<p>Since version 3.1.7, it has been possible to compile functions to their
internal format. It doesn't make the functions run any faster, it just
reduces their loading time; the shell just has to bring the function
into memory, then it `runs it as it does any other function. On many
modern computers, therefore, you don't gain a great deal from this. I
have to admit I don't use it, but there are other definite advantages.</p>
<p>Note that when I say `compiled' I don't mean the way a C compiler, say,
would take a file and turn it into the executable code which the
processor understands; here, it's simply the format that the shell
happens to use internally --- it's useless without a suitable version of
zsh to run it. Also, it's no use thinking you can hide your code from
prying eyes this way, like you can to some extent with an ordinary
compiler (disassembling anything non-trivial from scratch being a
time-consuming job): first of all, ordinary command lines appear inside
the compiled files, except in slightly processed form, and secondly
running `<code>functions</code>' on a compiled function which has been loaded will
show you just as much as it would if the function had been loaded
normally.</p>
<p>One other advantage is that you can create `digest' files, which are
sets of functions stored in a single file. If you often use a large
fraction of those files, or they are small, or you like the function
itself to appear when you run `functions' rather than a message saying
it hasn't been loaded, then this works well. In fact, you can compile
all the functions in a single directory in one go. You might think this
uses a lot of memory, but often zsh will simply `memory map' the file,
which means rather than reserving extra main memory for it and reading
it in --- the obvious way of reading files --- it will tell the
operating system to make the file available as if it were memory, and
the system will bring it into memory piece by piece, `paging' the file
as it is needed. This is a very efficient way of doing it. Actually, zsh
supports both this method and the obvious method of simply reading in
the file (as long as your operating system does); this is described
later on.</p>
<p>A little extra, in case you're interested: if you read in a file
normally, the system will usually reserve space on a disk for it, the
`swap', and do paging from there. So in this case you still get the
saving of main memory --- this is standard in all modern operating
systems. However, it's not <em>as</em> efficient: first of all, you had to read
the file in in the first place. Secondly it eats up swap space, which is
usually a fixed amount of disk, although if you've got enough main
memory, the system probably won't bother allocating swap. Thirdly ---
this is probably the clincher for standard zsh functions on a large
system --- if the file is directly mapped read-only, as it is in this
case, the system only needs one place in main memory, plus the single
original file on disk, to keep the function, which is very much more
efficient. With the other method, you would get multiple copies in both
main memory and (where necessary) swap. This is how the system treats
directly executable programmes like the shell itself --- the data is
specific to each process, but the programme itself can be shared because
it doesn't need to be altered when it's running.</p>
<p>Here's a simple example.</p>
<pre><code> % echo 'echo hello, world' >hw
% zcompile hw
% ls
hw hw.zwc
% rm hw
% fpath=(. $fpath)
% autoload hw
% hw
hello, world
</code></pre>
<p>We created a simple `hello, world' function, and compiled it. This
produces a file called `<code>hw.zwc</code>'. The extension stands for `Z-shell
Word Code', because it's based on the format of words (integers longer
than a single byte) used internally by the shell. Then we made sure the
current directory was in our <code>$fpath</code>, and autoloaded the function,
which ran as expected. We deleted the original file for demonstration
purposes, but as long as the `<code>.zwc</code>' file is newer, that will be used,
so you don't need to remove the originals in normal use. In fact, you
shouldn't, because you will lose any comments and formatting information
in it; you can regenerate the function itself with the `<code>functions</code>'
command (try it here), but the shell only remembers the information
actually needed to run the commands. Note that the function was in the
zsh autoload format, not the ksh one, in this case (but see below).</p>
<p><strong>And there's more</strong></p>
<p>Now some bells and whistles. Remember the <code>KSH_AUTOLOAD</code> thing? When you
compile a function, you can specify which format --- native zsh or ksh
emulation --- will be used for loading it next time, by using the option
<code>-k</code> or <code>-z</code>, instead of the default, which is to examine the option (as
would happen if you were autoloading directly from the file). Then you
don't need to worry about that option. So, for example, you could
compile all the standard zsh functions using `<code>zcompile -z</code>' and save
people the trouble of making sure they are autoloaded correctly.</p>
<p>You can also specify that aliases shouldn't be expanded when the files
are compiled by using <code>-U</code>: this has roughly the same effect as saying
<code>autoload -U</code>, since when the shell comes to load a compiled file, it
will never expand aliases, because the internal format assumes that all
processing of that kind has already been done. The difference in this
case is if you <em>don't</em> specify <code>-U</code>: then the aliases found when you
compile the file, not when you load the function from it, will be used.</p>
<p>Now digest files. Here's one convenient way of doing it.</p>
<pre><code> % ls ~/tmp/fns
hw1 hw2
% fpath=(~/tmp/fns $fpath)
% cd ~/tmp
% zcompile fns fns/*
% ls
fns fns.zwc
</code></pre>
<p>We've made a directory to put functions in, <code>~/tmp/fns</code>, and stuck some
random files in it. The <code>zcompile</code> command, this time, was given several
arguments: a filename to use for the compiled functions, and then a list
of functions to compile into it. The new file, <code>fns.zwc</code>, sits in the
same directory where the directory <code>fns</code>, found in <code>$fpath</code>, is. The
shell will actually search the digest file instead of the directory.
More precisely, it will search both, and see which is the more recent,
and use that as the function. So now</p>
<pre><code> % autoload hw1
% hw1
echo hello, first world
</code></pre>
<p>You can test what's in the digest file with:</p>
<pre><code> % zcompile -t fns
zwc file (read) for zsh-3.1.9-dev-3
fns/hw1
fns/hw2
</code></pre>
<p>Note that the names appear as you gave them on the command line, i.e.
with <code>fns/</code> in front. Only the basenames are important for autoloading
functions. The note `<code>(read)</code>' in the first line means that zsh has
marked the functions to be read into the shell, rather than memory
mapped as discussed above; this is easier for small functions,
particularly if you are liable to remove or alter a file which is
mapped, which will confuse the shell. It usually decides which method to
use based on size; you can force memory mapping by giving the <code>-M</code>
option. Memory mapping doesn't work on all systems (currently including
Cygwin).</p>
<p>I showed this for compiling files, but you can actually tell the shell
to output compiled functions --- in other words, it will look along
<code>$fpath</code> and compile the functions you specify. I find compiling files
easier, when I do it at all, since then I can use patterns to find them
as I did above. But if you want to do it the other way, you should note
two other options: <code>-a</code> will compile files by looking along <code>$fpath</code>,
while <code>-c</code> will output any functions already loaded by the shell (you
can combine the two to use either). The former is recommended, because
then you don't lose any information which was present in the autoload
file, but not in the function stored in memory ---- this is what would
happen if the file defined some extra widgets (in the non-technical
sense) which weren't part of the function called subsequently.</p>
<p>If you're perfectly happy with the shell <em>only</em> searching a digest file,
and not comparing the datestamp with files in the directory, you can put
that directly into your <code>$fpath</code>, i.e. <code>~/tmp/fns.zwc</code> in this case.
Then you can get rid of the original directory, or archive it somewhere
for reuse.</p>
<p>You can compile scripts, too. Since these are in the same format as a
zsh autoload file, you don't need to do anything different from
compiling a single function. You then run (say) <code>script.zwc</code> by typing
`<code>zsh script</code>' --- note that you should omit the <code>.zwc</code>, as zsh decides
if there's a compiled version of a script by explicitly appending the
suffix. What's more, you can run it using `<code>.</code>' or `<code>source</code>' in just
the same way (`<code>. script</code>') --- this means you can compile your startup
files if you find they take too long to run through; the shell will spot
a <code>~/.zshrc.zwc</code> as it would any other sourceable file. It doesn't make
much sense to use the memory mapping method in this case, since once
you've sourced the files you never want to run them again, so you might
as well specify `<code>zcompile -R</code>' to use the reading (non-memory-mapping)
method explicitly.</p>
<p>If you ever look inside a <code>.zwc</code> file, you will see that the information
is actually included twice. That's because systems differ about the
order in which numbers are stored: some have the least significant byte
first (notably Intel and some versions of Mips) and some the most
significant (notably SPARC and Cambridge Consultants' XAP processor,
which is notable here mainly because I spend my working hours
programming for it --- you can't run zsh on it). Since zsh uses integers
a great deal in the compiled code, it saves them in both possible orders
for ease of use. Why not just save it for the machine where you compiled
it? Then you wouldn't be able to share the files across a heterogeneous
network --- or even worse, if you made a distribution of compiled files,
they would work on some machines, and not on others. Think how Emacs
users would complain if the <code>.elc</code> files that arrived weren't the right
ones. (Worse, think how the vi users would laugh.) The shell never reads
or maps in the version it doesn't use, however; only extra disk space is
used.</p>
<p><strong>A little -Xtra help</strong></p>
<p>There are two final autoloading issues you might want to know about. In
versions of zsh since 3.1.7, you will see that when you run <code>functions</code>
on a function which is marked for autoload but hasn't yet been loaded,
you get:</p>
<pre><code>afunctionmarkedforautoloadwhichhasntbeenloaded () {
# undefined
builtin autoload -XU
}
</code></pre>
<p>The `<code># undefined</code>' is just printed to alert you that this was a
function marked as autoloadable by the <code>autoload</code> command: you can tell,
because it's the only time <code>functions</code> will emit a comment (though there
might be other `<code>#</code>' characters around). What's interesting is the
<code>autoload</code> command with the <code>-X</code> option. That option means `Mark me for
autoloading and run me straight away'. You can actually put it in a
function yourself, and it will have the same effect as running
`<code>autoload</code>' on a not-yet-existent function. Obviously, the <code>autoload</code>
command will disappear as soon as you do run it, to be replaced by the
real contents. If you put this inside a file to be autoloaded, the shell
will complain --- the alternative is rather more unpalatable.</p>
<p>Note also the <code>-U</code> option was set in that example: that simply means
that I used <code>autoload</code> with the <code>-U</code> option when I originally told the
shell to autoload the function.</p>
<p>There's another option, <code>+X</code>, the complete opposite of <code>-X</code>. This one
can <em>only</em> be used with autoload outside the function you're loading,
just as <code>-X</code> was only meaningful inside. It means `load the file
immediately, but don't run it', so it's a more active (or, as they say
nowadays, since they like unnecessarily long words, proactive) form of
<code>autoload</code>. It's useful if you want to be able to run the <code>functions</code>
command to see the function, but don't want to run the function itself.</p>
<p><strong>Special functions</strong></p>
<p>I'm in danger of simply quoting the manual, but there are various
functions with a special meaning to the shell (apart from <code>TRAP...</code>
functions, which I've already covered). That is, the functions
themselves are perfectly normal, but the shell will run them
automatically on certain occasions if they happen to exist, and silently
skip them if they don't.</p>
<p>The two most frequently used are <code>chpwd</code> and <code>precmd</code>. The former is
called whenever the directory changes, either via <code>cd</code>, or <code>pushd</code>, or
an <code>AUTO_CD</code> --- you could turn the first two into functions, and avoid
needing <code>chpwd</code> but not the last. Here's how to force an xterm, or a
similar windowing terminal, to put the current directory into the title
bar.</p>
<pre><code> chpwd() {
[[ -t 1 ]] || return
case $TERM in
(sun-cmd) print -Pn "\e]l%~\e\\"
;;
(*xterm*|rxvt|(dt|k|E)term) print -Pn "\e]2;%~\a"
;;
esac
}
</code></pre>
<p>The first line tests that standard output is really a terminal --- you
don't want to print the string in the middle of a script which is
directing its output to a file. Then we look to see if we have a
<code>sun-cmd</code> terminal, which has its own <em>sui generis</em> sequence for putting
a string into the title bar, or something which recognises xterm escape
sequences. In either case, the special sequences (a bit like termcap
sequences as discussed for <code>echotc</code>) are interpreted by the terminal,
and instead of being printed out cause it to put the string in the
middle into the title bar. The string here is `<code>%~</code>': I added the <code>-P</code>
option to <code>print</code> so it would expand prompt escapes. I could just have
used <code>$PWD</code>, but this way has the useful effect of shortening your home
directory, or any other named directory, into <code>~</code>-notation, which is a
bit more readable. Of course, you can put other stuff there if you like,
or, if you're really sophisticated, put in a parameter <code>$HEADER</code> and
define that elsewhere.</p>
<p>If programmes other than the shell alter what appears in the xterm title
bar, you might consider changing that <code>chwpd</code> function to <code>precmd</code>. The
function <code>precmd</code> is called just before every prompt; in this case it
will restore the title line after every command has run. Some people
make the mistake of using it to set up a prompt, but there are enough
ways of getting varying information into a fixed prompt string that you
shouldn't do that unless you have <em>very</em> odd things in your prompt. It's
a big nuisance having to redefine <code>precmd</code> to alter your prompt ---
especially if you don't know it's there, since then your prompt
apparently magically returns to the same format when you change it.
There are some good reasons for using <code>precmd</code>, too, but most of them
are fairly specialised. For example, on one system I use it to check if
there is new input from a programme which is sending data to the shell
asynchronously, and if so printing it out onto the terminal. This is
pretty much what happens with job control notification if you don't have
the <code>NOTIFY</code> option set.</p>
<p>The name <code>precmd</code> is a bit of a misnomer: <code>preprompt</code> would have been
better. It usurps the name more logically applied to the function
actually called <code>preexec</code>, which is run after you finished editing a
command line, but just before the line is executed. <code>preexec</code> has one
additional feature: the line about to be executed is passed down as an
argument. You can't alter what's going to be executed by editing the
parameter, however: that has been suggested as an upgrade, but it would
make it rather easy to get the shell into a state where you can't
execute any commands because <code>preexec</code> always messes them up. It's
better, where possible, to write function front-ends to specific
commands you want to handle specially. For example, here's my <code>ls</code>
function:</p>
<pre><code> local ls
if [[ -n $LS_COLORS ]]; then
ls=(ls --color=auto)
else
ls=(ls -F)
fi
command $ls $*
</code></pre>
<p>This handles GNU and non-GNU versions of ls. If <code>$LS_COLORS</code> is set, it
assumes we are using GNU ls, and hence colouring (or colorizing, in
geekspeak) is available. Otherwise, it uses the standard option <code>-F</code> to
show directories and links with a special symbol. Then it uses <code>command</code>
to run the real <code>ls</code> --- this is a key thing to remember any time you
use a function front-end to a command. I could have done this another
way: test in my initialisation files which version of <code>ls</code> I was using,
then alias <code>ls</code> to one of the two forms. But I didn't.</p>
<p>Apart from the trap functions, there is one remaining special function.
It is <code>periodic</code>, which is executed before a prompt, like <code>precmd</code>, but
only every now and then, in fact every <code>$PERIOD</code> seconds; it's up to you
to set <code>$PERIOD</code> when you defined <code>periodic</code>. If <code>$PERIOD</code> isn't set, or
is zero, nothing happens. Don't get <code>$PERIOD</code> confused with <code>$SECONDS</code>,
which just counts up from 0 when the shell starts.</p>
<p><span id="l52"></span></p>
<h2 id="34-aliases"><a class="header" href="#34-aliases">3.4: Aliases</a></h2>
<p>Aliases are much simpler than functions. In the C shell and its
derivatives, there are no functions, so aliases take their place and can
have arguments, which involve expressions rather like those which
extract elements of previous history lines with `<code>!</code>'. Zsh's aliases,
like ksh's, don't take arguments; you have to use functions for that.
However, there are things aliases can do which functions can't, so
sometimes you end up using both, for example</p>
<pre><code> zfget() {
# function to retrieve a file by FTP,
# using globbing on the remote host
}
alias zfget='noglob zfget'
</code></pre>
<p>The function here does the hard work; this is a function from the zftp
function suite, supplied with the shell, which retrieves a file or set
of files from another machine. The function allows patterns, so you can
retrieve an entire directory with `<code>zfget *</code>'. However, you need to
avoid the `<code>*</code>' being expanded into the set of files in the current
directory on the machine you're logged into; this is where the alias
comes in, supplying the `<code>noglob</code>' in front of the function. There's no
way of doing this with the function alone; by the time the function is
called, the `<code>*</code>' would already have been expanded. Of course you could
quote it, but that's what we're trying to avoid. This is a common reason
for using the alias/function combination.</p>
<p>Remember to include the `<code>=</code>' in alias definition, necessary in zsh,
unlike csh and friends. If you do:</p>
<pre><code> alias zfget noglob zfget
</code></pre>
<p>they are treated as a list of aliases. Since none has the `<code>=</code>' and a
definition, the shell thinks you want to list the definitions of the
listed words; I get the output</p>
<pre><code> zfget='noglob zfget'
zfget='noglob zfget'
</code></pre>
<p>since <code>zfget</code> was aliased as before, but <code>noglob</code> wasn't aliased and was
skipped, although the failed alias lookup caused status 1 to be
returned. Remember that the <code>alias</code> command takes as many arguments as
you like; any with `<code>=</code>' is a definition, any without is a request to
print the current definition.</p>
<p>Aliases can in fact be allowed to expand to almost anything the shell
understands, not just sets of words. That's because the text retrieved
from the alias is put back into the input, and reread more or less as if
you'd typed it. That means you can get away with strange combinations
like</p>
<pre><code> alias tripe="echo foo | sed 's/foo/bar/' |"
tripe cat
</code></pre>
<p>which is interpreted exactly the same way as</p>
<pre><code> echo foo | sed 's/foo/bar/' | cat
</code></pre>
<p>where the word `<code>foo</code>' is sent to the stream editor, which alters it to
`<code>bar</code>' (`<code>s/old/new/</code>' is <code>sed</code>'s syntax for a substitution), and
passes it on to `<code>cat</code>', which simply dumps the output. It's useless,
of course, but it does show what can lurk behind an apparently simple
command if it happens to be an alias. It is usually not a good idea to
do this, due to the potential confusion.</p>
<p>As the manual entry explains, you can prevent an alias from being
expanded by quoting it. This isn't like quoting any other expansion,
though; there's no particular important character which has to be
interpreted literally to stop the expansion. The point is that because
aliases are expanded early on in processing of the command line, looking
up an alias is done on a string without quotes removed. So if you have
an alias `<code>drivel</code>', none of the strings `<code>\drivel</code>', `<code>'d'rivel</code>',
or `<code>drivel""</code>' will be expanded as the alias: they all would have the
same effect as proper commands, after the quotes are removed, but as
aliases they appear different. The manual entry also notes that you can
actually make aliases for any of these special forms, e.g. `<code>alias '\drivel'=...</code>' (note the quotes, since you need the backslash to be
passed down to the alias command). You would need a pretty good reason
to do so.</p>
<p>Although my `<code>tripe</code>' example was silly, you know from the existence of
`precommand modifiers' that it's sometimes useful to have a special
command which precedes a command line, like <code>noglob</code> or the non-shell
command <code>nice</code>. Since they have commands following, you would probably
expect aliases to be expanded there, too. But this doesn't work:</p>
<pre><code> % alias foo='echo an alias for foo'
% noglob foo
zsh: command not found: foo
</code></pre>
<p>because the <code>foo</code> wasn't in command position. The way round this is to
use a special feature: aliases whose definitions end in a space force
the next word along to be looked up as a possible alias, too:</p>
<pre><code> % alias noglob='noglob '
% noglob foo
an alias for foo
</code></pre>
<p>which is useful for any command which can take a command line after it.
This also shows another feature of aliases: unlike functions, they
remember that you have already called an alias of a particular name, and
don't look it up again. So the `<code>noglob</code>' which comes from expanding
the alias is not treated as an alias, but as the ordinary precommand
modifier.</p>
<p>You may be a little mystified about this difference. A simple answer is
that it's useful that way. It's sometimes useful for functions to call
themselves; for example if you are handling a directory hierarchy in one
go you might get a function to examine a directory, do something for
every ordinary file, and for every directory file call itself with the
new directory name tacked on. Aliases are too simple for this to be a
useful feature. Another answer is that it's particularly easy to mark
aliases as being `in use' while they are being expanded, because it
happens while the strings inside them are being examined, before any
commands are called, where things start to get complicated.</p>
<p>Lastly, there are `global aliases'. If aliases can get you into a lot
of trouble, global aliases can get you into a lot of a lot of trouble.
They are defined with the option <code>-g</code> and are expanded not just in
command position, but anywhere on the command line.</p>
<pre><code> alias -g L='| less'
echo foo L
</code></pre>
<p>This turns into `<code>echo foo | less</code>'. It's a neat trick if you don't
mind your command lines having only a minimal amount to do with what is
actually executed.</p>
<p>I already pointed out that alias lookups are done so early that aliases
are expanded when you define functions:</p>
<pre><code> % alias hello='echo I have been expanded'
% fn() {
function> hello
function> }
% which fn
fn () {
echo I have been expanded
}
</code></pre>
<p>You can't stop this when typing in functions directly, except by quoting
part of the name you type. When autoloading, the <code>-U</code> option is
available, and recommended for use with any non-trivial function.</p>
<p>A brief word about that `<code>function></code>' which appears to prompt you while
you are editing a function; I mentioned this in the previous chapter but
here I want to be clearer about what's going on. While you are being
prompted like that, the shell is not actually executing the commands you
are typing in. Only when it is satisfied that it has a complete set of
commands will it go away and execute them (in this case, defining the
function). That means that it won't always spot errors until right at
the end. Luckily, zsh has multi-line editing, so if you got it wrong you
should just be able to hit up-arrow and edit what you typed; hitting
return will execute the whole thing in one go. If you have redefined
<code>$PS2</code> (or <code>$PROMPT2</code>), or you have an old version of the shell, you may
not see the full prompt, but you will usually see something ending in
`<code>></code>' which means the same.</p>
<p><span id="l53"></span></p>
<h2 id="35-command-summary"><a class="header" href="#35-command-summary">3.5: Command summary</a></h2>
<p>As a reminder, the shell looks up commands in this order:</p>
<ul>
<li>aliases, which will immediately be interpreted again as texts for
commands, possible even other aliases; they can be deleted with
`<code>unalias</code>',</li>
<li>reserved words, those special to the shell which often need to be
interpreted differently from ordinary commands due to the syntax,
although they can be disabled if you really need to,</li>
<li>functions; these can also be disabled, although it's usually easier
to `<code>unfunction</code>' them,</li>
<li>builtin commands, which can be disabled, or called as a builtin by
putting `<code>builtin</code>' in front,</li>
<li>external commands, which can be called as such, even if the name
clashes with one of the above types, by putting `<code>command</code>' in
front.</li>
</ul>
<p><span id="l54"></span></p>
<h2 id="36-expansions-and-quotes"><a class="header" href="#36-expansions-and-quotes">3.6: Expansions and quotes</a></h2>
<p>As I keep advertising, there will be a whole chapter dedicated to the
subject of shell expansions and what to do with them. However, it's a
rather basic subject, which definitely comes under the heading of basic
shell syntax, so I shall here list all the forms of expansion. As given
in the manual, there are five stages.</p>
<p><span id="l55"></span></p>
<h3 id="361-history-expansion"><a class="header" href="#361-history-expansion">3.6.1: History expansion</a></h3>
<p>This is the earliest, and is only done on an interactive command line,
and only if you have not set <code>NO_BANG_HIST</code>. It was described in the
section `<em>The history mechanism; types of history</em>' in the previous
chapter. It is almost independent of the shell's processing of the
command line; it takes place as the command line is read in, not when
the commands are interpreted. However, in zsh it is done late enough
that the `<code>!</code>'s can be quoted by putting them in single quotes:</p>
<pre><code> echo 'Hello!!'
</code></pre>
<p>doesn't insert the previous line at that point, but</p>
<pre><code> echo "Hello!!"
</code></pre>
<p>does. You can always quote active `<code>!</code>'s with a backslash, so</p>
<pre><code> echo "Hello\!\!"
</code></pre>
<p>works, with or without the double quotes. Amusingly, since single quotes
aren't special in double quotes, if you set the <code>HIST_VERIFY</code> option,
which puts the expanded history line back on the command line for
possible further editing, and try the first two of the three
possibilities above in order, then keep hitting return, you will find
ever increasing command lines:</p>
<pre><code> % echo 'Hello!!'
Hello!!
% echo "Hello!!"
% echo "Helloecho 'Hello!!'"
% echo "Helloecho 'Helloecho 'Hello!!''"
% echo "Helloecho 'Helloecho 'Helloecho 'Hello!!'''"
</code></pre>
<p>and if you understand why, you have a good grasp of how quotes work.</p>
<p>There's another way of quoting exclamation marks in a line: put a
`<code>!"</code>' in it. It can appear anywhere (as long as it's not in single
quotes) and will be removed from the line, but it has the effect of
disabling any subsequent exclamation marks till the end of the line.
This is the only time quote marks which are significant to the shell
(i.e. are not themselves quoted) don't have to occur in a matching pair.</p>
<p>Note that as exclamation marks aren't active in any text read
non-interactively --- and this includes autoloaded functions and sourced
files, such as startup files, read inside interactive shells --- it is
an error to quote any `<code>!</code>'s in double quotes in files. This will
simply pass on the backslashes to the next level of parsing. Other forms
of quoting are all right: `<code>\!</code>', because any character quoted with a
backslash is treated as itself, and <code>'!'</code> because single quotes can
quote anything anyway.</p>
<p><span id="l56"></span></p>
<h3 id="362-alias-expansion"><a class="header" href="#362-alias-expansion">3.6.2: Alias expansion</a></h3>
<p>As discussed above, alias expansion also goes on as the command line is
read, so is to a certain extent similar to history expansion. However,
while a history expansion may produce an alias for expansion, `<code>!</code>'s in
the text resulting from alias expansions are normal characters, so it
can be thought of as a later phase (and indeed it's implemented that
way).</p>
<p><span id="l57"></span></p>
<h3 id="363-process-parameter-command-arithmetic-and-brace-expansion"><a class="header" href="#363-process-parameter-command-arithmetic-and-brace-expansion">3.6.3: Process, parameter, command, arithmetic and brace expansion</a></h3>
<p>There are a whole group of expansions which are done together, just by
looking at the line constructed from the input after history and alias
expansion and reading it from left to right, picking up any active
expansions as the line is examined. Whenever a complete piece of
expandable text is found, it is expanded; the text is not re-examined,
except in the case of brace expansion, so none of these types of
expansion is performed on any resulting text. Whether later forms of
expansion --- in other words, filename generation and filename expansion
are performed --- is another matter, depending largely on the
<code>GLOB_SUBST</code> option as discussed in the previous chapter. Here's a brief
summary of the different types.</p>
<p><strong>Process substitution</strong></p>
<p>There are three forms that result in a command line argument which
refers to a file from or to which input or output is taken:
`<code><</code>(<em>process</em>)' runs the process which is expected to generate output
which can be used as input by a command; `<code>></code>(<em>process</em>)' runs the
process which will take input to it; and `<code>=</code>(<em>process</em>)' acts like the
first one, but it is guaranteed that the file is a plain file.</p>
<p>This probably sounds like gobbledygook. Here are some simple examples.</p>
<pre><code> cat < <(echo This is output)
</code></pre>
<p>(There are people in the world with nothing better to do than compile
lists of dummy uses of the `<code>cat</code>' command, as in that example, and
pour scorn on them, but I'll just have to brave it out.) What happens is
that the command `<code>echo This is output</code>' is run, with the obvious
result. That output is <em>not</em> put straight into the command line, as it
would be with command substitution, to be described shortly. Instead,
the command line is given a filename which, when read, gets that output.
So it's more like:</p>
<pre><code> echo This is output >tmpfile
cat < tmpfile
rm tmpfile
</code></pre>
<p>(note that the temporary file is cleaned up automatically), except that
it's more compact. In this example I could have missed out the remaining
`<code><</code>', since <code>cat</code> does the right thing with a filename, but I put it
there to emphasise the fact that if you want to redirect input from the
process substitution you need an <em>extra</em> `<code><</code>', over and above the one
in the substitution syntax.</p>
<p>Here's an example for the corresponding output substitution:</p>
<pre><code> echo This is output > \
>(sed 's/output/rubbish/' >outfile)
</code></pre>
<p>which is a perfectly foul example, but works essentially like:</p>
<pre><code> echo This is output >tmpfile
sed 's/output/rubbish/' <tmpfile >outfile
</code></pre>
<p>There's an obvious relationship to pipes here, and in fact this example
could be better written,</p>
<pre><code> echo This is output | sed 's/output/rubbish/' >outfile
</code></pre>
<p>A good example of an occasion where the output process substitution
can't be replaced by a pipe is when it's on the error output, and
standard output is being piped:</p>
<pre><code> ./myscript 2> >(grep -v idiot >error.log) |
process-output >output.log
</code></pre>
<p>a little abstract, but here the main point of the script `myscript' is
to produce some output which undergoes further processing on the
right-hand side of the pipe. However, we want to process the error
output here, by filtering out occurrences of lines which use the word
`idiot', before dumping those errors into a file <code>error.log</code>. So we get
an effect similar to having two pipelines at once, one for output and
one for error. Note again the <em>two</em> `<code>></code>' signs present next to one
another to get that effect.</p>
<p>Finally, the `<code>=</code>(<em>process</em>)' form. Why do we need this as well as the
one with `<code><</code>'? To understand that, you need to know a little of how
zsh tries to implement the latter type efficiently. Most modern
UNIX-like systems have `named pipes', which are essentially files that
behave like the `<code>|</code>' on the command line: one process writes to the
file, another reads from it, and the effect is essentially that data
goes straight through. If your system has them, you will usually find
the following demonstration works:</p>
<pre><code> % mknod tmpfile p
% echo This is output >tmpfile &
[2] 1507
% read line <tmpfile
%
[2] + 1507 done echo This is output >> tmpfile
% print -- $line
This is output
%
</code></pre>
<p>The syntax to create a named pipe is that rather strange `<code>mknod</code>'
command, with `<code>p</code>' for pipe. We stick this in the background, because
it won't do anything yet: you can't write to the pipe when there's
no-one to read it (a fundamental rule of pipes which isn't <em>quite</em> as
obvious as it may seem, since it <em>is</em> possible for data to lurk in the
pipe, buffered, before the process reading from it extracts it), so we
put that in the background to wait for action. This comes in the next
line, where we read from the pipe: that allows the <code>echo</code> to complete
and exit. Then we print out the line we've read.</p>
<p>The problem with pipes is that they are just temporary storage spaces
for data on the way through. In particular, you can't go back to the
beginning (in C-speak, `you can't seek backwards on a pipe') and
re-read what was there. Sometimes this doesn't matter, but some
commands, such as editors, need that facility. As the `<code><</code>' process
substitution is implemented with named pipes (well, maybe), there is
also the `<code>=</code>' form, which produces a real, live temporary file,
probably in the `<code>/tmp</code>' directory, containing the output from the
file, and then puts the name of that file on the command line. The
manual notes, unusually helpfully, that this is useful with the
`<code>diff</code>' command for comparing the output of two processes:</p>
<pre><code> diff =(./myscript1) =(./myscript2)
</code></pre>
<p>where, presumably, the two scripts produce similar, but not identical,
output which you want to compare.</p>
<p>I said `well, maybe' in that paragraph because there's another way zsh
can do `<code><</code>' process substitutions. Many modern systems allow you to
access a file with a name like `<code>/dev/fd/0</code>' which corresponds to file
descriptor 0, in this case standard input: to anticipate the section on
redirection, a `file descriptor' is a number assigned to a particular
input or output stream. This method allows you to access it as a file;
and if this facility is available, zsh will use it to pass the name of
the file in process substitution instead of using a named pipe, since in
this case it doesn't have to create a temporary file; the system does
everything. Now, if you are really on the ball, you will realise that
this doesn't get around the problem of pipes --- where is data on this
file descriptor going to come from? The answer is that it will either
have to come from a real temporary file --- which is pointless, because
that's what we wanted to avoid --- or from a pipe opened from some
process --- which is equivalent to the named pipe method, except with
just a file descriptor instead of a name. So even if zsh does it this
way, you still need the `<code>=</code>' form for programmes which need to go
backwards in what they're reading.</p>
<p><strong>Parameter substitution</strong></p>
<p>You've seen enough of this already. This comes from a `<code>$</code>' followed
either by something in braces, or by alphanumeric characters forming the
name of the parameter: `<code>$foo</code>' or `<code>${foo}</code>', where the second form
protects the expansion from any other strings at the ends and also
allows a veritable host of extra things to appear inside the braces to
modify the substitution. More detail will be held over to till <a href="zshguide05.html#subst">chapter
5</a>; there's a lot of it.</p>
<p><strong>Command substitution</strong></p>
<p>This has two forms, <code>$</code>(<em>process</em>) and <code>`</code><em>process</em><code>`</code>. They
function identically; the first form has two advantages: substitutions
can be nested, since the end character is different from the start
character, and (because it uses a `<code>$</code>') it reminds you that, like
parameter substitutions, command substitutions can take place inside
double-quoted strings. In that case, like most other things in quotes,
the result will be a single word; otherwise, the result is split into
words on any field separators you have defined, usually whitespace or
the null character. I'll use the <code>args</code> function again:</p>
<pre><code> % args() { print $# $*; }
% args $(echo two words)
2 two words
% args "$(echo one word)"
1 one word
</code></pre>
<p>The first form will split on newlines, not just spaces, so an equivalent
is</p>
<pre><code> % args $(echo two; echo words)
2 two words
</code></pre>
<p>Thus entire screens of text will be flattened out into a single line of
single-word command arguments. By contrast, with the double quotes no
processing is done whatsoever; the entire output is put verbatim into
one command argument, with newlines intact. This means that the quite
common case of wanting a single complete line from a file per command
argument has to be handled by trickery; zsh has such trickery, but
that's the stuff of <a href="zshguide05.html#subst">chapter 5</a>.</p>
<p>Note the difference from process substitution: no intermediate file name
is involved, the output itself goes straight onto the command line. This
form of substitution is considerably more common, and, unlike the other,
is available in all UNIX shells, though not in all shells with the more
modern form `<code>$</code>(<code>...</code>)'.</p>
<p>The rule that the command line is evaluated only once, left to right, is
adhered to here, but it's a little more complicated in this case since
the expression being substituted is scanned <em>as a complete command
line</em>, so can include anything a command usually can, with all the rules
of quoting and expansion being applied. So if you get confused about
what a command substitution is actually up to, you should extract the
commands from it and think of them as a command line in their own right.
When you've worked out what that's doing, decide what it's output will
be, and that's the result of the substitution. You can ignore any error
output; that isn't captured, so will go straight to the terminal. If you
want to ignore it, use the standard trick (see below) `<code>2>/dev/null</code>'
<em>inside</em> the command substitution --- not on the main command line,
where it won't work because substitutions are performed before
redirection of the main command line, and in any case that will have the
obvious side effect of changing the error output from the command line
itself.</p>
<p>The only real catch with command substitution is that, as it is run as a
separate process --- even if it only involves shell builtins --- no
effects other than the output will percolate back to the main shell:</p>
<pre><code> % print $(bar=value; print bar is $bar)
bar is value
% print bar is $bar
bar is
</code></pre>
<p>There is maybe room for a form of substitution that runs inside the
shell, instead; however, with modern computers the overhead in starting
the extra process is pretty small --- and in any case we seem to have
run out of new forms of syntax.</p>
<p>Once you know and are comfortable with command substitution, you will
probably start using it all the time, so there is one good habit to get
into straight away. A particularly common use is simply to put the
contents of a file onto the command line.</p>
<pre><code> # Don't do this, do the other.
process_cmd `cat file_arguments`
</code></pre>
<p>But there's a shortcut.</p>
<pre><code> # Do do this, don't do the other
process_cmd $(<file_arguments)
</code></pre>
<p>It's not only less writing, it's more efficient: zsh spots the special
syntax, with the <code><</code> immediately inside the parentheses, reads the file
directly without bothering to start `<code>cat</code>', and inserts its contents:
no external process is involved. You shouldn't confuse this with `null
redirections' as described below: the syntax is awfully similar,
unfortunately, but the feature shown here is not dependent on that other
feature being enabled or set up in a particular way. In fact, this
feature works in ksh, which doesn't have zsh's null redirections.</p>
<p>You can quote the file-reading form too, of course: in that case, the
contents of the file `<code>cmd_arguments</code>' would be passed as just one
argument, with newlines and spaces intact.</p>
<p>Sometimes, the rule about splitting the result of a command substitution
can get you into trouble:</p>
<pre><code> % typeset foo=`echo words words`
% print $foo
words
</code></pre>
<p>You probably expected the command substitution <em>not</em> to be split here.
but it was, and the shell executed typeset with the arguments
`<code>foo=words</code>' and `words'. That's because in zsh arguments to
<code>typeset</code> are treated pretty much normally, except for some jiggery
pokery with tildes described below. Other shells do this differently,
and zsh (from 4.0.2 and 4.1.1) provides a compatibility option,
<code>KSH_TYPESET</code>. In earlier versions you need to use quotes:</p>
<pre><code> % typeset foo="`echo words words`"
% print $foo
words words
</code></pre>
<p>A really rather technical afterword: using `<code>$(cat file_arguments)</code>',
you might have counted two extra processes to be started, one being the
usual one for a command substitution, and another the `<code>cat</code>' process,
since that's an external command itself. That would indeed be the
obvious way of doing it, but in fact zsh has an optimisation in cases
like this: if it knows the shell is about to exit --- in this case, the
forked process which is just interpreting the command line for the
substitution --- it will not bother to start a new process for the last
command, and here just replaces itself with the <code>cat</code>. So actually
there's only one extra process here. Obviously, an interactive shell is
never replaced in this way, since clairvoyance is not yet a feature of
the shell.</p>
<p><strong>Arithmetic substitution</strong></p>
<p>Arithmetic substitution is easy to explain: everything I told you about
the <code>(( ... ))</code> command under numerical parameters, above, applies to
arithmetic substitution. You simply bang a `<code>$</code>' in front, and it
becomes an expansion.</p>
<pre><code> % print $(( 32 + 2 * 5 ))
42
</code></pre>
<p>You can perform everything inside arithmetic substitution that you can
inside the builtin, including assignments; the only difference is that
the status is not set, instead the value is put directly onto the
command line in place of the original expression. As in C, the value of
an assignment is the value being assigned, `<code>$(( param = 3 + 2))</code>'
substitutes the value 5 as well as assigning it to <code>$param</code>.</p>
<p>By the way, there's an extra level of substitution involved in all
arithmetic expansions, since scalar parameters are subject to arithmetic
expansion when they're read in. This is simple if they only contain
numbers, but less obvious if they contain complete expressions:</p>
<pre><code> % foo=3+5
% print $(( foo + 2))
10
</code></pre>
<p>The foo was evaluated into 8 before it was substituted in. Note this
means there were two evaluations: this doesn't work:</p>
<pre><code> % foo=3+
% print $(( foo 2 ))
zsh: bad math expression: operand expected at `'
</code></pre>
<p>--- the complaint here is about the missing operand after the `<code>+</code>' in
the <code>$foo</code>. However the following <em>does</em> work:</p>
<pre><code> % foo=3+
% print $(( $foo 2 ))
5
</code></pre>
<p>That's because the scalar <code>$foo</code> is turned into <code>3+</code> first. This is more
logical than you might think: with the rule about left to right
evaluation, the <code>$foo</code> is picked up inside the <code>$((...))</code> and expanded
as an ordinary parameter substitution while the argument of <code>$((...))</code>
is being scanned. Then the complete argument `<code>3+ 2</code>' is expanded as an
arithmetical expression. (Unfortunately, zsh isn't always this logical;
there could easily be cases where we haven't thought it through --- you
should feel free to bring these to our attention.)</p>
<p>There's an older form with single square brackets instead of double
parentheses; there is now no reason to use it, as it's non-standard, but
you may sometimes still meet it.</p>
<p><strong>Brace expansion</strong></p>
<p>Brace expansion is a feature acquired from the C shell and it's
relatives, although some versions of ksh have it, as it's a compile time
option there. It's a useful way of saving you from typing the same thing
twice on a single command line:</p>
<pre><code> % print -l {foo,bar}' is used far too often in examples'
foo is used far too often in examples
bar is used far too often in examples
</code></pre>
<p>`<code>print</code>' is given two arguments which it is told to print out one per
line. The text in quotes is common to both, but one has `<code>foo</code>' in
front, while the other has `<code>bar</code>' in front. The brace expression can
equally be in the middle of an argument: for example, a common use of
this among programmers is for similarly named source files:</p>
<pre><code> % print zle_{tricky,vi,word}.c
zle_tricky.c zle_vi.c zle_word.c
</code></pre>
<p>As you see, you're not limited to two; you can have any number. You can
quote a comma if you need a real one:</p>
<pre><code> % print -l \`{\,,.}\'' is a punctuation character'
`,' is a punctuation character
`.' is a punctuation character
</code></pre>
<p>The quotes needed quoting with a backslash to get them into the output.
The second comma is the active one for the braces.</p>
<p>You can nest braces. Once again, this is done left to right. In</p>
<pre><code> print {now,th{en,ere{,abouts}}}
</code></pre>
<p>the first argument of the outer brace is `<code>now</code>', and the second is
`<code>th{en,ere{,abouts}}</code>'. This brace expands to `<code>then</code>' and then the
expansion of `<code>there{,abouts}</code>', which is `<code>there thereabouts</code>' ---
there's nothing to stop you having an empty argument. Putting this all
together, we have</p>
<pre><code> print now then there thereabouts
</code></pre>
<p>There's more to know about brace expansion, which will appear in
<a href="zshguide05.html#subst">chapter 5</a> on clever expansions.</p>
<p><span id="l58"></span></p>
<h3 id="364-filename-expansion"><a class="header" href="#364-filename-expansion">3.6.4: Filename Expansion</a></h3>
<p>It's a shame the names `filename expansion' and `filename generation'
sound so similar, but most people just refer to `<code>~</code> and <code>=</code> expansion'
and `globbing' respectively, which is all that is meant by the two. The
first is by far the simpler. The rule is: unquoted `<code>~</code>'s at the
beginning of words perform expansion of named directories, which may be
your home directory:</p>
<pre><code> % print ~
/home/pws
</code></pre>
<p>some user's home directory:</p>
<pre><code> % print ~root
/root
</code></pre>
<p>(that may turn up `<code>/</code>' on your system), a directory named directly by
you:</p>
<pre><code> % t=/tmp
% print ~t
/tmp
</code></pre>
<p>a directory you've recently visited:</p>
<pre><code> % pwd
/home/pws/zsh/projects/zshguide
% print ~+
/home/pws/zsh/projects/zshguide
% cd /tmp
% print ~-
/home/pws/zsh/projects/zshguide
</code></pre>
<p>or a directory in your directory stack:</p>
<pre><code> % pushd /tmp
% pushd ~
% pushd /var/tmp
% print ~2
/tmp
</code></pre>
<p>These forms were discussed above. There are various extra rules. You can
add a `<code>/</code>' after any of them, and the expansions still take place, so
you can use them to specify just the first part of a longer expression
(as you almost certainly have done with a simple `<code>~</code>'). If you quote
the `<code>~</code>' in any of the ways quoting normally takes place, the
expansion doesn't happen.</p>
<p>A <code>~</code> in the middle of the word means something completely different, if
you have the <code>EXTENDED_GLOB</code> option set; if you don't, it doesn't mean
anything. There are a few exceptions here; assignments are a fairly
natural one:</p>
<pre><code> % foo=~pws
% print $foo
/home/pws
</code></pre>
<p>(note that the `<code>~pws</code>', being unquoted, was expanded straight away at
the assignment, not at the print statement). But the following works
too:</p>
<pre><code> % PATH=$PATH:~pws/bin
</code></pre>
<p>because colons are special in assignments. Note that this happens even
if the variable isn't a colon-separated path; the shell doesn't know
what use you're going to make of all the different variables.</p>
<p>The companion of `<code>~</code>' is `<code>=</code>', which again has to occur at the start
of a word or assignment to be special. The remainder of the word (here
the <em>entire</em> remainder, because directory paths aren't useful) is taken
as the name of an external command, and the word is expanded to the
complete path to that command, using <code>$PATH</code> just as if the command were
to be executed:</p>
<pre><code> % print =ls
/bin/ls
</code></pre>
<p>and, slightly confusingly,</p>
<pre><code> % foo==ls
% print $foo
/bin/ls
</code></pre>
<p>where the two `<code>=</code>'s have two different meanings. This form is useful
in a number of cases. For example, you might want to look at or edit a
script which you know is in your path; the form</p>
<pre><code> % vi =scriptname
</code></pre>
<p>is more convenient than the more traditional</p>
<pre><code> % vi `whence -p ls`
</code></pre>
<p>where I put the `<code>-p</code>' in to force <code>whence</code> to follow the path,
ignoring builtins, functions, etc. This brings us to another use for
`<code>=</code>' expansion,</p>
<pre><code> % =ls
</code></pre>
<p>is a neat and extremely short way of referring to an external command
when <code>ls</code> is usually a function. It has some of the same effect as
`<code>command ls</code>', but is easier to type.</p>
<p>In versions up to and including <code>4.0</code>, this syntax will also expand
aliases, so you need to be a bit careful if you really want a path to an
external command:</p>
<pre><code> % alias foo='ls -F'
% print =foo
ls -F
</code></pre>
<p>(Path expansion is done in preference, so you are safe if you use <code>ls</code>,
unless your <code>$PATH</code> is strange.) Putting `<code>=foo</code>' at the start of the
command line doesn't work, and the reason why bears examination:
<code>=</code>-expansion occurs quite late on, after ordinary alias expansion and
word splitting, so that the result is the single word `<code>ls -F</code>', where
the space is part of the word, which probably doesn't mean anything (and
if it does, don't lend me your computer when I need something done in a
hurry). It's probably already obvious that alias expansion here is more
trouble than it's worth. A less-than-exhaustive search failed to find
anyone who liked this feature, and it has been removed from the shell
from 4.1, so that `<code>=</code>'-expansion now only expands paths to external
commands.</p>
<p>If you don't like <code>=</code>-expansion, you can turn it off by setting the
option <code>NO_EQUALS</code>. One catch, which might make you want to do that, is
that the commands <code>mmv</code>, <code>mcp</code> and <code>mln</code>, which are a commonly used
though non-standard piece of free software, use `<code>=</code>' followed by a
number to replace a pattern, for example</p>
<pre><code> mmv '*.c' '=1.old.c'
</code></pre>
<p>renames all files ending with <code>.c</code> to end with <code>.old.c</code>. If you were not
alert, you might forget to quote the second word. Otherwise, however,
<code>=</code>' isn't very common at the start of a word, so you're probably fairly
safe. For a way to do that with zsh patterns, see the discussion of the
function <code>zmv</code> below (the answer is `<code>zmv '(*).c' '$1.old.c'</code>').</p>
<p>Note that zsh is smart enough to complete the names of commands after an
`<code>=</code>' of the expandable sort when you hit TAB.</p>
<p><span id="l59"></span></p>
<h3 id="365-filename-generation"><a class="header" href="#365-filename-generation">3.6.5: Filename Generation</a></h3>
<p>Filename generation is exactly the same as `globbing': the expanding of
any unquoted wildcards to match files. This is only done in one
directory at a time. So for example</p>
<pre><code> print *.c
</code></pre>
<p>won't match files in a subdirectory ending in `<code>.c</code>'. However, it <em>is</em>
done on all parts of a path, so</p>
<pre><code> print */*.c
</code></pre>
<p>will match all `<code>.c</code>' files in all immediate subdirectories of the
current directory. Furthermore, zsh has an extension --- one of its most
commonly used special features --- to match files in any subdirectory at
any depth, including the current directory: use two `<code>*</code>'s as part of
the path:</p>
<pre><code> print **/*.c
</code></pre>
<p>will match `<code>prog.c</code>', `<code>version1/prog.c</code>', `<code>version2/test/prog.c</code>',
`<code>oldversion/working/saved/prog.c</code>', and so on. I will talk about
filename generation and other uses of zsh's extremely powerful patterns
at much greater length in <a href="zshguide05.html#subst">chapter 5</a>. My main
thrust here is to fit it into other forms of expansion; the main thing
to remember is that it comes last, after everything has already been
done.</p>
<p>So although you would certainly expect this to work,</p>
<pre><code> print ~/*
</code></pre>
<p>generating all files in your home directory, you now know why: it is
first expanded to `<code>/home/pws/*</code>' (or wherever), then the shell scans
down the path until it finds a pattern, and looks in the directory it
has reached (<code>/home/pws</code>) for matching files. Furthermore,</p>
<pre><code> foo=~/
print $foo*
</code></pre>
<p>works. However, as I explained in the last chapter, you need to be
careful with</p>
<pre><code> foo=*
print ~/$foo
</code></pre>
<p>This just prints `<code>/home/pws/*</code>'. To get the `<code>*</code>' from the parameter
to be a wildcard, you need to tell the shell explicitly that's what you
want:</p>
<pre><code> foo=*
print ~/${~foo}
</code></pre>
<p>As also noted, other shells do expand the <code>*</code> as a wildcard anyway. The
zsh attitude here, as with word splitting, is that parameters should do
exactly what they're told rather than waltz off generating extra words
or expansions.</p>
<p>Be even more careful with arrays:</p>
<pre><code> foo=(*)
</code></pre>
<p>will expand the <code>*</code> immediately, in the current directory --- the
elements of the array assignment are expanded exactly like a normal
command line glob. This is often very useful, but note the difference
from scalar assignments, which do other forms of expansion, but not
globbing.</p>
<p>I'll mention a few possible traps for the unwary, which might confuse
you until you are a zsh globbing guru. Firstly, parentheses actually
have two uses. Consider:</p>
<pre><code> print (foo|bar)(.)
</code></pre>
<p>The first set of parentheses means `match either <code>foo</code> or <code>bar</code>'. If
you've used <code>egrep</code>, you will probably be familiar with this. The
second, however, simply means `match only regular files'. The `<code>(.)</code>'
is called a `globbing qualifier', because it limits the scope of any
matches so far found. For example, if either or both of <code>foo</code> and <code>bar</code>
were found, but were directories, they would not now be matched. There
are many other possibilities for globbing qualifiers. For now, the
easiest way to tell if something at the end is <em>not</em> a globbing
qualifier is if it contains a `<code>|</code>'.</p>
<p>The second point is about forms like this:</p>
<pre><code> print file-<1-10>.dat
</code></pre>
<p>The `<code><</code>' and `<code>></code>' smell of redirection, as described next, but
actually the form `<code><</code>', optional start number, `<code>-</code>', optional finish
number, `<code>></code>' means match any positive integer in the range between the
two numbers, inclusive; if either is omitted, there is no limit on that
end, hence the cryptic but common `<code><-></code>' to match any positive integer
--- in other words, any group of decimal digits (bases other than ten
are not handled by this notation). Older versions of the shell allowed
the form `<code><></code>' as a shorthand to match any number, but the overlap
with redirection was too great, as you'll see, so this doesn't work any
more.</p>
<p>Another two cryptic symbols are the two that do negation. These only
work with the option `<code>EXTENDED_GLOB</code>' set: this is necessary to get
the most out of zsh's patterns, but it can be a trap for the unwary by
turning otherwise innocuous characters into patterns:</p>
<pre><code> print ^foo
</code></pre>
<p>This means any file in the current directory <em>except</em> the file <code>foo</code>.
One way of coming unstuck with `<code>^</code>' is something like</p>
<pre><code> stty kill ^u
</code></pre>
<p>where you would hope `<code>^u</code>' means control with `<code>u</code>', i.e. ASCII
character 21. But it doesn't, if <code>EXTENDED_GLOB</code> is set: it means `any
file in the current directory except one called `<code>u</code>' ', which is
definitely a different thing. The other negation operator isn't usually
so fraught, but it can look confusing:</p>
<pre><code> print *.c~f*
</code></pre>
<p>is a pattern of two halves; the shell tries to match `<code>*.c</code>', but
rejects any matches which also match `<code>f*</code>'. Luckily, a `<code>~</code>' right at
the end isn't special, so</p>
<pre><code> rm *.c~
</code></pre>
<p>removes all files ending in `<code>.c~</code>' --- it wouldn't be very nice if it
matched all files ending in `<code>.c</code>' and treated the final `<code>~</code>' as an
instruction not to reject any, so it doesn't. The most likely case I can
think of where you might have problems is with Emacs' numeric backup
files, which can have a `<code>~</code>' in the middle which you should quote.
There is no confusion with the directory use of `<code>~</code>', however: that
only occurs at the beginning of a word, and this use only occurs in the
middle.</p>
<p>The final oddments that don't fit into normal shell globbing are forms
with `<code>#</code>'. These also require that <code>EXTENDED_GLOB</code> be set. In the
simplest use, a `<code>#</code>' after a pattern says `match this zero or more
times'. So `<code>(foo|bar)#.c</code>' matches <code>foo.c</code>, <code>bar.c</code>, <code>foofoo.c</code>,
<code>barbar.c</code>, <code>foobarfoo.c</code>, ... With an extra <code>#</code>, the pattern before (or
single character, if it has no special meaning) must match at least
once. The other use of `<code>#</code>' is in a facility called `globbing flags',
which look like `<code>(#X)</code>' where `<code>X</code>' is some letter, possibly followed
by digits. These turn on special features from that point in the pattern
and are one of the newest features of zsh patterns; they will receive
much more space in <a href="zshguide05.html#subst">chapter 5</a>.</p>
<p><span id="l60"></span></p>
<h2 id="37-redirection-greater-thans-and-less-thans"><a class="header" href="#37-redirection-greater-thans-and-less-thans">3.7: Redirection: greater-thans and less-thans</a></h2>
<p>Redirection means retrieving input from some other file than the usual
one, or sending output to some other file than the usual one. The
simplest examples of these are `<code><</code>' and `<code>></code>', respectively.</p>
<pre><code> % echo 'This is an announcement' >tempfile
% cat <tempfile >newfile
% cat newfile
This is an announcement
</code></pre>
<p>Here, <code>echo</code> sends its output to the file <code>tempfile</code>; <code>cat</code> took its
input from that file and sent its output --- the same as its input ---
to the file <code>newfile</code>; the second <code>cat</code> takes its input from <code>newfile</code>
and, since its output wasn't redirected, it appeared on the terminal.</p>
<p>The other basic form of redirection is a pipe, using `<code>|</code>'. Some people
loosely refer to all redirections as pipes, but that's rather confusing.
The input and output of a pipe are <em>both</em> programmes, unlike the case
above where one end was a file. You've seen lots of examples already:</p>
<pre><code> echo foo | sed 's/foo/bar/'
</code></pre>
<p>Here, <code>echo</code> sends its output to the programme <code>sed</code>, which substitutes
foo by bar, and sends its own output to standard output. You can chain
together as many pipes as you like; once you've grasped the basic
behaviour of a single pipe, it should be obvious how that works:</p>
<pre><code> echo foo is a word |
sed 's/foo/bar/' |
sed 's/a word/an unword/'
</code></pre>
<p>runs another <code>sed</code> on the output of the first one. (You can actually
type it like that, by the way; the shell knows a pipe symbol can't be at
the end of a command.) In fact, a single <code>sed</code> will suffice:</p>
<pre><code> echo foo is a word |
sed -e 's/foo/bar/' -e 's/a word/an unword/'
</code></pre>
<p>has the same effect in this case.</p>
<p>Obviously, all three forms of redirection only work if the programme in
question expects input from standard input, and sends output to standard
output. You can't do:</p>
<pre><code> echo 'edit me' | vi
</code></pre>
<p>to edit input, since <code>vi</code> doesn't use the input sent to it; it always
deals with files. Most simple UNIX commands can be made to deal with
standard input and output, however. This is a big difference from other
operating systems, where getting programmes to talk to each other in an
automated fashion can be a major headache.</p>
<p><span id="l61"></span></p>
<h3 id="371-clobber"><a class="header" href="#371-clobber">3.7.1: Clobber</a></h3>
<p>The word `clobber', as in the option <code>NO_CLOBBER</code> which I mentioned in
the previous chapter, may be unfamiliar to people who don't use English
as their first language. Its basic meaning is `hit' or `defeat' or
`destroy', as in `Itchy and Scratchy clobbered each other with
mallets'. If you do:</p>
<pre><code> % echo first go >file
% echo second go >file
</code></pre>
<p>then <code>file</code> will contain only the words `second go'. The first thing
you put into the file, `first go', has been clobbered. Hence the
<code>NO_CLOBBER</code> option: if this is set, the shell will complain when you
try to overwrite the file. You can use `<code>>|file</code>' or `<code>>! file</code>' to
override this. You usually can't use `<code>>!file</code>' because history
expansion will try to expand `<code>!file</code>' before the shell parses the
line; hence the form with the vertical bar tends to be more useful.</p>
<p><span id="l62"></span></p>
<h3 id="372-file-descriptors"><a class="header" href="#372-file-descriptors">3.7.2: File descriptors</a></h3>
<p>UNIX-like systems refer to different channels such as input, output and
error by `file descriptors', which are small integers. Usually three
are special: 0, standard input; 1, standard output; and 2, standard
error. Bourne-like shells (but not csh-like shells) allow you to refer
to a particular file descriptor, instead of standard input or output, by
putting the integer immediately before the `<code><</code>' or `<code>></code>' (no space is
allowed). What's more, if the `<code><</code>' or `<code>></code>' is followed immediately
by `<code>&</code>', a file descriptor can follow the redirection (the one before
is optional as usual). A common use is:</p>
<pre><code> % echo This message will go to standard error >&2
</code></pre>
<p>The command sends its message to standard output, file descriptor 1. As
usual, `<code>></code>' redirects standard output. This time, however, it is
redirected not to a file, but to file descriptor 2, which is standard
error. Normally this is the same device as standard output, but it can
be redirected completely separately. So:</p>
<pre><code> % { echo A message
cursh> echo An error >&2 } >file
An error
% cat file
A message
</code></pre>
<p>Apologies for the slightly unclear use of the continuation prompt
`<code>cursh></code>': this guide goes into a lot of different formats, and some
are a bit finicky about long lines in preformatted text. As pointed out
above, the `<code>>file</code>' here will redirect all output from the stuff in
braces, just as if it were a single command. However, the `<code>>&2</code>'
inside redirects the output of the second <code>echo</code> to standard error.
Since this wasn't redirected, it goes straight to the terminal.</p>
<p>Note the form in braces in the previous example --- I'm going to use
that in a few more examples. It simply sends something to standard
output, and something else to standard error; that's its only use. Apart
from that, you can treat the bit in braces as a black box --- anything
which can produce both sorts of output.</p>
<p>Sometimes you want to redirect both at once. The standard Bourne-like
way of doing this is:</p>
<pre><code> % { echo A message
cursh> echo An error >&2 } >file 2>&1
</code></pre>
<p>The `<code>>file</code>' redirects standard output from the <code>{</code><em>...</em><code>}</code> to the
file; the following <code>2>&1</code> redirects standard error to wherever standard
output happens to be at that point, which is the same file. This allows
you to copy two file descriptors to the same place. Note that the order
is important; if you swapped the two around, `<code>2>&1</code>' would copy
standard error to the initial destination of standard output, which is
the terminal, before it got around to redirecting standard output.</p>
<p>Zsh has a shorthand for this borrowed from csh-like shells:</p>
<pre><code> % { echo A message
cursh> echo An error >&2 } >&file
</code></pre>
<p>is exactly equivalent to the form in the previous paragraph, copying
standard output and standard error to the same file. There is obviously
a clash of syntax with the descriptor-copying mechanism, but if you
don't have files whose names are numbers you won't run into it. Note
that csh-like shells don't have the descriptor-copying mechanism: the
simple `<code>>&</code>' and the same thing with pipes are the only uses of `<code>&</code>'
for redirections, and it's not possible there to refer to particular
file descriptors.</p>
<p>To copy standard error to a pipe, there are also two forms:</p>
<pre><code> % { echo A message
cursh> echo An error >&2 } 2>&1 | sed -e 's/A/I/'
I message
In error
% { echo A message
cursh> echo An error >&2 } |& sed -e 's/A/I/'
I message
In error
</code></pre>
<p>In the first case, note that the pipe is opened before the other
redirection, so that `<code>2>&1</code>' copies standard error to the pipe, not
the original standard output; you couldn't put that after the pipe in
any case, since it would refer to the `<code>sed</code>' command's output. The
second way is like csh; unfortunately, `<code>|&</code>' has a different meaning
in ksh (start a coprocess), so zsh is incompatible with ksh in this
respect.</p>
<p>You can also close a file descriptor you don't need: the form `<code>2<&-</code>'
will close standard error for the command where it appears.</p>
<p>One thing not always appreciated about redirections is that they can
occur anywhere on the command line, not just at the end.</p>
<pre><code> % >file echo foo
% cat file
foo
</code></pre>
<p><span id="l63"></span></p>
<h3 id="373-appending-here-documents-here-strings-read-write"><a class="header" href="#373-appending-here-documents-here-strings-read-write">3.7.3: Appending, here documents, here strings, read write</a></h3>
<p>There are various other forms which use multiple `<code>></code>'s and `<code><</code>'s.
First,</p>
<pre><code> % echo foo >file
% echo bar >>file
% cat file
foo
bar
</code></pre>
<p>The `<code>>``></code>' appends to the file instead of overwriting it. Note,
however, that if you use this a lot you may find there are neater ways
of doing the same thing. In this example,</p>
<pre><code> % { echo foo
cursh> echo bar } >file
% cat file
foo
bar
</code></pre>
<p>Here, `<code>cursh></code>' is a prompt from the shell that it is waiting for you
to close the `<code>{</code>' construct which executes a set of commands in the
current shell. This construct can have a redirection applied to the
entire sequence of commands: `<code>>file</code>' after the closing brace
therefore redirects the output from both <code>echo</code>s.</p>
<p>In the case of input, doubling the sign has a totally different effect.
The word after the <code><``<</code> is not a file, but a string which will be used
to mark in the end of input. Input is read until a line with only this
string is found:</p>
<pre><code> % sed -e 's/foo/bar/' <<HERE
heredoc> This line has foo in it.
heredoc> There is another foo in this one.
heredoc> HERE
This line has a bar in it.
There is another bar in this one.
</code></pre>
<p>The shell prompts you with `<code>heredoc></code>' to tell you it is reading a
`here document', which is how this feature is referred to. When it
finds the final string, in this case `<code>HERE</code>', it passes everything you
have typed as input to the command as if it came from a file. The
command in this case is the stream editor, which has been told to
replace the first `<code>foo</code>' on each line with a `<code>bar</code>'. (Replacing
things with a bar is a familiar experience from the city centre of my
home town, Newcastle upon Tyne.)</p>
<p>So far, the features are standard in Bourne-like shells, but zsh has an
extension to here documents, sometimes referred to as `here strings'.</p>
<pre><code> % sed -e 's/string/nonsense/' \
> <<<'This string is the entire document.'
This nonsense is the entire document.
</code></pre>
<p>Note that `<code>></code>' on the second line is a continuation prompt, not part
of the command line; it was just too long for the TeX version of this
document if I didn't split it. This is a shorthand form of `here'
document if you just want to pass a single string to standard input.</p>
<p>The final form uses both symbols: `<code><>file</code>' opens the file for reading
and writing --- but only on standard input. In other words, a programme
can now both read from and write to standard input. This isn't used all
that often, and when you do use it you should remember that you need to
open standard output explicitly to the same file:</p>
<pre><code> % echo test >/tmp/redirtest
% sed 's/e/Z/g' <>/tmp/redirtest 1>&0
% cat /tmp/redirtest
tZtst
</code></pre>
<p>As standard input (the 0) was opened for writing, you can perform the
unusual trick of copying standard output (the 1) into it. This is
generally not a particularly safe way of doing in-place editing,
however, though it seems to work fine with sed. Note that in older
versions of zsh, `<code><></code>' was equivalent to `<code><-></code>', which is a pattern
that matches any number; this was changed quite some time ago.</p>
<p><span id="l64"></span></p>
<h3 id="374-clever-tricks-exec-and-other-file-descriptors"><a class="header" href="#374-clever-tricks-exec-and-other-file-descriptors">3.7.4: Clever tricks: exec and other file descriptors</a></h3>
<p>All Bourne-like shells have two other features. First, the `command'
<code>exec</code>, which I described above as being used to replace the shell with
the command you give after it, can be used with only redirections after
it. These redirections then apply permanently to the shell itself,
rather than temporarily to a single command. So</p>
<pre><code> exec >file
</code></pre>
<p>makes <code>file</code> the destination for standard output from that point on.
This is most useful in scripts, where it's quite common to want to
change the destination of all output.</p>
<p>The second feature is that you can use file descriptors which haven't
even been opened yet, as long as they are single digits --- in other
words, you can use numbers 3 to 9 for your own purposes. This can be
combined with the previous feature for some quite clever effects:</p>
<pre><code> exec 3>&1
# 3 refers to stdout
exec >file
# stdout goes to `file', 3 untouched
# random commands output to `file'
exec 1>&3
# stdout is now back where it was
exec 3>&-
# file descriptor 3 closed to tidy up
</code></pre>
<p>Here, file descriptor 3 has been used simply as a placeholder to
remember where standard output was while we temporarily divert it. This
is an alternative to the `<code>{</code><em>...</em><code>} >file</code>' trick. Note that you can
put more than one redirection on the <code>exec</code> line: `<code>exec 3>&1 >file</code>'
also works, as long as you keep the order the same.</p>
<p><span id="l65"></span></p>
<h3 id="375-multios"><a class="header" href="#375-multios">3.7.5: Multios</a></h3>
<p>Multios allow you to do an implicit `<code>cat</code>' (concatenate files) on
input and `<code>tee</code>' (send the same data to different files) on output.
They depend on the option <code>MULTIOS</code> being set, which it is by default. I
described this in the last chapter in discussing whether or not you
should have the option set, so you can look at the examples there.</p>
<p>Here's one fact I didn't mention. You use output multios like this:</p>
<pre><code> command-generating-output >file1 >file2
</code></pre>
<p>where the command's output is copied to both files. This is done by a
process forked off by the shell: it simply sits waiting for input, then
copies it to all the files in its list. There's a problem in all
versions of the shell to date (currently 4.0.6): this process is
asynchronous, so you can't rely on it having finished when the shell
starts executing the next command. In other words, if you look at
<code>file1</code> or <code>file2</code> immediately after the command has finished, they may
not yet contain all the output because the forked process hasn't
finished writing to it.</p>
<p>This is really a bug, but for the time being you will have to live with
it as it's quite complicated to fix in all cases. Multios are most
useful as a shorthand in interactive use, like so much of zsh; in a
script or function it is safer to use <code>tee</code>,</p>
<pre><code> command-generating-output | tee file1 file2
</code></pre>
<p>which does the same thing, but as <code>tee</code> is handled as a synchronous
process <code>file1</code> and <code>file2</code> are guaranteed to be complete when the
pipeline exits.</p>
<p><span id="l66"></span></p>
<h2 id="38-shell-syntax-loops-subshells-and-so-on"><a class="header" href="#38-shell-syntax-loops-subshells-and-so-on">3.8: Shell syntax: loops, (sub)shells and so on</a></h2>
<p><span id="l67"></span></p>
<h3 id="381-logical-command-connectors"><a class="header" href="#381-logical-command-connectors">3.8.1: Logical command connectors</a></h3>
<p>I have been rather cavalier in using a couple of elements of syntax
without explaining them:</p>
<pre><code> true && print Previous command returned true
false || print Previous command returned false
</code></pre>
<p>The relationship between `<code>&&</code>' and `<code>||</code>' and tests is fairly
obvious, but in this case they connect complete commands, not test
arguments. The `<code>&&</code>' executes the following command if the one before
succeeded, and the `<code>||</code>' executes the following command if the one
before failed. In other words, the first is equivalent to</p>
<pre><code> if true; then
print Previous command returned true
fi
</code></pre>
<p>but is more compact.</p>
<p>There is a perennial argument about whether to use these or not. In the
comp.unix.shell newsgroup on Usenet, you see people arguing that the
`<code>&&</code>' syntax is unreadable, and only an idiot would use it, while
other people argue that the full `<code>if</code>' syntax is slower and clumsier,
and only an idiot would use that for a simple test; but Usenet is like
that, and both answers are a bit simplistic. On the one hand, the
difference in speed between the two forms is minute, probably measurable
in microseconds rather than milliseconds on a modern computer; the
scheduling of the shell process running the script by the operating
system is likely to make more difference if these are embedded inside a
much longer script or function, as they will be. And on the other hand,
the connection between `<code>&&</code>' and a logical `and' is so strong in the
minds of many programmers that to anyone with moderate shell experience
they are perfectly readable. So it's up to you. I find I use the `<code>&&</code>'
and `<code>||</code>' forms for a pair of simple commands, but use `<code>if</code>' for
anything more complicated.</p>
<p>I would certainly advise you to avoid chains like:</p>
<pre><code> true || print foo && print bar || false
</code></pre>
<p>If you try that, you will see `<code>bar</code>' but not `<code>foo</code>', which is not
what a C programmer might expect. Using the usual rules of precedence,
you would parse it as: either <code>true</code> must be true; or both the <code>print</code>
statements must be true; or the false must be true. However, the shell
parses it differently, using these rules:</p>
<ul>
<li>If you encounter an `<code>&&</code>',
<ul>
<li>if the command before it (really the complete pipeline)
succeeded, execute the command immediately after, and execute
what follows normally</li>
<li>else if the command failed, skip the next command and any others
until an `<code>||</code>' is encountered, or until the group of commands
is ended by a newline, a semicolon, or the end of an enclosing
group. Then execute whatever follows in the normal way.</li>
</ul>
</li>
<li>If you encounter an `<code>||</code>',
<ul>
<li>if the command before it succeeded, skip the next command and
any others until an `<code>&&</code>' is encountered, or until the end of
the group, and execute what follows normally</li>
<li>else if the command failed, execute the command immediately
after the `<code>||</code>'.</li>
</ul>
</li>
</ul>
<p>If that's hard to follow, just note that the rule is completely
symmetric; a simple summary is that the logical connectors don't
remember their past state. So in the example shown, the `<code>true</code>'
succeeds, we skip `<code>print foo</code>' but execute `<code>print bar</code>' and then
skip <code>false</code>. The expression returns status zero because the last thing
it executed did so. Oddly enough, this is completely standard behaviour
for shells. This is a roundabout way of saying `don't use combined
chains of `<code>&&</code>'s and `<code>||</code>'s unless you think Gödel's theorem is for
sissies'.</p>
<p>Strictly speaking, the and's and or's come in a hierarchy of things
which connect commands. They are above pipelines, which explains my
remark above --- an expression like `<code>echo $ZSH_VERSION | sed '/dev//'</code>' is treated as a single command between any logical connectors
--- and they are below newlines and semicolons --- an expression like
`<code>true && print yes; false || print no</code>' is parsed as two distinct sets
of logically connected command sequences. In the manual, a list is a
complete set of commands executed in one go:</p>
<pre><code> echo foo; echo bar
echo small furry animals
</code></pre>
<p>--- a shell function is basically a glorified list with arguments and a
name. A sublist is a set of commands up to a newline or semicolon, in
other words a complete expression possibly involving the logical
connectors:</p>
<pre><code> show -nomoreproc |
grep -q foo &&
print The word '`foo'\' occurs.
</code></pre>
<p>A pipeline is a chain of one or more commands connected by `<code>|</code>', for
example both individual parts of the previous sublist,</p>
<pre><code> show -nomoreproc | grep -q foo
</code></pre>
<p>and</p>
<pre><code> print The word '`foo'\' occurs.
</code></pre>
<p>count as pipelines. A simple command is one single unit of execution
with a command name, so to use the same example that includes all three
of the following,</p>
<pre><code> show -nomoreproc
grep -q foo
print The word '`foo'\' occurs.
</code></pre>
<p>This means that in something like</p>
<pre><code> print foo
</code></pre>
<p>where the command is terminated by a newline and then executed in one
go, the expression is all of the above --- list, sublist, pipeline and
simple command. Mostly I won't need to make the formal distinction; it
sometimes helps when you need to break down a complicated set of
commands. It's a good idea, and usually possible, to write in such a way
that it's obvious how the commands break down. It's not too important to
know the details, as long as you've got a feel for how the shell finds
the next command.</p>
<p><span id="l68"></span></p>
<h3 id="382-structures"><a class="header" href="#382-structures">3.8.2: Structures</a></h3>
<p>I've shown plenty of examples of one sort of shell structure already,
the <code>if</code> statement:</p>
<pre><code> if [[ black = white ]]; then
print Yellow is no colour.
fi
</code></pre>
<p>The main points are: the `<code>if</code>' itself is followed by some command
whose return status is tested; a `<code>then</code>' follows as a new command; any
number of commands may follow, as complex as you like; the whole
sequence is ended by a `<code>fi</code>' as a command on its own. You can write
the `<code>then</code>' on a new line if you like, I just happen to find it neater
to stick it where it is. If you follow the form here, remember the
semicolon before it; the <code>then</code> must start a separate command. (You can
put another command immediately after the <code>then</code> without a newline or
semicolon, though, although people tend not to.)</p>
<p>The double-bracketed test is by far the most common thing to put here in
zsh, as in ksh, but any command will do; only the status is important.</p>
<pre><code> if true; then
print This always gets executed
fi
if false; then
print This never gets executed
fi
</code></pre>
<p>Here, <code>true</code> always returns true (status 0), while <code>false</code> always
returns false (status 1 in zsh, although some versions return status 255
--- anything nonzero will do). So the statements following the <code>print</code>s
are correct.</p>
<p>The <code>if</code> construct can be extended by `<code>elif</code>' and `<code>else</code>':</p>
<pre><code> read var
if [[ $var = yes ]]; then
print Read yes
elif [[ $var = no ]]; then
print Read no
else
print Read something else
fi
</code></pre>
<p>The extension is pretty straightforward. You can have as many `<code>elif</code>'s
with different tests as you like; the code following the first test to
succeed is executed. If no test succeeded, and there is an `<code>else</code>'
(there doesn't need to be), the code following that is executed. Note
that the form of the `<code>elif</code>' is identical to that of `<code>if</code>',
including the `<code>then</code>', while the else just appears on its own.</p>
<p>The <code>while</code>-loop is quite similar to <code>if</code>. There are two differences:
the syntax uses <code>while</code>, <code>do</code> and <code>done</code> instead of <code>if</code>, <code>then</code> and
<code>fi</code>, and after the loop body is executed (if it is), the test is
evaluated again. The process stops as soon as the test is false. So</p>
<pre><code> i=0
while (( i++ < 3 )); do
print $i
done
</code></pre>
<p>prints 1, then 2, then 3. As with <code>if</code>, the commands in the middle can
be any set of zsh commands, so</p>
<pre><code> i=0
while (( i++ < 3 )); do
if (( i & 1 )); then
print $i is odd
else
print $i is even
fi
done
</code></pre>
<p>tells you that 1 and 3 are odd while 2 is even. Remember that the
indentation is irrelevant; it is purely there to make the structures
more easy to understand. You can write the code on a single line by
replacing all the newlines with semicolons.</p>
<p>There is also an <code>until</code> loop, which is identical to the <code>while</code> loop
except that the loop is executed until the test is true. `<code>until [[</code><em>...</em>' is equivalent to `<code>while ! [[</code><em>...</em>'.</p>
<p>Next comes the <code>for</code> loop. The normal case can best be demonstrated by
another example:</p>
<pre><code> for f in one two three; do
print $f
done
</code></pre>
<p>which prints out `<code>one</code>' on the first iteration, then `<code>two</code>', then
`<code>three</code>'. The <code>f</code> is set to each of the three words in turn, and the
body of the loop executed for each. It is very useful that the words
after the `<code>in</code>' may be anything you would normally have on a shell
command line. So `<code>for f in *; do</code>' will execute the body of the loop
once for each file in the current directory, with the file available as
<code>$f</code>, and you can use arrays or command substitutions or any other kind
of substitution to generate the words to loop over.</p>
<p>The <code>for</code> loop is so useful that the shell allows a shorthand that you
can use on the command line: try</p>
<pre><code> for f in *; print $f
</code></pre>
<p>and you will see the files in the current directory printed out, one per
line. This form, without the <code>do</code> and the <code>done</code>, involves less typing,
but is also less clear, so it is recommended that you only use it
interactively, not in scripts or functions. You can turn the feature off
with <code>NO_SHORT_LOOPS</code>.</p>
<p>The <code>case</code> statement is used to test a pattern against a series of
possibilities until one succeeds. It is really a short way of doing a
series of <code>if</code> and <code>elif</code> tests on the same pattern:</p>
<pre><code> read var
case $var in
(yes) print Read yes
;;
(no) print Read no
;;
(*) print Read something else
;;
esac
</code></pre>
<p>is identical to the <code>if</code>/<code>elif</code>/<code>else</code> example above. The <code>$var</code> is
compared against each pattern in turn; if one matches, the code
following that is executed --- then the statement is exited; no further
matches are looked for. Hence the `<code>*</code>' at the end, which can match
anything, acts like the `<code>else</code>' of an <code>if</code> statement.</p>
<p>Note the quirks of the syntax: the pattern to test must appear in
parentheses. For historical reasons, you can miss out the left
parenthesis before the pattern. I haven't done that mainly because
unbalanced parentheses confuse the system I am using for writing this
guide. Also, note the double semicolon: this is the only use of double
semicolons in the shell. That explains the fact that if you type `<code>;;</code>'
on its own the shell will report a `parse error'; it couldn't find a
<code>case</code> to associate it with.</p>
<p>You can also use alternative patterns by separating them with a vertical
bar. Zsh allows alternatives with extended globbing anyway; but this is
actually a separate feature, which is present in other shells which
don't have zsh's extended globbing feature; it doesn't depend on the
<code>EXTENDED_GLOB</code> option:</p>
<pre><code> read var
case $var in
(yes|true|1) print Reply was affirmative
;;
(no|false|0) print Reply was negative
;;
(*) print Reply was cobblers
;;
esac
</code></pre>
<p>The first `<code>print</code>' is used if the value of <code>$var</code> read in was
`<code>yes</code>', `<code>true</code>' or `<code>1</code>', and so on. Each of the separate items can
be a pattern, with any of the special characters allowed by zsh, this
time depending on the setting of the option <code>EXTENDED_GLOB</code>.</p>
<p>The <code>select</code> loop is not used all that often, in my experience. It is
only useful with interactive input (though the code may certainly appear
in a script or function):</p>
<pre><code> select var in earth air fire water; do
print You selected $var
done
</code></pre>
<p>This prints a menu; you must type 1, 2, 3 or 4 to select the
corresponding item; then the body of the loop is executed with <code>$var</code>
set to the value in the list corresponding to the number. To exit the
loop hit the break key (usually <code>^G</code>) or end of file (usually <code>^D</code>: the
feature is so infrequently used that currently there is a bug in the
shell that this tells you to use `<code>exit</code>' to exit, which is nonsense).
If the user entered a bogus value, then the loop is executed with <code>$var</code>
set to the empty string, though the actual input can be retrieved from
<code>$REPLY</code>. Note that the prompt printed for the user input is <code>$PROMPT3</code>,
the only use of this parameter in the shell: all normal prompt
substitutions are available.</p>
<p>There is one final type of loop which is special to zsh, unlike the
others above. This is `<code>repeat</code>'. It can be used two ways:</p>
<pre><code> % repeat 3 print Hip Hip Hooray
Hip Hip Hooray
Hip Hip Hooray
Hip Hip Hooray
</code></pre>
<p>Here, the first word after <code>repeat</code> is a count, which could be a
variable as normal substitutions are performed. The rest of the line (or
until the first semicolon) is a command to repeat; it is executed
identically each time.</p>
<p>The second form is a fully fledged loop, just like <code>while</code>:</p>
<pre><code> % repeat 3; do
repeat> print Hip Hip Hooray
repeat> done
Hip Hip Hooray
Hip Hip Hooray
Hip Hip Hooray
</code></pre>
<p>which has the identical effect to the previous one. The `<code>repeat></code>' is
the shell's prompt to show you that it is parsing the contents of a
`<code>repeat</code>' loop.</p>
<p><span id="l69"></span></p>
<h3 id="383-subshells-and-current-shell-constructs"><a class="header" href="#383-subshells-and-current-shell-constructs">3.8.3: Subshells and current shell constructs</a></h3>
<p>More catching up with stuff you've already seen. The expression in
parentheses here:</p>
<pre><code> % (cd ~; ls)
<all the files in my home directory>
% pwd
<where I was before, not necessarily ~>
</code></pre>
<p>is run in a subshell, as if it were a script. The main difference is
that the shell inherits almost everything from the main shell in which
you are typing, including options settings, functions and parameters.
The most important thing it doesn't inherit is probably information
about jobs: if you run <code>jobs</code> in a subshell, you will get no output; you
can't use <code>fg</code> to resume a job in a subshell; you can't use `<code>kill %</code><em>n</em>' to kill a job (though you can still use the process ID); and so
on. By now you should have some feel for the effect of running in a
separate process. Running a command, or set of commands, in a different
directory, as in this example, is one quite common use for this
construct. (In zsh 4.1, you can use <code>jobs</code> in a subshell; it lists the
jobs running in the parent shell; this is because it is very useful to
be able to pipe the output of jobs into some processing loop.)</p>
<p>On the other hand, the expression in braces here:</p>
<pre><code> % {cd ~; ls}
<all the files in my home directory>
% pwd
/home/pws
</code></pre>
<p>is run in the current shell. This is what I was blathering on about in
the section on redirection. Indeed, unless you need some special effect
like redirecting a whole set of commands, you won't use the
current-shell construct. The example here would behave just the same way
if the braces were missing.</p>
<p>As you might expect, the syntax of the subshell and current-shell forms
is very similar. You can use redirection with both, just as with simple
commands, and they can appear in most places where a simple command can
appear:</p>
<pre><code> [[ $test = true ]] && {
print Hello.
print Well, this is exciting.
}
</code></pre>
<p>That would be much clearer using an `<code>if</code>', but it works. For some
reason, you often find expressions of this form in system start-up files
located in the directory <code>/etc/rc.d</code> or, on older systems, in files
whose names begin with `<code>/etc/rc.</code>'. You can even do:</p>
<pre><code> if { foo=bar; [[ $foo = bar ]] }; then
print yes
fi
</code></pre>
<p>but that's also pretty gross.</p>
<p>One use for <code>{</code><em>...</em><code>}</code> is to make sure a whole set of commands is
executed at once. For example, if you copy a set of commands from a
script in one window and want them to be run in one go in a shell in
another window, you can do:</p>
<pre><code> % {
cursh> # now paste your commands in here...
...
cursh> }
</code></pre>
<p>and the commands will only be executed when you hit return after the
final `<code>}</code>'. This is also a workaround for some systems where cut and
paste has slightly odd effects due to the way different states of the
terminal are handled. The current-shell construct is a little bit like
an anonymous function, although it doesn't have any of the usual
features of functions --- you can't pass it arguments, and variables
declared inside aren't local to that section of code.</p>
<p><span id="l70"></span></p>
<h3 id="384-subshells-and-current-shells"><a class="header" href="#384-subshells-and-current-shells">3.8.4: Subshells and current shells</a></h3>
<p>In case you're confused about what happens in the current shell and what
happens in a subshell, here's a summary.</p>
<p>The following are run in the current shell.</p>
<ol>
<li>All shell builtins and anything which looks like one, such as a
precommand modifier and tests with `<code>[[</code>'.</li>
<li>All complex statements and loops such as <code>if</code> and <code>while</code>. Tests and
code inside the block must both be considered separately.</li>
<li>All shell functions.</li>
<li>All files run by `<code>source</code>' or `<code>.</code>' as well as startup files.</li>
<li>The code inside a `<code>{</code><em>...</em><code>}</code>'.</li>
<li>The right hand side of a pipeline: this is guaranteed in zsh, but
don't rely on it for other shells.</li>
<li>All forms of substitution except <code>`</code><em>...</em><code>`</code>, <code>$</code>(<em>...</em>),
<code>=</code>(<em>...</em>), <code><</code>(<em>...</em>) and <code>></code>(<em>...</em>).</li>
</ol>
<p>The following are run in a subshell.</p>
<ol>
<li>All external commands.</li>
<li>Anything on the left of a pipe, i.e. all sections of a pipeline but
the last.</li>
<li>The code inside a `<code> </code>(<em>...</em>)'.</li>
<li>Substitutions involving execution of code, i.e. <code>`</code><em>...</em><code>`</code>,
<code>$</code>(<em>...</em>), <code>=</code>(<em>...</em>), <code><</code>(<em>...</em>) and <code>></code>(<em>...</em>). (TCL fans note
that this is different from the `<code>[</code><em>...</em><code>]</code>' command substitution
in that language.)</li>
<li>Anything started in the background with `<code>&</code>' at the end.</li>
<li>Anything which has ever been suspended. This is a little subtle:
suppose you execute a set of commands in the current shell and
suspend it with <code>^Z</code>. Since the shell needs to return you to the
prompt, it forks a subshell to remember the commands it was
executing when you interrupted it. If you use <code>fg</code> or <code>bg</code> to
restart, the commands will stay in the subshell. This is a special
feature of zsh; most shells won't let you interrupt anything in the
current shell like that, though you can still abort it with <code>^C</code>.</li>
</ol>
<p>With an alias, you can't tell where it will be executed --- you need to
find out what it expands too first. The expansion naturally takes place
in the current shell.</p>
<p>Of course, if for some reason the current set of commands is already
running in a subshell, it doesn't get magically returned to the current
shell --- so a shell builtin on the left hand side of a pipeline is
running in a subshell. However, it doesn't get an extra subshell, as an
external command would. What I mean is:</p>
<pre><code> { print Hello; cat file } |
while read line; print $line; done
</code></pre>
<p>The shell forks, producing a subshell, to execute the left hand side of
the pipeline, and that subshell forks to execute the <code>cat</code> external
command, but nothing else in that set of commands will cause a new
subshell to be created.</p>
<p>(For the curious only: actually, that's not quite true, and I already
pointed this out when I talked about command substitutions: the shell
keeps track of occasions when it is in a subshell and has no more
commands to execute. In this case it will not bother forking to create a
new process for the <code>cat</code>, it will simply replace the subshell which is
not needed any more. This can only happen in simple cases where the
shell has no clearing up to do.)</p>
<p><span id="l71"></span></p>
<h2 id="39-emulation-and-portability"><a class="header" href="#39-emulation-and-portability">3.9: Emulation and portability</a></h2>
<p>I described the options you need to set for compatibility with ksh in
the previous chapter. Here I'm more interested in the best way of
running ksh scripts and functions.</p>
<p>First, you should remember that because of all zsh's options you can't
assume that a piece of zsh code will simply run a piece of sh or ksh
code without any extra changes. Our old friend <code>SH_WORD_SPLIT</code> is the
most common problem, but there are plenty of others. In addition to
options, there are other differences which simply need to be worked
around. I will list some of them a bit later. Generally speaking, Bourne
shell is simple enough that zsh emulates it pretty well --- although
beware in case you are using bash extensions, since to many Linux users
bash is the nearest approximation to the Bourne shell they ever come
across. Zsh makes no attempt to emulate bash, even though some of bash's
features have been incorporated.</p>
<p>To make zsh emulate ksh or sh as closely as it knows how, there are
various things you can do.</p>
<ol>
<li>
<p>Invoke zsh under the name sh or ksh, as appropriate. You can do this
by creating a symbolic link from zsh to sh or ksh. Then when zsh
starts up all the options will be set appropriately. If you are
starting that shell from another zsh, you can use the feature of zsh
that tricks a programme into thinking it has a different name:
`<code>ARGV0=sh zsh</code>' runs zsh under the name sh, just like the symbolic
link method.</p>
</li>
<li>
<p>Use `<code>emulate ksh</code>' at the top of the script or function you want
to run. In the case of a function, it is better to run `<code>emulate -L ksh</code>' since this makes sure the normal options will be restored when
the function exits; this is irrelevant for a script as the options
cannot be propagated to the process which ran the script. You can
also use the option `<code>-R</code>' after <code>emulate</code>, which forces more
options to be like ksh; these extra options are generally for user
convenience and not relevant to basic syntax, but in some cases you
may want the extra cover provided.</p>
<p>If it's possible the script may already be running under ksh, you
can instead use</p>
<pre><code> [[ -z $ZSH_VERSION ]] && emulate ksh
</code></pre>
<p>or for sh, using the simpler test command there,</p>
<pre><code> [ x$ZSH_VERSION = x ] && emulate sh
</code></pre>
</li>
</ol>
<p>Both these methods have drawbacks, and if you plan to be a heavy zsh
user there's no substitute for simply getting used to zsh's own basic
syntax. If you think there is some useful element of emulation we
missed, however, you should certainly tell the zsh-workers mailing list
about it.</p>
<p>Emulation of ksh88 is much better than emulation of ksh93. Support for
the latter is gradually being added, but only patchily.</p>
<p>There is no easy way of converting code written for any csh-like shell;
you will just have to convert it by hand. See the FAQ for some hints on
converting aliases to functions.</p>
<p><span id="l72"></span></p>
<h3 id="391-differences-in-detail"><a class="header" href="#391-differences-in-detail">3.9.1: Differences in detail</a></h3>
<p>Here are some differences from ksh88 which might prove significant for
ksh programmers. This is lifted straight from the corresponding section
of the FAQ; it is not complete, and indeed some of the `differences'
could be interpreted as bugs. Those marked `*' perform in a ksh-like
manner if the shell is invoked with the name `ksh', or if `emulate
ksh' is in effect.</p>
<ul>
<li>Syntax:
<ul>
<li>* Shell word splitting.</li>
<li>* Arrays are (by default) more csh-like than ksh-like:
subscripts start at 1, not 0; <code>array[0]</code> refers to <code>array[1]</code>;
<code>$array</code> refers to the whole array, not <code>$array[0]</code>; braces are
unnecessary: <code>$a[1] == ${a[1]}</code>, etc. The <code>KSH_ARRAYS</code> option is
now available.</li>
<li>Coprocesses are established by <code>coproc</code>; <code>|&</code> behaves like csh.
Handling of coprocess file descriptors is also different.</li>
<li>In <code>cmd1 && cmd2 &</code>, only <code>cmd2</code> instead of the whole expression
is run in the background in zsh. The manual implies this is a
bug. Use <code>{ cmd1 && cmd2 } &</code> as a workaround.</li>
</ul>
</li>
<li>Command line substitutions, globbing etc.:
<ul>
<li>
<p>* Failure to match a globbing pattern causes an error (use
<code>NO_NOMATCH</code>).</p>
</li>
<li>
<p>* The results of parameter substitutions are treated as plain
text: <code>foo="*"; print $foo</code> prints all files in ksh but <code>*</code> in
zsh (unset <code>GLOB_SUBST</code>).</p>
</li>
<li>
<p>* <code>$PSn</code> do not do parameter substitution by default (use
<code>PROMPT_SUBST</code>).</p>
</li>
<li>
<p>* Standard globbing does not allow ksh-style `pattern-lists'.
See <a href="zshguide05.html#subst">chapter 5</a> for a list of equivalent
zsh forms. The <code>^</code>, <code>~</code> and <code>#</code> (but not <code>|</code>) forms require
<code>EXTENDED_GLOB</code>. From version 3.1.3, the ksh forms are fully
supported when the option <code>KSH_GLOB</code> is in effect.</p>
<p>[1] Note that <code>~</code> is the only globbing operator to have a
lower precedence than <code>/</code>. For example, <code>**/foo~*bar*</code> matches
any file in a subdirectory called <code>foo</code>, except where <code>bar</code>
occurred somewhere in the path (e.g. <code>users/barstaff/foo</code> will
be excluded by the <code>~</code> operator). As the <code>**</code> operator cannot be
grouped (inside parentheses it is treated as <code>*</code>), this is the
way to exclude some subdirectories from matching a <code>**</code>.</p>
</li>
<li>
<p>Unquoted assignments do file expansion after colons (intended
for PATHs).</p>
</li>
<li>
<p><code>integer</code> does not allow <code>-i</code>.</p>
</li>
<li>
<p><code>typeset</code> and <code>integer</code> have special behaviour for assignments
in ksh, but not in zsh. For example, this doesn't work in zsh:</p>
<pre><code> integer k=$(wc -l ~/.zshrc)
</code></pre>
<p>because the return value from <code>wc</code> includes leading whitespace
which causes wordsplitting. Ksh handles the assignment specially
as a single word.</p>
</li>
</ul>
</li>
<li>Command execution:
<ul>
<li>* There is no <code>$ENV</code> variable (use <code>/etc/zshrc</code>, <code>~/.zshrc</code>;
note also <code>$ZDOTDIR</code>).</li>
<li><code>$PATH</code> is not searched for commands specified at invocation
without -c.</li>
</ul>
</li>
<li>Aliases and functions:
<ul>
<li>The order in which aliases and functions are defined is
significant: function definitions with () expand aliases.</li>
<li>Aliases and functions cannot be exported.</li>
<li>There are no tracked aliases: command hashing replaces these.</li>
<li>The use of aliases for key bindings is replaced by `bindkey'.</li>
<li>* Options are not local to functions (use LOCAL_OPTIONS; note
this may always be unset locally to propagate options settings
from a function to the calling level).</li>
</ul>
</li>
<li>Traps and signals:
<ul>
<li>* Traps are not local to functions. The option LOCAL_TRAPS is
available from 3.1.6.</li>
<li>TRAPERR has become TRAPZERR (this was forced by UNICOS which has
SIGERR).</li>
</ul>
</li>
<li>Editing:
<ul>
<li>The options <code>emacs</code>, <code>gmacs</code>, <code>viraw</code> are not supported. Use
bindkey to change the editing behaviour: <code>set -o {emacs,vi}</code>
becomes <code>bindkey -{e,v}</code>; for gmacs, go to emacs mode and use
<code>bindkey \^t gosmacs-transpose-characters</code>.</li>
<li>The <code>keyword</code> option does not exist and <code>-k</code> is instead
interactivecomments. (<code>keyword</code> will not be in the next ksh
release either.)</li>
<li>Management of histories in multiple shells is different: the
history list is not saved and restored after each command. The
option <code>SHARE_HISTORY</code> appeared in 3.1.6 and is set in ksh
compatibility mode to remedy this.</li>
<li><code>\</code> does not escape editing chars (use <code>^V</code>).</li>
<li>Not all ksh bindings are set (e.g. <code><ESC>#</code>; try <code><ESC>q</code>).</li>
<li>* <code>#</code> in an interactive shell is not treated as a comment by
default.</li>
</ul>
</li>
<li>Built-in commands:
<ul>
<li>Some built-ins (<code>r</code>, <code>autoload</code>, <code>history</code>, <code>integer</code> ...) were
aliases in ksh.</li>
<li>There is no built-in command newgrp: use e.g. <code>alias newgrp="exec newgrp"</code></li>
<li><code>jobs</code> has no <code>-n</code> flag.</li>
<li><code>read</code> has no <code>-s</code> flag.</li>
</ul>
</li>
<li>Other idiosyncrasies:
<ul>
<li><code>select</code> always redisplays the list of selections on each loop.</li>
</ul>
</li>
</ul>
<p><span id="l73"></span></p>
<h3 id="392-making-your-own-scripts-and-functions-portable"><a class="header" href="#392-making-your-own-scripts-and-functions-portable">3.9.2: Making your own scripts and functions portable</a></h3>
<p>There are also problems in making your own scripts and functions
available to other people, who may have different options set.</p>
<p>In the case of functions, it is always best to put `<code>emulate -L zsh</code>'
at the top of the function, which will reset the options to the default
zsh values, and then set any other necessary options. It doesn't take
the shell a great deal of time to process these commands, so try and get
into the habit of putting them any function you think may be used by
other people. (Completion functions are a special case as the
environment is already standardised --- see <a href="zshguide06.html#comp">chapter
6</a> for this.)</p>
<p>The same applies to scripts, since if you run the script without using
the option `<code>-f</code>' to zsh the user's non-interactive startup files will
be run, and in any case the file <code>/etc/zshenv</code> will be run. We urge
system administrators not to set options unconditionally in that file
unless absolutely necessary; but they don't always listen. Hence an
<code>emulate</code> can still save a lot of grief.</p>
<p><span id="l74"></span></p>
<h2 id="310-running-scripts"><a class="header" href="#310-running-scripts">3.10: Running scripts</a></h2>
<p>Here are some final comments on running scripts: they apply regardless
of the problems of portability, but you should certainly also be aware
of what I was saying in the previous section.</p>
<p>You may be aware that you can force the operating system to run a script
using a particular interpreter by putting `<code>#!</code>' and the path to the
interpreter at the top of the script. For example, a zsh script could
start with</p>
<pre><code> #!/usr/local/bin/zsh
print The arguments are $*
</code></pre>
<p>assuming that zsh lives in the directory <code>/usr/local/bin</code>. Then you can
run the script under its name as if it were an ordinary command. Suppose
the script were called `<code>scriptfile</code>' and in the current directory, and
you want to run it with the arguments `<code>one two forty-three</code>'. First
you must make sure the script is executable:</p>
<pre><code> % chmod +x scriptfile
</code></pre>
<p>and then you can run it with the arguments:</p>
<pre><code> % ./scriptfile one two forty-three
The arguments are one two forty-three
</code></pre>
<p>The shell treats the first line as a comment, since it begins with a
`<code>#</code>', but note it still gets evaluated by the shell; the system simply
looks inside the file to see if what's there, it doesn't change it just
because the first line tells it to execute the shell.</p>
<p>I put the `<code>./</code>' in front to refer to the current directory because I
don't usually have that in my path --- this is for safety, to avoid
running things which happen to have names like commands simply because
they were in the current directory. But many people aren't so paranoid,
and if `<code>.</code>' is in your path, you can omit the `<code>./</code>'. Hence,
obviously, it can be anywhere else in your path: it is searched for as
an ordinary executable.</p>
<p>The shell actually provides this mechanism even on operating systems
(now few and far between in the UNIX world) that don't have the feature
built into them. The way this works is that if the shell found the file,
and it was executable, but running it didn't work, then it will look for
the <code>#!</code>, extract the name following and run (in this example)
`<code>/usr/local/bin/zsh</code> <em><path></em>/scriptfile <code>one two forty-three</code>',
where <em><path></em> is the path where the file was found. This is, in fact,
pretty much what the system does if it handles it itself.</p>
<p>Some shells search for scripts using the path when they are given as
filenames at invocation, but zsh happens not to. In other words, `<code>zsh scriptfile</code>' only runs <code>scriptfile</code> in the current directory.</p>
<p>There are two other features you may want to be aware of. Both are down
to the operating system, if that is what is responsible for the `<code>#!</code>'
trick (true of all the most common UNIX-like systems at the moment).
First, you are usually allowed to supply one, but only one, argument or
option in the `<code>#!</code>' line, thus:</p>
<pre><code> #!/usr/local/bin/zsh -f
print other stuff here
</code></pre>
<p>which stops startup files other than <code>/etc/zshenv</code> from being run, but
otherwise works the same as before. If you need more options, you should
combine them in the same word. However, it's usually clearer, for
anything apart from <code>-f</code>, <code>-i</code> (which forces the shell into interactive
mode) and a few other options which need to take effect immediately, to
put a `<code>setopt</code>' line at the start of the body of the script. In a few
versions of zsh, there was an unexpected consequence of the fact that
the line would only be split once: if you accidentally left some spaces
at the end of the line (e.g. `<code>#!/usr/local/bin/zsh -f </code>') they would
be passed down to the shell, which would report an error, which was hard
to interpret. The spaces will still usually be passed down, but the
shell is now smart enough to ignore spaces in an option list.</p>
<p>The second point is that the length of the `<code>#!</code>' line which will be
evaluated is limited. Often the limit is 32 characters, in total, That
means if your path to zsh is long, e.g.
`<code>/home/users/psychology/research/dreams/freud/solaris_2.5/bin/zsh</code>'
the system won't be able to find the shell. Your only recourse is to
find a shorter path, or execute the shell directly, or some sneakier
trick such as running the script under <code>/bin/sh</code> and making that start
zsh when it detects that zsh isn't running yet. That's a fairly nasty
way of doing it, but just in case you find it necessary, here's an
example:</p>
<pre><code> #!/bin/sh
if [ x$ZSH_VERSION = x ]; then
# Put the right path in here ---
# or just rely on finding zsh in
# $path, since `exec' handles that.
exec /usr/local/bin/zsh $0 "$@"
fi
print $ZSH_VERSION
print Hello, this is $0
print with arguments $*.
</code></pre>
<p>Note that first `<code>$0</code>', which passes down the name of the script that
was originally executed. Running this as `<code>testexec foo bar</code>' gives me</p>
<pre><code> 3.1.9-dev-8
Hello, this is /home/pws/tmp/testexec
with arguments foo bar.
</code></pre>
<p>I hope you won't have to resort to that. By the way, really,
excruciatingly old versions of zsh didn't have <code>$ZSH_VERSION</code>. Rather
than fix the script, I suggest you upgrade the shell. Also, on some old
Bourne shells you might need to replace <code>"$@"</code> with <code>${1+"$@"}</code>, which
is more careful about only putting in arguments if there were any (this
is the sort of thing we'll see in <a href="zshguide05.html#subst">chapter 5</a>).
Usually this isn't necessary.</p>
<p>You can use the same trick on ancient versions of UNIX which didn't
handle `<code>#!</code>'. On some such systems, anything with a `<code>:</code>' as the
first character is run with the Bourne shell, so this serves as an
alternative to `<code>#!/bin/sh</code>', while on some Berkeley systems, a plain
`<code>#</code>' caused csh to be used. In the second case, you will need to
change the syntax of the first test to be understood by both zsh and
csh. I'll leave that as an exercise for the reader. If you have perl
(very probable these days) you can look at the <code>perlrun</code> manual page,
which discusses the corresponding problem of starting perl scripts from
a shell, for some ideas.</p>
<p>There's one other glitch you may come across. Sometimes if you type the
name of a script which you know is in your path and is executable, the
shell may tell you `<code>file not found</code>', or some equivalent message. What
this usually means is that the <em>interpreter</em> wasn't found, because you
mistyped the line after the `<code>#!</code>'. This confusing message isn't the
shell's fault: a lot of operating systems return the same system error
in this case as if the script were really not found. It's not worth the
shell searching the path to see if the script is there, because in the
vast majority of cases the error refers to the programme in the
execution path. If the operating system returned the more natural error,
`<code>exec format error</code>', then the shell would know that there was
something wrong with the file, and could investigate; but unfortunately
life's not that simple.</p>
<div id="chapter_begin" style="break-before: page; page-break-before: always;"></div><!-- START doctoc generated TOC please keep comment here to allow auto update -->
<!-- DON'T EDIT THIS SECTION, INSTEAD RE-RUN doctoc TO UPDATE -->
<p><strong>Table of Contents</strong> <em>generated with <a href="https://github.com/thlorenz/doctoc">DocToc</a></em></p>
<ul>
<li><a href="zshguide04.html#chapter-4-the-z-shell-line-editor">Chapter 4: The Z-Shell Line Editor</a>
<ul>
<li><a href="zshguide04.html#41-introducing-zle">4.1: Introducing zle</a>
<ul>
<li><a href="zshguide04.html#411-the-simple-facts">4.1.1: The simple facts</a></li>
<li><a href="zshguide04.html#412-vi-mode">4.1.2: Vi mode</a></li>
</ul>
</li>
<li><a href="zshguide04.html#42-basic-editing">4.2: Basic editing</a>
<ul>
<li><a href="zshguide04.html#421-moving">4.2.1: Moving</a></li>
<li><a href="zshguide04.html#422-deleting">4.2.2: Deleting</a></li>
<li><a href="zshguide04.html#423-more-deletion">4.2.3: More deletion</a></li>
</ul>
</li>
<li><a href="zshguide04.html#43-fancier-editing">4.3: Fancier editing</a>
<ul>
<li><a href="zshguide04.html#431-options-controlling-zle">4.3.1: Options controlling zle</a></li>
<li><a href="zshguide04.html#432-the-minibuffer-and-extended-commands">4.3.2: The minibuffer and extended commands</a></li>
<li><a href="zshguide04.html#433-prefix-digit-arguments">4.3.3: Prefix (digit) arguments</a></li>
<li><a href="zshguide04.html#434-words-regions-and-marks">4.3.4: Words, regions and marks</a></li>
<li><a href="zshguide04.html#435-regions-and-marks">4.3.5: Regions and marks</a></li>
</ul>
</li>
<li><a href="zshguide04.html#44-history-and-searching">4.4: History and searching</a>
<ul>
<li><a href="zshguide04.html#441-moving-through-the-history">4.4.1: Moving through the history</a></li>
<li><a href="zshguide04.html#442-searching-through-the-history">4.4.2: Searching through the history</a></li>
<li><a href="zshguide04.html#443-extracting-words-from-the-history">4.4.3: Extracting words from the history</a></li>
</ul>
</li>
<li><a href="zshguide04.html#45-binding-keys-and-handling-keymaps">4.5: Binding keys and handling keymaps</a>
<ul>
<li><a href="zshguide04.html#451-simple-key-bindings">4.5.1: Simple key bindings</a></li>
<li><a href="zshguide04.html#452-removing-key-bindings">4.5.2: Removing key bindings</a></li>
<li><a href="zshguide04.html#453-function-keys-and-so-on">4.5.3: Function keys and so on</a></li>
<li><a href="zshguide04.html#454-binding-strings-instead-of-commands">4.5.4: Binding strings instead of commands</a></li>
<li><a href="zshguide04.html#455-keymaps">4.5.5: Keymaps</a></li>
</ul>
</li>
<li><a href="zshguide04.html#46-advanced-editing">4.6: Advanced editing</a>
<ul>
<li><a href="zshguide04.html#461-multi-line-editing">4.6.1: Multi-line editing</a></li>
<li><a href="zshguide04.html#462-the-builtin-vared-and-the-function-zed">4.6.2: The builtin vared and the function zed</a></li>
<li><a href="zshguide04.html#463-the-buffer-stack">4.6.3: The buffer stack</a></li>
</ul>
</li>
<li><a href="zshguide04.html#47-extending-zle">4.7: Extending zle</a>
<ul>
<li><a href="zshguide04.html#471-widgets">4.7.1: Widgets</a></li>
<li><a href="zshguide04.html#472-executing-other-widgets">4.7.2: Executing other widgets</a></li>
<li><a href="zshguide04.html#473-some-special-builtin-widgets-and-their-uses">4.7.3: Some special builtin widgets and their uses</a></li>
<li><a href="zshguide04.html#474-special-parameters-normal-text">4.7.4: Special parameters: normal text</a></li>
<li><a href="zshguide04.html#475-other-special-parameters">4.7.5: Other special parameters</a></li>
<li><a href="zshguide04.html#476-reading-keys-and-using-the-minibuffer">4.7.6: Reading keys and using the minibuffer</a></li>
<li><a href="zshguide04.html#477-examples">4.7.7: Examples</a></li>
</ul>
</li>
</ul>
</li>
</ul>
<!-- END doctoc generated TOC please keep comment here to allow auto update -->
<p><span id="zle"></span><span id="l75"></span></p>
<h1 id="chapter-4-the-z-shell-line-editor"><a class="header" href="#chapter-4-the-z-shell-line-editor">Chapter 4: The Z-Shell Line Editor</a></h1>
<p>The zsh line editor is probably the first part of the shell you ever
used, when you started typing in commands. Even the most basic shells,
such as sh, provide some kind of editing capability, although in that
case probably just what the system itself does --- enter characters,
delete the last character, delete the entire line. Most shells you're
likely to use nowadays do quite a lot more. With zsh you can even extend
the set of editor commands using shell functions.</p>
<p><span id="l76"></span></p>
<h2 id="41-introducing-zle"><a class="header" href="#41-introducing-zle">4.1: Introducing zle</a></h2>
<p>The zsh line editor is usually abbreviated to `zle'. Normally it fires
itself up for any interative shell; you don't have to do anything
special until you decide you need to change its behaviour. If everything
looks OK and you're not interested in how zle is started up, skip to the
next subsection.</p>
<p>Nowadays, zle lives in its own loadable module, <code>zsh/zle</code>, which saves
all the overhead of having an editor if the shell isn't interactive.
However, you normally won't need to worry about that; I'll say more
about modules in <a href="zshguide07.html#ragbag">chapter 7</a>, but the shell
knows when you need zle and gives you it automatically. Usually the
module is in a directory with a name like
`<code>/usr/local/lib/zsh/4.0.4/zsh/zle.so</code>', where the `<code>4.0.4</code>' is the
shell's version number, the same as the value of the parameter
<code>$ZSH_VERSION</code>, and everything after that apart from the suffix `<code>.so</code>'
is the module name. The suffix may be `<code>.sl</code>' (HP-UX) or `<code>.dll</code>'
(Cygwin), but `<code>.so</code>' is by far the most common form. It differs
because zsh keeps the same convention for dynamically loadable
libraries, or `shared objects' in UNIX-speak, as the operating system.</p>
<p>If the shell is badly installed, you sometimes see error messages that
it, or a command such as `bindkey', couldn't be loaded. That means the
shell couldn't find `<code>zsh/zle</code>' anywhere in the module load path, the
array <code>$module_path</code>. Then you need to complain to your system
administrator. If you've just compiled zsh and are having this problem,
it's because you have to install the modules before you run the shell,
even if the shell itself isn't installed. You can do that by saying
`<code>make install.modules</code>'. Then the compiled zsh should run from where
it is.</p>
<p>Note that unlike bash's line editor, readline, which is an entirely
separate library, zle is an integral part of the shell. Hence you
configure it by sticking commands in your <code>.zshrc</code> --- as it's only
useful for an interactive shell, only <code>/etc/zshrc</code> and <code>.zshrc</code> make
sense for this purpose.</p>
<p>One tip if you're looking at the zsh manual using info, either with the
command of that name or <code>\C-h i</code> within Emacs, which I find the most
convenient way: the entry for zle is called `Zsh Line Editor', in full,
not just `Zle'. Have fun looking for `Shell Builtin Commands' (not
`Builtins') while you're at it.</p>
<p><span id="l77"></span></p>
<h3 id="411-the-simple-facts"><a class="header" href="#411-the-simple-facts">4.1.1: The simple facts</a></h3>
<p>As with any editor later than <code>ed</code>, you can move around the line and
change it using various `keystrokes', in other words one or more sets
of keys you type at once. For example, the keystroke to move back a word
is (maybe) <code>ESC b</code>. This means you first hit the escape key; nothing
happens yet. Then you hit `b', and the cursor instantly jumps back to
the start of the word. (I'll have more to say on what zle thinks is a
`word' --- it's not necessarily the same as what the rest of the shell
thinks is a word.)</p>
<p>It will probably help if I introduce the shell's way of describing
keystrokes right away; then when you need to enter them you can just
copy them straight in. The escape key is `<code>\e</code>', so that keystroke
would be `<code>\eb</code>'. Other common keystrokes include holding down the
control key, probably near the bottom left of the keyboard, and typing
another key at the same time. The simplest way of indicate a control key
is just to put `<code>^</code>' in front; so for example `<code>^x^x</code>' means hold down
control, and press `x' twice with control still held down. It has
exactly the same effect as `<code>^X^X</code>'. (You may find each time you do
that it takes you to the start of the line and back to where you were.)</p>
<p>I've already introduced the weasel word `maybe' to try to avoid lying.
This is because actually zle has two modes of operation, one (the
default) like Emacs, the other like vi. If you don't know either of
those venerable UNIX editors, I suggest you stick to Emacs mode, since
it tends to interfere a little less with what you're doing, and
furthermore completion is a little easier. Completion is an offshoot of
zle behaviour which is described in <a href="zshguide06.html#comp">chapter 6</a>
(which, you will notice, is longer than this one).</p>
<p>If you normally use vi, you may have one or both of the environment
variables <code>$EDITOR</code> or <code>$VISUAL</code> set to `<code>vi</code>'. (It all works the same
way if you use `<code>vim</code>' instead, or any editor that happens to contain
`<code>vi</code>' such as `<code>elvis</code>'.) In that case, zle will start up in its
`<code>vi</code>' mode, where the keystrokes are rather different. That's why you
might have found that `<code>\eb</code>' didn't do what I said, even though you
had made no attempt to configure zle. You can make zle always use either
emacs or vi mode by putting either</p>
<pre><code> bindkey -e
</code></pre>
<p>or</p>
<pre><code> bindkey -v
</code></pre>
<p>in your <code>.zshrc</code>. This is just one of many uses of <code>bindkey</code>.</p>
<p>If you're not familiar with this use of the word `bind', it just means
`make a keystroke execute a particular editor command'. Commands have
long-winded names with hyphens which give you quite a good description
of what they do, such as `<code>backward-delete-char</code>'. Normal keys which
correspond to printable characters are usually `bound to
<code>self-insert</code>', a perverse way of saying they do what you expect and
show up the character you typed. However, you can actually bind them to
something else. In vi command mode, this is perfectly normal.</p>
<p>Actually, if you use a windowing system, you might want to say
`<code>bindkey -me</code>', which binds a whole set of `meta' keys. In X Windows,
one of the keys on your keyboard, possibly <code>ALT</code>, may be designated a
`meta' key which has a special effect similar to the control key.
Bindings with the meta key held down are described a bit like they are
in Emacs, `<code>\M-b</code>'. (You can specify control keys similarly, in fact,
like `<code>\C-x</code>', but `<code>^x</code>' is shorter.) Using the `<code>-m</code>' option to
<code>bindkey</code> tells zsh that wherever it binds an escape sequence like
`<code>\eb</code>', it should also bind the corresponding meta sequence like
`<code>\M-b</code>'. Emacs always ties these together, but zsh doesn't --- you can
rebind them separately. and if you want both sequences to be bound to a
new command, you have to bind them both explicitly.</p>
<p>You need to be careful with `<code>bindkey -m</code>', however; the shell can't
tell whether you are typing a character with the top bit set, or
executing a command. This is likely to become worse as the UTF-8
encoding for characters becomes more popular, since a non-ASCII
character then consists of a whole series of bytes with the top bit set.</p>
<p>If you are interested in binding function keys, you may already have
found the key sequences they send apparently don't make any sense; see
<a href="zshguide04.html#fkeys">the section below</a> for more information. This
will introduce the function called <code>zkbd</code> which can make the process
less painful. The function also helps with `meta' and `ALT' keys.</p>
<p><span id="l78"></span></p>
<h3 id="412-vi-mode"><a class="header" href="#412-vi-mode">4.1.2: Vi mode</a></h3>
<p>I'm going to concentrate on Emacs mode for various reasons: firstly,
because I use it myself; secondly, because the most likely reason for
you using vi mode is that you are already familiar with vi and don't
need to be told how it works; thirdly, because most of the commands are
the same in both modes, just bound differently; and finally because if
you <em>don't</em> already know vi, you will quite likely find vi editing mode
rather counterintuitive and difficult to use.</p>
<p>However, here are a few remarks on it just to get it out of the way.
Like the real vi editor, there are two basic modes, insert mode, where
you type in text, and command mode, where the same keystrokes which
insert characters are bound instead to editing commands. Unlike the real
vi, the line editor starts in insert mode on every new command you edit.
This means you can often simply type a line up to the `return' at the
end and forget you are in vi mode at all.</p>
<p>To enter command mode, you hit `escape', again just like normal vi. At
this point you are in the magic world of vi commands, where typing an
ordinary character can have any effect whatsoever. However, the bindings
are similar to normal vi, so `<code>h</code>' and `<code>l</code>' move left and right. When
you want to insert more text, you can use any of the normal vi commands
which allow you to do that, such as `<code>i</code>' (<code>vi-insert</code>) or `<code>a</code>'
(<code>vi-add-next</code>).</p>
<p>Apart from the separate command and insert modes and the completely
different set of key bindings, there is no basic difference between
Emacs mode and vi mode. You can bind keys in both the vi modes --- they
don't <em>have</em> to correspond to <code>self-insert</code> in insert mode. Below, I'll
describe `keymaps', a complete set of descriptions for what all the
keys will do (less impressive than it sounds, since a lot of keys may be
set to `<code>undefined-key</code>, which means they don't do anything useful),
and you will see how to change the behaviour in both modes.</p>
<p><span id="l79"></span></p>
<h2 id="42-basic-editing"><a class="header" href="#42-basic-editing">4.2: Basic editing</a></h2>
<p>If you know Emacs or vi, you will very soon find out how to do simple
commands like moving the cursor, going up and down the history list, and
deleting and copying words. If you don't, you should read the <code>zshzle</code>
manual page for a concise description of what it can do. Here is a
summary for Emacs mode.</p>
<p><span id="l80"></span></p>
<h3 id="421-moving"><a class="header" href="#421-moving">4.2.1: Moving</a></h3>
<p>You can move forwards and backwards along the line using the cursor
keys. The are a variety of different conventions as to what keystrokes
the cursor keys produce. You might naively expect that pressing, say,
cursor right, sends a signal along the lines of `cursor right' to the
application. Unfortunately, there is no such character in the ASCII
character set, so programmes which read input as a string of characters
like zsh have to be given an arbitrary string of characters. (It's
different for programmes which understand other forms of input, like
windowing systems.)</p>
<p>The two most common conventions for cursor keys are where the up key
sends `<code>\e[A</code>' and the other three the same with <code>B</code>, <code>C</code> and <code>D</code> at
the end, and the convention where the `<code>[</code>' is replaced by an `<code>O</code>'
(uppercase letter `O'). In old versions of zsh, the only convention
supported was the first of those two. The second, and any other
convention, were not supported at all and you had to bind the keys
yourself. This was done by something like:</p>
<pre><code> bindkey "\eOA" up-line-or-history
bindkey "\eOB" down-line-or-history
bindkey "\eOC" forward-char
bindkey "\eOD" backward-char
</code></pre>
<p>The shell tries harder now, and provided your system has the correct
information about your terminal (zsh uses an old system called
`termcap' which has largely been superseded by another called
`terminfo') you should be lucky. If the shell thinks your keys are too
perverse --- in particular, if the keystroke it wants to bind the
function too is already defined by zsh --- you will still have to do it
by hand. The list above should serve as a template.</p>
<p>Instead of the cursor keys, traditional Emacs keys are available: <code>^b</code>
and <code>^f</code> for backward and forward, <code>^p</code> and <code>^n</code> for previous line and
next line, so you can continue even if the cursor keys don't work.</p>
<p>Moving longer distances is done by <code>\eb</code> and <code>\ef</code> for a word backwards
or forwards (or, as you saw, <code>\M-b</code> and <code>\M-f</code>), and <code>^a</code> and <code>^e</code> for
the start and the end of the line. That just about exhausts the ones you
will use the most frequently.</p>
<p><span id="l81"></span></p>
<h3 id="422-deleting"><a class="header" href="#422-deleting">4.2.2: Deleting</a></h3>
<p>For deleting, backspace or the delete key will delete backwards. There
is an eternal battle over these keys owing to the fact that on PC
keyboards the key at the top left of the central keyboard section is
`backspace' which is the character <code>^h</code>, while on traditional UNIX
keyboards it is `delete' which is the character 127, often written as
<code>^?</code> (which zsh also understands). When you are in the system's own
primitive line editing mode, as with sh (unless your sh is really bash),
only one of these is `bound', although it's not really a key binding,
it's a translation made by the system's terminal driver, and it's
usually the wrong one. Hence you often find the system prints `<code>^h</code>' on
the screen when you want it to delete. You can change the key using</p>
<pre><code> stty erase '^h'
</code></pre>
<p>but zsh protects you from all that --- both <code>^h</code> (backspace) and <code>^?</code>
(delete) will delete backwards one character. Note, by the way, that zsh
doesn't understand smart names for any keystrokes -- if you try to bind
a key called `<code>backspace</code>' zsh will bind a command to that sequence of
characters, not a key of that name. See comments on `<code>bindkey -s</code>' for
when something like this might even be useful.</p>
<p>To confuse matters further, the key often marked as `Delete' on a 101-
or 102-key PC keyboard in the group of 6 above the cursor keys is
completely different again, and probably doesn't send either of those
sequences. On my keyboard it sends the sequence `<code>\e[3~</code>'. I find it
convenient to have this delete the next charater, which is its tradional
role in the PC world, which I do by</p>
<pre><code> bindkey '\e[3~' delete-char
</code></pre>
<p>However, the tradtional <em>Emacs</em> way of deleting the next character is to
use `<code>^d</code>', which zsh binds for you by default. If you look at the
binding, which you can do by not giving bindkey an editor command to
bind,</p>
<pre><code> % bindkey '^d'
delete-char-or-list
</code></pre>
<p>you'll see it doesn't <em>quite</em> do what I suggested. The `<code>-or-list</code>'
part is for completion, and you'll find out about it in the next
chapter. The first shell I know of to have this odd combination was
tcsh.</p>
<p>Since I enjoy confusion, I might as well point out that usually <code>^d</code> has
another use, which is to tell the terminal driver you've reached the end
of a file. In the case of, say, a file on a disk, the system knows this
by itself, but if you are supplying a stream of characters, the only way
of telling it is to send a special character. The default is usually
<code>^d</code>. You will notice that if you type `<code>^d</code>' at the start of the line,
you see the message</p>
<pre><code> zsh: use 'exit' to exit.
</code></pre>
<p>That's because zsh recognises the <code>^d</code> as end-of-file in that position.
By default the shell warns you; you can turn this off by setting the
option <code>IGNORE_EOF</code>. You can tell the system you don't ever want to send
an end-of-file in this way with <code>stty</code>, again: the following are
equivalent in Linux but your system way want one or the other:</p>
<pre><code> stty eof '^-'
stty eof undef
</code></pre>
<p>Remember that <code>stty</code> is not part of the shell; it's a way of controlling
the state of the system's terminal driver. This means it survives as
long as the terminal or terminal window is still connected, even if you
start a new shell or exit one that isn't the login shell.</p>
<p>By the way, if you need to refer to a character by its number, the
easiest way is probably to use the syntax `<code>\x??</code>', where the `<code>??</code>'
are the two hex digits for the key. In the case of delete, it is
`<code>\x7f</code>'. You can confirm this by:</p>
<pre><code> % bindkey '\x7f'
"^?" backward-delete-char
</code></pre>
<p><span id="l82"></span></p>
<h3 id="423-more-deletion"><a class="header" href="#423-more-deletion">4.2.3: More deletion</a></h3>
<p>You can delete larger areas with `<code>\ed</code>' to delete the next word and
`<code>\e^h</code>' or `<code>\e^?</code>' (escape followed by delete backwards) to delete
the previous word. `<code>^u</code>' usually removes the entire line, before and
after the cursor --- this is not like Emacs, where <code>^u</code> introduces digit
arguments as I will describe in the next subsection. It is, however,
like another of those primitive editing commands the terminal driver
itself provides, this one being known to <code>stty</code> as `<code>kill</code>'. The most
common use of this outside zsh is for deleting your password when you
login, when you know you've typed it wrong but can't see how many
!@?*! characters you've typed, and maybe can't rely on the terminal
agreeing with you as to which of <code>^h</code> or <code>^?</code> will delete a single one.</p>
<p>Strictly speaking, all the keystrokes in the previous paragraph perform
a `kill' (zsh-speak, not to be confused with the <code>stty</code> `kill') rather
than a `delete' (or deletion, as we used to say when we had a distinct
between nouning and verbing). The difference is the same as in Emacs ---
`killed' text is saved for later `yanking' back somewhere else, which
you do with the <code>^y</code> key, whereas `deleted' text as with <code>^?</code> and <code>^d</code>
is gone forever. This is what everyone not brought up under Emacs calls
`cut' and `paste' (although since Emacs dates back to the seventies,
it could be everyone else that's wrong). Another feature borrowed from
Emacs is that if you do multiple `kills' without any other editing in
between, the killed text is joined together and you can yank it all back
in one go. I will say more when I talk about point and mark (another
Emacs idea).</p>
<p>Actually, even deleted text isn't gone forever: zsh has an Emacs-like
editing history, and you can undo the previous commands on the line.
This is usually bound to <code>^xu</code> and <code>^x^u</code>, and there is shorter binding
which is described rather confusingly as `<code>^_</code>' --- confusingly,
because on all not-completely-spaced-out keyboards I've ever used you
actually generate that sequence by holding down control and pressing the
`<code>/</code>' key. Zsh doesn't use <code>^z</code> by default and, if you are used to
Windows, that is another suitable binding for <code>undo</code>.</p>
<p>Zsh scores in one way over Emacs --- it also has `<code>redo</code>', not bound by
default. This means that if you undo to much, you can put back what you
just undid by repeatedly using the <code>redo</code> command.</p>
<p><span id="l83"></span></p>
<h2 id="43-fancier-editing"><a class="header" href="#43-fancier-editing">4.3: Fancier editing</a></h2>
<p><span id="l84"></span></p>
<h3 id="431-options-controlling-zle"><a class="header" href="#431-options-controlling-zle">4.3.1: Options controlling zle</a></h3>
<p>Unlike completion, <code>zle</code> doesn't have many options associated with it;
most of the control is done by key bindings and builtin commands. Only
two are really useful; both control beeps. The option <code>beep</code> can be
unset to tell the shell never to make a noise on an error; the option
<code>histbeep</code> can be unset to disable beeps only in the case of trying to
go back before the first or forward after the last history entry.</p>
<p>The not-very-useful options are <code>zle</code> and <code>singlelinezle</code>. The former
controls whether zle is active at all and isn't that useful because it's
usually on automatically whenever you need it, in other words in
interative shells, and off whenever you don't. It's sometimes useful to
test via `<code>[[ -o zle ]]</code>', however; this lets you make a function do
something cleverer in an interative shell.</p>
<p>The option <code>singlelinezle</code> restricts editing to one line; if it gets too
long, it will be truncated and a `<code>$</code>' printed where the missing bits
are. It's only there for compatibility with ksh and as a safeguard if
your terminal is really screwed up, though even in that case zsh tries
to guess whether everything it needs is available.</p>
<p>Other functions that affect zle include the history functions. These
were described back in <a href="zshguide02.html#init">chapter 2</a>; once you've
set it off, searching through the history works basically the same way
in zle as with the `<code>!</code>' history commands.</p>
<p><span id="l85"></span></p>
<h3 id="432-the-minibuffer-and-extended-commands"><a class="header" href="#432-the-minibuffer-and-extended-commands">4.3.2: The minibuffer and extended commands</a></h3>
<p>The `minibuffer' is yet another Emacs concept; it is a prompt that
appears just under the command line for you to enter some edit required
by the editor itself. Usually, it comes and goes as it pleases and you
don't need to think about it. The most common uses are entering text for
searches, and entering a command which isn't bound to a string. That's
yet another Emacs feature: <code>\ex</code> prompts you to enter the name of a
command. Luckily, since the names tend to be rather long, completion is
available. So typing `<code>echo foo<ESC>xba<TAB>w<TAB></code>' ends up with:</p>
<pre><code> % echo foo
execute: backward-word
</code></pre>
<p>and hitting return executes that function, taking you to the start of
the <code>foo</code>; you might be able to think of easier ways of doing that. This
does provide a way of running commands you don't often use.</p>
<p>(I hope my notation isn't too confusing. I write things like <code><TAB></code>
when I'm showing a single character you hit, to make it stand out from
the surrounding text. However, when I'm not showing text being entered,
I would write that as `<code>\t</code>', which is how you would enter the
character into a key sequence to be bound, or a string to be printed.)</p>
<p>The minibuffer only handles a very limited set of editing commands.
Typing one it doesn't understand usually exits whatever you were trying
to do with the minibuffer, then executes the keystroke. However, in this
particular case, it won't let you exit until you have finished typing a
command; your only other option is to abort. The usual zle abort
character is <code>^g</code>, `<code>send-break</code>'. This is different from the more
drastic <code>^c</code>, which sends the shell itself an interrupt signal. Quite
often they have the same effect in zle, however. (You'll notice <code>^c</code> is
actually `bound to <code>undefined-key</code>', in other words zle doesn't
consider it does anything. However, the terminal driver probably causes
it to send an interrupt, and zle does respond to that.)</p>
<p>Another feature useful with rare commands is `<code>where-is</code>'. Surprise!
it's not bound by default, so typing `<code><ESC>xwhere-is</code>' is then the way
of runing it. Then you type another editor command at the `<code>Where is:</code>'
prompt, and the shell will tell you what keystrokes, if any, are bound
to it. You can also simply use <code>grep</code> on the output of <code>bindkey</code>, which,
with no arguments, lists all bindings.</p>
<p><span id="l86"></span></p>
<h3 id="433-prefix-digit-arguments"><a class="header" href="#433-prefix-digit-arguments">4.3.3: Prefix (digit) arguments</a></h3>
<p>Many commands can be repeated by giving them a numeric prefix or digit
argument. For example, at the end of a long line of text, type
`<code><ESC>4<ESC>b</code>'. The `<code><ESC>b</code>' on its own would take you one word
backwards. The `<code><ESC>4</code>' passes it the number four and it moves four
words backwards. Generally speaking, this works any time it make sense
to repeat a command. It works for <code>self-insert</code>, too, just repeatedly
inserting the character. If it doesn't work, the prefix argument is
simply ignored.</p>
<p>You can build up long or negative arguments by repeating both the <code>\e</code>
and the digit or `<code>-</code>' after it; for example, `<code><ESC>-<ESC>1<ESC>0</code>'
specifies minus ten. It varies from command to command how useful
negative numbers are, but they generally switch from backwards to
forwards or similar: `<code><ESC>-<ESC>4<ESC>\f</code>' is a pointless way of
executing the same as `<code><ESC>4<ESC>b</code>'.</p>
<p>The shell also has Emacs' `<code>universal-argument</code>' feature, but it's not
bound by default --- in Emacs it is <code>\C-u</code>, but as we've seen that's
already in use. This is an alternative to all those escapes. If you bind
the command to a keystroke (it's absolutely pointless as a shortcut
otherwise), and type that key, then an option minus followed by any
digits are remembered as a prefix. The next keystroke which is not one
of those is then executed as a command, with the prefix formed by the
number typed after <code>universal-argument</code>.</p>
<p>For example, on my keyboard, the key <code>F12</code> sends the key sequence
`<code>\e[[24~</code>' --- see below for how to find out what functions keys send.
Hence I use</p>
<pre><code> bindkey '\e[[24~' universal-argument
</code></pre>
<p>Then if I hit the characters <code>F12</code>, <code>4</code>, <code>0</code>, <code>a</code>, a row of forty `a's
is inserted onto the command line. I'm not claiming this example is
particularly useful.</p>
<p><span id="l87"></span></p>
<h3 id="434-words-regions-and-marks"><a class="header" href="#434-words-regions-and-marks">4.3.4: Words, regions and marks</a></h3>
<p>Words are handled a bit differently in zsh from the way they are in most
editors. First, there is a difference between what Emacs mode and vi
mode consider words. That is to say, there is a difference between the
functions bound by default in those modes; you can use the same
functions in either mode by rebinding the keys.</p>
<p>In both vi and Emacs modes, the same logic about words applies whether
you are moving forward or backward a number of words, or deleting or
killing them; the same amount of text is removed when killing as the
cursor would move in the other case.</p>
<p>In vi mode, words are basically the same as what vi considers words to
be: a sequence of alphanumeric characters together with underscores ---
essentially, characters that can occur in identifiers, and in fact
that's how zsh internally recognises vi `word characters'. There is one
slight oddity about vi's wordwise behaviour, however, which you can
easily see if you type `<code>/a/filename/path/</code>', leave insert mode with
<code>ESC</code>, and use `<code>w</code>' or `<code>b</code>' to go forward or backward by words over
it. It alternates between moving over the characters in a word, and the
characters in the separator `<code>/</code>'.</p>
<p>In Emacs, however, it is done a bit differently. The vi `word
characters' are always considered parts of a word, but there is a
parameter <code>$WORDCHARS</code> which gives a string of characters which are
<em>also</em> part of a word. This is perhaps opposite to what you would
expect; given that alphanumerics are always part of a word, you might
expect there to be a parameter to which you add characters you <em>don't</em>
want to be part of a word. But it's not like that.</p>
<p>Also unlike vi, jumping a word always means jumping to a word character
at the start of a word. There is no extra `turn' used up in jumping
over the non-word characters.</p>
<p>The default value for <code>$WORDCHARS</code> is</p>
<pre><code> *?_-.[]~=/&;!#$%^(){}<>
</code></pre>
<p>i.e. pretty much everything and the kitchen sink. Usually, therefore,
you will want to remove characters which you don't want to be considered
parts of words; `<code>-</code>', `<code>/</code>' and `<code>.</code>' are particularly likely
possibilities. If you want to remove individual characters, you can do
it with some pattern matching trickery (next chapter):</p>
<pre><code> % WORDCHARS=${WORDCHARS//[&.;]}
% print $WORDCHARS
*?_-[]~=/!#$%^(){}<>
</code></pre>
<p>shows that the operation has removed those three characters in the
group, i.e. `<code>&</code>', `<code>.</code>' and `<code>;</code>', from <code>$WORDCHARS</code>. The `<code>//</code>'
indicates a global substitution: any of the characters in the square
brackets is replaced by nothing.</p>
<p>Many other line editors, even those like <code>readline</code> with Emacs bindings,
behave as if only identifier characters were part of a word, i.e. as if
<code>$WORDCHARS</code> was empty. This is very easy to do with a zle shell
function. Recent versions of zsh supply the functions
`<code>bash-forward-word</code>', `<code>bash-kill-word</code>', and a set of other similar
ones, for you to bind to keys in order to have that behaviour.</p>
<p>Other behaviours are also possible by writing functions; for example,
you can jump over real shell words (i.e. individual command arguments)
by using some more substitution trickery, or you can consider only
space-delimited words (though that's not so far from what you get with
<code>$WORDCHARS</code> by adding `<code>"`'\@</code>').</p>
<p><span id="l88"></span></p>
<h3 id="435-regions-and-marks"><a class="header" href="#435-regions-and-marks">4.3.5: Regions and marks</a></h3>
<p>Another useful concept from Emacs is that of regions and marks. In
Emacs-speak `point' is where the cursor is and `mark' is somewhere
where you leave a mark to come back to later. The command to set the
mark at the current point is `<code>^@</code>' as in Emacs, a hieroglyphic which
usually means holding down the control key and pressing the space key.
On some systems, such as the limited version of <code>telnet</code> provided with a
well-known non-UNIX-based windowing system, you can't send this
sequence, and you need to bind a different sequence to
<code>set-mark-command</code>. One possibility is `<code>\e </code>' (escape followed by
space), as in MicroEMACS. (Some X Windows configurations don't allow
<code>^@</code> to work in an xterm, either, though that is usually fixable.)</p>
<p>To continue with Emacs language, the region between point and mark is
described simply as `the region'. In zsh, you can't have this
highlighted, as you might be used to with editors running directly under
windowing systems, so the easiest way to find out the ends of the region
is with <code>^x^x</code>, <code>exchange-point-and-mark</code>, which I mentioned before ---
mark, by default, is left at the beginning of the line, hence the
behaviour you saw above.</p>
<p>Various editing commands --- usually those with `<code>region</code>' in the name
--- operate on this. The most usual are those which kill or copy the
region. Annoyingly, <code>kill-region</code> isn't bound --- in Emacs, it's <code>^w</code>,
but zsh follows the tradition of having that bound to
<code>backward-kill-word</code>, even though that's also available as the
traditional Emacs binding <code>\e^?</code>. So it's probably useful to rebind it.
To copy the region, the usual binding `<code>\ew</code>' works.</p>
<p>You then `yank' back the text copied or killed at another point with
`<code>^y</code>'. The shell implements the `kill ring' feature, which means if
you perform a yank, then type `<code><ESC>y</code>' (<code>yank-pop</code>) repeatedly, the
shell cycles back through previously killed or copied text, so that you
have more available than just the last one.</p>
<p><span id="l89"></span></p>
<h2 id="44-history-and-searching"><a class="header" href="#44-history-and-searching">4.4: History and searching</a></h2>
<p>Zle has access to the lines saved in the shell's history, as described
in `Setting up history' in <a href="zshguide02.html#init">chapter 2</a>. There are
essentially three ways of retrieving bits of the history: moving back
through it line by line, searching back for a matching line, and
extracting individual words from the history. In fact, the first two are
pretty similar, and there are hybrid commands which allow you to move
back step by step but still matching only particular lines.</p>
<p><span id="l90"></span></p>
<h3 id="441-moving-through-the-history"><a class="header" href="#441-moving-through-the-history">4.4.1: Moving through the history</a></h3>
<p>The simplest behaviour is what you get with the normal cursor key
bindings, `<code>up-line-or-history</code>' and `<code>down-line-or-history</code>'. If you
are in text which fits into a single line (which may be a continuation
line, i.e. it has a new prompt in the form given by <code>$PS2</code> at the start
of the line), this replaces the entire line with the line before or
after in the history. The history is not circular, it has a beginning
and an end. The beginning is the first line still remembered by the
shell (i.e. <code>$HISTSIZE</code> lines back, taking into account that the actual
number of lines present will be modified by the effects of any special
history options you have set to remove unwanted lines); the end is the
line you are typing. You can use <code>\e<</code> and <code>\e></code> to go to the first and
last line in the history.</p>
<p>The last phrase sounds trivial but isn't quite. Type `<code>echo This is the last line</code>', go back a few lines with the up arrow, and then back down
to the end, and you will see what I mean --- the shell has remembered
the line you were typing, even though it hadn't been entered, so you can
scroll up and down the history and still come back to it.</p>
<p>Of course, you can edit any of the earlier history lines, and hit
`return' so that they are executed --- that's the whole point of being
able to scroll back through the history. What is maybe not so obvious is
that the shell will remember changes you make to these lines, too, until
you hit `return'.</p>
<p>For example, type `<code>echo this is the last line</code>' at a new shell prompt,
but don't hit return. Now hit the up arrow once, and edit the previous
line to say `<code>echo this is the previous line</code>'. If you scroll down and
up, you will see that the shell has kept both of those lines. When you
decide which one to use and hit return, that line is executed and added
to the end of the history, and any changes to previous lines in the
history are forgotten.</p>
<p>Sometimes you don't want to add a new line to history, instead
re-execute a whole series of earlier commands one by one. This can be
done with <code>^o</code>, <code>accept-line-and-down-history</code>. When you hit <code>^o</code> on a
line in the history, that is executed, and the line after it in the
history is shown. So you just need to keep hitting it to keep executing
the commands.</p>
<p>There are two other similar commands I don't use as much,
<code>infer-next-history</code>, bound to <code>^x^n</code>, and
<code>accept-and-infer-next-history</code>, not bound by default. `Inferring' the
next history means that the shell looks at what is in the current line,
whatever its provenance --- you might just have typed it, for example
--- and looks back in the history for a matching line; the `inferred'
next history line is the one following that line. In the first case, you
are simply shown that line; in the second case, the current line is
executed first, then you are shown the inferred line. Feel free to drop
me a line if you find this is the best thing since sliced bread.</p>
<p>One slight confusion about the history is that it can be hard to
remember quite where you are in it, for example, if you were editing a
line and had to scroll back to look for something else. In cases like
this, <code>\e></code> is your friend, as it takes you the last line. Also,
whenever you hit return, you are guaranteed to be at the end of the
history, even if you were editing a line some back in the history,
unlike certain other systems (though <code>accept-line-and-down-history</code> can
emulate those). So it's usually not too hard to stay unconfused about
what you're editing.</p>
<p><span id="l91"></span></p>
<h3 id="442-searching-through-the-history"><a class="header" href="#442-searching-through-the-history">4.4.2: Searching through the history</a></h3>
<p>Zsh has the commands you would expect to search through the history,
i.e. where you hit a search key and then type in the words to search
for. However, it also has other features, probably more used by the zsh
community, where the search is based on some feature of the current
line, in particular the first word or the line up to the cursor
position. These typically enable you to search backwards more quickly,
since you don't need to tell the shell what you are looking for.</p>
<p><strong>Ordinary searching</strong></p>
<p>The standard search commands, by which I mean the ones your are probably
most familiar with from ordinary text editors (if either Emacs or vi can
be so called), are designed to make Emacs and vi users feel at home.</p>
<p>In Emacs mode, you have incremental search: <code>^r</code> to search backwards ---
this is usually what you want, since you usually start at the end ---
and <code>^s</code> to search forwards. Note that <code>^s</code> is another keystroke which
is often intercepted by the terminal driver; in this case, it usually
freezes output to the terminal until you type <code>^q</code> to turn it back on.
If you don't like this, you can either use `<code>stty stop</code>' and `<code>stty start</code>' to change the characters, or simply `<code>unsetopt flowcontrol</code>' to
turn that feature off altogether. However, the command bound to <code>^s</code>,
<code>history-incremental-search-forward</code>, is also bound to <code>^xs</code>, so you can
use that instead.</p>
<p>As in Emacs, for each character that you type, incremental search takes
you to the nearest history entry that matches all the characters, until
the match fails. Typing the search keystroke again at any point takes
you to the next match for the characters in the minibuffer.</p>
<p>In vi command mode, the keystrokes available by default are the familiar
`<code>/</code>' and `<code>?</code>'. There are various differences from vi, however. First
of all, it is `<code>/</code>' that searches backwards --- this is the one you
will use more often. Secondly, you can't search for regular expressions
(patterns); the only exception is that the character `<code>^</code>' at the start
anchors the search to the start of a line. Everything else is just a
plain string.</p>
<p>The other two standard vi search keystrokes are also present: `<code>n</code>'
searches for the next match of the current string, and `<code>N</code>' does the
same but reverses the direction of the search.</p>
<p><strong>Search for the first word</strong></p>
<p>The next sort of search is probably the most commonly used, but is only
bound in Emacs mode: <code>\ep</code> and <code>\en</code> search forward or backward for the
next history line with the same first word as the current line. So often
to reuse a command you will type just the command name itself, and hit
<code>\ep</code> until the command line you want appears. These commands are called
simply `<code>history-search-backward</code>' and `<code>history-search-forward</code>'; the
name doesn't really describe the function all that well.</p>
<p><strong>Prefix searching</strong></p>
<p>Finally, you can search backwards for a line which has the entire
starting point up to the cursor position the same as the current line.
This gives you a little more control than <code>history-search-</code><em>direction</em>.
The corresponding commands, <code>history-beginning-search-backward</code> and
<code>history-beginning-search-forward</code>, are not bound by default. I find it
useful to have them bound to <code>^xp</code> and <code>^xn</code> since this is similar to
the initial-word searches:</p>
<pre><code> bindkey '^xp' history-beginning-search-backward
bindkey '^xn' history-beginning-search-forward
</code></pre>
<p><strong>Other search commands based on functions</strong></p>
<p>Search commands are one of the types most often customised by writing
shell functions. Some are supplied with the latest versions of the
shell; have a look in the ZLE section of the <code>zshcontrib</code> manual page.
You should find the functions themselves installed somewhere in your
<code>$fpath</code>, typically</p>
<pre><code> /usr/local/share/zsh/$ZSH_VERSION/functions
</code></pre>
<p>or in the subdirectory <code>Zle</code> of that directory, depending how your
version of zsh was installed. If the shell was pre-installed, the most
likely location is</p>
<pre><code> /usr/share/zsh/$ZSH_VERSION/functions/Zle
</code></pre>
<p>These should guide you in writing your own.</p>
<p>One point to note is that when called from a function the
<code>history-search-</code><em>direction</em> and
<code>history-incremental-search-</code><em>direction</em> can take a string argument
specifying what to search for. In the first case, this is just a one off
search, while in the second, you remain in incremental search and the
string is used to prime the minibuffer, so you can edit it. I will later
say much more about writing zle functions, but calling a search command
from a user-defined editing function is as simple as:</p>
<pre><code> zle history-search-backward search-string
</code></pre>
<p>and you can test the return status to see if the search succeeded.</p>
<p><span id="l92"></span></p>
<h3 id="443-extracting-words-from-the-history"><a class="header" href="#443-extracting-words-from-the-history">4.4.3: Extracting words from the history</a></h3>
<p>Sometimes instead of editing a previous line you just want to extract a
word from it into the current line. This is particularly easy if the
word is the last on the line, and the line isn't far back in the
history: just hit <code>\e.</code> repeatedly, and the shell will cycle through the
last word on previous lines. You can give this a prefix argument to pick
the <em>N</em>th from last word on the line just above the last line you picked
a word from. As you can tell from the description, this gets a little
hairy; version 4.1 of the shell will probably provide a slightly more
flexible version.</p>
<p>Although it's strictly not to do with the history, you can copy the
previous word on the current line with <code>copy-prev-word</code>, which for some
reason is bound to <code>\e^_</code>, escape followed (probably) by control and
slash. I have this bound to <code>\e=</code> instead (in some versions of ksh that
key sequence is taken by the equivalent of <code>list-choices</code>). This copies
words delimited by whitespace, but you can copy what the shell would see
as the previous complete argument by using <code>copy-prev-shell-word</code>
instead. This isn't bound by default, as it is newer than the other one,
but it is arguably more useful.</p>
<p>Sometimes you want to complete a word from the history; this is possible
using the completion system, described in the next chapter.</p>
<p><span id="l93"></span></p>
<h2 id="45-binding-keys-and-handling-keymaps"><a class="header" href="#45-binding-keys-and-handling-keymaps">4.5: Binding keys and handling keymaps</a></h2>
<p>There are two topics to cover under the heading of key bindings: first,
how to bind keys themselves, and secondly, keymaps and how to use them.
Manipulating both key bindings and keymaps is done with the <code>bindkey</code>
command. The first topic is the more immediately useful, so I'll start
with that.</p>
<p><span id="l94"></span></p>
<h3 id="451-simple-key-bindings"><a class="header" href="#451-simple-key-bindings">4.5.1: Simple key bindings</a></h3>
<p>You've already seen basic use of <code>bindkey</code> to link editing commands to a
particular sequence of keys. You've seen the shorthand for naming keys,
with <code>\e</code> being escape and <code>^x</code> the character <code>x</code> pressed while the
control key is held down. I even said something about `meta' key
bindings.</p>
<p>Let me now launch into a little more detail. When you bind a key
sequence, which you do with `<code>bindkey</code> <em>key-sequence</em>
<em>editor-command</em>', the <em>key-sequence</em> can consist of as many characters
as you like. It doesn't even matter (much) if some initial set of the
key sequence is already bound. For example, you can do,</p>
<pre><code> bindkey '\eA' backward-word
bindkey '\eAA' beginning-of-line
</code></pre>
<p>Here, I'll follow the shell documentation in referring to <code>\eA</code> as the
prefix of <code>\eAA</code>.</p>
<p>This introduces two points. First, note that the binding for <code>\eA</code> is
distinct from that for <code>\ea</code>; you will see the latter still does
<code>accept-and-hold</code> (in Emacs mode), which means it excutes the current
line, then gives it back to you for editing --- useful for doing a lot
of quite similar tasks. Meanwhile, <code>\eA</code> takes you back a word.</p>
<p>This case sensitivity only applies to alphabetic characters which are a
complete key in their own right, not to those characters with the
control key held down --- <code>^x</code> and <code>^X</code> are identical. (You may have
found there are ways to bind both separately in Emacs when running under
a windowing system, since the windowing system can tell Emacs if the
shift key is held down with the others; it's not that simple if you are
using an ordinary terminal.)</p>
<p>If you entered both those <code>bindkey</code> commands, you may notice that there
is a short pause before <code>\eA</code> takes effect. That's because it's waiting
to see if you type another <code>A</code>. If you do type the extra <code>A</code> during that
pause, you will be taken to the beginning of the line instead. That
pause is how the shell decides whether to execute the prefix on its own.</p>
<p>The time it waits is configurable and is given by the parameter
<code>$KEYTIMEOUT</code>, which is the delay in hundredths of a second. The default
is 40, i.e. four tenths of a second. Its use is usually down to personal
preference; if you don't type very fast, you might want to increase it,
at the expense of a longer delay when you are waiting for the prefix to
be executed. If you are editing on a remote machine over a very slow
link, you also may need to increase it to be able to get full key
sequences which have such a prefix to work at all.</p>
<p>However, the shell only has this ambivalent behaviour if a prefix is
bound in its own right; if the initial key or keys don't mean anything
on their own, it will wait as long as you like for you to type a full
sequence which is bound. This is by far the normal case. The only common
example of a separately bound prefix is in vi insert mode, where <code><ESC></code>
takes you back to command mode, while there may be other bindings
starting with <code>\e</code> such as the cursor keys. We'll see below how you can
remove those if they offend your sense of vi purity. (Don't laugh, vi
users are strange.)</p>
<p>Note that if the whole sequence is not bound, after all, the shell will
abort as soon as it reads a complete key sequence which is no longer a
prefix. For example, if you type `<code>\e[</code>', the chances are the shell is
waiting for more, but if you add a `<code>/</code>', say, it will probably decide
you are being silly and abort. The next key you type then starts a new
sequence.</p>
<p><span id="l95"></span></p>
<h3 id="452-removing-key-bindings"><a class="header" href="#452-removing-key-bindings">4.5.2: Removing key bindings</a></h3>
<p>If you want to remove a key binding, you can simply bind it to something
else. Near all uses of the <code>bindkey</code> and <code>zle</code> commands are smart about
removing dead wood in such cases. However you can also use `<code>bindkey -r</code> <em>key-sequence</em>' to remove the binding explicitly. You can also
simply bind the sequence to the command <code>undefined-key</code>; this has
exactly the same effect --- even down to pruning completely any bindings
for long sequences. For example, suppose you bind `<code>\e\C-x\C-x</code>' to a
command, then to <code>undefined-key</code>. All memory that `<code>\e\C-x\C-x</code>' was
ever bound is removed; <code>\e\C-x</code> will no longer be marked as a prefix
key, unless you had some other binding with that prefix.</p>
<p>You can remove all bindings starting with a given prefix by adding the
`<code>-p</code> option. The example given in the manual,</p>
<pre><code> bindkey -rpM viins '\e'
</code></pre>
<p>(except it uses the equivalent form `<code>^[</code>') is amongst the most useful,
as it will remove the annoying delay after you type `<code>\e</code>' to enter vi
command mode. The delay is there because the cursor keys usually also
start with <code>\e</code> and the shell is waiting to see if you actually typed
one of those. So if you can make do without cursor keys in vi insert
mode you may want to consider this.</p>
<p>Note that any binding for the prefix itself is not removed. In this
example, <code>\e</code> stays bound as it was in the <code>viins</code> keymap, presumably to
<code>vi-cmd-mode</code>.</p>
<p>All manipulations like this are specific to one particular keymap. You
need to repeat them with a different <code>-M</code> <em>...</em> option argument, which
is described below, to have the same effect in other keymaps.</p>
<p><span id="l96"></span></p>
<h3 id="453-function-keys-and-so-on"><a class="header" href="#453-function-keys-and-so-on">4.5.3: Function keys and so on</a></h3>
<p><span id="fkeys"></span></p>
<p>It's usually possible to bind the function keys on your keyboard,
including the specially named ones such as `Home' and `Page Up'. It
depends a good deal on how your windowing system or terminal driver
handles them, but these days it's nearly always the case that a well
set-up system will allow the function keys to send a string of
characters to the terminal. To bind the keys you need to find out what
that string is.</p>
<p>Luckily, you are usually aided by the fact that only the first character
of the string is `funny', i.e. does something other than insert a
character. So there is a trick for finding out what the sequence is. In
a shell window, hit <code>^v</code> (if you are using vi bindings, you will need to
be in insert mode), then the function key in question. You will probably
see a string like `<code>^[OP</code>' --- this is what I get from the F1 key. A
note in my <code>.zshrc</code> suggests I used to get `<code>\e[11~</code>', so be prepared
for something different, even if, like me, you are using a standard
xterm terminal emulator. A quick poll of terminal emulators on this
Linux/GNU/XFree86 system suggests these two possibilities are by far the
most popular.</p>
<p>You may even be able to get different sequences by holding down shift or
control as well (after pressing <code>^v</code>, of course). On my keyboard,
combining F1 with shift gives me `<code>^[O2P</code>', with control `<code>^[O5P</code>' and
with both `<code>^[O6P</code>'. Again, your system may do something completely
different.</p>
<p>If you move the cursor back over that `<code>^[</code>', you'll find it's a single
character --- you can position the cursor over the `<code>^</code>', but not the
`<code>[</code>'. This is zsh's way of inserting a real, live escape character
into the line. In fact, if you type</p>
<pre><code> bindkey '
</code></pre>
<p>then <code>^v</code>, the function key, and the other single quote, you have a
perfectly acceptable way of binding the key on the command line. Zsh is
generally quite relaxed about your use of unprintable characters; they
may not show up correctly on your terminal, but the shell is able to
handle all single-byte characters. It doesn't yet have support for those
longer than a single byte, however.</p>
<p>You can also do the same in your <code>.zshrc</code>; the shell will handle strange
characters in input without a murmur. You can also use the two
characters `<code>^[</code>', which is just another way of entering an escape key.
However, the kosher thing to do is to turn it into `<code>\e</code>'. For example,</p>
<pre><code> bindkey '\e[OP' where-is # F1
bindkey '\e[O2P' universal-argument # shift F1
</code></pre>
<p>and so on. Using this, you can give sensible meanings to `Home',
`End', etc. Note the windowing system's sensible way of avoiding the
problem with prefixes --- any extra characters are inserted before the
final character, so the shell can easily tell when the sequence is
complete without having to wait and see if there is more to follow.</p>
<p>There is a utility supplied with zsh called <code>zkbd</code> which can help with
all of this by finding out and remembering the definitions for you. You
can probably use it simply by autoloading it and running it, as it is
usually installed with the other functions. It should be reasonably
self-explanatory, else consult the <code>zshcontrib</code> manual.</p>
<p>If you are using X Windows and are educated enough, you can tinker with
your <code>.Xdefaults</code> file to tweak how keys are interpreted by the terminal
emulator. For example, the following turns the backspace key into a
delete key in anything using a `VT100 widget', which is the basis of
xterm's usual mode of operation:</p>
<pre><code> *VT100.Translations: #override \
<Key>BackSpace: string(0x7F)
</code></pre>
<p>Part of the reason for showing this is that it makes zsh's key binding
system look wonderfully streamlined by comparison. However, tinkering
around at this level gives you very much more control over the use of
key modifiers (shift, alt, meta, control, and maybe even super and hyper
if you're lucky). This is far beyond the scope of this guide --- which I
say, as you probably realise by now, to cover up for not knowing much
about it. Here's another example from Oliver Kiddle, though; it uses
control with the left-cursor key to send an escape sequence: insert</p>
<pre><code> Ctrl<Key>Left: string(0x1b) string("[159q") \n\
</code></pre>
<p>into the middle of the example above --- this shows how multiple
definitions are handled. Modern xterms already send special escape
sequences which you can investigate and bind to as I've described.</p>
<p><span id="l97"></span></p>
<h3 id="454-binding-strings-instead-of-commands"><a class="header" href="#454-binding-strings-instead-of-commands">4.5.4: Binding strings instead of commands</a></h3>
<p>It's possible to assign an arbitrary string of characters to a key
sequence instead of an editor command by giving <code>bindkey</code> the option
<code>-s</code>. One of the good things about this is that the string of characters
are reinterpreted by zle, so they can contain active key sequences. In
the old days, this was quite often used as a basic form of macro, to
string together editor commands. For example, the following is a simple
way of moving backward two words by repeating the Emacs mode bindings.
I've used my F1 binding again; yours may be completely different.</p>
<pre><code> bindkey -s '\e[OP' '\eb\eb'
</code></pre>
<p>It's not a good idea to bind a key sequence to another string which
includes itself.</p>
<p>This method has the obvious drawback that if someone comes along and
rebinds `<code>\eb</code>', then F1 will stop working, too. Nowadays, this sort of
task can be done much more flexibly and clearly by writing a
user-defined widget, which is described in a later section. So bindings
of this sort are going a little out of fashion. However, they do provide
quick shortcuts. Two from Oliver Kiddle:</p>
<pre><code> bindkey -s '^[[072q' '^V^I' # Ctrl-Tab
bindkey -s "\C-x\C-z" "\eqsuspend\n"
</code></pre>
<p>You can also quite easily do some of the things you can do with global
aliases.</p>
<p>Remember that `ordinary' characters can be rebound, too; they just
usually happen to have a binding which makes them be inserted directly.
As a particularly pointless example, consider:</p>
<pre><code> bindkey -s secret 'Oh no!'
</code></pre>
<p>If you type `<code>secret</code>' fast enough the letters are swallowed up and
`<code>Oh no!</code>' appears instead. If you pause long enough anywhere in the
middle, the word is inserted just as normal. That's because all parts of
it can be interpreted as prefixes in their own right, so <code>$KEYTIMEOUT</code>
applies at every intervening stage. Less pointlessly, you could use this
as a way of defining abbreviations.</p>
<p><span id="l98"></span></p>
<h3 id="455-keymaps"><a class="header" href="#455-keymaps">4.5.5: Keymaps</a></h3>
<p>So far, all I've said about keymaps is that there are three standard
ones, one for Emacs mode and two for vi mode, and that `<code>bindkey -e</code>'
and `<code>bindkey -v</code>' pick Emacs or vi insert mode bindings. There's no
simple way of picking vi command mode bindings, since that's not usually
directly available but entered by the <code>vi-cmd-mode</code> command, usually
bound to <code>\e</code>, in vi insert mode. (There is a `<code>bindkey -a</code>', but that
doesn't pick the keymap for normal use; it's equivalent to, but less
clear than, `<code>bindkey -M vicmd</code>'.)</p>
<p>Most handling of keymaps is done through <code>bindkey</code>. The keymaps have
short names, <code>emacs</code>, <code>viins</code> and <code>vicmd</code>, for use with <code>bindkey</code>. There
is also a keymap <code>.safe</code> which you don't usually need but which never
changes, so can be used if your experimentation has completely ruined
every other keymap. It only has bindings for <code>self-insert</code> (most keys)
and <code>accept-line</code> (<code>^j</code> and <code>^m</code>), but that's enough to enter commands.</p>
<p>The names are most useful in two places. First, you can use `<code>bindkey -M</code> <em>keymap</em>' to define keys in a particular map:</p>
<pre><code> bindkey -M vicmd "\e[OA" up-line-or-history
</code></pre>
<p>binds the usual up-cursor key in <code>vicmd</code> mode, whatever keymap is
currently set. Actually, any version of the shell which understands the
<code>-M</code> option probably has that bound already.</p>
<p>Secondly, you can force zle to use a particular keymap. This is done in
a slightly non-obvious way: zle always uses the keymap <code>main</code> as the
current keymap (except when it's off in vi command mode, which is
handled a bit specially). To use your own, you need to make <code>main</code> an
alias for that with `<code>bindkey -A</code>'. The order after this is the same as
that after <code>ln</code>: the existing keymap you want to refer to comes first,
then what you want to make an alias for it, in this case <code>main</code>. This
means that</p>
<pre><code> bindkey -A emacs main
</code></pre>
<p>has the same effect as</p>
<pre><code> bindkey -e
</code></pre>
<p>but is more explicit, if a little more baroque. Don't link <code>vicmd</code> to
main, since then you can't use <code>viins</code>, which is bad. Note that
`<code>bindkey -M emacs</code>' doesn't have this effect; it simply lists the
bindings in the <code>emacs</code> keymap.</p>
<p>You can create your own keymaps, too. The easiest way is to copy an
existing keymap, such as</p>
<pre><code> bindkey -N mymap emacs
</code></pre>
<p>which creates (or replaces) <code>mymap</code> and initialises it with the bindings
from <code>emacs</code>. Now you can use <code>mymap</code> just like <code>emacs</code>. The bindings in
each are completely separate. If you finish with a keymap, you can
remove it with `<code>bindkey -D keymap</code>', although you'd better make sure
it's not linked to <code>main</code> first.</p>
<p>You can omit `<code>emacs</code>' to create an empty keymap; this might be
appropriate if your keymap is only going to be used in certain special
places and you want complete control on what goes into it. Currently the
shell isn't very good at letting you apply your own keymaps just in
certain places, however.</p>
<p>There are various other keymaps you might encounter, used in special
circumstances. If you list all keymaps, which is done by `<code>bindkey -l</code>', you may see <code>listscroll</code> and <code>menuselect</code>. These are used by the
new completion system, so if that isn't active, you probably won't see
them. They reside in the module <code>zsh/complist</code>. There will be more about
their effects in <a href="zshguide06.html#comp">chapter 6</a>; <code>listscroll</code> allows
you to move up and down completion lists which more than fill the
terminal window, and <code>menuselect</code> allows you to select items
interactively from a displayed list. You can bind keys in them as with
any other keymap.</p>
<p><span id="l99"></span></p>
<h2 id="46-advanced-editing"><a class="header" href="#46-advanced-editing">4.6: Advanced editing</a></h2>
<p>(In physics, the `advanced wave' is a hypothetical wave which moves
backwards in time. Unfortunately, however useful it would be to meet
deadlines, that's not what I meant by `advanced editing'.)</p>
<p>Here are are a few bits and pieces which go beyond ordinary line editing
of shell commands. Although they haven't been widespread in shells up to
now, I use all of them every day, so they aren't just for the
postgraduate zsh scholar.</p>
<p><span id="l100"></span></p>
<h3 id="461-multi-line-editing"><a class="header" href="#461-multi-line-editing">4.6.1: Multi-line editing</a></h3>
<p>All Bourne-like shells allow you to edit continuation lines; that is, if
the shell can work out for sure that you haven't finished typing, it
will show you a new prompt, given by <code>$PS2</code>, and allow you to continue
where you left off from the previous line. In zsh, you can even see what
it is the shell is waiting for. For a simple example, type
`<code>array=</code>(<code>first</code>' and then `return'. The shell is waiting for the
final parenthesis of the array, and prints `<code>array></code>' at you, unless
you have altered <code>$PS2</code>. You can continue to add elements to the array
until you close the parentheses.</p>
<p>Shells derived from csh are less happy about continuation lines;
historically, this is because they try to evaluate everything in one go,
and became confused if they couldn't. The original csh didn't have a
particularly sophisticated parser. For once, zsh doesn't have an option
to match the csh behaviour; you just have to get used to the idea that
things work in zsh.</p>
<p>Where zsh improves over other shells is that you aren't just limited to
editing a single continuation line; you can actually edit a whole block
of lines on screen as you would in a full screen editor --- although you
can't scroll off the chunk of lines you're editing, which wouldn't make
sense.</p>
<p>The easiest way of doing this is to hit escape before you type a newline
at the point where you haven't finished typing. Actually, you can do
this any time, even if the line so far is complete. For example,</p>
<pre><code> % print This is line one<ESC><RET>
print This is line two
</code></pre>
<p>where those angle brackets at the end of the line means you type escape,
then return. Nothing happens, and there is no new prompt; you just type
blithely on. Hit return, unescaped this time, and both lines will be
executed. Note there is no implicit backslash, or anything like that;
when zsh reads the whole caboodle, that escaped carriage return became a
real carriage return, just as the shell would have read it from a
script.</p>
<p>This works because `<code>\e\r</code>' is actually bound to the command
<code>self-insert-unmeta</code>' which means `insert the character you get by
stripping the escape or top bit from what I just typed' --- in other
words, a literal carriage return. You would have got exactly the same
effect by typing <code>^v^j</code>, since the <code>^v</code> likewise escapes the <code>^j</code>
(newline), as it does any other character.</p>
<p>(Aside for the terminally curious only: Why newline here and not
carriage return --- the `enter' key --- as you might expect? That's a
rather grotesque story. It turns out that for mostly historical reasons
UNIX terminal drivers like to swap newline and carriage return, so when
you type carriage return (sent both by that key and by <code>^m</code>, which is
the same as the character represented by <code>\r</code>), it comes out as newline
(on most keyboards, just sent by <code>^j</code>, which is the same as the
character represented by <code>\n</code>). It is the newline character which is the
one you `see' at the end of the line (by virtue of the fact it is the
end of the line). However, <code>^v</code> sees through this and if you type <code>^m</code>
after it, it inserts a literal <code>^m</code>, which just looks like a <code>^m</code>
because that's how zsh outputs it. So that's why that doesn't work.
Actually, <code>self-insert-unmeta</code> would see the <code>^m</code>, too, because that's
what you get when you strip off the <code>\e</code>, but it has a little extra code
to make UNIX users feel at home, and behaves as if it were a newline.
Normally, <code>^j</code> and <code>^m</code> are treated the same way (<code>accept-line</code>), but
the literal characters have different behaviours. If you're now very
confused, just be thankful I haven't told you about the additional
contortions which go on when outputting a newline.)</p>
<p>It probably doesn't seem particularly useful yet, because all you've
done is miss out a new prompt. What makes it so is that you can now go
up and down between the two (or more) lines just using the cursor keys.
I'm assuming you haven't rebound the cursor keys, your terminal isn't a
dumb one which doesn't support cursor up, and the option <code>singlelinezle</code>
isn't in effect --- just unset it if it is, you'll be grateful later.</p>
<p>So for example, type</p>
<pre><code> % if [[ true = false ]]; then<ESC><RET>
print Fuzzy logic rules<ESC><RET>
fi
</code></pre>
<p>where I indented that second line just with spaces, because I usually do
inside an `if'. There are no continuation prompts here, just the
original <code>$PS1</code>; that's not a misprint. Now, before hitting return, move
up two lines, and edit <code>false</code> to <code>true</code>. You can see how this can be
useful. Entering functions at the command line is probably a more
typical example.</p>
<p>Suppose you've already gone through a few continuation lines in the
normal way with <code>$PS2</code>'s? You can't scroll back then, even though the
block hasn't yet been edited. There's a magic way of turning all those
continuation lines into a single block: the editor command
<code>push-line-or-edit</code>. If you're not on a continuation line, it acts like
the normal <code>push-line</code> command, which we'll meet below, but for present
purpose you use it when you are on a continuation line. You are
presented with a seamless block of text from the (redrawn) prompt to the
end which you can edit as one. It's quite reasonable to bind
<code>push-line-or-edit</code> instead of <code>push-line</code>, to either <code>^q</code> or <code>\eq</code> (in
Emacs mode, which I will assume, as usual). Be careful with <code>^q</code>, though
--- if the option <code>flowcontrol</code> is set it will probably be swallowed up
by the terminal driver and not get through to the shell, the same
problem I mentioned above for <code>^s</code>.</p>
<p><span id="l101"></span></p>
<h3 id="462-the-builtin-vared-and-the-function-zed"><a class="header" href="#462-the-builtin-vared-and-the-function-zed">4.6.2: The builtin vared and the function zed</a></h3>
<p>I mentioned the <code>vared</code> command in <a href="zshguide03.html#syntax">chapter 3</a>;
it uses the normal line editor to editor a variable, typically a long
one you don't want to have to type in completely like <code>$path</code>, although
you need to remember <em>not</em> to put the `<code>$</code>' in front or the shell will
substitute it before <code>vared</code> is run. However, since it's just a piece of
text like any other input, this, too, can have multiple lines, which you
enter in the same way --- and since a shell parameter can contain
anything at all, you have a pretty general purpose editor. The shell
function `<code>zed</code>' is supplied with the shell and allows you to edit a
file using all the now-familiar commands. Since when editing files you
don't expect a carriage return to dump you out of the editor, just to
insert a new line, zed rebinds carriage return to <code>self-insert-unmeta</code>
(the `<code>-unmeta</code>' here is just to get the swapping behaviour of turning
the carriage return into a newline). To save and exit, you can type
<code>^j</code>, or, if your terminal does something odd with that, you can also
use <code>^x^w</code>, which is designed to look like Emacs' way of writing a file.</p>
<p>If you look at <code>zed</code>, you will see it has a few bells and whistles ---
for example, `<code>zed -f</code>' allows you to edit a function --- but the code
to read a file into a parameter, edit the parameter, and write the
parameter back to the file is extremely simple; all the hard editing
code is already handled within <code>vared</code>. Indeed, <code>zed</code> is essentially a
completely general purpose editor, though it quickly becomes inefficient
with long files, particularly if they are larger than a single screen;
as you would expect, zle was written to cope efficiently with short
chunks of text.</p>
<p>It would probably be nice if you could make key bindings that only
applied within vared by using a special keymap. That may happen one day.</p>
<p>By the way, note that you can edit arrays with vared and it will handle
the different elements sensibly. As usual, whitespace separates
elements; when it presents you with an array which contains whitespace
within elements, vared will precede it with a backslash to show it isn't
a separator. You can insert quoted spaces with backslashes yourself.
Only whitespace characters need this quoting, and only backslashes work.</p>
<p>For example,</p>
<pre><code> array=('one word' 'two or more words')
vared array
</code></pre>
<p>presents you with `<code>one\ word two\ or\ more\ words</code>'. If you add `<code> and\ some\ more.</code>', hit return, and type `<code>print -l $array</code>' to show
one element per line you will see</p>
<pre><code> one word
two or more words
and some more.
</code></pre>
<p>Some older versions of the shell were less careful about spaces within
elements.</p>
<p><span id="l102"></span></p>
<h3 id="463-the-buffer-stack"><a class="header" href="#463-the-buffer-stack">4.6.3: The buffer stack</a></h3>
<p>The mysterious other use for <code>push-line-or-edit</code> will now be explained.
Let's stick to <code>push-line</code>, in fact, since I've already dealt with the
<code>-or-edit</code> bit.</p>
<p>Type</p>
<pre><code> print I was just in the directory
</code></pre>
<p>(no newline). Oh dear, which directory were you just in? You don't want
to interrupt the flow of text to find out. Hit `<code>\eq</code>'; the line you've
been typing disappears --- but don't worry, it hasn't gone. Now type</p>
<pre><code> dirs
</code></pre>
<p>Two things happen: that last line is executed, of course, showing the
list of directories on the directory stack (your use of <code>pushd</code> and
<code>popd</code>), but also the line you got rid of before has reappeared, so you
can continue to edit it.</p>
<p>You may not realise straight away quite how useful this is, but I used
it several times just while I was writing the previous paragraph. For
example, I was alternating directories between the zle source code and
the directory where I keep this guide, and I started typing a `<code>grep</code>'
command before realising I was in the wrong directory. All I need to do
is type <code>\eq</code>, then <code>pushd</code>, to put me where I want to be, and finish
off the <code>grep</code>.</p>
<p>The `buffer stack', which is the jargon for this mechanism, can go as
deep as you like. It's a last-in-first-out (LIFO) stack, so the line
pushed onto it by the most recently typed <code>\eq</code> will reappear first,
followed by the back numbers in reverse order. You can even prime the
buffer stack from a function --- not necessarily a zle function, though
that works too --- with `<code>print -z</code> <em>command-line</em>'.</p>
<p>You can pull something explicitly off the stack, if you want, by typing
<code>\eg</code>, but that has the same effect as clearing the current line and
hitting return. You can of course push the same line multiple times: if
you need to do a whole series of things before executing it, just hit
<code>\eq</code> again each time the line pops back up.</p>
<p>I lied a little bit, to avoid confusion. The cleverness of
<code>push-line-or-edit</code> about multi-line buffers extends to the this case,
too. If you do a normal <code>push-line</code> on a multi-line buffer, only the
current single line is pushed; the command to push the whole lot, which
is probably what you want, is <code>push-input</code>. But if you have
<code>push-line-or-edit</code> bound, you can forget the distinction, since it will
do that for you. If you've been paying attention you can work out the
following sequence (assuming <code>\eq</code> has been rebound to
<code>push-line-or-edit</code>):</p>
<pre><code> % if [[ no = yes ]]; then
then> print<ESC>q<ESC>q
</code></pre>
<p>The first <code>\eq</code> turns the two lines into a single buffer, then the
second pushes the whole lot onto the buffer stack. This saves a lot of
thinking about bindings. Hence I would recommend users of Emacs mode add</p>
<pre><code> bindkey '\eq' push-line-or-edit
</code></pre>
<p>to their <code>.zshrc</code> and forget the distinctions.</p>
<p><span id="l103"></span></p>
<h2 id="47-extending-zle"><a class="header" href="#47-extending-zle">4.7: Extending zle</a></h2>
<p>We now come to the newest and most flexible part of zle, the ability to
create new editing commands, as complicated as you like, using shell
functions. This was originally introduced by Andrew Main (`Zefram') in
zsh 3.1 and so is standard in all versions of zsh 4, although work goes
on.</p>
<p><span id="l104"></span></p>
<h3 id="471-widgets"><a class="header" href="#471-widgets">4.7.1: Widgets</a></h3>
<p>If you don't speak English as you first language, first of all,
congratulations for getting this far. Secondly, you may think of
`widget' only as a technical word applied to the object which realises
some computational idea, like the thing that implements text editing in
a window system, for example. However, to most English speakers,
`widget' is a humorous word for an object, a bit like
`whatyoumacallit' or `thingummybob', as in `where's that clever
widget that undoes the foil and takes out the cork in one go'. Zsh's use
has always seemed to me closer to the second, non-technical version, but
I may be biased by the fact that the internal object introduced by
Zefram to represent a widget, and never seen by the user, is called a
`thingy', which I won't refer to again since you don't need to know.</p>
<p>Anyway, a `widget' is essentially what I've been calling an editor
command up to now, something you can bind to a key sequence. The reason
the more precise terminology is useful is that as soon as you have shell
functions flying around, the word `command' is hopelessly non-specific,
since functions are full of commands which may or may not be widgets. So
I make no apology for using the word.</p>
<p>So now we are introducing a second type of widget: one which, instead of
something handled by code built into the shell, is handled by a function
written by the user. They are completely equivalent; <code>bindkey</code> and
company don't care which it is. All you need to do to create a widget is</p>
<pre><code> zle -N widget-name function-name
</code></pre>
<p>then <em>widget-name</em> can be used in <code>bindkey</code>, or <code>execute-named-cmd</code>, and
the function <em>function-name</em> will be run. If the <code>widget-name</code> and
<code>function-name</code> are the same, which is often the simplest thing to do,
you just need one of them.</p>
<p>You can list the existing widgets by using `<code>zle -l</code>', although often
`<code>zle -lL</code>' is a better choice since the output format is then the same
as the form you would use to define the widget. If you see lots of
`<code>zle -C</code>' widgets when you do that, ignore them for now; they are
completion widgets, handled a bit differently and described in <a href="zshguide06.html#comp">chapter
6</a>.</p>
<p>Now you need to know what should go into the function.</p>
<p><span id="l105"></span></p>
<h3 id="472-executing-other-widgets"><a class="header" href="#472-executing-other-widgets">4.7.2: Executing other widgets</a></h3>
<p>The simplest thing you can do inside a function implementing a widget is
call an existing function. So,</p>
<pre><code> my-widget() {
zle backward-word
}
zle -N my-widget
</code></pre>
<p>creates a widget called <code>my-widget</code> which behaves in every respect
(except speed) like the builtin widget <code>backward-word</code>. You can even
give it a prefix argument, which is passed down; <code>\e3</code> then whatever you
bound the widget to (or <code>\exmy-widget</code>) will go backward three words.</p>
<p>Suppose you wanted to pass your own prefix argument to <code>backward-word</code>,
instead of what the user typed? Or suppose you want to take account of
the prefix argument, but do something different with it? Both are
possible.</p>
<p>Let's take the first of those. You can supply a prefix argument for this
command alone by putting <code>-n</code> <em>argument</em> after the widget name (note
this is not where most options go).</p>
<pre><code> my-widget() {
zle backward-word -n 2
}
</code></pre>
<p>This always goes backwards two words, overriding any numeric argument
given by the user. (You can redefine the function without telling zle
about it, by the way; zle just calls whatever function happens to be
defined when the widget is run.) If you put just <code>-N</code> after the name
instead, it will cancel out any prefix given by the user, without
introducing a new one.</p>
<p>The other part of prefix handling --- intercepting the one the user
specified and maybe modifying it --- introduces one of the most
important parts of user-defined widgets. Zle provides various parameters
which can be read and often written to alter the behaviour of the editor
or even the text being edited. In this case, the parameter is <code>$PREFIX</code>.
For example,</p>
<pre><code> my-widget() {
zle backward-word -n $(( ${NUMERIC:-1} * 2 ))
}
</code></pre>
<p>This uses an arithmetic substitution to provide an argument to
<code>backward-word</code> which is twice what the user gave. Note that
<code>${NUMERIC:-1}</code> notation, which is important: most of the time, you
don't give a numeric argument to a command at all, and in that case zle
naturally enough treats <code>$NUMERIC</code> as if it wasn't set. This would mess
up the arithmetic substitution.</p>
<p>By the way, if you do make an error in a shell function, you won't see
it; you'll just get a beep, unless you've turned that off with <code>setopt nobeep</code>. The output from such functions is junked, since it would mess
up the display. So you should do any basic debugging before turning the
function into a widget, for example, stick a <code>print</code> in front and run it
directly --- you can't execute widgets from outside the editor.</p>
<p>The following also works:</p>
<pre><code> my-widget() {
(( NUMERIC = ${NUMERIC:-1} * 2 ))
zle backward-word
}
</code></pre>
<p>because you can alter <code>$NUMERIC</code> directly, and unless overridden by the
<code>-n</code> argument it is used by any widgets called from the function. If you
called more widgets inside the function --- and you can call as many as
you like --- the same argument would apply to all the ones that didn't
have an explicit <code>-n</code> or <code>-N</code>.</p>
<p>Some widgets allow you to specify non-numeric arguments. At the moment
these are mainly search functions, which you can give an explicit search
string. Usually, however, you want to specify a new search string each
time. The most useful way of using this I can see is to provide an
initial argument for incremental search commands. Later, I'll show you
how you can read in characters in a similar fashion to Emacs mode's <code>^r</code>
binding, <code>history-incremental-search-backwards</code>.</p>
<p><span id="l106"></span></p>
<h3 id="473-some-special-builtin-widgets-and-their-uses"><a class="header" href="#473-some-special-builtin-widgets-and-their-uses">4.7.3: Some special builtin widgets and their uses</a></h3>
<p>There are some things you might want to do with the editor in a zle
function which wouldn't be useful executed directly from zle. One is to
cause an error in the same way as a normal widget does. You can do that
with `<code>zle beep</code>'. However, this doesn't automatically stop your
function at that point; it's up to you to return from it.</p>
<p>It's possible to redefine a builtin widget just by declaring it with
`<code>zle -N</code>' and defining the corresponding function. From now on, all
existing bindings which refer to that widget will cause yours to be run
instead of the builtin one. This happens because zle doesn't actually
care what a widget does until it is run. You can see this by using
<code>bindkey</code> to define a key sequence to call an undefined widget such as
<code>any-old-string</code>. The shell doesn't complain until you actually hit the
key sequence.</p>
<p>Sometimes, however, you want to be sure to call the builtin widget, even
if the behaviour has been redefined. You can do this by putting a `<code>.</code>'
in front of the name of the widget; `<code>zle .up-line-or-history</code>' always
calls the builtin widget usually referred to as <code>up-line-or-history</code>,
even if the latter has been redefined. One use for this is to rebind
`<code>accept-line</code>' to do something whenever zle is about to pass a line up
to the shell, but to accept the line anyway: you write your own widget
<code>accept-line</code>, make sure it calls `<code>zle .accept-line</code> just before it
finishes, and then use `<code>zle -N accept-line</code>. Here's a trivial but not
entirely stupid example:</p>
<pre><code> accept-line() {
print -n "\e]2;Executing $BUFFER\a"
zle .accept-line
}
zle -N accept-line
</code></pre>
<p>Now every time you hit return to execute a command, that <code>print</code> command
will be executed first. As written, it puts `<code>Executing</code>' and then the
contents of the command line (see below) into the title of your xterm
window, assuming it understands the usual xterm escape sequences. In
fact, this particular example is usually handled with the special shell
function (not zle function) `<code>preexec</code>' which is passed a command line
about to be executed as an argument instead of in <code>$BUFFER</code>. There seems
to be a side effect of rebinding <code>accept-line</code> that the return key stops
working in the minibuffer under some circumstances.</p>
<p>Note that to undo the fact that return executes your new widget, you
need to alias <code>accept-line</code> back to <code>.accept-line</code>:</p>
<pre><code> zle -A .accept-line accept-line
</code></pre>
<p>If you have trouble remembering the order, as with most alias or rename
commands in zsh and UNIX generally, including <code>ln</code> and <code>bindkey -A</code>, the
existing command, the one whose properties you want to keep, comes
first, while the new name for it comes second. Also, as with those
commands, it doesn't matter if the second name on the line currently
means something else; that will be replaced by the new meaning.
Afterwards, you don't need to worry about your own <code>accept-line</code> widget;
zle handles the details of removing widgets when they're no longer
referred to. The function's still there, however, since as far as the
rest of the shell is concerned it's just an ordinary shell function
which you need to `<code>unfunction</code>' to remove.</p>
<p>Do remember, however, not to delete a widget which redefines a basic
internal widget by the obvious command</p>
<pre><code> # Noooo!
zle -D accept-line
</code></pre>
<p>which stops the return key having any effect other than complaining
there's no such widget. If you get into real trouble,
`<code>\ex.accept-line</code>' should work, as you can use the `<code>.</code>'-widgets
anywhere you can use any other except where they would redefine or
delete a `<code>.</code>' widget. Use the `<code>zle -A</code>' command above with the
extended-command form of `<code>.accept-line</code>' to return to normality. If
you try to redefine or delete a `<code>.</code>' widget, zle will tell you it's
protected. You can remove any other widget in this way, however, even if
it is still bound to a key sequence; you will then see an error if you
type that sequence.</p>
<p>One point to note about <code>accept-line</code> is that the line isn't passed up
to zsh instantly, only when your own function exits. This is pretty
obvious when you think about it; zle is called from the main shell, and
if your own zle widget hasn't finished executing, the main shell hasn't
got control back yet. But it does mean, for example, that if you modify
the command line after a call to <code>accept-line</code> or <code>.accept-line</code>, those
changes are reflected in the line passed up to the shell:</p>
<pre><code> # Noooo! to this one too.
accept-line() {
zle .accept-line
BUFFER='Ha ha!'
}
</code></pre>
<p>This always returns the string `<code>Ha ha!</code>' to the main shell. This is
not particularly useful unless you are constructing a Samuel Beckett
shell for display at an installation in a Parisian art gallery.</p>
<p><span id="l107"></span></p>
<h3 id="474-special-parameters-normal-text"><a class="header" href="#474-special-parameters-normal-text">4.7.4: Special parameters: normal text</a></h3>
<p>The shell makes various parameters available for easy manipulation of
the command line. You've already seen <code>$NUMERIC</code>. You may wonder what
happens if you have your own parmeter called <code>$NUMERIC</code>; after all, it's
a fairly simple string to use as a name. The good news is you don't need
to worry; when the shell runs a zle function, it simply hides any
existing occurrences of a parameter and makes its special parameters
available. Then when it exits, the original parameter is reenabled. So
all you have to worry about is making sure you don't use these special
parameters for anything else while you are inside a zle widget.</p>
<p>There are four particularly common zle parameters.</p>
<p>First, there are three ways of referring to the text on the command
line: <code>$BUFFER</code> is the entire line as a string, <code>$LBUFFER</code> is the line
left of the cursor position, and <code>$RBUFFER</code> is the line after it
including the character under the cursor, so that the division is always
at the point where the next inserted character would go. Any or all of
these may be empty, and <code>$BUFFER</code> is always the string
<code>$LBUFFER$RBUFFER</code>.</p>
<p>The necessary counterpart to these is <code>$CURSOR</code>, which is the cursor
position with 1 being the first character. If you know how the shell
handles substrings in parameter substitutions, you will be able to see
that <code>$LBUFFER</code> is <code>$BUFFER[1,$CURSOR-1]</code>, while <code>$RBUFFER</code> is
<code>$BUFFER[$CURSOR,-1]</code> (unless you are using the option <code>KSH_ARRAYS</code> for
compatibility of indexes with ksh --- this isn't recommended for
implementing zle or completion widgets as it causes confusion with the
ones supplied with the shell).</p>
<p>The really useful thing about these is that they are modifiable. If you
modify <code>$LBUFFER</code> or <code>$RBUFFER</code>, then <code>$BUFFER</code> and <code>$CURSOR</code> will be
modified appropriately; lengthening or shortening <code>$LBUFFER</code> increases
or decreases <code>$CURSOR</code>. If you modify <code>$BUFFER</code>, you may need to set
<code>$CURSOR</code> yourself as the shell can't tell for sure where the cursor
should be. If you alter <code>$CURSOR</code>, characters will be moved between
<code>$LBUFFER</code> and <code>$RBUFFER</code>, but <code>$BUFFER</code> will remain the same.</p>
<p>This makes tasks along the lines of basic movement and deletion commands
extremely simple, often just a matter of pattern matching. However, it
definitely pays to know about zsh's more sophisticated pattern matching
and parameter substitution features, described in the next chapter. For
example, if you start a widget function with</p>
<pre><code> emulate -L zsh
setopt extendedglob
LBUFFER=${LBUFFER%%[^[:blank:]]##}
</code></pre>
<p>then <code>$LBUFFER</code> contains the line left of the cursor stripped of all the
non-blank characters (usually anything except space or tab) immediately
to the left of the cursor.</p>
<p>This function uses the parameter substitution feature
`<code>${</code><em>param</em><code>%%</code><em>pattern</em><code>}</code>' which removes the longest match of
<em>pattern</em> from the end of <code>$</code><em>param</em>. The `<code>emulate -L zsh</code>' ensures
the shell options are set appropriately for the function and makes all
option settings local, and `<code>setopt extendedglob</code>' which turns on the
extended pattern matching features; it is this that makes the sequence
`<code>##</code>' appearing in the pattern mean `at least one repetition of the
previous pattern element'. The previous pattern element is `anything
except a blank character'. Hence, all occurrences of non-blank
characters are removed from the end of <code>$LBUFFER</code>.</p>
<p>If you want to move the cursor over those characters, you can tweak the
function slightly:</p>
<pre><code> emulate -L zsh
setopt extendedglob
chars=${(M)LBUFFER%%[^[:blank:]]##}
(( CURSOR -= ${#chars} ))
</code></pre>
<p>The string `<code>(M)</code>' has appeared at the start of the parameter
substitution. This is part of zsh's unique system of parameter flags;
this one means `insert the matched portion of the substitution'. In
other words, instead of returning <code>$LBUFFER</code> stripped of non-blank
characters at the end, the substitution returns those very characters
which it would have stripped. To skip over them is now a simple matter
of decreasing <code>$CURSOR</code> by the length of that string.</p>
<p>You'll find if you try these examples that they probably don't do quite
what you want. In particular, they don't handle any blank characters
found next to the non-blank ones which normal word-orientated functions
do. However, you now have enough information to add tests for that
yourself.</p>
<p>If you get more sophisticated, you can then add handling for <code>$NUMERIC</code>.
Remember this isn't set unless the user gave it explicitly, so it's up
to you to treat it as 1 in that case.</p>
<p><span id="l108"></span></p>
<h3 id="475-other-special-parameters"><a class="header" href="#475-other-special-parameters">4.7.5: Other special parameters</a></h3>
<p>A large fraction of what you are likely to want to do can be done with
the parameters we've already met. Here are some hints as to how you
might want to use some of the other parameters available. As always, for
a complete list with rather less in the way of hints see the manual.</p>
<p><code>$KEYS</code> tells you the keys which were used to call the widget; it's a
string of those raw characters, not turned into the <code>bindkey</code> format. In
other words, if it was a single key (including possibly a control key or
a meta key), <code>$KEYS</code> will just contain a single character. So you can
change the widget's behaviour for different keys. Here's a very (very)
simple function like <code>self-insert</code>:</p>
<pre><code> LBUFFER=$LBUFFER$KEYS
</code></pre>
<p>Note this doesn't work very well with <code>\ex</code> extended command handling;
you just get the <code>^m</code> from the end of the line. You need to make sure
any widgets which use <code>$KEYS</code> are sensibly bound. This also doesn't
handle numeric arguments to repeat characters; it's a fairly simple
exercise (particularly given zsh's `<code>repeat</code>' loop) to add that.</p>
<p><code>$WIDGET</code> and <code>$LASTWIDGET</code> tell you the name of the current widget
being executed and the one before that. These don't sound all that
useful at first hearing. However, you can use <code>$WIDGET</code> together with
the fact that a widget doesn't need to have the same name as the
function that defines it. You can define</p>
<pre><code> zle -N this-widget function
zle -N that-widget function
</code></pre>
<p>and test <code>$WIDGET</code> inside <code>function</code> to see if it contains <code>this-widget</code>
or <code>that-widget</code>. If these have a lot of shared code, that is a
considerable simplification without having to write extra functions.</p>
<p><code>$LASTWIDGET</code> tends to be used for a slightly different purpose:
checking whether the last command to be executed was the same as the
current one, or maybe was just friendly with it. Here are edited
highlights of the function <code>up-line-or-beginning-search</code>, a sort of
cross between <code>up-line-or-search</code> and
<code>history-beginning-search-backward</code> which has been added to the shell
distribution for <code>4.1</code>. If there are previous lines in the buffer, it
moves up through them; else if it's the first in a sequence of calls to
this function it remembers the cursor position and looks backwards for a
line with the same text from the start up to that point, and puts the
cursor at the end of the line; else if the same widget has just been
executed, it uses the old cursor position to search for another match
further back in the history.</p>
<pre><code> if [[ $LBUFFER == *$'\n'* ]]; then
zle .up-line-or-history
__searching=''
else
if [[ $LASTWIDGET = $__searching ]]; then
CURSOR=$__savecursor
else
__savecursor=$CURSOR
fi
__searching=$WIDGET
zle .history-beginning-search-backward
zle .end-of-line
fi
</code></pre>
<p>We test <code>$__searching</code> instead of <code>$WIDGET</code> directly to be able to tell
the case when we are moving lines instead of searching. <code>$__savecursor</code>
gives the position for the backward search, after which we put the
cursor at the end of the line. The parameters beginning `<code>__</code>' aren't
local to the function, because we need to test them from the previous
execution, so they have been given underscores in front to try to
distinguish them from other parameters which might be around.</p>
<p>You'll see that the actual function supplied in the distribution is a
little more complicated than this; for one thing, it uses styles set by
the user to decide it's behaviour. Styles are described for use with
completion widgets in <a href="zshguide06.html#comp">chapter 6</a>, but you can use
them exactly the same way in zle functions.</p>
<p>The full version of <code>up-line-or-beginning-search</code> uses another
parameter, <code>$PREBUFFER</code>. This contains any text already absorbed by
<code>zle</code> which you can no longer edit --- in other words, text read in
before the shell prompted with <code>$PS2</code> for the remainder. Testing `<code>[[ -n $PREBUFFER ]]</code>' therefore effectively tests whether you are at the
<code>$PS2</code>. You can use this to implement behaviour after the fashion of
<code>push-line-or-edit</code>.</p>
<p><span id="l109"></span></p>
<h3 id="476-reading-keys-and-using-the-minibuffer"><a class="header" href="#476-reading-keys-and-using-the-minibuffer">4.7.6: Reading keys and using the minibuffer</a></h3>
<p>Every now and then you want the editor to do a sequence of operations
with user input in the middle. This is usually done by a combination of
two commands.</p>
<p>First, you may need to prompt the user in the minibuffer, just like
<code>\ex</code> does. You can do this with `<code>zle -R</code>'. Its basic function is to
redisplay the command line, flushing all the changes you have made in
your function so far, but you can give it a string argument which
appears in the minibuffer, just below the command line. You can give it
a list of other strings after that, which appear in a similar way to
lists of possible completions, but have no special significance to zle
in this case.</p>
<p>To get input back from the user, you can use `<code>read -k</code>' which reads a
single key (not a sequence; no lookup takes place). This command is
always available in the shell, but in this case it is handled by zle
itself. The key is returned as a raw byte. Two facilities of arithmetic
evaluation are useful for handling this key: `<code>#key</code>' returns the ASCII
code for the first character of <code>$key</code>, while `<code>##</code><em>key</em>' returns the
ASCII code for <em>key</em>, which is in the form that <code>bindkey</code> would
understand. For example,</p>
<pre><code> read -k key
if (( #key == ##\C-g )); then
...
</code></pre>
<p>makes the use of arithmetic evaluation. The form on the left turns the
first character in <code>$key</code> into a number, the second turns the literal
bindkey-style string <code>\C-g</code> into a number (ASCII 7, since 1 to 26 are
just <code>\C-a</code> to <code>\C-z</code>). Don't confuse either of these forms with
`<code>$#key</code>', which is the length of the string in the parameter, in this
case almost certainly 1 for a single byte; this form works both inside
and outside arithmetic substitution, the other forms only inside. The
`<code>(( ... ))</code>' form is recommended for arithmetic substitutions whenever
possibly; you can do it with the basic `<code>[[ ... ]]</code>' form, since
`<code>-eq</code>' and similar tests treat both sides as arithmetic, though you
may need extra quoting; however, the only good reason I know for doing
that is to avoid using two types of condition syntax in the same complex
test.</p>
<p>These tricks are only really useful for quite complicated functions. For
an example, look at the function <code>incremental-complete-word</code> supplied
with the zsh source distribution. This function doesn't add to clarity
by using the form `<code>#\\C-g</code>' instead of `<code>##\C-g</code>'; it does the same
thing but the double backslash is very confusing, which is why the other
form was introduced.</p>
<p><span id="l110"></span></p>
<h3 id="477-examples"><a class="header" href="#477-examples">4.7.7: Examples</a></h3>
<p><strong>transpose-words-about-point</strong></p>
<p>This function is a variant on <code>transpose-words</code>. It has various twists.
First, the words in question are always space-delimited, neither shell
words nor words in the <code>$WORDCHARS</code> sense. This makes it fairly
predictable.</p>
<p>Second, it will transpose words about the current point (hence its name)
even if the character under the cursor is not a whitespace character. I
find this useful because I am eternally typing compound words a bit like
`<code>function_name</code>' only to find that what I should have typed was
`<code>name_function</code>'. Now I just position the cursor over the underscore
and execute this widget.</p>
<pre><code> emulate -L zsh
setopt extendedglob
local match mbegin mend pat1 pat2 word1 word2 ws1 ws2
pat1=${LBUFFER%%(#b)([^[:blank:]]##)([[:blank:]]#)}
word1=$match[1]
ws1=$match[2]
match=()
pat2=${RBUFFER##(#b)(?[[:blank:]]#)([^[:blank:]]##)}
ws2=$match[1]
word2=$match[2]
if [[ -n $word1 && -n $word2 ]]; then
LBUFFER="$pat1$word2$ws1"
RBUFFER="$ws2$word1$pat2"
else
zle beep
fi
</code></pre>
<p>The only clever stuff here is the pattern matching. It makes a great
deal of use of `backreferences' an extended globbing feature which is
used in all forms of pattern matching including, as in this case,
parameter substitution. It will be described fully in the next chapter.
The key things to look for are the `<code>(#b)</code>', which activates
backreferences if the option <code>EXTENDED_GLOB</code> is turned on, the
parentheses following that, which mark out the bits you want to refer
to, and the references to elements of the array <code>$match</code>, which store
those bits. The shell also sets <code>$mbegin</code> and <code>$mend</code> to give the
positions of the start and end of those matches, which is why those
parameters are made local; we want to preserve them from being seen
outside the function even though we don't actually use them.</p>
<p>You might also need to know about the `<code>#</code>' characters: one after a
pattern means `zero or more repetitions', and two mean `one or more
repetitions'. Finally, `<code>[:blank:]</code>' in a character class refers to any
blank character; when negated, as in the character class
`<code>[^[:blank:]]</code>', it means any non-blank character. With the `<code>#</code>'s we
match a series blank or non-blank characters. Given that, you can work
out the rest of what's going on.</p>
<p>Here's a more sophisticated version of that. If you found the previous
one heavy going. you probably don't want to look too closely at this.</p>
<pre><code> emulate -L zsh
setopt extendedglob
local wordstyle blankpat wordpat1 wordpat2
local match mbegin mend pat1 pat2 word1 word2 ws1 ws2
zstyle -s ':zle:transpose-words-about-point' word-style wordstyle
case $wordstyle in
(shell) local bufwords
# This splits the line into words as the shell understands them.
bufwords=(${(z)LBUFFER})
wordpat1="${(q)bufwords[-1]}"
# Take substring of RBUFFER to skip over first character,
# which is the one under the cursor.
bufwords=(${(z)RBUFFER[2,-1]})
wordpat2="${(q)bufwords[1]}"
blankpat='[[:blank:]]#'
;;
(space) blankpat='[[:blank:]]#'
wordpat1='[^[:blank:]]##'
wordpat2=$wordpat1
;;
(*) local wc=$WORDCHARS
if [[ $wc = (#b)(?*)-(*) ]]; then
# We need to bring any `-' to the front to avoid confusing
# character classes... we get away with `]' since in zsh
# this isn't a pattern character if it's quoted.
wc=-$match[1]$match[2]
fi
# A blank is anything not in the character class consisting
# of alphanumerics and the characters in $wc.
# Quote $wc where necessary, because we don't want those
# characters to be considered as pattern characters later on.
blankpat="[^${(q)wc}a-zA-Z0-9]#"
# and a word character is anything else.
wordpat1="[${(q)wc}a-zA-Z0-9]##"
wordpat2=$wordpat1
;;
esac
# The eval makes any special characters in the parameters active.
# In particular, we need the surrounding `[' s to be `real'.
# This is why we quoted the wordpats in the `shell' option, where
# they have to be treated as literal strings at this point.
eval pat1='${LBUFFER%%(#b)('${wordpat1}')('${blankpat}')}'
word1=$match[1]
ws1=$match[2]
match=()
eval pat2='${RBUFFER##(#b)(?'${blankpat}')('${wordpat2}')}'
ws2=$match[1]
word2=$match[2]
if [[ -n $word1 && -n $word2 ]]; then
LBUFFER="$pat1$word2$ws1"
RBUFFER="$ws2$word1$pat2"
else
zle beep
fi
</code></pre>
<p>What has been added is the ability to use a style to define how the
shell finds a `word'. By default, words are the same as what the shell
usually thinks of as a word; this is handled by the branch of the case
statement which uses `<code>$WORDCHARS</code>' and a little extra trickery to get
a pattern which matches the set of characters considered parts of a
word. We used the <code>eval</code>'s because it allowed us to have some bits of
<code>$wordpat1</code> and friends active as pattern characters while others were
quoted.</p>
<p>This introduces two types of parameter expansion flags:
<code>${(q)</code><em>param</em><code>}</code> adds backslashes to quote special characters in
<code>$</code><em>param</em>, so that when the parameter appears after <code>eval</code> the result
is just the original string. <code>${(z)</code><em>param</em><code>}</code> splits the parameter just
as if it were a shell command line being split into a command and words,
so the result is an array; `<code>z</code>' stands for zsh-splitting or just
zplitting as you fancy.</p>
<p>If you set</p>
<pre><code> zstyle ':zle:*' word-style space
</code></pre>
<p>you get back to the behaviour of the original function.</p>
<p>Finally, if you replace `<code>space</code>' with `<code>shell</code>' in that <code>zstyle</code>
command, you will get words as they are split for normal use within the
shell; for example try</p>
<pre><code> echo execute the widget 'between these' 'two quoted expressions'
</code></pre>
<p>and the entire quoted expressions will be transposed. You may find that
if you do this in the middle of a quoted expression, you don't get a
sensible result; that's because the <code>(z)</code>-splitting doesn't know what to
do with the improperly completed quotes to its left and right. Some
versions of the shell have a bug (fixed in 4.0.5) that the expressions
which couldn't be split properly, because the quotes weren't complete,
have an extra space character at the end.</p>
<p><strong>insert-numeric</strong></p>
<p>Here's a widget which allows you to insert an ASCII character which you
know by number. I can't for the life of me remember where it came from,
but it's been lying around apparently for two and a half years (please
do email me if you think you wrote it, otherwise I'll assume I did). You
can give it a numeric prefix (that's the easy part of the function),
else it will prompt you for a number. If you type `<code>x</code>' or `<code>o</code>' the
number is treated as hexadecimal or octal, respectively, else as
decimal.</p>
<pre><code> # Set up standard options.
# Important for portability.
emulate -L zsh
# x must display in hexadecimal
typeset -i 16 x
if (( ${+NUMERIC} )); then
# Numeric prefix given; just use that.
x=$NUMERIC
else
# We need to read the ASCII code.
local msg modes key mode=dec code char
# Prompt for and read a base.
integer base=10
zle -R "ASCII code (o -> oct, x -> hex) [$mode]: "
read -k key
case $key in
(o) base=8 mode=oct
zle -R "ASCII code [$mode]: "
read -k key
;;
(x) base=16 mode=hex
zle -R "ASCII code [$mode]: "
read -k key
;;
esac
# Now we are looking for numbers in that base.
# Loop until newline or return.
while [[ '#key' -ne '##\n' && '#key' -ne '##\r' ]]; do
if [[ '#key' -eq '##^?' || '#key' -eq '##^h' ]]; then
# Delete a character
[[ -n $code ]] && code=${code[1,-2]}
elif [[ ($mode == hex && $key != [0-9a-fA-f]) ||
($mode == dec && $key != [0-9]) ||
($mode == oct && $key != [0-7]) ]]; then
# Character not in range, beep
zle beep
elif [[ '#key' -eq '##\C-g' ]]; then
# Abort: returning 1 signals to zle that this
# is an abnormal termination.
return 1
else
code="${code}${key}"
fi
char=
if [[ -n $code ]]; then
# Work out the character using the
# numbers typed so far.
(( x = ${base}#${code} ))
if (( x > 255 )); then
zle beep
code=${code[1,-2]}
[[ -n $code ]] && (( x = ${base}#${code} ))
fi
[[ -n $code ]] && eval char=\$\'\\x${x##???}\'
fi
# Prompt for any more digits, showing
# the character as it would be inserted.
zle -R "ASCII code [$mode]: $code${char:+ = $char}"
read -k key || return 1
done
# If aborted with no code, return
[[ -z $code ]] && return 0
# Now we have the ASCII code.
(( x = ${base}#${code} ))
fi
# Finally, if we have a single-byte character,
# insert it to the left of the cursor
if (( x < 0 || x > 255 )); then
return 1
else
eval LBUFFER=\$LBUFFER\$\'\\x${x##???}\'
fi
</code></pre>
<p>This shows how to do interactive input. The `<code>zle -R</code>'s prompt the
user, while the `<code>read -k</code>'s accept a character at a time. As an extra
feature, while you are typing the number, the character that would be
inserted if you hit return is shown. The widget also handles deletion
with backspace or the (UNIX-style, not PC-style) delete key.</p>
<p>One blight on this is the way of turning the number in x into a
character, which is done by all those <code>eval</code>s and backslashes. It uses
the feature that e.g. <code>$'\x41'</code> is the character <code>0x41</code> (an ASCII `A').
To use this, we must make sure the character (stored in x) appears as
hexadecimal, and following ksh zsh outputs hexadecimal numbers as
`<code>16#41</code>' or similar. (The new option <code>C_BASES</code> shows hexadecimal
numbers as 0x41 and similar, but here we need the plain number in any
case.) Hence we strip the `<code>16#</code>' and construct our <code>$'\x41'</code>. Now we
need to persuade the shell to interpret this as a quoted string by
passing it to <code>eval</code> with the special characters (<code>$</code>, <code>\</code>, <code>'</code>) quoted
with a backslash so that they aren't interpreted too early.</p>
<p>By the way, note that zsh only handles ordinary 8-bit characters at the
moment. It doesn't matter if some do-gooder on your system has set
things up to use UTF-8 (a UNIX-friendly version of the international
standard for multi-byte characters, Unicode) to appeal to the
international market, I'm afraid zsh is stuck with ISO 8859 and similar
character sets for now.</p>
<div id="chapter_begin" style="break-before: page; page-break-before: always;"></div><!-- START doctoc generated TOC please keep comment here to allow auto update -->
<!-- DON'T EDIT THIS SECTION, INSTEAD RE-RUN doctoc TO UPDATE -->
<p><strong>Table of Contents</strong> <em>generated with <a href="https://github.com/thlorenz/doctoc">DocToc</a></em></p>
<ul>
<li><a href="zshguide05.html#chapter-5-substitutions">Chapter 5: Substitutions</a>
<ul>
<li><a href="zshguide05.html#51-quoting">5.1: Quoting</a>
<ul>
<li><a href="zshguide05.html#511-backslashes">5.1.1: Backslashes</a></li>
<li><a href="zshguide05.html#512-single-quotes">5.1.2: Single quotes</a></li>
<li><a href="zshguide05.html#513-posix-quotes">5.1.3: POSIX quotes</a></li>
<li><a href="zshguide05.html#514-double-quotes">5.1.4: Double quotes</a></li>
<li><a href="zshguide05.html#515-backquotes">5.1.5: Backquotes</a></li>
</ul>
</li>
<li><a href="zshguide05.html#52-modifiers-and-what-they-modify">5.2: Modifiers and what they modify</a></li>
<li><a href="zshguide05.html#53-process-substitution">5.3: Process Substitution</a></li>
<li><a href="zshguide05.html#54-parameter-substitution">5.4: Parameter substitution</a>
<ul>
<li><a href="zshguide05.html#541-using-arrays">5.4.1: Using arrays</a></li>
<li><a href="zshguide05.html#542-using-associative-arrays">5.4.2: Using associative arrays</a></li>
<li><a href="zshguide05.html#543-substituted-substitutions-top--and-tailing-etc">5.4.3: Substituted substitutions, top- and tailing, etc.</a></li>
<li><a href="zshguide05.html#544-flags-for-options-splitting-and-joining">5.4.4: Flags for options: splitting and joining</a></li>
<li><a href="zshguide05.html#545-flags-for-options-glob_subst-and-rc_expand_param">5.4.5: Flags for options: <code>GLOB_SUBST</code> and <code>RC_EXPAND_PARAM</code></a></li>
<li><a href="zshguide05.html#546-yet-more-parameter-flags">5.4.6: Yet more parameter flags</a></li>
<li><a href="zshguide05.html#547-a-couple-of-parameter-substitution-tricks">5.4.7: A couple of parameter substitution tricks</a></li>
<li><a href="zshguide05.html#548-nested-parameter-substitutions">5.4.8: Nested parameter substitutions</a></li>
</ul>
</li>
<li><a href="zshguide05.html#55-that-substitution-again">5.5: That substitution again</a></li>
<li><a href="zshguide05.html#56-arithmetic-expansion">5.6: Arithmetic Expansion</a>
<ul>
<li><a href="zshguide05.html#561-entering-and-outputting-bases">5.6.1: Entering and outputting bases</a></li>
<li><a href="zshguide05.html#562-parameter-typing">5.6.2: Parameter typing</a></li>
</ul>
</li>
<li><a href="zshguide05.html#57-brace-expansion-and-arrays">5.7: Brace Expansion and Arrays</a></li>
<li><a href="zshguide05.html#58-filename-expansion">5.8: Filename Expansion</a></li>
<li><a href="zshguide05.html#59-filename-generation-and-pattern-matching">5.9: Filename Generation and Pattern Matching</a>
<ul>
<li><a href="zshguide05.html#591-comparing-patterns-and-regular-expressions">5.9.1: Comparing patterns and regular expressions</a></li>
<li><a href="zshguide05.html#592-standard-features">5.9.2: Standard features</a></li>
<li><a href="zshguide05.html#593-extensions-usually-available">5.9.3: Extensions usually available</a></li>
<li><a href="zshguide05.html#594-extensions-requiring-extended_glob">5.9.4: Extensions requiring <code>EXTENDED_GLOB</code></a></li>
<li><a href="zshguide05.html#595-recursive-globbing">5.9.5: Recursive globbing</a></li>
<li><a href="zshguide05.html#596-glob-qualifiers">5.9.6: Glob qualifiers</a></li>
<li><a href="zshguide05.html#597-globbing-flags-alter-the-behaviour-of-matches">5.9.7: Globbing flags: alter the behaviour of matches</a></li>
<li><a href="zshguide05.html#598-the-function-zmv">5.9.8: The function <code>zmv</code></a></li>
</ul>
</li>
</ul>
</li>
</ul>
<!-- END doctoc generated TOC please keep comment here to allow auto update -->
<p><span id="subst"></span><span id="l111"></span></p>
<h1 id="chapter-5-substitutions"><a class="header" href="#chapter-5-substitutions">Chapter 5: Substitutions</a></h1>
<p>This chapter will appeal above all to people who are excited by the fact
that</p>
<pre><code> print ${array[(r)${(l.${#${(O@)array//?/X}[1]}..?.)}]}
</code></pre>
<p>prints out the longest element of the array <code>$array</code>. For the
overwhelming majority that forms the rest of the population, however,
there should be plenty that is useful before we reach that stage.
Anyway, it should be immediately apparent why there is no obfuscated zsh
code competition.</p>
<p>For those who don't do a lot of function writing and spend most of the
time at the shell prompt, the most useful section of this chapter is
probably that on filename generation (i.e. globbing) at the end of the
chapter. This will teach you how to avoid wasting your time with <code>find</code>
and the like when you want to select files for a command.</p>
<p><span id="l112"></span></p>
<h2 id="51-quoting"><a class="header" href="#51-quoting">5.1: Quoting</a></h2>
<p>I've been using quotes of some sort throughout this guide, but I've
never gone into the detail. It's about time I did, since using quotes is
an important part of controlling the effects of the shell's various
substitutions. Here are the basic quoting types.</p>
<p><span id="l113"></span></p>
<h3 id="511-backslashes"><a class="header" href="#511-backslashes">5.1.1: Backslashes</a></h3>
<p>The main point to make about backslashes is that they are really
trivial. You can quote any character whatsoever from the shell with a
backslash, even if it didn't mean anything unquoted; so if the worst
comes to the worst, you can take any old string at all, whatever it has
in it --- random collections of quotes, backslashes, unprintable
characters --- quote every single character with a backslash, and the
shell will treat it as a plain string:</p>
<pre><code> print \T\h\i\s\ \i\s\ \*\p\o\i\n\t\l\e\s\s\*\ \
\-\ \b\u\t\ \v\a\l\i\d\!
</code></pre>
<p>Remember, too that, this means you need an extra layer of quotation to
pass a `<code>\n</code>', or whatever, down to <code>print</code>.</p>
<p>However, zsh has an easier way of making sure everything is quoted with
a backslash when that's needed. It's a special form of parameter
substitution, just one of many tricks you can do by supplying flags in
parentheses:</p>
<pre><code> % read string
This is a *string* with various `special' characters
% print -r -- ${(q)string}
This\ is\ a\ \*string\*\ with\ various\ \`special\'\ characters
</code></pre>
<p>The <code>read</code> builtin didn't do anything to what you typed, so <code>$string</code>
contains just those characters. The <code>-r</code> flag to print told it to print
out what came after it in raw fashion, and here's the special part:
<code>${(q)string}</code> tells the shell to output the parameter with backslashes
where needed to prevent special characters being interpreted. All
parameter flags are specific to zsh; no other shell has them.</p>
<p>The flag is not very useful there, because zsh usually (remember the
<code>GLOB_SUBST</code> option?) doesn't do anything special to characters from
substitutions anyway. Where it <em>is</em> extremely useful is if you are going
to re-evaluate the text in the substitution but still want it treated as
a plain string. So after the above,</p>
<pre><code> % eval print -r -- ${(q)string}
This is a *string* with various `special' characters
</code></pre>
<p>and you get back what you started with, because at the <code>eval</code> of the
command line the backslashes put in by the <code>(q)</code> flag meant that the
value was treated as a plain string.</p>
<p>You can strip off quotes in parameters, too; the flag <code>(Q)</code> does this.
It doesn't care whether backslashes or single or double quotes are used,
it treats them all the way the shell's parser would. You only need this
when the parameter has somehow acquired quotes in its value. One way
this can happen is if you try reading a file containing shell commands,
and for this there's another trick: the <code>(z)</code> flag splits a line into an
array in the same way as if the line had been read in and was, say,
being assigned to an array. Here's an example:</p>
<pre><code> % cat file
print 'a quoted string' and\ another\ argument
% read -r line <file
% for word in ${(z)line}; do
for> print -r "quoted: $word"
for> print -r "unquoted: ${(Q)word}"
for> done
quoted: print
unquoted: print
quoted: 'a quoted string'
unquoted: a quoted string
quoted: and\ another\ argument
unquoted: and another argument
</code></pre>
<p>You will notice that the <code>(z)</code> doesn't remove any of the quotes from the
words read in, but the <code>(Q)</code> flag does. Note the <code>-r</code> flags to both
<code>read</code> and <code>print</code>: the first prevents the backslashes being absorbed by
<code>read</code>, and the second prevents them being absorbed by <code>print</code>. I'm
afraid backslashes can be a bit of a pain in the neck.</p>
<p><span id="l114"></span></p>
<h3 id="512-single-quotes"><a class="header" href="#512-single-quotes">5.1.2: Single quotes</a></h3>
<p>The only thing you can't quote with single quotes is another single
quote. However, there's an option <code>RC_QUOTES</code>, where two single quotes
inside a single-quoted string are turned into one. Apparently `<code>RC</code>'
refers to the shell <code>rc</code> which appeared in plan9; it seems to be one of
those programmes that some people get fanatically worked up about while
the rest of us can't quite work out why. Zsh users may sympathise. (This
was corrected by Oliver Kiddle and Bart Schaefer after I guessed
incorrectly that <code>RC</code> stood for recursive, although you're welcome to
think of it that way anyway. It doesn't really work for
<code>RC_EXPAND_PARAM</code>, however, which is definitely from the <code>rc</code> shell, and
if you look at the source code you will find a variable called
`<code>plan9</code>' which is tested to see if that option is in effect.)</p>
<p>You might remember something like this from BASIC, although in that case
with double quotes --- in zsh, it works only with single quotes, for
some reason. So,</p>
<pre><code> print -r 'A ''quoted'' string'
</code></pre>
<p>would usually give you the output `<code>A quoted string</code>', but with the
option set it prints `<code>A 'quoted' string</code>'. The <code>-r</code> option to <code>print</code>
doesn't do anything here, it's just to show I'm not hiding anything.
This is usually a useful and harmless option to have set, since there's
no other good reason for having two quotes together within quotes.</p>
<p>The standard way of quoting single quotes is to end the quote, insert a
backslashed single quote, and restart quotes again:</p>
<pre><code> print -r 'A '\''quoted'\'' string'
</code></pre>
<p>which is unaffected by the option setting, since the quotes immediately
after the backslashes are always treated as an ordinary printable
character. What you <em>can't</em> ever do is use backslashes as a way of
quoting characters inside single quotes; they are just treated as
ordinary characters there.</p>
<p>You can make parameter flags produce strings quoted with single quotes
instead of backslashes by doubling the `<code>q</code>': `<code>${(qq)param}</code>' instead
of `<code>${(q)param}</code>'. The main use for this is that the result is shorter
if you know there are a lot of special characters in the string, and
it's also a bit more easy to read for humans rather than machines, but
usually it gains nothing over the other form. It can tell whether you
have <code>RC_QUOTES</code> set and uses that to make the string even shorter, so
be careful if you might use the resulting string somewhere where the
option isn't set.</p>
<p><span id="l115"></span></p>
<h3 id="513-posix-quotes"><a class="header" href="#513-posix-quotes">5.1.3: POSIX quotes</a></h3>
<p>There's a relative of single quotes which uses the syntax <code>$'</code> to
introduce a quoted string and <code>'</code> to end it; I refer to them as `POSIX
quotes' because they appear in the POSIX standard and I don't know what
else to call them; `string quotes' is one possibility, but sounds a bit
vague (what else would you quote?) The difference from single quotes is
that they understand the same backslash sequences as the print builtin.
Hence you can have the convenience of using `<code>\n</code>' for newline, `<code>\e</code>'
for escape, `<code>\xFF</code>' for an arbitrary character in hexadecimal, and so
on, for any command:</p>
<pre><code> % cat <<<$'Line\tone\nLine\ttwo'
Line one
Line two
</code></pre>
<p>Remember the `here string' notation `<code><<<</code>', which supplies standard
input for the command. Hence the output shows exactly how the quoted
string is being interpreted. It is the same as</p>
<pre><code> % print 'Line\tone\n\Line\ttwo'
Line one
Line two
</code></pre>
<p>but there the interpretation is done inside <code>print</code>, which isn't always
convenient. POSIX quotes are currently rather underused.</p>
<p>This is as good a point as any to mention that the shell is completely
`eight-bit clean', which means you can have any of the 256 possible
characters anywhere in your string. For example, <code>$'foo\000bar'</code> has an
embedded ASCII NUL in it (that's not a misprint --- officially, ASCII
non-printing characters have two- or three-letter abbreviations).
Usually this terminates a string, but the shell works around this when
you are using it internally; when you try and pass it as an argument to
an external programme, however, all bets are off. Almost certainly the
first NUL in that case will cause the programme to think the string is
finished, because no information about the length of arguments is passed
down and there's nothing the shell can do about it. Hence, for example:</p>
<pre><code> % echo $'foo\000bar'
foobar
% /bin/echo $'foo\000bar'
foo
</code></pre>
<p>The shell's <code>echo</code> knows about the shell's 8-bit conventions, and prints
out the NUL, which the terminal doesn't show, then the remainder of the
string. The external version of <code>echo</code> didn't know any better than to
stop when it reached the NUL.</p>
<p>There are actually uses for embedded NULs: some versions of <code>find</code> and
<code>xargs</code>, for example, will put or accept NULs instead of newlines
between their bits of input and output (as distinct from command line
arguments), which is much safer if there's a chance the input or output
can contain a live newline. Using <code>$'\000'</code> allows the shell to fit in
very comfortably with these. If you want to try this, the corresponding
options are <code>-print0</code> for <code>find</code> (print with a NUL terminator instead of
newline) and <code>-0</code> for <code>xargs</code> (read input assuming a NUL terminator).</p>
<p>In older versions of the shell, characters with the top bit set, such as
those from non-English character sets found in ISO 8859 fonts, could
cause problems, since the shell also uses such characters internally to
represent its own special characters, but recent versions of the shell
(from about 3.0) side-step this problem in the same way as for NULs. Any
remaining problems --- it's quite tricky to handle this completely
consistently --- are bugs and should be reported.</p>
<p>You can force parameters to be quoted with POSIX quotes by the somewhat
absurd expedient of making the <code>q</code> in the quote flag appear a total of
four times. I can't think why you would ever want to do that, except
that it will turn newlines into `<code>\n</code>' and hence the result will fit on
a single (maybe rather long) line. Plus you get the replacement of funny
characters with escape sequences.</p>
<p><span id="l116"></span></p>
<h3 id="514-double-quotes"><a class="header" href="#514-double-quotes">5.1.4: Double quotes</a></h3>
<p>Double quotes allow some, but not all, forms of substitution inside.
More specifically, they allow parameter expansion, command substitution
and arithmetic substitution, but not any of the others: process
substitution doesn't happen, braces and initial tildes and equals signs
are not expanded and patterns are not special. Here's a table; each
expression on the left is some command line argument, and the results
show what is substituted if it appears outside quotes, or in double
quotes.</p>
<pre><code> Expression Outside quotes In double quotes
------------------------------------------------
=(echo hi mum) /tmp/zshTiqpL =(echo hi mum)
$ZSH_VERSION 4.0.1 4.0.1
$(echo hi mum) hi mum hi mum
$((6**2 + 6)) 42 42
{a,b}cd acd bcd {a,b}cd
~/foo /home/pws/foo ~/foo
.zl* .zlogin .zlogout .zl*
</code></pre>
<p>That `<code>/tmp/zshTiqpL</code>' could be any temporary filename, and indeed
several of the other substitutions will be different in your case.</p>
<p>You might already have guessed that `<code>${(qqq)string}</code>' forces <code>$string</code>
to use double quotes to quote its special characters. As with the other
forms, this is all properly handled --- the shell knows just which
characters need quoting inside double quotes, and which don't.</p>
<p><strong>Word-splitting in double quotes</strong></p>
<p>Where the substitutions are allowed, the (almost) invariable side effect
of double quotes is that word-splitting is suppressed. You can see this
using `<code>print -l</code>', which prints one argument per line:</p>
<pre><code> % array=(one two)
% print -l $(echo foo bar) $array
foo
bar
one
two
% print -l "$(echo foo bar) $array"
foo bar one two
</code></pre>
<p>The reason this is `almost' invariable is that parameter substitution
allows you to specify that normal word-splitting will occur. There are
two ways of doing this; both use the symbol `<code>@</code>'. You probably
remember this from the parameter `<code>$@</code>' which has just that effect when
it appears in double quotes: the arguments to the script or function are
split into words like a normal array, except that empty arguments are
not removed. I covered this at some length in <a href="zshguide03.html#syntax">chapter
3</a>.</p>
<p>This is extended for other parameters in the following way:</p>
<pre><code> % array=(one two three)
% print -l "${array[@]}"
one
two
three
</code></pre>
<p>and more generally for all forms of substitution using another flag,
<code>(@)</code>:</p>
<pre><code> % print -l "${(@)array}"
one
two
three
</code></pre>
<p><strong>Digression on subscripts</strong></p>
<p>The version with flags is perhaps less clear than the other, but it can
appear in lots of different places. For example, here is how you pick a
slice of an array in zsh:</p>
<pre><code> % print -l ${array[2,-1]}
two
three
</code></pre>
<p>where negative numbers count from the end of the array. The numbers in
square brackets are referred to as subscripts. This can get the <code>(@)</code>
treatment, too:</p>
<pre><code> % print -l "${(@)array[2,-1]}"
two
three
</code></pre>
<p>Although it's probably not obvious, you can use the other notation in
this case:</p>
<pre><code> % print -l "${array[@][2,-1]}"
two
three
</code></pre>
<p>The shell will actually handle arbitrary numbers of subscripts in
parameter substitutions, not just one; each applies to the result of the
previous one:</p>
<pre><code> % print -l "${array[@][2,-1][1]}"
two
</code></pre>
<p>What you have to watch out for is that that last subscript selected a
single word. You can continue to apply subscripts, but they will apply
only on the <em>characters</em> in that word, not on array elements:</p>
<pre><code> % print -l "${array[@][2,1][1][2,-1]}"
wo
</code></pre>
<p>We've now strayed severely off topic: the subscripts will of course work
quite independently from whether the word is being split or appears in
double quotes. Despite the joining of words that occurs in double
quotes, subscripts of arrays still select array elements. This is a
consequence of the order in which the rules of parameter expansion
apply. There is a long, involved section on this in the <code>zshexpn</code> manual
entry (look for the heading `Rules' there or in the `Parameter
Expansion' node of the corresponding Info or HTML file).</p>
<p><strong>Word-splitting of quoted command substitutions</strong></p>
<p>Zsh has the useful feature that you can force the shell to apply the
rules of parameter expansion to the result of a command substitution. To
see where that might be useful, consider the case of the special
`command substitution' (although it's handled entirely in the shell,
not by running an external command) which puts the contents of a file on
the command line:</p>
<pre><code> % args() { print $#; } # report number of arguments
% cat file
Words on line one
Words on line two
% args $(<file)
8
% args "$(<file)"
1
</code></pre>
<p>The unquoted substitution split the file into individual words; the
quoted substitution didn't split it at all. These are the standard shell
rules.</p>
<p>It's very common, however, that you want one line per argument, not
splitting on spaces within the line. This is where parameter expansion
can come in. There is a flag <code>(f)</code> which says `split the result of the
expansion, one word per line'. Here's how to use it in this case:</p>
<pre><code> % args "${(f)$(<file)}"
2
</code></pre>
<p>Where you would usually put the name of a parameter, you put the command
substitution instead, and the shell operates on the result of that (note
that it does not treat the result as the name of a parameter, but as a
value --- this is discussed in more detail below). The double quotes
were necessary because otherwise the file would already have been split
into individual words by the time the parameter substitution came to
look at the result. You can easily verify that the two arguments are the
individual lines of the file. I don't remember what the `<code>f</code>' stands
for, but we were already using up flag codes quite fast when it came
along; Bart Schaefer believes it stands for `fold', which might at
least help you remember it.</p>
<p><span id="l117"></span></p>
<h3 id="515-backquotes"><a class="header" href="#515-backquotes">5.1.5: Backquotes</a></h3>
<p>The main thing to say about backquotes is that you should use the other
form of command substitution instead. There are two good reasons.</p>
<p>First, the other form can be nested:</p>
<pre><code> % print $(print $(print a word))
a word
</code></pre>
<p>Obviously that's a silly example, but the main point is that the only
time parentheses should occur unquoted in the shell is in pairs (the
patterns in case statements are an exception, but pairs of parentheses
around patterns are valid, too, and I have used that form in this
guide). Thus you can be confident that any piece of well-formatted shell
code can appear inside the command substitution.</p>
<p>This is clearly not true with <code>`...`</code>, even though the basic effect
is the same. Any unquoted <code>`</code> which happens to appear in a chunk of
code within the backquotes will be treated as the end of the quotes.</p>
<p>The second reason, which is closely related, is that it can be quite
difficult to decide how many levels of quotes are required inside a
backquoted expression. Consider:</p>
<pre><code> % print "`echo \"hello\"`"
hello
% print "$(echo \"hello\")"
"hello"
</code></pre>
<p>It's hard to explain quite what the difference here is without waving my
hands, which prevents me from typing, but the essential point is really
the same one about nesting: you can't do it with backquotes, because the
start and end symbols are the same, but you can do it with parentheses.
So in the second case there is no doubt that the embedded command line,
`<code>echo \"hello\"</code>', is to be treated exactly as if that had appeared
outside the command substitution; whereas in the first place, the quotes
within quotes had to be, um, quoted.</p>
<p>As a consequence, in</p>
<pre><code> % print "$(echo "hello")"
hello
</code></pre>
<p>you need to be careful: at first glance, the pairs of double quotes
surround `<code>$</code>(<code>echo </code>' and `)', but they don't, they are nested by
virtue of the substitution. You see the same thing with parameter
substitution:</p>
<pre><code> % unset foo
% print "${foo:-"a string"}"
a string
</code></pre>
<p>A third, less good, reason for using the form with parentheses is that
your more sophisticated friends will laugh at you otherwise. Peer
pressure is so important in this complex world.</p>
<p>That's all I have to say about command substitution, since I already
said a lot about it when I discussed the basic syntax in <a href="zshguide03.html#syntax">chapter
3</a>.</p>
<p><span id="l118"></span></p>
<h2 id="52-modifiers-and-what-they-modify"><a class="header" href="#52-modifiers-and-what-they-modify">5.2: Modifiers and what they modify</a></h2>
<p>Modifiers were introduced in <a href="zshguide02.html#init">chapter 2</a> when I
talked about `bang history', since that's where they came from. In zsh,
however, they can be used in a couple of other places. They have the
same form in each case: a colon, followed by a letter which is the code
for what the modifier does, possibly (in the case of substitutions)
followed by some other string. So, to jog your memory, unless you have
<code>NO_BANG_HIST</code> set:</p>
<pre><code> % print ~/file
/home/pws/file
% print !-1:t
file
</code></pre>
<p>where `<code>:t</code>' takes the tail (non-directory part) of the filename.</p>
<p>The second use is in parameters. This follows on very naturally. Note
that neither this nor any of the later uses of modifiers rely on the
<code>NO_BANG_HIST</code> option; that's purely for history.</p>
<pre><code> % param=~/file
% print ${param:t}
file
</code></pre>
<p>Normally you can miss out the braces in the parameter substitution, but
I tend to use them with modifiers for the sake of clarity. The fact that
the same parts of the shell are used for modifiers wherever they come
from has certain consequences:</p>
<pre><code> % print foo
foo
% ^foo^bar
bar
% param='this sentence contains a foo.'
% print ${param:&}
this sentence contains a bar.
</code></pre>
<p>The ampersand repeats the last substitution, which is the same for
parameter modifiers as for history modifiers. I find parameter modifiers
even more useful than history ones; extracting the head or tail of a
path is a very common operation on parameters.</p>
<p>Modifiers are also smart enough to handle arrays in a useful fashion.
Note this is not true of sets of arguments in history expansions;
`<code>:t</code>' will only extract one tail in that case, which may not be quite
what you're expecting:</p>
<pre><code> % print a sentence with a /real/live/bogus/path in it.
% print !!:t
path in it.
</code></pre>
<p>However, arrays <em>are</em> handled the way you might hope:</p>
<pre><code> % array=(~/.zshenv ~/.zshrc ~/.zlogout)
% print ${array:t}
.zshenv .zshrc .zlogout
</code></pre>
<p>The same logic is applied with substitutions. This means that the first
match in every element of the array is replaced:</p>
<pre><code> % array=('a bar of chocolate' 'a bar of barflies'
array> 'a barrier of barns')
% print ${array:s/bar/car/}
a car of chocolate a car of barflies a carrier of barns
</code></pre>
<p>unless, of course, you do a global replacement:</p>
<pre><code> % print ${array:gs/bar/car/}
a car of chocolate a car of carflies a carrier of carns
</code></pre>
<p>Note, however, that parameter substitution has its own <em>much</em> more
powerful equivalent, which does pattern matching, partial replacement of
modified parts of the original string, and so on. We'll come to this all
in good time.</p>
<p>The final use of modifiers is in filename generation, i.e. globbing.
Since this usually works by having special characters on the command
line, and modifiers just consist of ordinary characters, the syntax is a
little different:</p>
<pre><code> % print *.c
parser.c lexer.c input.c output.c
% print *.c(:r)
parser lexer input output
</code></pre>
<p>so you need parentheses around them. This is a special case of `glob
qualifiers' which you'll meet below; you can mix them, but the modifiers
must appear at the end. For example,</p>
<pre><code> % print -l ~/stuff/*
/home/pws/stuff/onefile.c
/home/pws/stuff/twofile.c
/home/pws/stuff/subdir
% print ~/stuff/*(.:r:t)
onefile twofile
</code></pre>
<p>The globbing qualifier `<code>.</code>' specifies that files must be regular, i.e.
not directories nor some form of special file. The `<code>:r</code>' removes the
suffix from the result, and the `<code>:t</code>' takes away the directory part.
Consequently, filename modifiers will be turned off if you set the
option <code>NO_BARE_GLOB_QUAL</code>.</p>
<p>Two final points to note about modifiers with filenames. First, it is
the only form of globbing where the result is no longer a filename; it
is always performed right at the end, after all normal filename
generation. Presumably, in the examples above, the word which was
inserted into the command line doesn't actually correspond to a real
file any more.</p>
<p>Second, although it <em>does</em> work if the word on the command line isn't a
pattern but an ordinary word with a modifier tacked on, it <em>doesn't</em>
work if that pattern, before modification, doesn't correspond to a real
file. So `<code>foo.c(:r)</code>' will only strip off the suffix if <code>foo.c</code> is
there in the current directory. This is perfectly logical given that the
attempt to match a file kicks the globbing system, including modifiers,
into action. If this is a problem for you, there are ways round; for
example, insert the right value by hand in a simple case like this, or
more realistically store the value in a parameter and apply the modifier
to that.</p>
<p><span id="l119"></span></p>
<h2 id="53-process-substitution"><a class="header" href="#53-process-substitution">5.3: Process Substitution</a></h2>
<p>I don't have much new to say on process substitution, but I do have an
example of where I find it useful. If you use the pager `less' you may
know it has the facility to preprocess the files you look at, for
example uncompressing files temporarily via the environment variable
<code>$LESSOPEN</code> (and maybe <code>$LESSCLOSE</code>). Zsh can very easily and, to my
thoroughly unbiased way of looking, more conveniently do the same thing.
Here's a subset of my zsh function front-end to less --- or indeed any
pager, which is given here by the standard environment variable <code>$PAGER</code>
with the default <code>less</code>. You can hard-wire any file-displaying command
at that point if you prefer.</p>
<pre><code> integer i=1
local args arg
args=($*)
for arg in $*; do
case $arg in
(*.bz2) args[$i]="=(bunzip2 -c ${(q)arg})"
;;
# this assumes your zcat is the one installed with gzip:
(*.(gz|Z)) args[$i]="=(zcat ${(q)arg})"
;;
(*) args=${(q)arg}
;;
esac
(( i++ ))
done
eval command ${PAGER:-less} $args
</code></pre>
<p>The main pieces of interest is how elements of the array <code>$args</code> were
replaced. The reason each argument was given an extra layer of quotes
via <code>(q)</code> is the <code>eval</code> at the end; <code>$args</code> is turned into an array of
literal characters first, which hence need quoting to protect special
characters. Without that, filenames with spaces or asterisks or whatever
wouldn't be shown properly.</p>
<p>The reason the <code>eval</code> is there is so that the process substitutions are
evaluated on the command line when the pager is run, and not before.
They are assigned back to elements of <code>$args</code> in quotes, so don't get
evaluated at that point. The effect will be to turn:</p>
<pre><code> less file.gz file.txt
</code></pre>
<p>into</p>
<pre><code> less =(zcat file.gz) file.txt
</code></pre>
<p>The `<code>command</code>' at the end of the function is there just in case the
function has the same name as the pager (i.e. `less' in this example);
it forces the external command to be called rather than the function.
The process substitution is ideal in this context; it provides <code>less</code>
with the name of a file to which the decompressed contents of <code>file.gz</code>
have been sent, and it deletes the file after the command exits.
Furthermore, the substitution happens in such a way that you can still
specify multiple files on the command line as you usually can with less.
The only problem is that the filename that appears in the `<code>less</code>'
prompt is meaningless.</p>
<p>In case you haven't come across it, <code>bzip2</code> is a programme very similar
to <code>gzip</code>, and it is used almost identically, but it provides better
compression.</p>
<p>There's an infelicity in output process substitutions, just as there is
with multios.</p>
<pre><code> echo hello > >(sed s/hello/goodbye)
</code></pre>
<p>The shell spawns the <code>sed</code> process to handle the output from the command
line --- and then forgets about it. It does not wait for it (at least,
not until after it exits, when it will use the <code>wait</code> system call to
tidy up). So it is dangerous to rely on the result of the process being
available in the next command. If you try it interactively, in fact, you
may well find that the next prompt is printed before the output from
<code>sed</code> shows up on the terminal. This can probably be considered a bug,
but it is quite difficult to fix.</p>
<p><span id="l120"></span></p>
<h2 id="54-parameter-substitution"><a class="header" href="#54-parameter-substitution">5.4: Parameter substitution</a></h2>
<p>You can probably see from the above that parameter substitutions are at
the heart of much of the power available to transform zsh command lines.
What's more, we haven't covered even a significant fraction of what's on
offer.</p>
<p><span id="l121"></span></p>
<h3 id="541-using-arrays"><a class="header" href="#541-using-arrays">5.4.1: Using arrays</a></h3>
<p>The array syntax in zsh is quite powerful (surprised?); just don't
expect it to be as efficient as, say, perl. Like other features of zsh,
it exists to make users' lives easier, not to make your computer run
blindingly fast.</p>
<p>I've covered, somewhat sporadically, how to set arrays, and how to
extract bits of them --- the following illustrates this:</p>
<pre><code> % array=(one two three four)
% print ${array}
one two three four
% print ${array[3]}
three
% print ${array[2,-1]}
two three four
</code></pre>
<p>Remember you need `<code>typeset</code>' or equivalent if you want the array to be
local to a function. The neat way is `<code>typeset -a</code>', which creates an
empty array, but as long as you assign to the array before trying to use
it any old <code>typeset</code> will do.</p>
<p>You can use the array index and array slice notations for assigning to
arrays, in other words on the left-hand side of an `<code>=</code>':</p>
<pre><code> % array=(what kind of fool am i)
% array[2]=species
% print $array
what species of fool am i
% array[2]=(a piece)
% print $array
what a piece of fool am i
% array[-3,-1]=(work is a man)
% print $array
what a piece of work is a man
</code></pre>
<p>So you can replace a single element of an array by a single element, or
by an array slice; likewise you can replace a slice in one go by a slice
of a different length --- only the bits you explicitly tell it to
replace are changed, the rest is left intact and maybe shifted along to
make way. This is similar to perl's `splice' command, only for once
maybe a bit more memorable. Note that you shouldn't supply any braces on
the left hand side. The appearance of the expression in an assignment is
enough to trigger the special behaviour of subscripts, even if
<code>KSH_ARRAYS</code> is in effect --- though you need to subtract one from your
subscripts in that case.</p>
<p>You can remove bits in the middle, too, but note you should use an empty
array:</p>
<pre><code> % array=(one two three four)
% print $#array
4
% array[2]=
% print $#array
4
% array[2]=()
% print $#array
3
</code></pre>
<p>The first assignment set element 2 to the empty string, it didn't remove
it. The second replaced the array element with an array of length zero,
which did remove it.</p>
<p>Just as parameter substitutions have flags for special purposes, so do
subscripts. You can force them to search through arrays, matching on the
values. You can return the value matched ((r)everse subscripting):</p>
<pre><code> % array=(se vuol ballare signor contino)
% print ${array[(r)s*]}
se
% print ${array[(R)s*]}
signor
</code></pre>
<p>The <code>(r)</code> flag takes a pattern and substitutes the first element of the
array matched, while the <code>(R)</code> flag does the same but starting from the
end of the array. If nothing matched, you get the empty string; as usual
with parameters, this will be omitted if it's the only thing in an
unquoted argument. Using our <code>args</code> function to count the arguments
passed to a command again:</p>
<pre><code> % array=(some words)
% args() { print $#; }
% args ${array[(r)s*]}
1
% args ${array[(r)X*]}
0
% args "${array[(r)X*]}"
1
</code></pre>
<p>where in the last case the empty string was quoted, and passed down as a
single, empty argument.</p>
<p>You can also return the index matched; <code>(i)</code> to start matching from the
beginning, and <code>(I)</code> to start from the end.</p>
<pre><code> % array=(se vuol venire nella mia scuola)
% print ${array[(i)v*]}
2
% print ${array[(I)v*]}
3
</code></pre>
<p>matching `vuol' the first time and `venire' the second. What happens
if they don't match may be a little unexpected, but is reasonably
logical: you get the next index along. In other words, failing to match
at the end gives you the length of the array plus one, and failing to
match at the beginning gives you zero, so:</p>
<pre><code> array=(three egregious words)
for pat in '*e*e*' '*a*a*'; do
if [[ ${array[(i)$pat]} -le ${#array} ]]; then
print "Pattern $pat matched in array: ${array[(r)$pat]}."
else
print "Pattern $pat failed to match in array"
fi
done
</code></pre>
<p>prints:</p>
<pre><code> Pattern *e*e* matched in array: three.
Pattern *a*a* failed to match in array
</code></pre>
<p>If you adapt that chunk of code, you'll see you get the indices 1 and 4
returned. Note that the characters in <code>$pat</code> were treated as a pattern
even though putting <code>$pat</code> on the command line would normally just
produce the characters themselves. Subscripts are special in that way;
trying to keep the syntax under control at this point is a little hairy.
There is a more detailed description of this in the manual in the
section `Subscript Parsing' of the <code>zshparam</code> manual page or the
`Array Parameters' info node; to quote the characters in <code>pat</code>, you
would actually have to supply the command line strings <code>'\*e\*e\*'</code> and
<code>'\*a\*a\*'</code>. Just go round mumbling `extra layer of pattern expansion'
and everyone will think you know what you're talking about (it works for
me, fitfully).</p>
<p>There is currently no way of extracting a complete set of matches from
an ordinary array with subscript flags. We'll see other ways of doing
that below, however.</p>
<p><span id="l122"></span></p>
<h3 id="542-using-associative-arrays"><a class="header" href="#542-using-associative-arrays">5.4.2: Using associative arrays</a></h3>
<p>Look back at <a href="zshguide03.html#syntax">chapter 3</a> if you've forgotten
about associative arrays. These take subscripts, like ordinary arrays
do, but here the subscripts are arbitrary strings (or keys) associated
with the value stored in the element of the array. Remember, you need to
use `<code>typeset -A</code>' to create one, or one of <code>typeset</code>'s relatives with
the same option. This means that if you created it inside a function it
will be limited to the local scope, so if you want to create a global
associative array you will need to give the <code>-g</code> flag as well. This is
particularly common with associative arrays, which are often used to
store global information such as configuration details.</p>
<p>Retrieving information from associative arrays can get you into some of
the problems already hinted at in the use of subscript flags with
arrays. However, since normal subscripting doesn't make patterns active,
there is a way round here: make the subscript into another parameter:</p>
<pre><code> % typeset -A assoc
% assoc=(key value Shlüssel Wert clavis valor)
% subscript='key'
% print ${assoc[$subscript]}
value
</code></pre>
<p>I used fairly boring keys here, but they can be any string of
characters:</p>
<pre><code> % assoc=(']' right\ square\ bracket '*' asterisk '@' at\ sign)
% subscript=']'
% print ${assoc[$subscript]}
right square bracket
</code></pre>
<p>and <em>that</em> is harder to get the other way. Nonetheless, if you define
your own keys you will often use simple words, and in that case they can
happily appear directly in the square brackets.</p>
<p>I introduced two parameter flags, <code>(k)</code> and <code>(v)</code> in <a href="zshguide03.html#syntax">chapter
3</a>:</p>
<pre><code> % print ${(k)assoc}
* ] @
</code></pre>
<p>prints out keys, while</p>
<pre><code> % print ${(kv)assoc}
* asterisk ] right square bracket @ at sign
</code></pre>
<p>and the remaining two possibilities do the same thing:</p>
<pre><code> % print ${(v)assoc}
asterisk right square bracket at sign
% print ${assoc}
asterisk right square bracket at sign
</code></pre>
<p>You now know these are part of a much larger family of tricks to apply
to substitutions. There's nothing to stop you combining flags:</p>
<pre><code> % print -r ${(qkv)assoc}
\* asterisk \] right\ square\ bracket @ at\ sign
</code></pre>
<p>which helps see the wordbreaks. Don't forget the `<code>print -l</code>' trick for
separating out different words, and hence elements of arrays and
associative arrays:</p>
<pre><code> % print -l ${(kv)assoc}
*
asterisk
]
right square bracket
@
at sign
</code></pre>
<p>which is quite a lot clearer. As always, this will fail if you engage in
un-zsh activities with <code>SH_WORD_SPLIT</code>, but judicious use of <code>@</code>,
whether as a flag or a subscript, and double quotes, will always work:</p>
<pre><code> % print -l "${(@kv)assoc}"
*
asterisk
]
right square bracket
@
at sign
</code></pre>
<p>regardless of the option setting.</p>
<p>Apart from the subscripts, the second major difference between
associative and ordinary arrays is that the former don't have any order
defined. This will be entirely familiar if you have used Perl; the
principle here is identical. However, zsh has no notion at all, even as
a convenience, of slices of associative arrays. You can assign
individual elements or whole associative arrays --- remembering that in
the second case the right hand side must consist of key/value pairs ---
but you can't assign subgroups. Any attempt to use the slice notation
with commas will be met by a stern error message.</p>
<p>What zsh does have, however, is extra subscript flags for you to match
and retrieve one or more elements. If instead of an ordinary subscript
you use a subscript preceded by the flag <code>(i)</code>, the shell will search
for a matching key (not value) with the pattern given and return that.
This is deliberately the same as searching an ordinary array to get its
key (which in that case is just a number, the index), but note this time
it doesn't match on the value, it really does match, as well as return,
the key:</p>
<pre><code> % typeset -A assoc
% assoc=(fred third\ man finnbar slip roger gully trevor long\ off)
% print ${assoc[(i)f*]}
fred
</code></pre>
<p>You can still use the parameter flags <code>(k)</code> and <code>(v)</code> to tell the shell
which part of the key and/or value to return:</p>
<pre><code> % print ${(kv)assoc[(i)f*]}
fred third man
</code></pre>
<p>Note the division of labour. The subscript flag tells the shell what to
match against, while the parameter flags tell it which bit of the
matched element(s) you actually want to see.</p>
<p>Because of the essentially random ordering of associative arrays, you
couldn't tell here whether fred or finnbar would be chosen. However, you
can use the capital form <code>(I)</code> to tell the shell to retrieve all
matches. This time, let's see the values of the elements for which the
keys were matched:</p>
<pre><code> % print -l ${(v)assoc[(I)f*]}
third man
slip
</code></pre>
<p>and here we also got the position occupied by <code>finnbar</code>. The same rules
about patterns apply as with <code>(r)</code> in ordinary arrays --- a subscript is
treated as a pattern even if it came from a parameter substitution
itself.</p>
<p>You probably aren't surprised to hear that the subscript flags <code>(r)</code> and
<code>(R)</code> try to match the values of the associative array rather than its
keys. These, too, print out the actual part matched, here the value,
unless you use the parameter flags.</p>
<pre><code> % print ${assoc[(r)*i*]}
third man
% print ${(k)assoc[(R)*i*]}
fred finnbar
</code></pre>
<p>There's one more pair of subscript flags of particular relevance to
associative arrays, <code>(k)</code> and <code>(K)</code>. These work a bit like a case
statement: the subscripts are treated as strings, and the keys of the
associative arrays as patterns, instead of the other way around. With
<code>(k)</code>, the value of the first key which matches the subscript is
substituted; with <code>(K)</code>, the values of all matching keys are substituted</p>
<pre><code> % typeset -A assoc
% assoc=('[0-9]' digit '[a-zA-Z]' letter '[^0-9a-zA-Z]' neither)
% print ${assoc[(k)0]}
digit
% print ${assoc[(k)_]}
neither
</code></pre>
<p>In case you're still confused, the `<code>0</code>' in the first subscript was
taken as a string and all the keys in <code>$assoc</code> were treated as patterns
in turn, a little like</p>
<pre><code> case 0 in
([0-9]) print digit
;;
([a-zA-Z]) print letter
;;
([^0-9a-zA-Z]) print neither
;;
esac
</code></pre>
<p>One important way in which this is <em>not</em> like the selection in a case
statement is that you can't rely on the order of the comparison, so you
can't rely on more general patterns being matched after more specific
ones. You just have to use keys which are sufficiently explicit to match
just the strings you want to match and no others. That's why we picked
the pattern `<code>[^0-9a-zA-Z]</code>' instead of just `<code>*</code>' as we would
probably have used in the case statement.</p>
<p>I said storing information about configuration was a common use of
associative arrays, but the shell has a more powerful way of doing that:
styles, which will figure prominently in the discussion of programmable
completion in the next chapter. The major advantage of styles over
associative arrays is that they can be made context-sensitive; you can
easily make the same style return the same value globally, or make it
have a default but with a different value in one particular context, or
give it a whole load of different values in different places. Each shell
application can decide what is meant by a `context'; you are not tied
to the same scheme as the completion system uses, or anything like it.
Use of hierarchical contexts in the manner of the completion system does
mean that it is easy to create sets of styles for different modules
which don't clash.</p>
<p>Here, finally, is a comparison of some of the uses of associative arrays
in perl and zsh.</p>
<pre><code> perl zsh
-----------------------------------------------------------------
%hash = qw(key value); typeset -A hash; hash=(key value)
$hash{key} ${hash[key]}
keys %hash ${(k)hash}
values %hash ${(v)hash}
%hash2 = %hash; typeset -A hash2; hash2=("${(@kv)hash}")
unset %hash; unset hash
if (exists $hash{key}) { if (( ${+hash[key]} )); then
... ...
} fi
</code></pre>
<p>One final reminder: if you are creating associative arrays inside a
function which need to last beyond the end of the function, you should
create them with `<code>typeset -gA</code>' which puts them into the surrounding
scope. The `<code>-g</code>' flag is of course useful with all types of parameter,
but the associative array is the only type that doesn't automatically
spring into existence when you assign to it in the right context; hence
the flag is particularly worthy of note here.</p>
<p><span id="l123"></span></p>
<h3 id="543-substituted-substitutions-top--and-tailing-etc"><a class="header" href="#543-substituted-substitutions-top--and-tailing-etc">5.4.3: Substituted substitutions, top- and tailing, etc.</a></h3>
<p>There are many transformations which you can do on the result of a
parameter substitution. The most powerful involve the use of patterns.
For this, the more you know about patterns, the better, so I will
reserve explanation of some of the whackiest until after I have gone
into more detail on patterns. In particular, it's useful if you know how
to tell the shell to mark subexpressions which it has matched for future
extraction. However, you can do some very useful things with just the
basic patterns common to all shells.</p>
<p><strong>Standard forms: lengths</strong></p>
<p>I'll separate out zsh-specific forms, and start off with some which
appear in all shells derived from the Bourne shell. A more compact
(read: terse) list is given in the manual, as always.</p>
<p>A few simple forms don't use patterns. First, the substitution
<code>${#</code><em>param</em><code>}</code> outputs the length of <code>$</code><em>param</em>. In zsh, you don't need
the braces here, though in most other shells with this feature you do.
Note that <code>${#}</code> on its own is the number of parameters in the command
line argument array, which is why explicit use of braces is clearer.</p>
<p><code>$#</code> works differently on scalar values and array values; in the former
case, it gives the length in characters, and in the latter case the
length in elements. Note that I said `values', not `parameters' ---
you have to work out whether the substitution is giving you a scalar or
an array:</p>
<pre><code> % print ${#path}
8
% print ${#path[1]}
13
</code></pre>
<p>The first result shows I have 8 directories in my path, the latter that
the first directory (actually `<code>/home/pws/bin</code>') has 13 characters. You
should bear this in mind with nested substitutions, as discussed below,
which can also return either an array or a scalar.</p>
<p>Earlier versions of zsh always returned a character count if the
expression was in double quotes, or anywhere the shell evalauted the
expression as a single word, but that doesn't happen any more; it
depends only on the type of the value. However, you can force the shell
to count characters by using the <code>(c)</code> flag, and to count words (even in
scalars, which it will split if necessary) by using <code>(w)</code>:</p>
<pre><code> % print ${#PATH}
84
% print ${(c)#path}
84
% foo="three scalar words"
% print ${(w)#foo}
3
</code></pre>
<p>Comparing the first two, you will see that character count with arrays
includes the space used for separating (equal to the number of colons
separating the elements in <code>$PATH</code>). There's a relative of <code>(w)</code> called
<code>(W)</code>, which treats multiple word separators as having zero-length words
in between:</p>
<pre><code> % foo="three well-spaced word"
% print ${(w)#foo}
3
% print ${(W)#foo}
5
</code></pre>
<p>giving two extra words over <code>(w)</code>, which treats the groups of spaces in
the same way as one. Being parameter flags, these modifications of the
syntax are specific to zsh.</p>
<p>Note that if you use lengths in an arithmetic context (inside <code>((...))</code>
or <code>$((...))</code>), you must include the leading `<code>$</code>', which you don't
need for substituting the parameters themselves. That's because
`<code>#foo</code>' means something different here --- the number in the ASCII
character set (or whatever extension of it you are using if it is an
extended character set) of the first character in <code>$foo</code>.</p>
<p><strong>Standard forms: conditional substitutions</strong></p>
<p>The next group of substitutions is a whole series where the parameter is
followed by an option colon and then `<code>-</code>', `<code>=</code>', `<code>+</code>' or `<code>?</code>'.
The colon has the same effect in each case: without a colon, the shell
tests whether the parameter is set before performing the operation,
while with the colon it tests whether the parameter has non-zero length.</p>
<p>The simplest is `<code>${</code><em>param</em><code>:-</code><em>value</em><code>}</code>'. If <code>$param</code> has non-zero
length (without the colon, if it is set at all), use its value, else use
the <em>value</em> supplied. Suppose <code>$foo</code> wasn't set at the start of the
following (however unlikely that may seem):</p>
<pre><code> % print ${foo-bar}
bar
% foo=''
% print ${foo-bar}
% print ${foo:-bar}
bar
% foo='please no anything but bar'
% print ${foo:-bar}
please no anything but bar
</code></pre>
<p>It's more usual to use the form with the colon. One reason for that is
that in functions you will often create the parameter with a <code>typeset</code>
before using it, in which case it always exists, initially with zero
length, so that the other form would never use the default value. I'll
use the colon for describing the other three types.</p>
<p>`<code>${</code><em>param</em><code>:=</code><em>value</em><code>}</code>' is similar to the previous type. but in
this case the shell will not only substitute <em>value</em> into the line, it
will assign it to <em>param</em> if (and only if) it does so. This leads to the
following common idiom in scripts and functions:</p>
<pre><code> : ${MYPARAM:=default} ${OTHERPARAM:=otherdefault}
</code></pre>
<p>If the user has already set <code>$MYPARAM</code>, nothing happens, otherwise it
will be set to `<code>default</code>', and similarly for <code>${OTHERPARAM}</code>. The
`<code>:</code>' command does nothing but return true after the command line has
been processed.</p>
<p>`<code>${</code><em>param</em><code>:+</code><em>value</em><code>}</code>' is the opposite of `<code>:-</code>', logically
enough: the <em>value</em> is substituted if the parameter <em>doesn't</em> have zero
length. In this case, <em>value</em> will often be another parameter
substitution:</p>
<pre><code> print ${value:+"the value of value is $value"}
</code></pre>
<p>prints the string only if <code>$#value</code> is greater than zero. Note that what
can appear after the `<code>+</code>' is pretty much any single word the shell can
parse; all the usual single-word substitutions (so globbing is excluded)
will be applied to it, and quotes will work just the same as usual. This
applies to the values after `<code>:-</code>' and `<code>:=</code>', too. One other commonly
seen trick might be worth mentioning:</p>
<pre><code> print ${1+"$@"}
</code></pre>
<p>substitutes all the positional parameters as they were passed if the
first one was set (here you don't want the colon). This was necessary in
some old shells because <code>"$@"</code> on its own gave you a single empty
argument instead of no arguments when no arguments were passed. This
workaround isn't necessary in zsh, nor in most modern Bourne-derived
shells. There's a bug in zsh's handling, however; see the section on
function parameters in chapter 3.</p>
<p>The final type isn't that often used (meaning I never have):
<code>${</code><em>param</em><code>?</code><em>message</em><code>}</code> tests if <em>param</em> is set (no colon), and if it
isn't, prints the message and exits the shell. An interactive shell
won't exit, but it will return you immediately to the prompt, skipping
anything else stored up for execution. It's a rudimentary safety
feature, a little bit like `assert' in C programmes; most shell
programmers seem to cover the case of missing parameter settings by more
verbose tests. It's quite neat in short shell functions for interactive
use:</p>
<pre><code> mless() { mtype ${@:?missing filename} | $PAGER }
</code></pre>
<p><strong>Standard forms: pattern removal</strong></p>
<p>Most of the more sophisticated Bourne-like shells define two pairs of
pattern operators, which I shall call `top and tail' operators. One
pair (using `<code>#</code>' and `<code>##</code>') removes a given pattern from the head of
the string, returning the rest, while the other pair (using `<code>%</code>' and
`<code>%%</code>') removes a pattern from the tail of the string. In each case,
the form with one symbol removes the shortest matching pattern, while
the one with two symbols removes the longest matching pattern. Two
typical uses are:</p>
<pre><code> % print $HOME
/home/pws
% print ${HOME##*/}
pws
% print ${HOME%/*}
/home
</code></pre>
<p>which here have the same effect of <code>${HOME:t}</code> and and <code>${HOME:h}</code>, and
in zsh you would be more likely to use the latter. However, as you can
see the pattern forms are much more general. Note the difference from:</p>
<pre><code> % print ${HOME#*/}
home/pws
% print ${HOME%%/*}
</code></pre>
<p>where the shortest match of `<code>*/</code>' at the head was just the first
slash, since `<code>*</code>' can match an empty string, while the longest match
of `<code>/*</code>' at the tail was the entire string, right back to the first
slash. Although these are standard forms, remember that the full power
of zsh patterns is available.</p>
<p>How do you remember which operator does what? The fact that the longer
form does the longer match is probably easy. Remembering that `<code>#</code>'
removes at the head and `<code>%</code>' at the tail is harder. Try to think of
`hash' and `head' (if you call it a `pound sign', when it's nothing
of the sort since a pound since looks like `£', you will get no
sympathy from me), and `percent' and `posterior'. It never worked for
me, but maybe I just don't have the mental discipline. Oliver Kiddle
points out that `<code>#</code>' is further to the left (head) on a standard US
keyboard. On my UK keyboard, `<code>#</code>' is right next to the return key,
unfortunately, although here the confusion with `pound sign' will jog
your memory.</p>
<p>The most important thing to remember is: this notation is not our fault.
Sorry, anyway. By the way, notice there's no funny business with colons
in the case of the pattern operators. (Well --- except for the zsh
variant noted below.)</p>
<p><strong>Zsh-specific parameter substitutions</strong></p>
<p>Now for some enhancements that zsh has for using the forms of parameter
substitution I've just given as well as some similar but different ones.</p>
<p>One simple enhancement is that in addition to
`<code>${</code><em>param</em><code>=</code><em>value</em><code>}</code>' and `<code>${</code><em>param</em><code>:=</code><em>value</em><code>}</code>', zsh has
`<code>${</code><em>param</em><code>::=</code><em>value</em><code>}</code>' which performs an unconditional assignment
as well as sticking the value on the command line. It's not really any
different from using a normal assignment, then a normal parameter
substitution, except that zsh users like densely packed code.</p>
<p>All the assignment types are affected by the parameter flags `<code>A</code>' and
`<code>AA</code>' which tell the shell to perform array and associative array
assignment (in the second case, you need pairs of key/value elements as
usual). You need to be a little bit careful with array elements and word
splitting, however:</p>
<pre><code> % print -l ${(A)foo::=one two three four}
one two three four
% print ${#foo}
1
</code></pre>
<p>That made <code>$foo</code> an array all right, but treated the argument as a
scalar value and assigned it to the first element. There's a way round
this:</p>
<pre><code> % print -l ${(A)=foo::=one two three four}
one
two
three
four
% print ${#foo}
4
</code></pre>
<p>Here, the `<code>=</code>' <em>before</em> the parameter name has a completely different
effect from the others: it turns on word-splitting, just as if the
option <code>SH_WORD_SPLIT</code> is in effect. You may remember I went into this
in appalling detail in the section `Function parameters' in <a href="zshguide03.html#syntax">chapter
3</a>.</p>
<p>You should be careful, however, as more sophisticated attempts at
putting arrays inside parameter values can easily lead you astray. It's
usually much easier to use the `<em>array</em><code>=</code>(<em>...</em>)' or `<code>set -A</code> <em>...</em>'
notations.</p>
<p>One extremely useful zsh enhancement is the notation `<code>${+foo}</code>' which
returns 1 if <code>$foo</code> is set and 0 if it isn't. You can use this in
arithmetic expressions. This is a much more flexible way of dealing with
possibly unset parameters than the more standard `<code>${foo?goodbye}</code>'
notation, and consequently is better used by zsh programmers. The
notation `plus foo' for `foo is set' should be fairly memorable, too.
A more standard way of doing this (noted by David Korn) is
`<code>0${foo+1}</code>', giving 0 if <code>$foo</code> is not set and 01 if it is.</p>
<p><strong>Parameter flags and pattern substitutions</strong></p>
<p>Zsh increases the usefulness of the `top and tail' operators with some
of its parameter flags. Usually these show you what's left after the
removal of some matched portion. However, with the flag <code>(M)</code> the shell
will instead show you the matched portion itself. The flag <code>(R)</code> is the
opposite and shows the rest: that's not all that useful in the normal
case, since you get that by default. It only starts being useful when
you combine it with other flags.</p>
<p>Next, zsh allows you to match on substrings, not just on the head or
tail. You can do this by giving the flag <code>(S)</code> with either of the `<code>#</code>'
or `<code>%</code>' pattern-matching forms. The difference here is whether the
shell starts searching for a matching substring at the start or end of
the full string. Let's take</p>
<pre><code> foo='where I was huge lizards walked here and there'
</code></pre>
<p>and see what we get matching on `<code>h*e</code>':</p>
<pre><code> % print -l ${(S)foo#h*e} ${(S)foo##h*e} ${(S)foo%h*e} ${(S)foo%%h*e}
wre I was huge lizards walked here and there
w
where I was huge lizards walked here and tre
where I was huge lizards walked here and t
</code></pre>
<p>There are some odd discrepancies at first sight, but here's what
happens. In the first case, `<code>#</code>' the shell looks forward until it
finds a match for `<code>h*e</code>', and takes the shortest, which is the `<code>he</code>'
in the first word. With `<code>##</code>', the match succeeds at the same point,
but the longest match extends to the `<code>e</code>' right at the end of the
string. With the other two forms, the shell starts scanning backwards
from the end, and stops as soon as it reaches a starting point which has
a match. For both `<code>%</code>' and `<code>%%</code>' this is the last `<code>h</code>', but the
former matches `<code>he</code>' and the latter matches `<code>here</code>'.</p>
<p>You can extend this by using the <code>(I)</code> flag to specify a numeric index.
The index needs to be delimited, conventionally, although not
necessarily, by colons. The shell will then scan forward or backward,
depending on the form used, until it has found the <code>(I)</code>'th match. Note
that it only ever counts a single match from each position, either the
longest or the shortest, so the <code>(I)</code>'th match starts from the <code>(I)</code>'th
position which has any match. Here's what happens when we remove all the
matches for `<code>#</code>' using the example above.</p>
<pre><code> % for (( i = 1; i <= 5; i++ )); do
for> print ${(SI:$i:)foo#h*e}
for> done
wre I was huge lizards walked here and there
where I was lizards walked here and there
where I was huge lizards walked re and there
where I was huge lizards walked here and tre
where I was huge lizards walked here and there
</code></pre>
<p>Each time we match and remove one of the possible `<code>h*e</code>' sets where
there is no `<code>e</code>' in the middle, moving from left to right. The last
time there was nothing left to match and the complete string was
returned. Note that the index we used was itself a parameter.</p>
<p>It's obvious what happens with `<code>##</code>': it will find matches at all the
same points, but they will all extend to the `<code>e</code>' at the end of the
string. It's probably less obvious what happens with `<code>%%</code>' and `<code>%</code>',
but if you try it you will find they produce just the same set of
matches as `<code>##</code>' and `<code>#</code>', respectively, but with the indices in the
reverse order (4 for 1, 3 for 2, etc.).</p>
<p>You can use the `<code>M</code>' flag to leave the matched portion rather than the
rest of the string, if you like. There are three other flags which let
you get the indices associated with the match instead of the string:
<code>(B)</code> for the beginning, using the usual zsh convention where the first
character is 1, <code>(E)</code> for the character <em>after</em> the end, and <code>(N)</code> for
the length, simply <code>B-E</code>. You can even have more than one of these; the
value substituted is a string with the given values with spaces between,
always in the order beginning, end, length.</p>
<p>There is a sort of opposite to the `<code>(S)</code>' flag, which instead of
matching substrings will only match the whole string; to do this, put a
colon before the `<code>#</code>'. Hence:</p>
<pre><code> % print ${foo:#w*g}
where I was huge lizards walked here and there
% print ${foo:#w*e}
%
</code></pre>
<p>The first one didn't match, because the `<code>g</code>' is not at the end; the
second one did, because there is an `<code>e</code>' at the end.</p>
<p><strong>Pattern replacement</strong></p>
<p>The most powerful of the parameter pattern-matching forms has been
borrowed from bash and ksh93; it doesn't occur in traditional Bourne
shells. Here, you use a pair of `<code>/</code>'s to indicate a pattern to be
replaced, and its replacement. Lets use the lizards again:</p>
<pre><code> % print ${foo/h*e/urgh}
wurgh
</code></pre>
<p>A bit incomprehensible: that's because like most pattern matchers it
takes the longest match unless told otherwise. In this case the <code>(S)</code>
flag has been pressed into service to mean not a substring (that's
automatic) but the shortest match:</p>
<pre><code> % print ${(S)foo/h*e/urgh}
wurghre I was huge lizards walked here and there
</code></pre>
<p>That only replace the first match. This is where `<code>//</code>' comes in; it
replaces every match:</p>
<pre><code> % print ${(S)foo//h*e/urgh}
wurghre I was urgh lizards walked urghre and turghre
</code></pre>
<p>(No doubt you're starting to feel like a typical anachronistic Hollywood
cave-dweller already.) Note the syntax: it's a little bit like
substitution in <code>sed</code> or perl, but there's no slash at the end, and with
`<code>//</code>' only the first slash is doubled. It's a bit confusing that with
the other pattern expressions the single and double forms mean the
shortest and longest match, while here it's the flag <code>(S)</code> that makes
the difference.</p>
<p>The index flag <code>(I)</code> is useful here, too. In the case of `<code>/</code>', it
tells the shell which single match to substitute, and in the case of
`<code>//</code>' it tells the shell at which match to start: all matches starting
from that are replaced.</p>
<p>Overlapping matches are never replaced by `<code>//</code>'; once it has put the
new text in for a match, that section is not considered further and the
text just to its right is examined for matches. This is probably
familiar from other substitution schemes.</p>
<p>You may well be thinking `wouldn't it be good to be able to use the
matched text, or some part of it, in the replacment text?' This is what
you can do in sed with `<code>\1</code>' or `<code>\&</code>' and in perl with `<code>$1</code>' and
`<code>$&</code>'. It turns out this <em>is</em> possible with zsh, due to part of the
more sophisticated pattern matching features. I'll talk about this when
we come on to patterns, since it's not really part of parameter
substitution, although it's designed to work well with that.</p>
<p><span id="l124"></span></p>
<h3 id="544-flags-for-options-splitting-and-joining"><a class="header" href="#544-flags-for-options-splitting-and-joining">5.4.4: Flags for options: splitting and joining</a></h3>
<p>There are three types of flag that don't look like flags, for historical
reasons; you've already seen them in <a href="zshguide03.html#syntax">chapter
3</a>. The first is the one that turns on the
<code>SH_WORD_SPLIT</code> option, <code>${=foo}</code>. Note that you can mix this with flags
that <em>do</em> look like flags, in parentheses, in which case the `<code>=</code>' must
come after the closing parenthesis. You can force the option to be
turned off for a single substitution by doubling the symbol:
`<code>${==foo}</code>'. However, you wouldn't do that unless the option was
already set, in which case you are probably trying to be compatible with
some other shell, and wouldn't want to use that form.</p>
<p>More control over splitting and joining is possible with three of the
more standard type of flags, <code>(s)</code>, <code>(j)</code> and <code>(z)</code>. These do splitting
on a given string, joining with a given string, and splitting just the
way the shell does it, respectively. In the first two cases, you need to
specify the string in the same way as you specified the index for the
<code>(I)</code> flag. So, for example, here's how to turn <code>$PATH</code> into an ordinary
array without using <code>$path</code>:</p>
<pre><code> % print -l ${(s.:.)PATH}
/home/pws/bin
/usr/local/bin
/usr/sbin
/sbin
/bin
/usr/bin
/usr/X11R6/bin
/usr/games
</code></pre>
<p>Any character can follow the <code>(s)</code> or <code>(j)</code>; the string argument lasts
until the matching character, here `<code>.</code>'. If the character is one of
the bracket-like characters including `<code><</code>', the `matching' character
is the corresponding right bracket, e.g. `<code>${(s<:>)PATH}</code>' and
`<code>${(s(:))PATH}</code>' are both valid. This applies to all flags that need
arguments, including <code>(I)</code>.</p>
<p>Although the split or join string isn't a pattern, it doesn't have to be
a single character:</p>
<pre><code> % foo=(array of words)
% print ${(j.**.)foo}
array**of**words
</code></pre>
<p>The <code>(z)</code> flag doesn't take an argument. As it handles splitting on the
full shell definition of a word, it goes naturally with quoted
expressions, and I discussed above its use with the <code>(Q)</code> flag for
extracting words from a line with the quotes removed.</p>
<p>It's possible for the same parameter expression to have both splitting
and joining applied to it. This always occurs in the same order,
regardless of how you specify the flags: joining first, then splitting.
This is described in the (rather hairy) complete set of rules in the
manual entry for parameter substitution. There are one or two occasions
where this can be a bit surprising. One is when you have <code>SH_WORD_SPLIT</code>
set and try to join a string:</p>
<pre><code> % setopt shwordsplit
% foo=('another array' of 'words with spaces')
% print -l ${(j.:.)foo}
another
array:of:words
with
spaces
</code></pre>
<p>You might not have noticed if you didn't use the `<code>-l</code> option to print,
but the spaces still caused word-spliting even though you asked for the
array to be joined with colons. To avoid this, either don't use
<code>SH_WORD_SPLIT</code> (my personal preference), or use quotes:</p>
<pre><code> % print -l "${(j.:.)foo}"
another array:of:words with spaces
</code></pre>
<p>The elements of an array would normally be joined by spaces in this
case, but the character specified by the <code>(j)</code> flag takes precedence. In
just the same way, if <code>SH_WORD_SPLIT</code> is in effect, any splitting string
given by <code>(s)</code> is used instead of the normal set of characters, which
are any characters that occur in the string <code>$IFS</code>, by default space,
tab, newline and NUL.</p>
<p>Specifying a split for a particular parameter substitution not only sets
the string to split on, but also ensures the split will take place even
if the expression is quoted:</p>
<pre><code> % array=('element one' 'element two' 'element three')
% print -l "${=array}"
element
one
element
two
element
three
</code></pre>
<p>To be clear about what's happening here: the quotes force the elements
to be joined with spaces, giving a single string, which is then split on
the original spaces as well as the one used to join the elements of the
array.</p>
<p>I will talk shortly about nested parameter substitution; you should also
note that splitting and joining will if necessary take place at all
levels of a nested substitution, not just the outermost one:</p>
<pre><code> % foo="three blind words"
% print ${#${(z)foo}}
3
</code></pre>
<p>This prints the length of the innermost expression; because of the
zplit, that has produced a three-element array.</p>
<p><span id="l125"></span></p>
<h3 id="545-flags-for-options-glob_subst-and-rc_expand_param"><a class="header" href="#545-flags-for-options-glob_subst-and-rc_expand_param">5.4.5: Flags for options: <code>GLOB_SUBST</code> and <code>RC_EXPAND_PARAM</code></a></h3>
<p>The other two flags that don't use parentheses affect options for single
substitutions, too. The second is the `<code>~</code>' flag that turns on
<code>GLOB_SUBST</code>, making the result of a parameter substitution eligible for
pattern matching. As the notation is supposed to indicate, it also makes
filename expansion possible, so</p>
<pre><code> % foo='~'
% print ${~foo}
/home/pws
</code></pre>
<p>It's that first `<code>~</code>' which is giving the home directory; the one in
the parameter expansion simply allows that to happen. If you have
<code>GLOB_SUBST</code> set, you can use `<code>${~~foo}</code>' to turn it off for one
substitution.</p>
<p>There's one other of these option flags: `<code>^</code>' forces on
<code>RC_EXPAND_PARAM</code> for the current substitution, and `<code>^^</code>' forces it
off. In <a href="zshguide03.html#syntax">chapter 3</a>, I showed how parameters
expanded with this option on fitted in with brace expansions.</p>
<p><span id="l126"></span></p>
<h3 id="546-yet-more-parameter-flags"><a class="header" href="#546-yet-more-parameter-flags">5.4.6: Yet more parameter flags</a></h3>
<p>Here are a few other parameter flags; I'm repeating some of these. A
very useful one is `<code>t</code>' to tell you the type of a parameter. This came
up in <a href="zshguide03.html#syntax">chapter 3</a> as well. It's most common use
is to test the basic type of the parameter before trying to use it:</p>
<pre><code> if [[ ${(t)myparam} != *assoc* ]]; then
# $myparam is not an associative array. Do something about it.
fi
</code></pre>
<p>Another very useful type is for left or right padding of a string, to a
specified length, and optionally with a specified fill string to use
instead of space; you can even specify a one-off string to go right next
to the string in question.</p>
<pre><code> foo='abcdefghij'
for (( i = 1; i <= 10; i++ )); do
goo=${foo[1,$i]}
print ${(l:10::X::Y:)goo} ${(r:10::X::Y:)goo}
done
</code></pre>
<p>prints out the rather pretty:</p>
<pre><code> XXXXXXXXYa aYXXXXXXXX
XXXXXXXYab abYXXXXXXX
XXXXXXYabc abcYXXXXXX
XXXXXYabcd abcdYXXXXX
XXXXYabcde abcdeYXXXX
XXXYabcdef abcdefYXXX
XXYabcdefg abcdefgYXX
XYabcdefgh abcdefghYX
Yabcdefghi abcdefghiY
abcdefghij abcdefghij
</code></pre>
<p>Note that those colons (which can be other characters, as I explained
for the <code>(s)</code> and <code>(j)</code> flags) always occur in pairs before and after
the argument, so that with three arguments, the colons in between are
doubled. You can miss out the `<code>:Y:</code>' part and the `<code>:X:</code>' part and
see what happens. The fill strings don't need to be single characters;
if they don't fit an exact number of times into the filler space, the
last repetition will be truncated on the end furthest from the parameter
argument being inserted.</p>
<p>Two parameters tell the shell that you want something special done with
the value of the parameter substitution. The <code>(P)</code> flag forces the value
to be treated as a parameter name, so that you get the effect of a
double substitution:</p>
<pre><code> % final=string
% intermediate=final
% print ${(P)intermediate}
string
</code></pre>
<p>This is a bit as if <code>$intermediate</code> were what in ksh is called a
`nameref', a parameter that is marked as a reference to another
parameter. Zsh may eventually have those, too; there are places where
they are a good deal more convenient than the `<code>(P)</code>' flag.</p>
<p>A more powerful flag is <code>(e)</code>, which forces the value to be rescanned
for all forms of single-word substitution. For example,</p>
<pre><code> % foo='$(print $ZSH_VERSION)'
% print ${(e)foo}
4.0.2
</code></pre>
<p>made the value of <code>$foo</code> be re-examined, at which point the command
substitution was found and executed.</p>
<p>The remaining flags are a few simple special formatting tricks: order
array elements in normal lexical (character) order with <code>(o)</code>, order in
reverse order with <code>(O)</code>, do the same case-independently with <code>(oi)</code> or
<code>(Oi)</code> respectively, expand prompt `<code>%</code>'-escapes with <code>(%)</code> (easy to
remember), expand backslash escapes as <code>print</code> does with <code>p</code>, force all
characters to uppercase with <code>(U)</code> or lowercase with <code>(L)</code>, capitalise
the first character of the string or each array element with <code>(C)</code>, show
up special characters as escape sequences with <code>(V)</code>. That should be
enough to be getting on with.</p>
<p><span id="l127"></span></p>
<h3 id="547-a-couple-of-parameter-substitution-tricks"><a class="header" href="#547-a-couple-of-parameter-substitution-tricks">5.4.7: A couple of parameter substitution tricks</a></h3>
<p>I can't resist describing a couple of extras.</p>
<p>Zsh can do so much on parameter expressions that sometimes it's useful
even without a parameter! For example, here's how to get the length of
a fixed string without needing to put it into a parameter:</p>
<pre><code> % print ${#:-abcdefghijklm}
13
</code></pre>
<p>If the parameter whose name you haven't given has a zero length (it
does, because there isn't one), use the string after the `<code>:-</code>'
instead, and take it's length. Note you need the colon, else you are
asking the shell to test whether a parameter is set, and it becomes
rather upset when it realises there isn't one to test. Other shells are
unlikely to tolerate any such syntactic outrages at all; the <code>#</code> in that
case is likely to be treated as <code>$#</code>, the number of shell arguments. But
zsh knows that's not going to have zero length, and assumes you know
what you're doing with the extra part; this is useful, but technically a
violation of the rules.</p>
<p>Sometimes you don't need anything more than the flags. The most useful
case is making the `fill' flags generate repeated words, with the
effect of perl's `<code>x</code>' operator (for those not familiar with perl, the
expression `<code>"string" x 3</code>' produces the string `stringstringstring'.
Here, you need to remember that the fill width you specify is the total
width, not the number of repetitions, so you need to multiply it by the
length of the string:</p>
<pre><code> % print ${(l.18..string.)}
stringstringstring
</code></pre>
<p><span id="l128"></span></p>
<h3 id="548-nested-parameter-substitutions"><a class="header" href="#548-nested-parameter-substitutions">5.4.8: Nested parameter substitutions</a></h3>
<p>Zsh has a system for multiple nested parameter substitutions. Whereas in
most shells or other scripting languages you would do something like:</p>
<pre><code> % p=/directory/file.ext
% p2=${p##*/} # remove longest match of */ from head
% print $p2
file.ext
% print ${p%.*} # remove shortest match of .* from tail
file
</code></pre>
<p>in zsh you can do this in one substitution:</p>
<pre><code> % p=/directory/file.ext
% print ${${p##*/}%.*}
file
</code></pre>
<p>saving the temporary parameter in the middle. (Again, you are more
likely to use <code>${p:t:r}</code> in this particular case.) Where this becomes a
major advantage is with arrays: if <code>$p</code> is an array, all the
substitutions are applied to every element of the array:</p>
<pre><code> % p=(/dir1/file1.ext1 /dir2/file2.ext2)
% print ${${p##*/}%.*}
file1 file2
</code></pre>
<p>This can result in some considerable reductions in the code for
processing arrays. It's a way of getting round the fact that an ordinary
command line interface like zsh, designed originally for direct
interaction with the user, doesn't have all the sophistication of a
non-interactive language like perl, whose `<code>map</code>' function would
probably be the neatest way of doing the same thing:</p>
<pre><code> # Perl code.
@p = qw(/dir1/file1.ext1 /dir2/file2.ext2);
@q = map { m%^(?:.*/)(.*?)(?:\.[^.]*|)$%; } @p;
print "@q\n";'
</code></pre>
<p>or numerous possible variants. In a shell, there's no way of putting
functions like that into the command line without complicating the basic
`command, arguments' syntax; so we resort to trickery with
substitutions. Note, however, that this degree of brevity makes for a
certain lack of readability even in Perl. Furthermore, zsh is so
optimised for common cases that</p>
<pre><code> print ${p:t:r}
</code></pre>
<p>will work for both arrays and scalars: the <code>:t</code> takes only the tail of
the filename, stripping the directories, and the <code>:r</code> removes the
suffix. These two operators could have slightly unexpected effects in
versions of zsh before 4.0.1, removing `suffixes' which contained
directory paths, for example (though this is what the pattern forms
taken separately do, too).</p>
<p>Note one feature of the nested substitution: you might have expected the
`<code>${...}</code>' inside the other one to do a full parameter substitution, so
that the outer one would act on the value of that --- that's what you'd
get if the substitution was on its own, after all. However, that's not
what happens: the `<code>${...}</code>' inside is simply a syntactic trick to say
`here come more operations on the parameter'. This means that</p>
<pre><code> bar='this doesn'\''t get substituted'
foo='bar'
print ${${foo}}
</code></pre>
<p>simply prints `<code>bar</code>', not the value of <code>$bar</code>. This is the same case
we had before but without any of the extra `<code>##</code>' and `<code>%</code>' bits. The
reason is historical: when the extremely useful nested substitution
feature was added, it was much simpler to have the leading `<code>$</code>'
indicate to the shell that it should call the substitution function
again than find another syntax. You can make the value be re-interpreted
as another parameter substitution, using the <code>(P)</code> substitution flag
described above. Just remember that <code>${${foo}}</code> and <code>${(P)foo}</code> are
different.</p>
<p><span id="l129"></span></p>
<h2 id="55-that-substitution-again"><a class="header" href="#55-that-substitution-again">5.5: That substitution again</a></h2>
<p>Finally, here is a brief explanation of how to read the expression at
the top of the chapter. This is for advanced students only (nutcases, if
you ask me). You can find all the bits in the manual, if you try hard
enough, even the ones I didn't get around to explaining above. As an
example, let's suppose the array contains</p>
<pre><code> array=(long longer longest short brief)
</code></pre>
<p>and see what</p>
<pre><code> print ${array[(r)${(l.${#${(O@)array//?/X}[1]}..?.)}]}
</code></pre>
<p>gives.</p>
<ol>
<li>
<p>Always start from the inside. The innermost expression here is</p>
<pre><code> ${(O@)array//?/X}
</code></pre>
<p>Not much clearer? Start from the inside again: there's the parameter
we're operating on, whose name is <code>array</code>. Before that there are two
flags in parenthesis: (<code>O</code>) says sort the result in descending
alphabetic order, (<code>@</code>) treat the result as an array, which is
necessary because this inner substitution occurs where a scalar
value (actually, an arithmetic expression) would usually occur, and
we need to take an array element. After the array name, `<code>//?/X</code>'
is a global substitution: take the pattern `<code>?</code>' (any character)
wherever it occurs, and replace it with the string `<code>X</code>'. The
result of this is an array like <code>$array</code>, but with all the elements
turned into strings consisting of `<code>X</code>'s in place of the original
characters, and with the longest first, because that's how reverse
alphabetic order works for strings with the same character. So</p>
<pre><code> long longer longest short brief
</code></pre>
<p>would have become</p>
<pre><code> XXXXXXX XXXXXX XXXXX XXXXX XXXX
</code></pre>
</li>
<li>
<p>Next, we have `<code>${#</code><em>result</em><code>[1]}</code>' wrapped around that. That means
that we take the first element of the array we arrived at above (the
`<code>[1]</code>': that's why we had to make sure it was treated as an
array), and then take the length of that (the `<code>#</code>'). We will end
up in this case with 7, the length of the first (and longest
element). We're finally getting somewhere.</p>
</li>
<li>
<p>The next step is the `<code>${</code>(<code>l.</code><em>result</em><code>..?.</code>)<code>}</code>'. Our previous
<em>result</em> appears as an argument to the `<code>(l)</code>' flag of the
substitution. That's a rather special case of nested substitution:
at this point, the shell expects an arithmetical expression, giving
the minimum length of a string to be filled on the left. The
previous substitution was evaluated because arithmetic expressions
undergo parameter substitution. So it is the result of that, 7,
which appears here, giving the more manageable</p>
<pre><code> ${(l.7..?.)}
</code></pre>
<p>The expression for the `<code>(l)</code>' flag in full says `fill the result
of this parameter substitution to a minimum width of 7 using the
fill character `<code>?</code>'. What is the substitution we are filling? It's
empty: zsh is smart enough to assume you know what you're doing when
you don't give a parameter name, and just puts in an empty string
instead. So the empty string is filled out to length 7 with question
marks, giving `<code>???????</code>'.</p>
</li>
<li>
<p>Now we have `<code>${array[(r)???????]}</code>'. It may not be obvious
(congratulations if the rest is), but the question marks are active
as a pattern. Subscripts are treated specially in this respect. The
subscript flag `<code>(r)</code>' means `reverse match', not reverse as in
backwards, but as in the opposite way round: search the array itself
for a matching value, rather than taking this as an index. The only
thing that will match this is a string of length 7. Bingo! that
must be the element `longest' in this case. If there were other
elements of the same length, you would only get the first of that
length; I haven't thought of a way of getting all the elements of
that length substituted by a single expression without turning
<code>$array</code> into an associative array, so if you have, you should feel
smug.</p>
</li>
</ol>
<p>After I wrote this, Sven Wischnowsky (who is responsible for a large
fraction of the similar hieroglyphics in the completion functions)
pointed out that a similar way of achieving this is:</p>
<pre><code> print ${(M)array:#${~${(O@)array//?/?}[1]}}
</code></pre>
<p>which does indeed show all the elements of the maximum length. A brief
summary of how this works is that the innermost expression produces an
array of `<code>?</code>' corresponding to the elements, longest first in the way
we did above, turning the `<code>?</code>' into pattern match characters. The next
expansion picks the longest. Finally, the outermost expansion goes
through <code>$array</code> to find elements which match the complete string of
`<code>?</code>' and selects out those that do match.</p>
<p>If you are wondering about how to do that in perl in a single
expression, probably sorting on length is the easiest:</p>
<pre><code> # Perl code
@array = qw(long longer longest short brief);
@array = sort { length $b <=> length $a } @array;
</code></pre>
<p>and taking out the first element or first few elements of <code>@array</code>.
However, in a highly-optimized scripting language you would almost
certainly do it some other way: for example, avoid sorting and just
remember the longest element:</p>
<pre><code> # Perl code
$elt = '';
$l = 0;
foreach (@array) {
$newl = length $_;
$elt = $_, $l = $newl if $l > $newl;
}
print $elt, "\n";
</code></pre>
<p>You can do just the same thing in zsh easily enough in this case;</p>
<pre><code> local val elt
integer l newl
for val in $array; do
newl=${#val}
if (( newl > l )); then
elt=$val
(( l = newl ))
fi
done
print $elt
</code></pre>
<p>so this probably isn't a particularly good use for nested substitution,
even though it illustrates its power.</p>
<p>If you enjoyed that expression, there are many more like it in the
completion function suite for you to goggle at.</p>
<p><span id="l130"></span></p>
<h2 id="56-arithmetic-expansion"><a class="header" href="#56-arithmetic-expansion">5.6: Arithmetic Expansion</a></h2>
<p>Performing mathematics within the shell was first described in <a href="zshguide03.html#syntax">chapter
3</a> where I showed how to create numeric
parameters with variants of `<code>typeset</code>', and said a little about
arithmetic substitution.</p>
<p>In addition to the math library, loadable with `<code>zmodload zsh/mathfunc</code>', zsh has essentially all the operators you expect from C
and other languages derived from it. In other words, things like</p>
<pre><code> (( foo = bar ? 3 : 1, ++brr ))
</code></pre>
<p>are accepted. The comma operator works just as in C; all the arguments
are evaluated, in this case `<code>foo = bar ? 3 : 1</code>' assigns 3 or 1 to
<code>$foo</code> depending whether or not <code>bar</code> is non-zero, and then <code>$brr</code> is
incremented by 1. The return status is determined by the final
expression, so if <code>$brr</code> is zero after increment the return status is
one, else it is zero (integers may be negative).</p>
<p>One extra operator has been borrowed from FORTRAN, or maybe Perl, the
exponentiation operator, `<code>**</code>'. This can take either integers or
floating point numbers, though a negative exponent will cause a floating
point number to be returned, so `<code>$(( 2 ** -1 ))</code>' gives you 0.5, not
rounded down to zero. This is why the standard library function <code>pow</code> is
missing from <code>zsh/mathfunc</code> --- it's already there in that other form.
Pure integer exponentiation, however, is done by repeated multiplication
--- up to arbitrary sizes, so instead of `<code>2 ** 100</code>', you should use
`<code>1 << 100</code>', and for powers of any other integer where you don't need
an exact result, you should use floating point numbers. For this
purpose, the <code>zsh/mathfunc</code> library makes `casts' available;
`<code>float</code>(<em>num</em>)' forces the expression <em>num</em> to interpreted as a
floating point number, whatever it would otherwise have given, although
the trick of adding `<code>0.0</code>' to a number works as well. Note that,
although this works like a cast in C, the syntax is that of an ordinary
function call. Likewise, `<code>int</code>(<em>num</em>)' causes the number to be
interpreted as an integer --- rounding towards zero; you can use <code>floor</code>
and <code>ceil</code> to round down or up, and <code>rint</code> to round to the nearest
integer, although these three actually produce floating point numbers.
They are standard C library functions.</p>
<p>For completeness, the assignment form of exponentiation `<code>**=</code>' also
works. I can't remember ever using it.</p>
<p>The range of integers depends on how zsh was configured on your machine.
The primary goal is to make sure integers are large enough to represent
indexes into files; on some systems where the hardware usually deals
with 32-bit integers, file sizes may be given by 64-bit integers, and
zsh will try to use 64-bit integers as well. However, zsh will test for
large integers even if no large file support is available; usually it
just requires that your compiler has some easy to recognise way of
defining 64-bit integers, such as `<code>long long</code>' which may be handled by
gcc even if it isn't by the native compiler. You can easily test; if
your zsh supports 64-bit integers, the largest available integer is:</p>
<pre><code> % print $(( 0x7FFFFFFFFFFFFFFF ))
9223372036854775807
</code></pre>
<p>and if you try adding something positive to that, you will get a
negative result due to two's complement arithmetic. This should be large
enough to count most things.</p>
<p>The range of floating point numbers is always that of a C `<code>double</code>',
which is usually also 64 bits, and internally the number is highly
likely to be in the IEEE standard form, which also affects the precision
and range you can get, though that's system specific, too. On most
systems, the math library functions handle <code>double</code>s rather than single
precision <code>float</code>s, so this is the natural choice. The cast function is
called `<code>float</code>' because, unlike C, the representation of a floating
point number is chosen for you, so the generic name is used.</p>
<p><span id="l131"></span></p>
<h3 id="561-entering-and-outputting-bases"><a class="header" href="#561-entering-and-outputting-bases">5.6.1: Entering and outputting bases</a></h3>
<p>I'll say a word or two about bases. I already said you could enter a
number with any small base in a form like `<code>2#101010</code>' or `<code>16#ffff</code>',
and that the latter could also be `<code>0xffff</code>' as in C. You can't,
however, enter octal numbers just by using a leading `<code>0</code>', which you
might expect from C. Here's an example of why not. Let's set:</p>
<pre><code> % foo=${(%):-%D}
% print $foo
01-08-06
</code></pre>
<p>The first line is another of those bogus parameter substitutions where
we gave it a literal string and a blank parameter. We also gave it the
flag `<code>(%)</code>', which forces prompt escapes to be expanded, and in
prompts `<code>(%D)</code>' is the date as <em>yy</em>-<em>mm</em>-<em>dd</em>. Let's write a short
program to find out what the date after <code>$foo</code> is. We have the luxury of
99 years to worry about the century wrapping, so we'll ignore it (and
the Gregorian calendar).</p>
<pre><code> mlens=(31 28 31 30 31 30 31 31 30 31 30 31)
date=(${(s.-.)foo}) # splits to array (01 08 23)
typeset -Z 2 incr
if (( ${date[3]} < ${mlens[${date[2]}]} )); then
# just increment day
(( incr = ${date[3]} + 1 ))
date[3]=$incr
else
# go to first of next month
date[3]=01
if (( ${date[2]} < 12 )); then
(( incr = ${date[2]} + 1 ))
date[2]=$incr
else
# happy new year
date[2]=01
(( incr = ${date[3]} + 1 ))
date[3]=$incr
fi
fi
print ${date[1]}-${date[2]}-${date[3]}
</code></pre>
<p>This will print `<code>01-08-07</code>'. Before I get to the point, various other
explanations. We forced <code>$foo</code> to be split on any `<code>-</code>' in it, giving a
three-part array. The next trick was `<code>typeset -Z 2 incr</code>', which tells
the shell that <code>$incr</code> is to be at least two characters, filled with
leading zeroes. That's how we got the `<code>07</code>' at the end, instead of
just `<code>7</code>'. There's another way of doing this: replace</p>
<pre><code>
typeset -Z 2 incr
(( incr = ${date[2]} + 1 ))
date[2]=$incr
</code></pre>
<p>with:</p>
<pre><code> date[2]=${(l.2..0.)$(( ${date[2]} + 1 ))}
</code></pre>
<p>This uses the <code>(l)</code> parameter flag to fill up to two characters with a
zero (the default is a space, so we need to specify the `<code>0</code>' this
time), using the fact that parameter operations can have a nested
<code>$</code>-substution. This second form is less standard, however.</p>
<p>Now, finally, the point. In that `$(( ${date[2]} + 1 ))', the
`<code>${date[2]}</code>' is simply the <em>scalar</em> `<code>08</code>' --- the result of
splitting an arbitrary string into an array. Suppose we used leading
zeroes to signify octal numbers. We would get something like:</p>
<pre><code> % print $(( ${date[2]} + 1 ))
zsh: bad math expression: operator expected at `8 + 1 '
</code></pre>
<p>because the expression in the substitution becomes `<code>08 + 1</code>' and an 8
can't appear in an octal number. So we would have to strip off any
otherwise harmless leading zeroes. Parsing dates, or indeed strings with
leading zeroes as padding, is a fairly common thing for a shell to do,
and octal arithmetic isn't. So by default leading zeroes don't have that
effect.</p>
<p>However, there is an option you can set, <code>OCTAL_ZEROES</code>; this is
required for compatibility with the POSIX standard. That's how I got the
error message in the previous paragraph, in fact.</p>
<p>Floating point numbers are never octal, always decimal:</p>
<pre><code> % setopt octalzeroes
% print $(( 077 ))
63
% print $(( 077.43 ))
77.430000000000007
</code></pre>
<p>The other option to do with bases is <code>C_BASES</code>, which makes hexadecimal
(and, if you have <code>OCTAL_ZEROES</code> set, octal) numbers appear in the form
that you would use as input to a C (or, once again, Perl) program.</p>
<p>How do you persuade the shell to print out numbers in a particular base
anyway? There are two ways. The first is to associate a base with a
parameter, which you do with an argument after the `<code>-i</code>' option to
typeset:</p>
<pre><code> % typeset -i 16 hexnum=32
% print $hexnum
16#20
</code></pre>
<p>This is the standard way. By the way, there's a slight catch with bases,
taken over from ksh: if you <em>don't</em> specify a base, the first assignment
will do the job for you.</p>
<pre><code> % integer anynum
% (( anynum = 16#20 ))
% print $anynum
16#20
</code></pre>
<p>Only constants with explicit bases in an expression produce this effect;
the first time `<code>anynum</code>' comes into contact with a `<em>base</em><code>#</code><em>num</em>',
or a hexadecimal or (where applicable) octal expression in the standard
C form, it will acquire a default output base. So you need to use
`<code>typeset -i 10</code>' if you don't like that.</p>
<p>Often, however, you just want to print out an expression in, say,
hexadecimal. Zsh has a shorthand for this, which is only in recent
versions (and not in other shells). Preceding an expression by
`<code>[#</code><em>base</em><code>]</code>' causes the default output base to be set to <code>base</code> with
the the usual prefix showing the base, and `<code>[##</code><em>base</em><code>]</code>' will do the
same but without the prefix, i.e. `<code>$(( [##16]255 ))</code>' is simply
`<code>FF</code>'. This has no effect on assignments to a parameter, not even on
the parameter's default output base, but it will affect the result of a
direct substitution using <code>$((...))</code>.</p>
<p><span id="l132"></span></p>
<h3 id="562-parameter-typing"><a class="header" href="#562-parameter-typing">5.6.2: Parameter typing</a></h3>
<p>Just as creating a parameter with an ordinary assignment makes it a
scalar, so creating it in an arithmetic substitution makes it either an
integer or a floating point parameter, according to the value assigned.
This is likely to be a floating point number if there was a floating
point number in the expression on the right hand side, and an integer
otherwise. However, there are reasons why a floating point number on the
right may not have this effect --- use of <code>int</code>, for example, since it
produces an integer.</p>
<p>However, relying on implicit typing in this fashion is bad. One of the
reasons is explained in the manual entry, and I can't do better than use
that example (since I wrote it):</p>
<pre><code> for (( f = 0; f < 1; f += 0.1 )); do
print $f
done
</code></pre>
<p>If you try this, and <code>$f</code> does not already exist, you will see an
endless stream of zeroes. What's happening is that the original
assignment creates <code>$f</code> as an integer to store the integer <code>0</code> in. After
printing this, <code>$f</code> is incremented by adding <code>0.1</code> to it. But once
created, <code>$f</code> remains an integer, so the resulting <code>0.1</code> is cast back to
an integer, and the resulting zero is stored back in <code>$f</code>. The result is
that <code>$f</code> is never incremented.</p>
<p>You could turn the first <code>0</code> into <code>0.0</code>, but a better way is to declare
`<code>float f</code>' before the loop. In a function, this also ensures <code>$f</code> is
local to the function.</p>
<p>If you use a scalar to store an integer or floating point, everything
will work. You don't have the problem just described, since although
<code>$f</code> contains what looks like an integer to start with, it has no
numeric type associated with it, and when you store <code>0.1</code> into <code>$f</code>, it
will happily overwrite the string `<code>0</code>'. It's a bit more inefficient to
use scalars, but actually not that much. You can't specify an output
base or precision, and in versions of zsh up to 4.0.x, there is a
problem when the parameter already has a string in it which doesn't make
sense as a numeric expression:</p>
<pre><code> % foo='/file/name'
% (( foo = 3 ))
zsh: bad math expression: operand expected at `/file/name'
</code></pre>
<p>The unexpected error comes because `<code>/file/name/</code>' is evaluated even
though the shell is about to overwrite the contents of <code>$foo</code>. Versions
of the shell from 4.1.1 have a fix for this, and the integer assignment
works as expected.</p>
<p>You need to be careful with scalars that might contain an empty string.
If you declare `<code>integer i</code>', it will immediately contain the value 0,
but if you declare `<code>typeset s</code>', the scalar <code>$s</code> will just contain the
empty string. You get away with this if you use the parameter without a
`<code>$</code>' in front:</p>
<pre><code> % typeset s
% print $(( 3 * s ))
0
</code></pre>
<p>because the math code tries to retrieve <code>$s</code>, and when it fails puts a
<code>0</code> there. However, if you explicitly use <code>$s</code>, the math code gets
confused:</p>
<pre><code> % print $(( 3 * $s ))
zsh: bad math expression: operand expected at `'
</code></pre>
<p>because `<code>$s</code>' evaluates to an empty string before the arithmetic
evaluation proper, which spoils the syntax. There's one common case
where you need to do that, and that's with positional parameters:</p>
<pre><code> % fn() { print "Twice $1 is $(( 2 * $1 ))"; }
% fn 3
Twice 3 is 6
% fn
fn: bad math expression: operand expected at `'
</code></pre>
<p>Obviously turning the `<code>$1</code>' into `<code>1</code>' means something completely
different. You can guard against this with default values:</p>
<pre><code> % fn() { print "Twice ${1:=0} is $(( 2 * $1 ))"; }
% fn
Twice 0 is 0
</code></pre>
<p>This assigns a default value for <code>$0</code> if one was not set. Since
parameter expansion is performed in one go from left to right, the
second reference to <code>$1</code> will pick up that value.</p>
<p>Note that you need to do this even if it doesn't look like the number
will be needed:</p>
<pre><code> % fn() { print $(( ${1:-0} ? $1 : 3 )); }
% fn
fn: bad math expression: operand expected at `: 3 '
</code></pre>
<p>The expression before the `<code>?</code>' evaluates to zero if <code>$1</code> is not
present, and you expect the expression after the colon to be used in
that case. But actually it's too late by then; the arithmetic expression
parser has received `<code>0 ? : 3</code>', which doesn't make sense to it, hence
the error. So you need to put in `<code>${1:-0}</code>' for the second <code>$1</code>, too
--- or <code>${1:-32}</code>, or any other number, since it won't be evaluated if
<code>$1</code> is empty, it just needs to be parsed.</p>
<p>You should note that just as you can put numbers into scalar parameters
without needing any special handling, you can also do all the usual
string-related tricks on numeric parameters, since there is automatic
conversion in the other direction, too:</p>
<pre><code> % float foo
% zmodload -i zsh/mathfunc
% (( foo = 4 * atan(1.0) ))
% print $foo
3.141592654e+00
% print ${foo%%.*}${foo##*.[0-9]##}
3e+00
</code></pre>
<p>The argument <code>-i</code> to <code>zmodload</code> tells it not to complain if the math
library is already loaded. This gives us access to <code>atan</code>. Remember,
`<code>float</code>' declares a parameter whose output includes an exponent ---
you can actually convert it to a fixed point format on the fly using
`<code>typeset -F foo</code>', which retains the value but alters the output type.
The substitution uses some <code>EXTENDED_GLOB</code> chicanery: the final
`<code>[0-9]##</code>' matches one or more occurrences of any decimal digit. So
the head of the string value of <code>$foo</code> up to the last digit after the
decimal point is removed, and the remainder appended to whatever appears
before the decimal point.</p>
<p>Starting from 4.1.1, a calculator function called <code>zcalc</code> is bundled
with the shell. You type a standard arithmetic expression and the shell
evaluates the formula and prints it out. Lines already entered are
prefixed by a number, and you can use the positional parameter
corresponding to that number to retrieve that result for use in a new
formula. The function uses <code>vared</code> to read the formulae, so the full
shell editing mechanism is available. It will also read in
<code>zsh/mathfunc</code> if that is present.</p>
<p><span id="l133"></span></p>
<h2 id="57-brace-expansion-and-arrays"><a class="header" href="#57-brace-expansion-and-arrays">5.7: Brace Expansion and Arrays</a></h2>
<p>Brace expansion, which you met in <a href="zshguide03.html#syntax">chapter 3</a>,
appears in all csh derivatives, in some versions of ksh, and in bash, so
is fairly standard. However, there are some features and aspects of it
which are only found in zsh, which I'll describe here.</p>
<p>A complication occurs when arrays are involved. Normally, unquoted
arrays are put into a command line as if there is a break between
arguments when there is a new element, so</p>
<pre><code> % array=(three separate words)
% print -l before${array}after
beforethree
separate
wordsafter
</code></pre>
<p>unless the <code>RC_EXPAND_PARAM</code> option is set, which combines the before
and after parts with <em>each</em> element, so you get:</p>
<pre><code> % print -l before${^array}after
beforethreeafter
beforeseparateafter
beforewordsafter
</code></pre>
<p>--- the `<code>^</code>' character turns on the option just for that expansion,
as `<code>=</code>' does with <code>SH_WORD_SPLIT</code>. If you think of the character as a
correction to a proof, meaning `insert a new word between the others
here', it might help you remember (this was suggested by Bart Schaefer).</p>
<p>These two ways of expanding arrays interact differently with braces; the
more useful version here is when the <code>RC_EXPAND_PARAM</code> option is on.
Here the array acts as sort of additional nesting:</p>
<pre><code> % array=(two three)
% print X{one,${^array}}Y
XoneY XtwoY XoneY XthreeY
</code></pre>
<p>with the <code>XoneY</code> tacked on each time, but because of the braces it
appears as a separate word, so there are four altogether.</p>
<p>If <code>RC_EXPAND_PARAM</code> is not set, you get something at first sight
slightly odd:</p>
<pre><code> % array=(two three)
% print X{one,$array}Y
X{one,two three}Y
</code></pre>
<p>What has happened here is that the <code>$array</code> has produced two words; the
first has `<code>X{one,</code>' tacked in front of the array's `<code>two</code>', while the
second likewise has `<code>}Y</code>' on the end of the array's `<code>three</code>'. So by
the time the shell comes to think about brace expansion, the braces are
in different words and don't do anything useful.</p>
<p>There's no obvious simple way of forcing the <code>$array</code> to be embedded in
the braces at the same level, instead of like an additional set of
braces. There are more complicated ways, of course.</p>
<pre><code> % array=(two three)
% print X${^=:-one $array}Y
XoneY XtwoY XthreeY
</code></pre>
<p>Yuk. We gave parameter substitution a string of words, the array with
<code>one</code> stuck in front, and told it to split them on spaces (this will
split on any extra spaces in elements of <code>$array</code>, unfortunately), while
setting <code>RC_EXPAND_PARAM</code>. The parameter flags are `<code>^=</code>'; the `<code>:-</code>'
is the usual `insert the following if the substitution has zero length'
operator. It's probably better just to create your own temporary array
and apply <code>RX_EXPAND_PARAM</code> to that. By the way, if you had
<code>RC_EXPAND_PARAM</code> set already, the last result would have been different
becuase the embedded <code>$array</code> would have been expanded together with the
`<code>one </code>' in front of it.</p>
<p>Braces allow numeric expressions; this works a little like in Perl:</p>
<pre><code> % print {1..10}a
1a 2a 3a 4a 5a 6a 7a 8a 9a 10a
</code></pre>
<p>and you can ask the numbers to be padded with zeroes:</p>
<pre><code> % print {01..10}b
01b 02b 03b 04b 05b 06b 07b 08b 09b 10b
</code></pre>
<p>or have them in descending order:</p>
<pre><code> % print {10..1}c
10c 9c 8c 7c 6c 5c 4c 3c 2c 1c
</code></pre>
<p>Nesting this within other braces works in the expected way, but you
can't have any extra braces inside: the syntax is fixed to number, two
dots, number, and the numbers must be positive.</p>
<p>There's also an option <code>BRACE_CCL</code> which, if the braces aren't in either
of the above forms, expands single letters and ranges of letters:</p>
<pre><code> % setopt braceccl
% print 1{abw-z}2
1a2 1b2 1w2 1x2 1y2 1z2
</code></pre>
<p>An important point to be made about braces is that they are <em>not</em> part
of filename generation; they have nothing to do with pattern matching at
all. The shell blindly generates all the arguments you specify. If you
want to generate only some arguments, depending on what files are
matched, you should use the alternative-match syntax. Compare:</p>
<pre><code> % ls
file1
% print file(1|2)
file1
% print file{1,2}
file1 file2
</code></pre>
<p>The first matches any of `<code>file1</code>' or `<code>file2</code>' it happens to find in
the directory (regardless of other files). The second doesn't look at
files in the directory at all; it simply expands the braces according to
the rules given above.</p>
<p>This point is particularly worthy of note if you have come from a
C-shell world, or use the <code>CSH_NULL_GLOB</code> option:</p>
<pre><code> csh% echo file{1,2}
file1 file2
csh% echo f*{1,2}
file1
</code></pre>
<p>(`<code>csh%</code>' is the prompt, to remind you if you're skipping through
without reading the text), where the difference occurs because in the
first case there was no pattern, so brace expansion was done on ordinary
words, while in the second case the `<code>*</code>' made pattern expansion
happen. In zsh, the sequence would be: `<code>f*{1,2}</code>' becomes `<code>f*1 f*2</code>'; the first becomes <code>file1</code> and the second fails to match. With
<code>CSH_NULL_GLOB</code> set, the failed match is simply removed; there is no
error because one pattern has succeeded in matching. This is presumably
the logic usually followed by the C shell. If you stick with
`<code>file(1|2)</code>' and `<code>f*(1|2)</code>' --- in this case you can simplify them
to `<code>file[12]</code>' and `<code>f*[12]</code>', but that's not true if you have more
than one character in either branch --- you are protected from this
difference.</p>
<p><span id="l134"></span></p>
<h2 id="58-filename-expansion"><a class="header" href="#58-filename-expansion">5.8: Filename Expansion</a></h2>
<p>Filename expansions consists of just `<code>~/...</code>', `<code>~user/...</code>',
`<code>~namedir/...</code>' and `<code>=prog</code>', where the `<code>~</code>' and `<code>=</code>' must be
the first character of a word, and the option <code>EQUALS</code> must be set (it
is by default) for the `<code>=</code>' to be special. I told you about all this
in <a href="zshguide03.html#syntax">chapter 3</a>.</p>
<p>There's really only one thing to add, and that's the behaviour of the
<code>MAGIC_EQUAL_SUBST</code> option. Assignments after <code>typeset</code> and similar
statements are handled as follows</p>
<pre><code> % typeset foo=~pws
% print $foo
/home/pws
% typeset PATH=$PATH:~pws/bin
% print ${path[-1]}
/home/pws/bin
</code></pre>
<p>It may not be obvious why this is not obvious. The point is that
`<code>typeset</code>' is an ordinary command which happens to be a shell builtin;
the arguments of ordinary commands are not assignments. However, a
special case is made here for <code>typeset</code> and its friends so that this
works, even though, as I've said repeatedly, array assignments can't be
done after <code>typeset</code>. The parameter <code>$PATH</code> isn't handled differently
from any other --- any colon in an assignment to any variable is special
in the way shown.</p>
<p>It's often useful to have this feature with commands of your own. There
is an option, <code>MAGIC_EQUAL_SUBST</code>, which spots the forms `<code>...=~...</code>'
and `<code>...=...:~...</code>' for any command at all and expands
<code>~</code>-expressions. Commands where this is particularly useful include
<code>make</code> and the GNU <code>configure</code> command used for setting up the
compilation of a software package from scratch.</p>
<p>A related new option appeared in version 4.0.2 when it became clear
there was an annoying difference between zsh and other shells such as
ksh and bash. Consider:</p>
<pre><code> export FOO=`echo hello there`
</code></pre>
<p>In ksh and bash, this exports <code>$foo</code> with the value `<code>hello there</code>'. In
zsh, however, an unquoted backquote expression forces wordsplitting, so
the line becomes</p>
<pre><code> export FOO=hello there
</code></pre>
<p>and exports <code>$FOO</code> with the value `<code>hello</code>', and <code>$there</code> with any
value it happens to have already or none if it didn't exist. This is
actually perfectly logical according to the rules, but you can set the
option <code>KSH_TYPESET</code> to have the other interpretation.</p>
<p>Normally, <code>KSH_TYPESET</code> applies only after parameter declaration
builtins, and then only in the values of an assignment. However, in
combination with <code>MAGIC_EQUAL_SUBST</code>, you will get the same behaviour
with any command argument that looks like an assignment --- actually,
anything following an `<code>=</code>' which wasn't at the start of the word, so
`<code>"hello mother, => I'm home "$(echo right now)</code>' qualifies.</p>
<p>It seems that bash behaves as if both <code>KSH_TYPESET</code> <em>and</em>
<code>MAGIC_EQUAL_SUBST</code> are always in effect.</p>
<p><span id="l135"></span></p>
<h2 id="59-filename-generation-and-pattern-matching"><a class="header" href="#59-filename-generation-and-pattern-matching">5.9: Filename Generation and Pattern Matching</a></h2>
<p>The final topic is perhaps the biggest, even richer than parameter
expansion. I'm finally going to explain the wonderful world of zsh
pattern matching. In addition to patterns as such, you will learn such
things as how to find all files in all subdirectories, searching
recursively, which have a given name, case insensitive, are at least 50
KB large, no more than a week old and owned by the root user, and
allowing up to a single error in the spelling of the name. In fact, the
required expression looks like this:</p>
<pre><code> **/(#ia1)name(LK+50mw-1u0)
</code></pre>
<p>which might appear, at first sight, a mite impenetrable. We'll work up
to it gradually.</p>
<p>To repeat: filename generation is just the same as globbing, only
longer. I use the terms interchangeably.</p>
<p><span id="l136"></span></p>
<h3 id="591-comparing-patterns-and-regular-expressions"><a class="header" href="#591-comparing-patterns-and-regular-expressions">5.9.1: Comparing patterns and regular expressions</a></h3>
<p>It can be confusing that there are two rather different sorts of pattern
around, those used for matching files on a command line as in zsh and
other shells, and those used for matching text inside files as in
<code>grep</code>, <code>sed</code>, <code>emacs</code>, <code>perl</code> and many other utilities, each of which,
typically, has a slightly different form for patterns (called in this
case `regular expressions', because UNIX was designed by computer
scientists). There are even some utilities like TCL which provide both
forms.</p>
<p>Zsh deals exclusively with the shell form, which I've been calling by
its colloquial name, `globbing', and consequently I won't talk about
regular expressions in any detail. Here are the two classic differences
to note. First, in a shell, `<code>*</code>' on its own matches any set of
characters, while in a regular expression it always refers to the
previous pattern, and says that that can be repeated any number of
times. Second, in a shell `<code>.</code>' is an ordinary (and much used)
character, while in a regular expression it means `any character',
which is specified by `<code>?</code>' in the shell. Put this together, and what a
shell calls `<code>*</code>' is given by `<code>.*</code>' in a regular expression. `<code>*</code>'
in the latter case is called a `Kleene closure': it's those computer
scientists again. In zsh, art rather than science tends to be in
evidence.</p>
<p>In fact, zsh does have many of the features available in regular
expressions, as well as some which aren't. Remember that anywhere in zsh
where you need a pattern, it's of the same form, whether it's matching
files on the command line or a string in a <code>case</code> statement. There are a
few features which only fit well into one or another use of patterns;
for example the feature that selects files by examining their type,
owner, age, etc. (the final parenthesis in the expression I showed
above) are no use in matching against a string.</p>
<p><span id="l137"></span></p>
<h3 id="592-standard-features"><a class="header" href="#592-standard-features">5.9.2: Standard features</a></h3>
<p>There is one thing to note about the simple pattern matching features
`<code>*</code>' and `<code>?</code>', which is that when matching file names (not in other
places patterns are used, however) they never match a leading `<code>.</code>'.
This is a convention in UNIX-like systems to hide certain files which
are not interesting to most users. You may have got the impression that
files begining with `<code>.</code>' are somehow special, but that's not so; only
the files `<code>.</code>' (the current directory) and `<code>..</code>' (the parent
directory, or the current directory in <code>/</code>) are special to the system.
Other files beginning with `<code>.</code>' only appear special because of a
conspiracy between the shell (the rule I've just given) and the command
<code>ls</code>, which, when it lists a directory, doesn't show files beginning
`<code>.</code>' unless you give the `<code>-a</code>' option. Otherwise `<code>.</code>'-files are
perfectly normal files.</p>
<p>You can suppress the special rule for an initial `<code>.</code>' by setting the
option <code>GLOB_DOTS</code>, in which case `<code>*</code>' will match every single file
and directory except for `<code>.</code>' and `<code>..</code>'.</p>
<p>In addition to `<code>*</code>' and `<code>?</code>', which are so basic that even DOS had
them (though I never <em>quite</em> worked out exactly what it was doing with
them a lot of the time), the pattern consisting of a set of characters
in square brackets appears in all shells. This feature happens to be
pretty much the same as in regular expressions. `<code>[abc]</code>' matches any
one of those three characters; `<code>[a-z]</code>' matches any character between
<code>a</code> and <code>z</code>, inclusive; `<code>[^a-z]</code>' matches any single character
<em>except</em> those 26 --- but notice it still matches a single character.</p>
<p>A recent common enhancement to character ranges features in zsh, which
is to specify types of characters instead of listing them; I'm just
repeating the manual entry here, which you should consult for more
detail. The special syntax is like `<code>[:</code><em>spec</em><code>:]</code>', where the square
brackets there are in addition to the ones specifying the range. If you
are familiar with the `ctype' macros use in C programmes, you will
probably recognise the things that <em>spec</em> can be: <code>alnum</code>, <code>alpha</code>,
<code>blank</code>, <code>cntrl</code>, <code>digit</code>, <code>graph</code>, <code>lower</code>, <code>print</code>, <code>punct</code>, <code>space</code>,
<code>upper</code>, <code>xdigit</code>. The similarity to C macros isn't just for show: the
shell really does call the macro (or function) `<code>isalpha</code>' to test for
<code>[:alpha:]</code>ness, and so on. On most modern systems which support
internationalization this means the shell can tell you whether a
character is, say, an alphabetic letter in the character set in use on
your machine. By the way, zsh doesn't use international character set
support for sorting matches --- this turned out to produce too many
unexpected effects.</p>
<p>So `<code>[^[:digit:]]</code>' matches any single character other than a decimal
digit. Standards say you should use `<code>!</code>' instead of `<code>^</code>' to signify
negation, but most people I know don't; also, this can clash with
history substitution. However, it is accepted by zsh anywhere where
history substitution doesn't get its hands on the `<code>!</code>' first (which
includes all scripts and autoloaded functions).</p>
<p><span id="l138"></span></p>
<h3 id="593-extensions-usually-available"><a class="header" href="#593-extensions-usually-available">5.9.3: Extensions usually available</a></h3>
<p>Now we reach the bits specific to zsh. I've divided these into two
parts, since some require the option `<code>EXTENDED_GLOB</code>' to be set ---
those which are most likely to clash with other uses of the characters
in question.</p>
<p><strong>Numeric ranges</strong></p>
<p>One possibility that is always available is the syntax for numeric
ranges in the form `<code><</code><em>num1</em><code>-</code><em>num2</em><code>></code>'. You can omit either <em>num1</em>,
which defaults to zero, or <em>num2</em>, which defaults to infinity, or both,
in which case any set of digits will be matched. Note that this really
<em>does</em> mean infinity, despite the finite range of integers; missing out
<em>num2</em> is treated as a special case and the shell will simply advance
over any number of digits. (In <em>very</em> old versions of zsh you had to use
`<code><></code>' to get that effect, but that has been removed and `<code><></code>' is now
a redirection operator, as in other shells; `<code><-></code>' is what you need
for any set of digits.)</p>
<p>I repeat another warning from the manual: this test</p>
<pre><code> [[ 342 = <1-30>* ]]
</code></pre>
<p>succeeds, even though the number isn't in the range 1 to 30. That's
because `<code><1-30></code>' matches `<code>3</code>' and `<code>*</code>' matches 42. There's no use
moaning, it's a consequence of the usual rule for patterns of all types
in shells or utilities: pattern operators are tried independently, and
each `uses up' the longest piece of the string it is matching without
causing the rest of the match to fail. We would have to break this
simple and well-tried rule to stop numeric ranges matching if there is
another digit left. You can test for that yourself, of course:</p>
<pre><code> [[ 342 = <1-30>(|[^[:digit:]]*) ]]
</code></pre>
<p>fails. I wrote it so that it would match any number between 1 and 30,
either not followed by anything, or followed by something which doesn't
start with a digit; I will explain what the parentheses and the vertical
bar are doing in the next section. By the way, leading zeroes are
correctly handled (and never force octal interpretation); so
`<code>00000003NaN</code>' would successfully match the pattern.</p>
<p>The numbers in the range are always positive integers; you need extra
pattern trickery to match floating point. Here's one attempt, which uses
<code>EXTENDED_GLOB</code> operators, so come back and look when you've read the
rest of this section if it doesn't make sense now:</p>
<pre><code> isfloat() {
setopt localoptions extendedglob
if [[ $1 = ([-+]|)([0-9]##.[0-9]#|[0-9]#.[0-9]##)\
([eE]([-+]|)[0-9]##|) ]]; then
print -r -- "$1 is a floating point number"
else
print -r -- "$1 is not a floating point number"
fi
}
</code></pre>
<p>I've split it over two lines to fit. The first parenthesis matches an
optional minus or plus sign --- careful with `<code>-</code>' in square brackets,
since if it occurs in the middle it's taken as a range, and if you want
it to match itself, it has to be at the start or end. The second
parenthesis contains an alternative because `<code>.</code>' isn't a floating
point number (at least, not in my book, and not in zsh's, either), but
both `<code>0.</code>' and `<code>.0</code>' <em>are</em> properly formed numbers. So we need at
least one digit, either before or after the decimal point; the `<code>##</code>'
means `at least one occurrence of the previous expression', while the
`<code>#</code>' means `zero or more occurrences of the previous expression'. The
expresion on the next line matches an exponent; here you need at least
one digit, too. So `<code>3.14159E+00</code>' is successfully matched, and indeed
you'll find that zsh's arithmetic operations handle it properly.</p>
<p>The range operator is the only special zsh operator that you can't turn
off with an option. This is usually not a problem, but in principle a
string like `<code><3-10></code>' is ambiguous, since in another shell it would be
read as `<code><3-10 ></code>', meaning `take input from file <code>3-10</code>, and send
output to the file formed by whatever comes after the expression'. It's
very unlikely you will run across this in practice, however, since shell
code writers nearly alwys put a space after the end of a file name for
redirection if something else follows on the command line, and that's
enough to differentiate it from a range operator.</p>
<p><strong>Parentheses</strong></p>
<p>Parentheses are quite natural in zsh if you've used extended regular
expressions. They are usually available, and only turned off if you set
the `<code>SH_GLOB</code>' option to ensure compatibility with shells that don't
have it. The key part of the expression is the vertical bar, which
specifies an alternative. It can occur as many times as necessary;
`<code>(a|b|c|d|e|f|g|h|i|j|k|l|m)</code>' is a rather idiosyncratic way of
writing `<code>[a-m]</code>'. If you don't include the vertical bar (we'll see
reasons for not doing so later), and you are generating filenames, you
should be careful that the expression doesn't occur at the end of the
pattern, else it would be taken as a `glob qualifier', as described
below. The rather unsightly hack of putting `<code>(|)</code>' (match the empty
string or the empty string --- guess what this matches?) right at the
end will get around that problem.</p>
<p>The vertical bar usually needs to be inside parentheses so that the
shell doesn't take it as a pipe, but in some contexts where this won't
happen, such as a case statement label, you can omit any parentheses
that would completely surround the pattern. So in</p>
<pre><code> case $foo in
(bar|rod|pipe) print "foo represents a piece of metal"
;;
(*) print "Are you trying to be different?"
;;
esac
</code></pre>
<p>the surrounding parentheses are the required syntax for <code>case</code>, rather
than pattern parentheses --- the same syntax works in other shells. Then
`<code>bar|rod</code>' is an ordinary zsh expression matching either <code>bar</code> or
<code>rod</code>, in a context where the `<code>|</code>' can't be mistaken for a pipe. In
fact, this whole example works with <code>ksh</code> --- but there the use of
`<code>|</code>' is a special case, while in zsh it fits in with the standard
pattern rules.</p>
<p>Indeed, ksh has slightly different ways of specifying patterns: to make
the use of parentheses less ambiguous, it requires a character before
the left parenthesis. The corresponding form for a simple alternative is
`<code>@(this|that)</code>'. The `<code>@</code>' can also be a `<code>?</code>', for zero or one
occurrences of what's in the parentheses; `<code>*</code>' for any number of
repetitions, for example `<code>thisthisthatthis</code>'; or `<code>!</code>' for anything
except what's in the parentheses. Zsh allows this syntax if you set the
option <code>KSH_GLOB</code>. Note that this is independent of the option
<code>SH_GLOB</code>; if you set <code>KSH_GLOB</code> but not <code>SH_GLOB</code>, you can actually use
both forms for pattern matching, with the ksh form taking precedence in
the case of ambiguities. This is probably to be avoided. In ksh
emulation, both options are set; this is the only sensible reason I know
of for using these options at all. I'll show some comparisons in the
next section.</p>
<p>An important thing to note is that when you are matching files, you
can't put directory separators inside parentheses:</p>
<pre><code> # Doesn't work!
print (foo/bar|bar/foo)/file.c
</code></pre>
<p>doesn't work. The reason is that it's simply too difficult to write;
pattern matching would be bound in a highly intricate way with searching
the directory hierarchy, with the end of a group sending you back up to
try another bit of the pattern on a directory you'd already visited.
It's probably not impossible, but the pattern code maintainer (me) isn't
all that enthusiastic about it.</p>
<p><span id="l139"></span></p>
<h3 id="594-extensions-requiring-extended_glob"><a class="header" href="#594-extensions-requiring-extended_glob">5.9.4: Extensions requiring <code>EXTENDED_GLOB</code></a></h3>
<p>Setting <code>EXTENDED_GLOB</code> makes three new types of operator available:
those which excluded a particular pattern from matching; those which
specify that a pattern may occur a repeated number of times; and a set
of `globbing flags', a little bit like parameter flags which I'll
describe in a later section since they are really the icing on the cake.</p>
<p><strong>Negative matches or exclusions</strong></p>
<p>The simpler of the two exclusions uses `<code>^</code>' to introduce a pattern
which must <em>not</em> be matched. So a trivial example (I will assume for
much of the rest of the chapter that the option <code>EXTENDED_GLOB</code> is set)
is:</p>
<pre><code> [[ foo = ^foo ]]
[[ bar = ^foo ]]
</code></pre>
<p>The first test fails, the second succeeds. It's important to realise
that that the pattern demands nothing else whatever about the relevant
part of the test string other than it doesn't match the pattern that
follows: it doesn't say what length the matched string should have, for
example. So</p>
<pre><code> [[ foo = *^foo ]]
</code></pre>
<p>actually <em>does</em> match: <code>*</code> swallows up the whole string, and the
remaining empty string successfully fails to be `<code>foo</code>'. Remember the
mantra: each part of the pattern matches the longest possible substring
that causes the remainder of the pattern not to fail (unless, of course,
failure is unavoidable).</p>
<p>Note that the <code>^</code> applies to the whole pattern to its right, either to
the end of the string, or to the end of the nearest enclosing
parenthesis. Here's a couple more examples:</p>
<pre><code> [[ foo = ^foo* ]]
</code></pre>
<p>Overall, this fails to match: the pattern `<code>foo*</code>' always matches the
string on the left, so negating means it always fails.</p>
<pre><code> [[ foo = (^foo)* ]]
</code></pre>
<p>This is similar to the last example but one. The expression in the
parenthesis first matches against <code>foo</code>; this causes the overall match
to fail because of the <code>^</code>, so it backs up one character and tries
again. Now `<code>fo</code>' is successfully matched by <code>^foo</code> and the remaining
`<code>o</code>' is matched by the <code>*</code>, so the overall match succeeds. When you
know about backreferences, you will be able to confirm that, indeed, the
expression in parentheses matches `<code>fo</code>'. This is a quite subtle point:
it's easy to imagine that `<code>^foo</code>' says `match any three letter string
except the one I've given you', but actually there is no requirement
that it match three letters, or indeed any.</p>
<p>In filename generation, the <code>^</code> has a lower precedence than a slash:</p>
<pre><code> % print /*/tmp
/data/tmp /home/tmp /usr/tmp /var/tmp
% print /^usr/tmp
/data/tmp /home/tmp /var/tmp
</code></pre>
<p>successfully caused the first level of directories to match anything but
`<code>usr</code>'. A typical use of this with files is `<code>^*.o</code>' to match
everything in a directory except files which end with `<code>.o</code>'.</p>
<p>Note one point mentioned in the FAQ --- probably indicating the reason
that `<code>^</code>' is only available with <code>EXTENDED_GLOB</code> switched on. Some
commands use an initial `<code>^</code>' to indicate a control character; in fact,
zsh's <code>bindkey</code> builtin does this:</p>
<pre><code> bindkey '^z' backward-delete-word
</code></pre>
<p>which attaches the given function to the keystroke <code>Ctrl-z</code>. You must
remember to quote that keystroke expression, otherwise it would expand
to a list of all files in the current directory not called `<code>z</code>', very
likely all of them.</p>
<p>There's another reason this isn't available by default: in some versions
of the Bourne shell, `<code>^</code>' was used for pipes since `<code>|</code>' was missing
on some keyboards.</p>
<p>The other exclusion operator is closely related. `<em>pat1</em><code>~</code><em>pat2</em>'
means `anything that matches <em>pat1</em> as long as it doesn't also match
<em>pat2</em>'. If <em>pat1</em> is <code>*</code>, you have the same effect as `<code>^</code>' --- in
fact, that's pretty much how `<code>^</code>' is currently implemented.</p>
<p>There's one significant difference between `<code>*~</code><em>pat</em>' and `<code>^</code><em>pat</em>':
the <code>~</code> has a <em>lower</em> precedence than `<code>/</code>' when matching against
filenames. What's more, the pattern on the right of the <code>~</code> is not
treated as a filename at all; it's simply matched against any filename
found on the left, to see if it should be rejected. This sounds like
black magic, but it's actually quite useful, particularly in combination
with the recursive globbing syntax:</p>
<pre><code> print **/*~*/CVS(/)
</code></pre>
<p>matches any subdirectory of the current directory to any depth, except
for directories called <code>CVS</code> --- the `<code>*</code>' on the right of the `<code>~</code>'
will match any character including `<code>/</code>'. The final `<code>(/)</code>' is a glob
qualifier indicating that only directories are to be allowed to match
--- note that it's a positive assertion, despite being after the `<code>~</code>'.
Glob qualifiers do not feel the effect of preceding exclusion operators.</p>
<p>Note that in that example, any subdirectory of a directory called <code>CVS</code>
would have matched successfully; you can see from the pattern that the
expression after the `<code>~</code>' wouldn't weed it out. Slightly less
obviously, the `<code>**/*</code>' matches files in the current directory, while
the `<code>*/CVS</code>' never matches a `<code>CVS</code>' in the current directory, so
that could appear. If you want to, you can fix that up like this:</p>
<pre><code> print **/*~(*/|)CVS(/*|)(/)
</code></pre>
<p>again relying on the fact that `<code>/</code>'s are not special after the `<code>~</code>'.
This will ruthlessly weed out any path with a directory component called
<code>CVS</code>. An easier, but less instructive, way is</p>
<pre><code> print ./**/*~*/CVS(/)
</code></pre>
<p>You can restrict the range of the tilde operator by putting it in
parentheses, so `<code>/(*~usr)/tmp</code>' is equivalent to `<code>/^usr/tmp</code>'.</p>
<p>A `<code>~</code>' at the beginning is never treated as excluding what follows; as
you already know, it has other uses. Also, a `<code>~</code>' at the end of a
pattern isn't special either; this is lucky, because Emacs produces
backup files by adding a `<code>~</code>' to the end of the file name. You may
have problems if you use Emacs's facility for numbered backup files,
however, since then there is a `<code>~</code>' in the middle of the file name,
which will need to be quoted when used in the shell.</p>
<p><strong>Closures or repeated matches</strong></p>
<p>The extended globbing symbols `<code>#</code>' and `<code>##</code>', when they occur in a
pattern, are equivalent to `<code>*</code>' and `<code>+</code>' in extended regular
expressions: `<code>#</code>' allows the previous pattern to match any number of
times, including zero, while with `<code>##</code>' it must match at least once.
Note that this pattern does not extend beyond two hashes --- there is no
special symbol `<code>###</code>', which is not recognised as a pattern at all.</p>
<p>The `previous pattern' is the smallest possible item which could be
considered a complete pattern. Very often it is something in
parentheses, but it could be a group in square or angle brackets, or a
single ordinary character. Note particularly that in</p>
<pre><code> # fails
[[ foofoo = foo# ]]
</code></pre>
<p>the test fails, because the `<code>#</code>' only refers to the final `<code>o</code>', not
the entire string. What you need is</p>
<pre><code> # succeeds
[[ foofoo = (foo)# ]]
</code></pre>
<p>It might worry you that `<code>#</code>' also introduces comments. Since a
well-formatted pattern never has `<code>#</code>' at the start, however, this
isn't a problem unless you expect comments to start in the middle of a
word. It turns out that doesn't even happen in other shells --- `<code>#</code>'
must be at the start of a line, or be unquoted and have space in front
of it, to be recognised as introducing a comment. So in fact there is no
clash at all here. There is, of course, a clash if you expect
`<code>.#foo.c.1.131</code>' (probably a file produced by the version control
system CVS while attempting to resolve a conflict) to be a plain string,
hence the dependence on the <code>EXTENDED_GLOB</code> option.</p>
<p>That's probably all you need to know; the `<code>#</code>' operators are generally
much easier to understand than the exclusion operators. Just in case you
are confused, I might as well point out that repeating a <em>pattern</em> is
not the same as repeating a <em>string</em>, so</p>
<pre><code> [[ onetwothreetwoone = (one|two|three)## ]]
</code></pre>
<p>successfully matches; the string is different for each repetition of the
pattern, but that doesn't matter.</p>
<p>We now have enough information to construct a list of correspondences
between zsh's normal pattern operators and the ksh ones, available with
<code>KSH_GLOB</code>. Be careful with `<code>!</code>(<em>...</em>)'; it seems to have a slightly
different behaviour to the zsh near-equivalent. The following table is
lifted directly from the zsh FAQ.</p>
<pre><code>----------------------------------------------------------------------
ksh zsh Meaning
------ ------ ---------
!(foo) ^foo Anything but foo.
or foo1~foo2 Anything matching foo1 but foo2.
@(foo1|foo2|...) (foo1|foo2|...) One of foo1 or foo2 or ...
?(foo) (foo|) Zero or one occurrences of foo.
*(foo) (foo)# Zero or more occurrences of foo.
+(foo) (foo)## One or more occurrences of foo.
----------------------------------------------------------------------
</code></pre>
<p>In both languages, the vertical bar for alternatives can appear inside
any set of parentheses. Beware of the precedences of <code>^foo</code> and
`<code>foo1~foo2</code>'; surround them with parentheses, too, if necessary.</p>
<p><span id="l140"></span></p>
<h3 id="595-recursive-globbing"><a class="header" href="#595-recursive-globbing">5.9.5: Recursive globbing</a></h3>
<p>One of the most used special features of zsh, and one I've already used
a couple of times in this section, is recursive globbing, the ability to
match any directory in an arbitrarily deep (or, as we say in English,
tall) tree of directories. There are two forms: `<code>**/</code>' matches a set
of directories to any depth, including the top directory, what you get
by replacing `<code>**/</code>' by `<code>./</code>, i.e. <code>**/foo</code> can match <code>foo</code> in the
current directory, but also <code>bar/foo</code>, <code>bar/bar/bar/foo</code>,
<code>bar/bar/bar/poor/little/lambs/foo</code> nad so on. `<code>***/</code>' does the same,
but follows symbolic links; this can land you in infinite loops if the
link points higher up in the same directory hierarchy --- an odd thing
to do, but it can happen.</p>
<p>The `<code>**/</code>' or `<code>***/</code>' can't appear in parentheses; there's no way of
specifying them as alternatives. As already noticed, however, the
precedence of the exclusion operator `<code>~</code>' provides a useful way of
removing matches you don't want. Remember, too, the recursion operators
don't need to be at the start of the pattern:</p>
<pre><code> print ~/**/*.txt
</code></pre>
<p>prints the name of all the files under your home directory ending with
`<code>.txt</code>'. Don't expect it to be particularly fast; it's not as well
optimised as the standard UNIX command <code>find</code>, although it is a whole
lot more convenient. The traditional way of searching a file which may
be anywhere in a directory tree is along the lines of:</p>
<pre><code> find ~/src -name '*.c' -print | xargs grep pattern
</code></pre>
<p>which is horrendously cumbersome. What's happening is that <code>find</code>
outputs a newline-separated list of all the files it finds, and <code>xargs</code>
assembles these as additional arguments to the command `<code>grep pattern</code>'. It simplifies in zsh to the much more natural</p>
<pre><code> grep pattern ~/src/**/*.c
</code></pre>
<p>In fact, strictly speaking you probably ought to use</p>
<pre><code> find ~/src -name '*.c' -print0 | xargs -0 grep pattern
</code></pre>
<p>for the other form --- this passes null-terminated strings around, which
is safer since any character other than a NUL or a slash can occur in a
filename. But of course you don't need that now.</p>
<p>Do remember that this includes the current directory in the search, so
in that last example `<code>foo.c</code>' in the directory where you typed the
command would be searched. This isn't completely obvious because of the
`<code>/</code>' in the pattern, which erroneously seems to suggest at least one
directory.</p>
<p>It's a little known fact that this is a special case of a more general
syntax, `(<em>pat</em><code>/</code>)<code>#</code>'. This syntax isn't perfect, either; it's the
only time where a `<code>/</code>' can usefully occur in parentheses. The pattern
<em>pat</em> is matched against each directory; if it succeeds, <em>pat</em> is
matched against each of the subdirectories, and so on, again to
arbitrary depth. As this uses the character `<code>#</code>', it requires the
<code>EXTENDED_GLOB</code> option, which the more common syntax doesn't, since
no-one would write two <code>*</code>'s in a row for any other reason.</p>
<p>You should consider the `<code>/</code>)' to be in effect a single pattern token;
for example in</p>
<pre><code> % print (F*|B*/)#*.txt
FOO/BAR/thingy.txt
</code></pre>
<p>both `<code>F*</code>' and `<code>B*</code>' are possible directory names, not just the
`<code>B*</code>' next to the slash. The difference between `<code>#</code>' and `<code>##</code>' is
respected here --- with the former, zero occurrences of the pattern may
be matched (i.e. `<code>*.txt</code>'), while with the latter, at least one level
of subdirectories is required. Thus `<code>(*/)##*.txt</code>' is equivalent to
`<code>*/**/*.txt</code>', except that the first `<code>*</code>' in the second pattern will
match a symbolic link to a directory; there's no way of forcing the
other syntax to follow symbolic links.</p>
<p>Fairly obviously, this syntax is only useful with files. Other uses of
patterns treat slashes as ordinary characters and `<code>**</code>' or `<code>***</code>'
the same as a single `<code>*</code>'. It's not an error to use multiple `<code>*</code>'s,
though, just pointless.</p>
<p><span id="l141"></span></p>
<h3 id="596-glob-qualifiers"><a class="header" href="#596-glob-qualifiers">5.9.6: Glob qualifiers</a></h3>
<p>Another very widely used zsh enhancement is the ability to select types
of file by using `glob qualifiers', a group of (rather terse) flags in
parentheses at the end of the pattern. Like recursive globbing, this
feature only applies for filename generation in the command line
(including an array assignment), not for other uses of patterns.</p>
<p>This feature requires the <code>BARE_GLOB_QUAL</code> option to be turned on, which
it usually is; the name implies that one day there may be another,
perhaps more ksh-like, way of doing the same thing with a more
indicative syntax than just a pair of parentheses.</p>
<p><strong>File types</strong></p>
<p>The simplest glob qualifiers are similar to what the completion system
appends at the end of file names when the <code>LIST_TYPES</code> option is on;
these are in turn similar to the indications used by `<code>ls -F</code>'. So</p>
<pre><code> % print *(.)
file1 file2 cmd1 cmd2
% print *(/)
dir1 dir2
% print *(*)
cmd1 cmd2
% print *(@)
symlink1 symlink2
</code></pre>
<p>where I've invented unlikely filenames with obvious types. <code>file1</code> and
<code>file2</code> were supposed to be just any old file; <code>(.)</code> picks up those but
also executable files. Sockets <code>(=)</code>, named pipes <code>(p)</code>, and device
files <code>(%)</code> including block <code>(%b)</code> and character <code>(%c)</code> special files
are the other types of file you can detect.</p>
<p>Associated with type, you can also specify the number of hard links to a
file: <code>(l2)</code> specifies exactly 2 links, <code>(l+3)</code> more than 3 links,
<code>(l-5)</code> fewer than 5.</p>
<p><strong>File permissions</strong></p>
<p>Actually, the <code>(*)</code> qualifier really applies to the file's permissions,
not it's type, although it does require the file to be an executable
non-special file, not a directory nor anything wackier. More basic
qualifiers which apply just to the permissions of the files are <code>(r)</code>,
<code>(w)</code> and <code>(x)</code> for files readable, writeable and executable by the
owner; <code>(R)</code>, <code>(W)</code> and <code>(X)</code> correspond to those for world permissions,
while <code>(A)</code>, <code>(I)</code> and <code>(E)</code> do the job for group permissions --- sorry,
the Latin alphabet doesn't have middle case. You can speciy permissions
more exactly with `<code>(f)</code>' for file permissions: the expression after
this can take various forms, but the easiest is probably a delimited
string, where the delimiters work just like the arguments for parameter
flags and the arguments, separated by commas, work just like symbolic
arguments to <code>chmod</code>; the example from the manual,</p>
<pre><code> print *(f:gu+w,o-rx:)
</code></pre>
<p>picks out files (of any type) which are writeable by the owner (`user')
and group, and neither readable nor executable by anyone else
(`other').</p>
<p><strong>File ownership</strong></p>
<p>You can match on the other three mode bits, setuid ((s)), setgid ((S))
and sticky ((t)), but I'm not going to go into what those are if you
don't know; your system's manual page for <code>chmod</code> may (or may not)
explain.</p>
<p>Next, you can pick out files by owner; <code>(U)</code> and <code>(G)</code> say that you or
your group, respectively, owns the file --- really the effective user or
group ID, which is usually who you are logged in as, but this may be
altered by tricks such as a programme running setuid or setgid (the
things I'm not going to explain). More generally, <code>u0</code> says that the
file is owned by root and <code>(u501)</code> says it is owned by user ID 501; you
can use names if you delimiit them, so <code>(u:pws:)</code> says that the owner
must be user <code>pws</code>; similarly for groups with <code>(g)</code>.</p>
<p><strong>File times</strong></p>
<p>You can also pick files by modification ((m)) or access ((a)) time,
either before ((-)), at, or after ((+)) a specific time, which may be
measured in days (the default), months ((M)), weeks ((w)), hours ((h)),
minutes ((m)) or seconds ((s)). These must appear in the order <code>m</code> or
<code>a</code>, optional unit, optional plus or minus, number. Hence:</p>
<pre><code> print *(m1)
</code></pre>
<p>Files that were modified one day ago --- i.e. less than 48 but more than
24 hours ago.</p>
<pre><code> print *(aw-1)
</code></pre>
<p>Files accessed within the last week, i.e. less than 7 days ago.</p>
<p>In addition to <code>(m)</code> and ((a)), there is also <code>(c)</code>, which is sometimes
said to refer to file creation, but it is actually something a bit less
useful, namely <em>inode</em> change. The inode is the structure on disk where
UNIX-like filing systems record the information about the location and
nature of the file. Information here can change when some aspect of the
file information, such as permissions, changes.</p>
<p><strong>File size</strong></p>
<p>The qualifier <code>(L)</code> refers to the file size (`L' is actually for
length), by default in bytes, but it can be in kilobytes <code>(k)</code>,
megabytes <code>(m)</code>, or 512-byte blocks <code>(p, unfortunately)</code>. Plus and minus
can be used in the same way as for times, so</p>
<pre><code> print *(Lk3)
</code></pre>
<p>gives files 3k large, i.e. larger than 2k but smaller than 4k, while</p>
<pre><code> print *(Lm+1)
</code></pre>
<p>gives files larger than a megabyte.</p>
<p>Note that file size applies to directories, too, although it's not very
useful. The size of directories is related to the number of slots for
files currently available inside the directory (at the highest level,
i.e. not counting children of children and deeper). This changes
automatically if necessary to make more space available.</p>
<p><strong>File matching properties</strong></p>
<p>There are a few qualifiers which affect option settings just for the
match in question: <code>(N)</code> turns on <code>NULL_GLOB</code>, so that the pattern
simply disappears from the command line if it fails to match; <code>(D)</code>
turns on <code>GLOB_DOTS</code>, to match even files beginning with a `<code>.</code>', as
described above; <code>(M)</code> or <code>(T)</code> turn on <code>MARK_DIRS</code> or <code>LIST_TYPES</code>, so
that the result has an extra character showing the type of a directory
only (in the first case) or of any special file (in the second); and
<code>(n)</code> turns on <code>NUMERIC_GLOB_SORT</code>, so that numbers in the filename are
sorted arithmetically --- so <code>10</code> comes after <code>1A</code>, because the 1 and 10
are compared before the next character is looked at.</p>
<p>Other than being local to the pattern qualified, there is no difference
in effect from setting the option itself.</p>
<p><strong>Combining qualifiers</strong></p>
<p>One of the reasons that some qualifiers have slightly obscure syntax is
that you can chain any number of them together, which requires that the
file has all of the given properties. In other words `<code>*(UWLk-10)</code>' are
files owned by you, world writeable and less than 10k in size.</p>
<p>You can negate a set of qualifiers by putting `<code>^</code>' in front of those,
so `<code>*(ULk-10^W)</code>' would specify the corresponding files which were not
world writeable. The `<code>^</code>' applies until the end of the flags, but you
can put in another one to toggle back to assertion instead of negation.</p>
<p>Also, you can specify alternatives; `<code>*(ULk-10,W)</code>' are files which
either are owned by you and are less than 10k, or are world writeable
--- note that the `and' has higher precedence than the `or'.</p>
<p>You can also toggle whether the assertions or negations made by
qualifiers apply to symbolic links, or the files found by following
symbolic links. The default is the former --- otherwise the <code>(@)</code>
qualifier wouldn't work on its own. By preceding qualifiers with <code>-</code>,
they will follow symbolic links. So <code>*(-/)</code> matches all directories,
including those reached by a symbolic link (or more than one symbolic
link, up to the limit allowed by your system). As with `<code>^</code>', you can
toggle this off again with another one `<code>-</code>'. To repeat what I said in
<a href="zshguide03.html#syntax">chapter 3</a>, you can't distinguish between the
other sort of links, hard links, and a real file entry, because a hard
link just supplies an alternative but equivalent name for a file.</p>
<p>There's a nice trick to find broken symlinks: the pattern `<code>**/*(-@)</code>'.
This is supposed to follow symlinks; but that `<code>@</code>' tells it to match
only on symlinks! There is only one case where this can succeed, namely
where the symlink is broken. (This was pointed out to me by Oliver
Kiddle.)</p>
<p><strong>Sorting and indexing qualifiers</strong></p>
<p>Normally the result of filename generation is sorted by alphabetic order
of filename. The globbing flags <code>(o)</code> and <code>(O)</code> allow you to sort in
normal or reverse order of other things: <code>n</code> is for names, so <code>(on)</code>
gives the default behaviour while <code>(On)</code> is reverse order; <code>L</code>, <code>l</code>,
<code>m</code>, <code>a</code> and <code>c</code> refer to the same thing as the normal flags with those
letters, i.e. file size, number of links, and modification, access and
inode change times. Finally, <code>d</code> refers to subdirectory depth; this is
useful with recursive globbing to show a file tree ordered depth-first
(subdirectory contents appear before files in any given directory) or
depth-last.</p>
<p>Note that time ordering produces the most recent first as the standard
ordering (<code>(om)</code>, etc.), and oldest first as the reverse ordering
<code>(OM)</code>, etc.). With size, smallest first is the normal ordering.</p>
<p>You can combine ordering criteria, with the most important coming first;
each criterion must be preceded by <code>o</code> or <code>O</code> to distinguish it from an
ordinary globbing flag. Obviously, <code>n</code> serves as a complete
discriminator, since no two different files can have the same name, so
this must appear on its own or last. But it's a good idea, when doing
depth-first ordering, to use <code>odon</code>, so that files at a particular depth
appear in alphabetical order of names. Try</p>
<pre><code> print **/*(odon)
</code></pre>
<p>to see the effect, preferably somewhere above a fairly shallow directory
tree or it will take a long time.</p>
<p>There's an extra trick you can play with ordered files, which is to
extract a subset of them by indexing. This works just like arrays, with
individual elements and slices.</p>
<pre><code> print *([1])
</code></pre>
<p>This selects a single file, the first in alphabetic order since we
haven't changed the default ordering.</p>
<pre><code> print *(om[1,5])
</code></pre>
<p>This selects the five most recently modified files (or all files, if
there are five or fewer). Negative indices are understood, too:</p>
<pre><code> print *(om[1,-2])
</code></pre>
<p>selects all files but the oldest, assuming there are at least two.</p>
<p>Finally, a reminder that you can stick modifiers after qualifiers, or
indeed in parentheses without any qualifiers:</p>
<pre><code> print **/*(On:t)
</code></pre>
<p>sorts files in subdirectories into reverse order of name, but then
strips off the directory part of that name. Modifiers are applied right
at the end, after all file selection tasks.</p>
<p><strong>Evaluating code as a test</strong></p>
<p>The most complicated effect is produced by the <code>(e)</code> qualifer. which is
followed by a string delimited in the now-familiar way by either
matching brackets of any of the four sorts or a pair of any other
characters. The string is evaluated as shell code; another layer of
quotes is stripped off, to make it easier to quote the code from
immediate expansion. The expression is evaulated separately for each
match found by the other parts of the pattern, with the parameter
<code>$REPLY</code> set to the filename found.</p>
<p>There are two ways to use <code>(e)</code>. First, you can simply rely on the
return code. So:</p>
<pre><code>
print *(e:'[[ -d $REPLY ]]':)
print *(/)
</code></pre>
<p>are equivalent. Note that quotes around the expression, which are
necessary in addition to the delimiters (here `<code>:</code>') for expressions
with special characters or whitespace. In particular, <code>$REPLY</code> would
have been evaluated too early --- before file generation took place ---
if it hadn't been quoted.</p>
<p>Secondly, the function can alter the value of <code>$REPLY</code> to alter the name
of the file. What's more, the expression can set <code>$reply</code> (which
overrides the use of <code>$REPLY</code>) to an array of files to be inserted into
the command line; it may be any size from zero items upward.</p>
<p>Here's the example in the manual:</p>
<pre><code> print *(e:'reply=(${REPLY}{1,2})':)
</code></pre>
<p>Note the string is delimited by colons <em>and</em> quoted. This takes each
file in the current directory, and for each returns a match which has
two entires, the filename with `<code>1</code>' appended and the filename with
`<code>2</code>' appended.</p>
<p>For anything more complicated than this, you should write a shell
function to use <code>$REPLY</code> and set that or <code>$reply</code>. Then you can replace
the whole expression in quotes with that name.</p>
<p><span id="l142"></span></p>
<h3 id="597-globbing-flags-alter-the-behaviour-of-matches"><a class="header" href="#597-globbing-flags-alter-the-behaviour-of-matches">5.9.7: Globbing flags: alter the behaviour of matches</a></h3>
<p>Another <code>EXTENDED_GLOB</code> features is `globbing flags'. These are a bit
like the flags that can appear in perl regular expressions; instead of
making an assertion about the type of the resulting match, like glob
qualifiers do, they affect the way the match is performed. Thus they are
available for all uses of pattern matching --- though some flags are not
particularly useful with filename generation.</p>
<p>The syntax is borrowed from perl, although it's not the same: it looks
like `<code>(#X)</code>', where <code>X</code> is a letter, possibily followed by an argument
(currently only a number and only if the letter is `<code>a</code>'). Perl
actually uses `<code>?</code>' instead of `<code>#</code>'; what these have in common is
that they can't appear as a valid pattern characters just after an open
parenthesis, since they apply to the pattern before. Zsh doesn't have
the rather technical flags that perl does (lookahead assertions and so
on); not surprisingly, its features are based around the shortcuts often
required by shell users.</p>
<p><strong>Mixed-case matches</strong></p>
<p>The simplest sort of globbing flag will serve as an example. You can
make a pattern, or a portion of a pattern, match case-insensitively with
the flag <code>(#i)</code>:</p>
<pre><code> [[ FOO = foo ]]
[[ FOO = (#i)foo ]]
</code></pre>
<p>Assuming you have <code>EXTENDED_GLOB</code> set so that the `<code>#</code>' is an active
pattern character, the first match fails while the second succeeds. I
mentioned portions of a pattern. You can put the flags at any point in
the pattern, and they last to the end either of the pattern or any
enclosing set of parentheses, so in</p>
<pre><code> [[ FOO = f(#i)oo ]]
[[ FOO = F(#i)oo ]]
</code></pre>
<p>once more the first match fails and the second succeeds. Alternatively,
you can put them in parentheses to limit their scope:</p>
<pre><code> [[ FOO = ((#i)fo)o ]]
[[ FOO = ((#i)fo)O ]]
</code></pre>
<p>gives a failure then a success again. Note that you need extra
parentheses; the ones around the flag just delimit that, and have no
grouping effect. This is different from Perl.</p>
<p>There are two flags which work in exactly the same way: <code>(#l)</code> says that
only lowercase letters in the pattern match case-insensitively;
uppercase letters in the pattern only match uppercase letters in the
test string. This is a little like Emacs' behaviour when searching case
insensitvely with the <code>case-fold-search</code> option variable set; if you
type an uppercase character, it will look only for an uppercase
character. However, Emacs has the additional feature that from that
point on the whole string becomes case-sensitive; zsh doesn't do that,
the flag applies strictly character by character.</p>
<p>The third flag is <code>(#I)</code>, which turns case-insensitive matching off from
that point on. You won't often need this, and you can get the same
effect with grouping --- unless you are applying the case-insensitive
flag to multiple directories, since groups can't span more than one
directory. So</p>
<pre><code> print (#i)/a*/b*/(#I)c*
</code></pre>
<p>is equivalent to</p>
<pre><code> print /[aA]*/[bB]*/c*
</code></pre>
<p>Note that case-insensitive searching only applies to characters not in a
special pattern of some sort. In particular, ranges are not
automatically made case-insensitive; instead of `<code>(#i)[ab]*</code>', you must
use `<code>[abAB]*</code>'. This may be unexpected, but it's consistent with how
other flags, notably approximation, work.</p>
<p>You should be careful with matching multiple directories
case-insensitively. First,</p>
<pre><code> print (#i)~/.Z*
</code></pre>
<p>doesn't work. This is due to the order of expansions: filename expansion
of the tilde happens before pattern matching is ever attempted, and the
`<code>~</code>' isn't at the start where filename expansion needs to find it.
It's interpreted as an empty string which doesn't match `<code>/.Z*</code>',
case-insensitively --- in other words, it will match any empty string.</p>
<p>Hence you should put `<code>(#i)</code>' and any other globbing flags after the
first slash --- unless, for some reason, you <em>really</em> want the
expression to match `<code>/Home/PWS/</code>' etc. as well as `<code>/home/pws</code>'.</p>
<p>Second,</p>
<pre><code> print (#i)$HOME/.Z*
</code></pre>
<p>does work --- prints all files beginning `<code>.Z</code>' or `<code>.z</code>' in your home
directory --- but is inefficient. Assume <code>$HOME</code> expands to my home
directory, <code>/home/pws</code>. Then you are telling the shell it can match in
the directories `<code>/Home/PWS/</code>', `<code>/HOME/pWs</code>' and so on. There's no
quick way of doing this --- the shell has to look at every single entry
first in `<code>/</code>' and then in `<code>/home</code>' (assuming that's the only match
at that level) to check for matches. In summary, it's a good idea to use
the fact that the flag doesn't have to be at the beginning, and write
this as:</p>
<pre><code> print ~/(#i).Z*
</code></pre>
<p>Of course,</p>
<pre><code> print ~/.[zZ]*
</code></pre>
<p>would be easier and more standard in this oversimplified example.</p>
<p>On <code>Cygwin</code>, a UNIX-like layer running on top of, uh, a well known
graphical user interface claiming to be an operating system, filenames
are usually case insensitive anyway. Unfortunately, while Cygwin itself
is wise to this fact, zsh isn't, so it will do all that extra searching
when you give it the <code>(#i)</code> flag with an otherwise explicit string.</p>
<p>A piece of good news, however, is that matching of uppercase and
lowercase characters will handle non-ASCII character sets, provided your
system handles locales, (or to use the standard hieroglyphics, `i18n'
--- count the letters between `i' and `n' in `internationalization',
which may not even be a word anyway, and wince). In that case you or
your system administrator or the shell environment supplied by your
operating system vendor needs to set <code>$LC_ALL</code> or <code>$LC_CTYPE</code> to the
appropriate locale -- C for the default, <code>en</code> for English, <code>uk</code> for
Ukrainian (which I remember because it's confusing in the United
Kingdom), and so on.</p>
<p><strong>`Backreferences'</strong></p>
<p>The feature labelled as `backreferences' in the manual isn't really
that at all, which is my fault. Many regular expression matchers allow
you to refer back to bits already matched. For example, in Perl the
regular expression `<code>([A-Z]{3})$1</code>' says `match three uppercase
characters followed by the same three characters again. The `<code>$1</code>' is a
backreference.</p>
<p>Zsh has a similar feature, but in fact you can't use it while matching a
single pattern; it just makes the characters matched by parentheses
available after a successful complete match. In this, it's a bit more
like Emacs's <code>match-beginning</code> and <code>match-end</code> functions.</p>
<p>You have to turn it on for each pattern with the globbing flag
`<code>(#b)</code>'. The reason for this is that it makes matches involving
parentheses a bit slower, and most of the time you use parentheses just
for ordinary filename generation where this feature isn't useful. Like
most of the other globbing flags, it can have a local effect: only
parentheses after the flag produce backreferences, and the effect is
local to enclosing parentheses (which don't feel the effect themselves).
You can also turn it off with `<code>(#B)</code>'.</p>
<p>What happens when a pattern with active parentheses matches is that the
elements of the array <code>$match</code>, <code>$mbegin</code> and <code>$mend</code> are set to reflect
each active parenthesis in turn --- names inspired by the corresponding
Emacs feature. The string matching the first pair of parentheses is
stored in the first element of <code>$match</code>, its start position in the
string is stored in the first element of <code>$mbegin</code>, and its end position
in the string <code>$mend</code>. The same happens for later matched parentheses.
The parentheses around any globbing flags do not count.</p>
<p><code>$mbegin</code> and <code>$mend</code> use the indexing convention currently in effect,
i.e. zero offset if <code>KSH_ARRAYS</code> is set, unit offset otherwise. This
means that if the string matched against is stored in the parameter
<code>$teststring</code>, then it will always be true that <code>${match[1]}</code> is the
same string as <code>${teststring[${mbegin[1]},${mend[1]}]}</code>. and so on. (I'm
assuming, as usual, that <code>KSH_ARRAYS</code> isn't set.) Unfortunately, this is
different from the way the <code>E</code> parameter flag works --- that substitutes
the character after the end of the matched substring. Sorry! It's my
fault for not following that older convention; I thought the string
subscripting convention was more relevant.</p>
<p>An obvious use for this is to match directory and non-directory parts of
a filename:</p>
<pre><code> local match mbegin mend
if [[ /a/file/name = (#b)(*)/([^/]##) ]]; then
print -l ${match[1]} ${match[2]}
fi
</code></pre>
<p>prints `<code>/a/file</code>' and `<code>name</code>'. The second parenthesis matches a
slash followed by any number of characters, but at least one, which are
not slashes, while the first matches anything --- remember slashes
aren't special in a pattern match of this form. Note that if this
appears in a function, it is a good idea to make the three parameters
local. You don't have to clear them, or even make them arrays. If the
match fails, they won't be touched.</p>
<p>There's a slightly simpler way of getting information about the match:
the flag <code>(#m)</code> puts the matched string, the start index, and the index
for the <em>whole</em> match into the scalars <code>$MATCH</code>, <code>$MBEGIN</code> and <code>$MEND</code>.
It may not be all that obvious why this is useful. Surely the whole
pattern always matches the whole string? Actually, you've already seen
cases where this isn't true for parameter substitutions:</p>
<pre><code> local MATCH MBEGIN MEND string
string=aLOha
: ${(S)string##(#m)([A-Z]##)}
</code></pre>
<p>You'll find this sets <code>$MATCH</code> to <code>LO</code>, <code>$MBEGIN</code> to 2 and <code>$MEND</code> to 3.
In the parameter expansion, the <code>(S)</code> is for matching substrings, so
that the `<code>##</code>' match isn't anchored to the start of <code>$string</code>. The
pattern is <code>(#m)([A-Z]##)</code>, which means: turn on full-match
backreferencing and match any number of capital letters, but at least
one. This matches <code>LO</code>. Then the match parameters let you see where in
the test parameter the match occurred.</p>
<p>There's nothing to stop you using both these types of backreferences at
once, and you can specify multiple globbing flags in the short form
`<code>(#bm)</code>'. This will work with any combination of flags, except that
some such as `<code>(#bB)</code>' are obviously silly.</p>
<p>Because ordinary globbing produces a list of files, rather than just
one, this feature isn't very useful and is turned off. However, it <em>is</em>
possible to use backreferences in global substitutions and substitutions
on arrays; here are both at once:</p>
<pre><code> % array=(mananan then in gone June)
% print ${array//(#m)?n/${(C)MATCH[1]}n}
mAnAnAn thEn In gOne JUne
</code></pre>
<p>The substitution occurs separately on each element of the array, and at
each match in each element <code>$MATCH</code> gets set to what was matched. We use
this to capitalize every character that is followed by a lowercase
`<code>n</code>'. This will work with the <code>(#b)</code> form, too. The perl equivalent of
this is:</p>
<pre><code> % perl -pe 's/.n/\u$&/g' <<<$array
mAnAnAn thEn In gOne JUne
</code></pre>
<p>(People sometimes say Perl has a difficult syntax to understand; I hope
I'm convincing you how naive that view is when you have zsh.)</p>
<p>Now I can convince you of one point I made about excluded matches above:</p>
<pre><code> % [[ foo = (#b)(^foo)* ]] && print $match
fo
</code></pre>
<p>As claimed, the process of making the longest possible match, then
backtracking from the end until the whole thing is successful, leads to
the `<code>(^foo)</code>' matching `<code>fo</code>'.</p>
<p><strong>Approximate matching</strong></p>
<p>To my knowledge, zsh is the first command line interpreter to make use
of approximate matching. This is very useful because it provides the
shell with an easy way of correcting what you've typed. First, some
basics about what I mean by `approximate matching'.</p>
<p>There are four ways you can make a mistake in typing. You can leave out
a letter which should be there; you can insert a letter which shouldn't;
you can type one letter instead of another; and you can transpose two
letters. The last one involves two different characters, so some systems
for making approximate matches count it as two different errors; but
it's a particularly common one when typing, and quite useful to be able
to handle as a single error. I know people who even have `<code>mkae</code>'
aliased to `<code>make</code>'.</p>
<p>You can tell zsh how many errors you are willing to allow in a pattern
match by using, for example <code>(#a1)</code>, which says only a single error
allowed. The rules for the flag are almost identical to those for
case-insensitive matching, in particular for scoping and the way
approximate matching is handled for a filename expansion with more than
one directory. The number of errors is global; if the shell manages to
match a directory in a path with an error, one fewer error is allowed
for the rest of the path. You can specify as many errors as you like;
the practical limit is that with too many allowed errors the pattern
will match far too many strings. The shell doesn't have a particularly
nifty way of handling approximate matching (unlike, for example, the
program <code>agrep</code>), but you are unlikely to encounter problems if the
number of matches stays in a useful range.</p>
<p>The fact that the error count applies to the whole of a filename path is
a bit of a headache, actually, because we have to make sure the shell
matches each directory with the minimum number of errors. With a single
pattern, the shell doesn't care as long as it doesn't use up all the
errors it has, while with multiple directories if it uses up the errors
early on, it may fail to match something it should match. But you don't
have to worry about that; this explanation is just to elicit sympathy.</p>
<p>So the pattern <code>(#a1)README</code> will match <code>README</code>, <code>READ.ME</code>, <code>READ_ME</code>,
<code>LEADME</code>, <code>REDME</code>, <code>READEM</code>, and so on. It will not match <code>_README_</code>,
<code>ReadMe</code>, <code>READ</code> or <code>AAREADME</code>. However, you can combine it with
case-insensitivity, for which the short form <code>(#ia1)README</code> is allowed,
and then it will match <code>ReadMe</code>, <code>Read.Me</code>, <code>read_me</code>, and so on. You
can consider filenames with multiple directories as single strings for
this purpose --- with one exception, that `<code>foo/bar</code>' and `<code>fo/obar</code>'
are two errors apart, not one. Because the errors are counted separately
in each directory, you can't transpose the `<code>/</code>' with another
character. This restriction doesn't apply in other forms of pattern
matching where <code>/</code> is not a special character.</p>
<p>Another common feature with case-insensitive matching is that only the
literal parts of the string are handled. So if you have `<code>[0-9]</code>' in a
pattern, that character must match a decimal digit even if approximation
is active. This is often useful to impose a particular form at key
points. The main difficulty, as with the `<code>/</code>' in a directory, is that
transposing with another character is not allowed, either. In other
words, `<code>(#a1)ab[0-9]</code>' will fail to match `<code>a1b</code>'; it will match with
two errors, by removing the `<code>b</code>' before the digit and inserting it
after.</p>
<p>As an example of what you can do with this feature, here is a simple
function to correct misspelled filenames.</p>
<pre><code> emulate -LR zsh
setopt extendedglob
local file trylist
integer approx max_approx=6
file=$1
if [[ -e $file ]]; then
# no correction necessary
print $file
return
fi
for (( approx = 1; approx <= max_approx; approx++ )); do
trylist=( (#a$approx)"$file"(N) )
(( $#trylist )) && break
done
(( $#trylist )) || return 1
print $trylist
</code></pre>
<p>The function tries to match a file with the minimum possible number of
errors, but in any case no more than 6. As soon as it finds a match, it
will print it out and exit. It's still possible there is more than one
match with that many errors, however, and in this case the complete list
is printed. The function doesn't handle `<code>~</code>' in the filename.</p>
<p>It does illustrate the fact that you can specify the number of
approximations as a parameter. This is purely a consequence of the fact
that filename generation happens right at the end of the expansion
sequence, after the parameters have already been substituted away. The
numbers and the letter in the globbing flag aren't special characters,
unlike the parentheses and the `<code>#</code>'; if you wanted those to be special
when substituted from a parameter, you would need to set the <code>KSH_GLOB</code>
flag, possibly by using the `<code>~</code>' parameter flag.</p>
<p>A more complicated version of that function is included with the shell
distribution in the file <code>Completion/Base/Widget/_correct_filename</code>.
This is designed to be used either on its own, or as part of the
completion system.</p>
<p>Indeed, the completion system described in the next chapter is where you
are most likely to come across approximate matching, buried inside
approximate completion and correction --- in the first case, you tell
the shell to complete what you have typed, trying to correct mistakes,
and in the second case, you tell the shell that you have finished typing
but possibly made some mistakes which it should correct. If you already
have the new completion system loaded, you can use <code>^Xc</code> to correct a
word on the command line; this is context-sensitive, so more
sophisticated than the function I showed.</p>
<p><strong>Anchors</strong></p>
<p>The last two globbing flags are probably the least used. They are there
really for completeness. They are <code>(#s)</code>, to match only at the start of
a string, and <code>(#e)</code>, to match only at the end. Unlike the other flags
they are purely local, just making a statement about the point where
they occur in the pattern.</p>
<p>They correspond to the much more commonly used `<code>^</code>' and `<code>$</code>' in
regular expressions. The difference is that shell patterns nearly always
match a complete string, so telling the pattern that a certain point is
the start or end isn't usually very useful. There are two occasions when
it is. The first is when the start or end is to be matched as an
alternative to something else. For example,</p>
<pre><code> [[ $file = *((#s)|/)dirpart((#e)|/)* ]]
</code></pre>
<p>succeeds if <code>dirpart</code> is a complete path segment of <code>$file</code> --- with a
slash or nothing at all before and after it. Remember, once again, that
slashes aren't special in pattern matches unless they're performing
filename generation. The effect of these two flags isn't altered at all
by their being inside another set of parentheses.</p>
<p>The second time these are useful is in parameter matches where the
pattern is not guaranteed to match a complete string. If you use <code>(#s)</code>
or <code>(#e)</code>, it will force that point to be the start or end despite the
operator in use. So <code>${</code><em>param</em><code>##</code><em>pattern</em><code>(#e)}</code> will remove
<em>pattern</em> from <code>$</code><em>param</em> only if it matches the entire string: the <code>##</code>
must match at the head, while the <code>(#e)</code> must match at the end.</p>
<p>You can get the effect with <code>${</code><em>param</em><code>:#</code><em>pattern</em><code>}</code>, and further
more this is rather faster. The <code>:#</code> operator has some global knowledge
about how to match; it knows that since <em>pattern</em> will match as far as
it can along the test string, it only needs to try the match once.
However, since `<code>##</code>' just needs to match at the head of the string, it
will backtrack along the pattern, trying to match <em>pattern</em><code>(#e)</code>,
entirely heedless of the fact that the pattern itself specifically won't
match if it doesn't extend to the end. So it's more efficient to use the
special parameter operators whenever they're available.</p>
<p><span id="l143"></span></p>
<h3 id="598-the-function-zmv"><a class="header" href="#598-the-function-zmv">5.9.8: The function <code>zmv</code></a></h3>
<p>The shell is supplied with a function <code>zmv</code>, which may have been
installed into the default <code>$fpath</code>, or can be found in the source tree
in the directory <code>Functions/Misc</code>. This provides a way of renaming,
copying and linking files based on patterns. The idea is that you give
two arguments, a pattern to match, and a string which uses that pattern.
The pattern to match has backreferences turned on; these are stored in
the positional parameters to make them easy to refer to. The function
tries to be safe: any file whose name is not changed is simply ignored,
and usually overwriting an existing file is an error, too. However, it
doesn't make sure that there is a one to one mapping from source to
target files; it doesn't know if the target file is supposed to be a
directory (though it could be smarter about that).</p>
<p>In the examples, I will use the option <code>-n</code>, which forces <code>zmv</code> to print
out what it will do without actually doing it. This is a good thing to
try first if you are unsure.</p>
<p>Here's a simple example.</p>
<pre><code> % ls
foo
% zmv -n '(*)' '${(U)1}'
mv -- foo FOO
</code></pre>
<p>The pattern matches anything in the current directory, excluding files
beginning with a `<code>.</code>' (the function starts with an `<code>emulate</code>', so
<code>GLOB_DOTS</code> is forced to be off). The complete string is stored as the
first backreference, which is in turn put into <code>$1</code>. Then the second
argument is used and <code>$1</code> in uppercase is substituted.</p>
<p><strong>Essentials of the function</strong></p>
<p>The basic code in <code>zmv</code> is very simple. It boils down to more or less
the following.</p>
<pre><code> setopt nobareglobqual extendedglob
local files pattern result f1 f2 match mbegin mend
pattern=$1
result=$2
for f1 in ${~pattern}; do
[[ $f1 = (#b)${~pattern} ]] || continue
set -- $match
f2=${(e)result}
mv -- $f1 $f2
done
</code></pre>
<p>Here's what's going on. We store the arguments as <code>$pattern</code> and
<code>$result</code>. We then expand the pattern to a list of files --- remember
that <code>${~pattern}</code> makes the characters in <code>$pattern</code> active for the
purposes of globbing. For each file we find, we match against the
pattern again, but this time with backreferences turned on, so that
parentheses are expanded into the array <code>$match</code>. If, for some reason,
the pattern match failed this time, we just skip the file. Then we store
<code>$match</code> in the positional parameters; the `<code>-``-</code>' for <code>set</code> and for
<code>mv</code> is in case <code>$match[1]</code> begins with a `<code>-</code>'.</p>
<p>Then we evaluate the result, assuming that it will refer to the
positional parameters. In our example, <code>$result</code> contains argument
`<code>${(U)1}</code>' and if we matched `<code>foo</code>', then <code>$1</code> contains foo. The
effect of `<code>${(e)result}</code>' is to perform an extra substitution on the
<code>${(U)1}</code>, so <code>$f2</code> will be set to <code>FOO</code>. Finally, we use the <code>mv</code>
command to do the actual renaming. The effect of the <code>-n</code> option isn't
shown, but it's essentially to put a `<code>print</code>' in front of the <code>mv</code>
command line.</p>
<p>Notice I set <code>nobareglobqual</code>, turning off the use of glob qualifiers.
That's necessary because of all those parentheses; otherwise, `<code>(*)</code>'
would have been interpreted as a qualifier. There is an option, <code>-Q</code>,
which will turn qualifiers back on, if you need them. That's still not
quite ideal, since the second pattern match, the one where we actually
use the backreferences, isn't filename generation, just a test against a
string, so doesn't handle glob qualifers. So in that case the code needs
to strip qualifiers off. It does this by a fairly simple pattern match
which will work in simple cases, though you can confuse it if you try
hard enough, particularly if you have extra parentheses in the glob
qualifier.</p>
<p>Note also the use of `<code>${(e)result}</code>' to force substitution of
parameters when <code>$result</code> is evaluated. This way of doing it safely
ignores other metacharacters which may be around: all <code>$</code>-expansions,
plus backquote expansion, are performed, but otherwise <code>$result</code> is left
alone.</p>
<p><strong>More complicated examples</strong></p>
<p><code>zmv</code> has some special handling for recursive globbing, but only with
the patterns <code>**/</code> and <code>***/</code>. If you put either of these in parentheses
in the pattern, they will be spotted and used in the normal way. Hence,</p>
<pre><code> % ls test
lonely
% zmv -n '(**/)lonely' '$1solitary'
mv -- test/lonely test/solitary
</code></pre>
<p>Note that, as with other uses of `<code>**/</code>', the slash is part of the
recursive match, so you don't need another one. You don't need to
separate <code>$1</code> from <code>solitary</code> either, since positional parameters are a
special case, but you could use `<code>${1}solitary</code>' for the look of it.
Like glob qualifiers, recursive matches are handled by some magic in the
function; in ordinary globbing you can't put these two forms inside
parentheses.</p>
<p>For the lazy, the option <code>-w</code> (which means `with wildcards') will tell
<code>zmv</code> to decide for itself where all the patterns are and automatically
add parentheses. The two examples so far become</p>
<pre><code> zmv -nw '*' '${(U)1}'
zmv -nw '***/lonely' '$1solitary'
</code></pre>
<p>with exactly the same effects.</p>
<p>Another way of getting useful effects is to use the `<code>${1//foo/bar}</code>'
substitution in the second argument. This gives you a general way of
substitution bits in filenames. Often, you can then get away with having
`<code>(*)</code>' as the first argument:</p>
<pre><code> zmv '(*)' '${1//(#m)[aeiou]/${(U)MATCH}}'
</code></pre>
<p>capitalises all the vowels in all filenames in the current directory.
You may be familiar with a perl script called <code>rename</code> which does tricks
like this (though there's another, less powerful, programme of the same
name which simply replaces strings).</p>
<p><strong>The effect of <code>zmv</code></strong></p>
<p>In addition to renaming, <code>zmv</code> can be made to copy or link files. If you
call it <code>zcp</code> or <code>zln</code> instead of <code>zmv</code>, it will have those effects, and
in the case of <code>zln</code> you can use the option <code>-s</code> to create symbolic
links, just as with <code>ln</code>. Beware the slightly confusing behaviour of
symbolic links containing relative directories, however.</p>
<p>Alternatively, you can force the behavour of <code>zmv</code>, <code>zcp</code> and <code>zln</code> just
by giving the options <code>-M</code>, <code>-C</code> or <code>-L</code> to the function, whatever it is
called. Or you can use `<code>-p</code> <em>prog</em>' to execute <code>prog</code> instead of <code>mv</code>,
<code>cp</code> or <code>ln</code>; <em>prog</em> should be able to be run as `<em>prog</em> <code>-``-</code>
<em>oldname</em> <em>newname</em>', whatever it does.</p>
<p>The option <code>-i</code> works a bit like the same option to the basic programmes
which <code>zmv</code> usually calls, prompting you before any action --- in this
case, not just overwriting, but any action at all. Likewise, <code>-f</code> tells
<code>zmv</code> to force overwriting of files, which it will usually refuse to do
because of the potential dangers. Although many versions of <code>mv</code> etc.
take this option, some don't, so it's not passed down; instead there's a
generic way of passing down options to the programmes executed, using
<code>-o</code> followed by a string. For example,</p>
<pre><code> % ls
foo
% zmv -np frud -o'-a -b' '(*)' '${(U)1}'
frud -a -b -- foo FOO
</code></pre>
<div id="chapter_begin" style="break-before: page; page-break-before: always;"></div><!-- START doctoc generated TOC please keep comment here to allow auto update -->
<!-- DON'T EDIT THIS SECTION, INSTEAD RE-RUN doctoc TO UPDATE -->
<p><strong>Table of Contents</strong> <em>generated with <a href="https://github.com/thlorenz/doctoc">DocToc</a></em></p>
<ul>
<li><a href="zshguide06.html#chapter-6-completion-old-and-new">Chapter 6: Completion, old and new</a>
<ul>
<li><a href="zshguide06.html#61-completion-and-expansion">6.1: Completion and expansion</a></li>
<li><a href="zshguide06.html#62-configuring-completion-using-shell-options">6.2: Configuring completion using shell options</a>
<ul>
<li><a href="zshguide06.html#621-ambiguous-completions">6.2.1: Ambiguous completions</a></li>
<li><a href="zshguide06.html#622-always_last_prompt">6.2.2: <code>ALWAYS_LAST_PROMPT</code></a></li>
<li><a href="zshguide06.html#623-menu-completion-and-menu-selection">6.2.3: Menu completion and menu selection</a></li>
<li><a href="zshguide06.html#624-other-ways-of-changing-completion-behaviour">6.2.4: Other ways of changing completion behaviour</a></li>
<li><a href="zshguide06.html#625-changing-the-way-completions-are-displayed">6.2.5: Changing the way completions are displayed</a></li>
</ul>
</li>
<li><a href="zshguide06.html#63-getting-started-with-new-completion">6.3: Getting started with new completion</a></li>
<li><a href="zshguide06.html#64-how-the-shell-finds-the-right-completions">6.4: How the shell finds the right completions</a>
<ul>
<li><a href="zshguide06.html#641-contexts">6.4.1: Contexts</a></li>
<li><a href="zshguide06.html#642-tags">6.4.2: Tags</a></li>
</ul>
</li>
<li><a href="zshguide06.html#65-configuring-completion-using-styles">6.5: Configuring completion using styles</a>
<ul>
<li><a href="zshguide06.html#651-specifying-completers-and-their-options">6.5.1: Specifying completers and their options</a></li>
<li><a href="zshguide06.html#652-changing-the-format-of-listings-groups-etc">6.5.2: Changing the format of listings: groups etc.</a></li>
<li><a href="zshguide06.html#653-styles-affecting-particular-completions">6.5.3: Styles affecting particular completions</a></li>
</ul>
</li>
<li><a href="zshguide06.html#66-command-widgets">6.6: Command widgets</a>
<ul>
<li><a href="zshguide06.html#661-_complete_help">6.6.1: <code>_complete_help</code></a></li>
<li><a href="zshguide06.html#662-_correct_word-_correct_filename-_expand_word">6.6.2: <code>_correct_word</code>, <code>_correct_filename</code>, <code>_expand_word</code></a></li>
<li><a href="zshguide06.html#663-_history_complete_word">6.6.3: <code>_history_complete_word</code></a></li>
<li><a href="zshguide06.html#664-_most_recent_file">6.6.4: <code>_most_recent_file</code></a></li>
<li><a href="zshguide06.html#665-_next_tags">6.6.5: <code>_next_tags</code></a></li>
<li><a href="zshguide06.html#666-_bash_completions">6.6.6: <code>_bash_completions</code></a></li>
<li><a href="zshguide06.html#667-_read_comp">6.6.7: <code>_read_comp</code></a></li>
<li><a href="zshguide06.html#668-_generic">6.6.8: <code>_generic</code></a></li>
<li><a href="zshguide06.html#669-predict-on-incremental-complete-word">6.6.9: <code>predict-on</code>, <code>incremental-complete-word</code></a></li>
</ul>
</li>
<li><a href="zshguide06.html#67-matching-control-and-controlling-where-things-are-inserted">6.7: Matching control and controlling where things are inserted</a>
<ul>
<li><a href="zshguide06.html#671-case-insensitive-matching">6.7.1: Case-insensitive matching</a></li>
<li><a href="zshguide06.html#672-matching-option-names">6.7.2: Matching option names</a></li>
<li><a href="zshguide06.html#673-partial-word-completion">6.7.3: Partial word completion</a></li>
<li><a href="zshguide06.html#674-substring-completion">6.7.4: Substring completion</a></li>
<li><a href="zshguide06.html#675-partial-words-with-capitals">6.7.5: Partial words with capitals</a></li>
<li><a href="zshguide06.html#676-final-notes">6.7.6: Final notes</a></li>
</ul>
</li>
<li><a href="zshguide06.html#68-tutorial">6.8: Tutorial</a>
<ul>
<li><a href="zshguide06.html#681-the-dispatcher">6.8.1: The dispatcher</a></li>
<li><a href="zshguide06.html#682-subcommand-completion-_arguments">6.8.2: Subcommand completion: <code>_arguments</code></a></li>
<li><a href="zshguide06.html#683-completing-particular-argument-types">6.8.3: Completing particular argument types</a></li>
<li><a href="zshguide06.html#684-the-rest">6.8.4: The rest</a></li>
</ul>
</li>
<li><a href="zshguide06.html#69-writing-new-completion-functions-and-widgets">6.9: Writing new completion functions and widgets</a>
<ul>
<li><a href="zshguide06.html#691-loading-completion-functions-compdef">6.9.1: Loading completion functions: <code>compdef</code></a></li>
<li><a href="zshguide06.html#692-adding-a-set-of-completions-compadd">6.9.2: Adding a set of completions: <code>compadd</code></a></li>
<li><a href="zshguide06.html#693-functions-for-generating-filenames-etc">6.9.3: Functions for generating filenames, etc.</a></li>
<li><a href="zshguide06.html#694-the-zshparameter-module">6.9.4: The <code>zsh/parameter</code> module</a></li>
<li><a href="zshguide06.html#695-special-completion-parameters-and-compset">6.9.5: Special completion parameters and <code>compset</code></a></li>
<li><a href="zshguide06.html#696-fancier-completion-using-the-tags-and-styles-mechanism">6.9.6: Fancier completion: using the tags and styles mechanism</a></li>
<li><a href="zshguide06.html#697-getting-the-work-done-for-you-handling-arguments-etc">6.9.7: Getting the work done for you: handling arguments etc.</a></li>
<li><a href="zshguide06.html#698-more-completion-utility-functions">6.9.8: More completion utility functions</a></li>
</ul>
</li>
<li><a href="zshguide06.html#610-finally">6.10: Finally</a></li>
</ul>
</li>
</ul>
<!-- END doctoc generated TOC please keep comment here to allow auto update -->
<p><span id="comp"></span><span id="l144"></span></p>
<h1 id="chapter-6-completion-old-and-new"><a class="header" href="#chapter-6-completion-old-and-new">Chapter 6: Completion, old and new</a></h1>
<p>Completion of command arguments is something zsh is particularly good
at. The simplest case is that you hit <code><TAB></code>, and the shell guesses
what has to go there and fills it in for you:</p>
<pre><code> % ls
myfile theirfile yourfile
% cat t<TAB>
</code></pre>
<p>expands the command line to</p>
<pre><code> % cat theirfile
</code></pre>
<p>and you only had to type the initial letter, then <code>TAB</code>.</p>
<p>In the early days when this feature appeared in the C shell, only
filenames could be completed; there were no clever tricks to help you if
the name was ambiguous, it simply printed the unambiguous part and
beeped so that you had to decide what to do next. You could also list
possible completions; for some reason this became attached to the <code>^D</code>
key in csh, which in later shells with Emacs-like bindings also deletes
the next character, so that history has endowed zsh, like other shells,
with the slightly odd combined behaviour:</p>
<pre><code> % cat yx
</code></pre>
<p>Now move the cursor back one character onto the x and hit ^D twice and
you see: <code>yourfile</code>. That doesn't work if you use vi-like bindings, or,
obviously, if you've rebound <code>^D</code>.</p>
<p>Next, it became possible to complete other items such as names of users,
commands or hosts. Then zsh weighed in with menu completion, so you
could keep on blindly hitting <code><TAB></code> until the right answer appeared,
and never had to type an extra character yourself.</p>
<p>The next development was tcsh's, and then zsh's, programmable completion
system; you could give instructions to the shell that in certain
contexts, only certain items should be completed; for example, after
<code>cd</code>, you would only want directories. In tcsh, there was a command
called <code>complete</code>; each `<code>complete ...</code>' statement defined the
completion for the arguments of a particular command, such as <code>cd</code>; the
equivalent in zsh is <code>compctl</code>, which was inspired by <code>complete</code> but is
different in virtually every significant detail. There is a perl script
<code>lete2ctl</code> in the <code>Misc</code> directory of the shell distribution to help you
convert from the tcsh to the zsh formats. You put a whole series of
<code>compctl</code> commands into <code>.zshrc</code>, and everything else is done by the
shell.</p>
<p>Zsh's system has become more and more sophisticated, and in version
3.1.6 a new completion system appeared which is supposed to do
everything for you: you simply call a function, <code>compinit</code>, from an
initialization file, after which zsh knows, for example, that <code>gunzip</code>
should be followed by files ending in <code>.gz</code>. The new system is based on
shell functions, an added bonus since they are extremely flexible and
you already know the syntax. However, given the complexity it's quite
difficult to get started writing your own completions now, and hard
enough to know what to do to change the settings the way you like. The
rest of the chapter should help.</p>
<p>I shall concentrate on the new completion system, which seems destined
to take over completely from the old one eventually, now that the 3.1
release series has become the 4.0 production release. The old
<strong>compctl</strong> command is still available, and old completion definitions
will remain working in future versions of zsh --- in fact, on most
operating systems which support dynamically linked libraries the old
completion system is in a different file, which the shell loads when
necessary, so there's very little overhead for this.</p>
<p>The big difference in the new system is that, instead of everything
being set up once and for all when the shell starts, various bits of
shell code are called after you hit <code><TAB></code>, to generate the completions
there and then. There's enough new in the shell that all those
unmemorable options to <code>compctl</code> (`<code>-f</code>' for files `<code>-v</code>' for
variables and so on) can be replaced by commands that produce the list
of completions directly; the key command in this case is called
`<code>compadd</code>', which is passed this list and decides what to use to
complete the word on the command line. So the simplest possible form of
new completion looks roughly like this:</p>
<pre><code> # tell the shell that the function mycompletion can do completion
# when called by the widget name my-completion-widget, and that
# it behaves like the existing widget complete-word
zle -C my-completion-widget .complete-word mycompletion
# define a key that calls the completion widget
bindkey '^x^i' my-completion-widget
# define the function that will be called
mycompletion() {
# add a list of completions
compadd alpha bravo charlie delta
}
</code></pre>
<p>That's very roughly what the completion system is doing, except that the
function is called <code>_main_complete</code> and calls a lot of other functions
to do its dirty work based on the context where completion was called
(all the things that <code>compctl</code> used to do), and the widgets are just the
old completion widgets (`<code>expand-or-complete</code>' etc.) redefined and
still bound to all the original keys. But, in case you hadn't guessed,
there's more to it than that.</p>
<p>Here's a plan for the sections of this chapter.</p>
<ol>
<li>A broad description of completion and expansion, applying equally to
old and new completion.</li>
<li>How to configure completion using shell options. Most of this
section applies to old completion, too, although I won't explicitly
flag up any differences. After this, I shall leave the <code>compctl</code>
world behind.</li>
<li>How to start up new completion.</li>
<li>The basics of how the new completion system works.</li>
<li>How to configure it using the new `<code>zstyle</code>' builtin.</li>
<li>Separate commands which do something other than the usual completion
system, as well as some other editing widgets that have to do with
completion.</li>
<li>Matching control, a powerful way of deciding such things as whether
to complete case-insensitively, to allow insertion of extra parts of
words before punctuation characters, or to ignore certain characters
in the word on the command line.</li>
<li>How to write your own completion functions; you won't need to have
too solid an understanding of all the foregoing just to do simple
completions, but I will gradually introduce the full baroque
splendour of how to make tags and styles work in your own functions,
and how to make completion do the work of handling command arguments
and options.</li>
<li>Ends the chapter gracefully, on the old `beginning, middle, end'
principle.</li>
</ol>
<p><span id="l145"></span></p>
<h2 id="61-completion-and-expansion"><a class="header" href="#61-completion-and-expansion">6.1: Completion and expansion</a></h2>
<p>More things than just completion happen when you hit tab. The first
thing that zsh tries to do is expand the line. Expansion was covered in
a previous chapter: basically all the things described there are
possible candidates for expanding in-line by the editor. In other words,
history substitutions with bangs, the various expansions using `<code>$</code>' or
backquote, and filename generation (globbing) can all take place, with
the result replacing what was there on the command line:</p>
<pre><code> % echo $PWD<TAB>
-> echo /home/pws/zsh/projects/zshguide
% echo `print $ZSH_VERSION`<TAB>
-> echo 3.1.7
% echo !!<TAB>
-> echo echo 3.1.7
% echo ~/.z*<TAB>
-> echo /home/pws/.zcompdump /home/pws/.zlogout
/home/pws/.zshenv /home/pws/.zshrc
</code></pre>
<p>Note that the `<code>~</code>' also gets expanded in this case.</p>
<p>This is often a good time to remember the `undo' key, `<code>^_</code>' or
`<code>^Xu</code>'; typing this will restore what was there before the expansion
if you don't like the result. Many keyboards have a quirk that what's
described as `<code>^_</code>' should be typed as control with slash, which you'd
write `<code>^/</code>' except unfortunately that does something else; this is not
zsh's fault. There's another half-exception, namely filename generation:
paths like `<code>~/file</code>' don't get expanded, because you usually know what
they refer to and it's usually convenient to leave them for use in
completion. However, the `<code>=cmdname</code>' form does get expanded, unless
you have <code>NO_EQUALS</code> set.</p>
<p>In fact, deciding whether expansion or completion takes place can
sometimes be tricky, since things that would be expanded if they were
complete, may need to be completed first; for example <code>$PAT</code> should
probably be completed to <code>$PATH</code>, but it's quite possible there is a
parameter <code>$PAT</code> too. You can decide which, if you prefer. First, the
commands <code>expand-word</code>, bound to `<code>^X*</code>', and the corresponding command
for listing what would be expanded, <code>list-expand</code>, bound to `<code>^Xg</code>', do
expansion only --- all possible forms except alias expansion, including
turning `<code>~/file</code>' into a full path.</p>
<p>From the other point of view, you can use commands other than
<code>expand-or-complete</code>, the one bound by default to <code><TAB></code>, to perform
only completion. The basic command for this is <code>complete-word</code>, which is
not bound by default. It is quite sensible to bind this to `<code>^I</code>' (i.e.
<code><TAB></code>) if you are happy to use the separate commands for expansion,
i.e.</p>
<pre><code> # Now tab does only completion, not expansion
bindkey '^i' complete-word
</code></pre>
<p>Furthermore, if you do this and use the new completion system, then as
we shall see there is a way of making the completion system perform
expansion --- see the description of the <code>_expand</code> completer below. In
this case you have much more control over what forms of expansion are
tried, and at what point, but you have to make sure you use
<code>complete-word</code>, not <code>expand-or-complete</code>, else the standard expansion
system will take over.</p>
<p>There's a close relative of <code>expand-or-complete</code>,
<code>expand-or-complete-prefix</code>, not bound by default. The only difference
is that it will ignore everything under and to the right of the cursor
when completing. It's as if there was a space where the cursor was, with
everything to be ignored shifted to the right (guess how it's
implemented). Use this if you habitually type new words in the line
before other words, and expect them to complete or expand on their own
even before you've typed the space after them. Some other shells work
this way all the time. To be more explicit:</p>
<pre><code> % ls
filename1
% ls filex
</code></pre>
<p>Move the cursor to the <code>x</code> and hit tab. With <code>expand-or-complete</code>
nothing happens; it's trying to complete a file called `<code>filex</code>' ---
or, with the option <code>COMPLETE_IN_WORD</code> set, it's trying to find a file
whose name starts with `<code>file</code>' and ends with `<code>x</code>'. If you do</p>
<pre><code> bindkey '^i' expand-or-complete-prefix
</code></pre>
<p>and try the same experiment, you will find the whole thing is completed
to `<code>filename1x</code>', so that the `<code>x</code>' was ignored, but not removed.</p>
<p>One possible trap is that the listing commands, both
<code>delete-char-or-list</code>, bound by default to `<code>^D</code>' in emacs mode, and
<code>list-options</code>, bound by default to `<code>^D</code>' in vi insert mode and the
basic command for listing completions as it doesn't have the
delete-character behaviour, do not show possible expansions, so with the
default bindings you can use `<code>^D</code>' to list, then hit <code><TAB></code> and find
that the line has been completely rewritten by some expansion. Using
<code>complete-word</code> instead of <code>expand-or-complete</code> will of course fix this.
If you know how to write new editor widgets (<a href="zshguide04.html#zle">chapter
4</a>), you can make up a function which tries
<code>list-expand</code>, and if that fails tries <code>list-options</code>.</p>
<p>There are four completion commands I haven't mentioned yet: three are
<code>menu-complete</code>, <code>menu-expand-or-complete</code> and <code>reverse-menu-complete</code>,
which perform menu completion, where you can cycle through all possible
completions by hitting the same key. The first two correspond to
<code>complete-word</code> and <code>expand-or-complete</code> respectively, while the third
has no real equivalent as it takes you backwards through a completion
list. The effect of the third can't be reached just by setting options
for menu completion, so it's a useful one to bind separately. I have it
bound to `<code>\M-\C-i</code>', i.e. tab with the Meta key pressed down, but it's
not bound by default.</p>
<p>The fourth is <code>menu-select</code>, which performs an enhanced form of menu
completion called `menu selection' which I'll describe below when I
talk about options. You have to make sure the <code>zsh/complist</code> module is
loaded to use this zle command. If you use the style, zsh should be able
to load this automatically when needed, as long as you have dynamic
loading, which you probably do these days. <span id="l146"></span></p>
<h2 id="62-configuring-completion-using-shell-options"><a class="header" href="#62-configuring-completion-using-shell-options">6.2: Configuring completion using shell options</a></h2>
<p>There are two main ways of altering the behaviour of completion without
writing or rewriting shell functions: shell options, as introduced in
<a href="zshguide02.html#init">chapter 2</a>, and styles, as introduced above. I
shall first discuss the shell options, although as you will see some of
these refer to the styles mechanism. Setting shell options affects every
single completion, unless special care has been taken (using a
corresponding style for the context, or setting an option locally) to
avoid that.</p>
<p>In addition to the options which directly affect the completion system,
completion is sensitive to various other options which describe shell
behaviour. For example, if the option <code>MAGIC_EQUAL_SUBST</code> is set, so
that arguments of all commands looking like `<code>foo=~/file</code>' have the
`<code>~</code>' expanded as if it was at the start of an argument, then the
default completion for arguments of commands not specially handled will
try to complete filenames after the `<code>=</code>'.</p>
<p>Needless to say, if you write completion functions you will need to
worry about a lot of other options which can affect shell syntax. The
main starting point for completion chosen by context (everything except
the commands for particular completions bound separately to keystrokes)
is the function <code>_main_complete</code>, which includes the effect of the
following lines to make sure that at least the basic options are set up
within completion functions:</p>
<pre><code> setopt glob bareglobqual nullglob rcexpandparam extendedglob unset
unsetopt markdirs globsubst shwordsplit shglob ksharrays cshnullglob
unsetopt allexport aliases errexit octalzeroes
</code></pre>
<p>but that by no means exhausts the possibilities. Actually, it doesn't
include those lines: the options to set are stored in the array
<code>$_comp_options</code>, with <code>NO_</code> in front if they are to be turned off. You
can modify this if you find you need to (and maybe tell the maintainers,
too).</p>
<p>By the way, if you are wondering whether you can re-use the function
<code>_main_complete</code>, by binding it to a different key with slightly
different completion definitions, look instead at the description of the
<code>_generic</code> command widget below. It's just a front-end to
<code>_main_complete</code> which allows you to have a different set of styles in
effect.</p>
<p><span id="l147"></span></p>
<h3 id="621-ambiguous-completions"><a class="header" href="#621-ambiguous-completions">6.2.1: Ambiguous completions</a></h3>
<p>The largest group of options deals with what happens when a completion
is ambiguous, in other words there is more than one possible completion.
The seven relevant options are as follows, as copied from the FAQ; many
different combinations are possible:</p>
<ul>
<li>with <code>NO_BEEP</code> set, that annoying beep goes away,</li>
<li>with <code>NO_LIST_BEEP</code>, beeping is only turned off for ambiguous
completions,</li>
<li>with <code>AUTO_LIST</code> set, when the completion is ambiguous you get a
list without having to type <code>^D</code>,</li>
<li>with <code>BASH_AUTO_LIST</code> set, the list only happens the second time you
hit tab on an ambiguous completion,</li>
<li>with <code>LIST_AMBIGUOUS</code>, this is modified so that nothing is listed if
there is an unambiguous prefix or suffix to be inserted --- this can
be combined with <code>BASH_AUTO_LIST</code>, so that where both are applicable
you need to hit tab three times for a listing,</li>
<li>with <code>REC_EXACT</code>, if the string on the command line exactly matches
one of the possible completions, it is accepted, even if there is
another completion (i.e. that string with something else added) that
also matches,</li>
<li>with <code>MENU_COMPLETE</code> set, one completion is always inserted
completely, then when you hit TAB it changes to the next, and so on
until you get back to where you started,</li>
<li>with <code>AUTO_MENU</code>, you only get the menu behaviour when you hit TAB
again on the ambiguous completion.</li>
</ul>
<p><span id="l148"></span></p>
<h3 id="622-always_last_prompt"><a class="header" href="#622-always_last_prompt">6.2.2: <code>ALWAYS_LAST_PROMPT</code></a></h3>
<p>The option <code>ALWAYS_LAST_PROMPT</code> is set by default, and has been since an
earlier 3.1 release of zsh; after listing a completion, the cursor is
taken back to the line it was on before, instead of reprinting it
underneath. The downside of this is that the listing will be obscured
when you execute the command or produce a different listing, so you may
want to unset the option. <code>ALWAYS_LAST_PROMPT</code> behaviour is required for
menu selection to work, which is why I mention it now instead of in the
ragbag below.</p>
<p>When you're writing your own editor functions which invoke completion,
you can actually cancel the effect of this with the widget
<code>end-of-list</code>, which you would call as <code>zle end-of-list</code> (it's a normal
editing function, not a completion function). You can also bind it to a
key to use to preserve the existing completion list. On the other hand,
if you want to control the behaviour within a completion function, i.e.
to decide whether completion will try to return to the prompt above the
list, you can manipulate it with the <code>last_prompt</code> element of the
<code>$compstate</code> associative array, so for example:</p>
<pre><code> compstate[last_prompt]=''
</code></pre>
<p>will turn off the behaviour for the completion in progress. <code>$compstate</code>
is the place to turn if you find yourself wanting to control completion
behaviour in this much detail; see the <code>zshcompwid</code> manual page.</p>
<p><span id="l149"></span></p>
<h3 id="623-menu-completion-and-menu-selection"><a class="header" href="#623-menu-completion-and-menu-selection">6.2.3: Menu completion and menu selection</a></h3>
<p>The most significant matter decided by the options above is whether or
not you are using menu completion. If you are not, you will need to type
the next character explicitly when completion is ambiguous; if you are,
you just need to keep hitting tab until the completion you want appears.
In the second case, of course, this works best if there are not too many
possibilities. Use of <code>AUTO_MENU</code> or binding the <code>menu-complete</code> widget
to a separate key-stroke gives you something of both worlds.</p>
<p>A new variant of menu completion appeared in 3.1.6; in fact, it deserves
the name menu completion rather more than the original form, but since
that name was taken it is called `menu selection'. This allows you to
move the cursor around the list of completions to select one. It is
implemented by a separate module, <code>zsh/complist</code>; you can make sure this
is loaded by putting `<code>zmodload -i zsh/complist</code>' in <code>.zshrc</code>, although
it should be loaded automatically when the style <code>menu</code> is set as below.
For it to be useful, you need two other things. The first is
<code>ALWAYS_LAST_PROMPT</code> behaviour; this is suppressed if the whole
completion list won't appear on the screen, since there's no line on the
screen to go back to. However, menu selection does still work, by
allowing you to scroll the list up and down. The second thing is that
you need to start menu completion in any of the usual ways; menu
selection is an addition to menu completion, not a replacement.</p>
<p>Now you should set the following style:</p>
<pre><code> zstyle ':completion:*' menu select=<NUM>
</code></pre>
<p>If an ambiguous completion produces at least <code><NUM></code> possibilities, menu
selection is started. You can understand this best by trying it. One of
the completions in the list, initially the top-leftmost, is highlighted
and inserted into the line. By moving the cursor in the obvious
directions (with wraparound at the edges), you change both the value
highlighted and the value inserted into the line. When you have the
value you want, hit return, which removes the list and leaves the
inserted value. Hitting <code>^G</code> (the editor function <code>send-break</code>) aborts
menu selection, removes the list and restores the command line.</p>
<p>Internally, zsh actually uses the parameter <code>$MENUSELECT</code> to supply the
number and hence start menu selection. However, this is always
initialised from the style as defined above, so you shouldn't set
<code>$MENUSELECT</code> directly (unless you are using <code>compctl</code>, which will
happily use menu selection). As with other styles, you can specify
different values for different contexts; the <code>default</code> tag is checked if
the current context does not produce a value for the style with whatever
the current tag is. Note that the <code>menu</code> style also allows you to
control whether menu completion is started at all, with or without
selection; in other words, it is a style corresponding to the
<code>MENU_COMPLETE</code> option.</p>
<p>There is one other additional feature when using menu selection. The zle
command <code>accept-and-infer-next-history</code> has a different meaning here; it
accepts a completion, and then tries to complete again using menu
selection. This is very useful with directory hierarchies, and in
combination with <code>undo</code> gives you a simple file browser. You need to
bind it in the special keymap <code>menuselect</code>; for example, I use</p>
<pre><code> bindkey -M menuselect '^o' accept-and-infer-next-history
</code></pre>
<p>because the behaviour reminds me of what is usually bound to <code>^O</code> in
emacs modes, namely <code>accept-line-and-down-history</code>. Binding it like this
has no effect on <code>^O</code> in the normal keymaps. Try it out by entering menu
selection on a set of files including directories, and typing <code>^O</code> on
one of the directories. You should immediately have the contents of that
directory presented for the next selection, while <code>undo</code> is smart enough
not only to remove that selection but return to completion on the parent
directory.</p>
<p>You can choose the manner in which the currently selected value in the
completion list is highlighted using exactly the same mechanism as for
specifying colours for particular types of matches; see the description
of the <code>list-colors</code> style below.</p>
<p><span id="l150"></span></p>
<h3 id="624-other-ways-of-changing-completion-behaviour"><a class="header" href="#624-other-ways-of-changing-completion-behaviour">6.2.4: Other ways of changing completion behaviour</a></h3>
<p><strong><code>COMPLETE_ALIASES</code></strong></p>
<p>If you set an alias such as</p>
<pre><code> alias pu=pushd
</code></pre>
<p>then the alias `<code>pu</code>' will be expanded when the completion system is
looking for the name of the command, so that it will instead find the
command name `<code>pushd</code>'. This is quite useful to avoid having to define
extra completions for all your aliases. However, it's possible you may
want to define something different for the alias than for the command it
expands to. In that case, you will need to set <code>COMPLETE_ALIASES</code>, and
to make arrangements for completing after every alias which does not
already match the name of a command. Hence `<code>alias zcat="myzcat -dc"</code>'
will work with the option set, even if you haven't told the system about
`<code>myzcat</code>', while `<code>alias myzcat="gzip -dc"</code>' will not work unless you
do define a completion for myzcat: here `<code>compdef _gzip myzcat</code>' would
probably be good enough. Without the option set, it would be the other
way around: the first alias would not work without the extra <code>compdef</code>,
but the second would.</p>
<p><strong><code>AUTO_REMOVE_SLASH</code></strong></p>
<p>This option is turned on by default. If you complete a directory name
and a slash is added --- which it usually is, both to tell you that you
have completed a directory and to allow you to complete files inside it
without adding a `<code>/</code>' by hand --- and the next thing you type is <em>not</em>
something which would insert or complete part of a file in that
directory, then the slash is removed. Hence:</p>
<pre><code> % rmdir my<TAB>
-> rmdir mydir/
% rmdir mydir/<RETURN>
-> `rmdir mydir' executed
</code></pre>
<p>This example shows why this behaviour was added: some versions of
`<code>rmdir</code>' baulk at having the slash after the directory name. On the
other hand, if you continued typing after the slash, or hit tab again to
complete inside <code>mydir</code>, then the slash would remain.</p>
<p>This is at worst harmless under most circumstances. However, you can
unset the option <code>AUTO_REMOVE_SLASH</code> if you don't like that behaviour.
One thing that may cause slight confusion, although it is the same as
with other suffixes (i.e. bits which get added automatically but aren't
part of the value being completed), is that the slash is added straight
away if the value is being inserted by menu completion. This might cause
you to think wrongly that the completion is finished, and hence is
unique when in fact it isn't.</p>
<p>Note that some forms of completion have this type of behaviour built in,
not necessarily with a slash, when completing lists of arguments. For
example, enter `<code>typeset ZSH_V<TAB></code>' and you will see
`<code>ZSH_VERSION=</code>' appear, in case you want to assign something to the
parameter; hitting space, which is not a possible value, makes the
`<code>=</code>' disappear. This is not controlled by the <code>AUTO_REMOVE_SLASH</code>
option, which applies only to directories inserted by the standard
filename completion system.</p>
<p><strong><code>AUTO_PARAM_SLASH</code>, <code>AUTO_PARAM_KEYS</code></strong></p>
<p>These options come into effect when completing expressions with
parameter substitutions. If <code>AUTO_PARAM_SLASH</code> is set, then any
parameter expression whose value is the name of a directory will have a
slash appended when completed, just as if the value itself had been
inserted by the completion system.</p>
<p>The behaviour for <code>AUTO_PARAM_KEYS</code> is a bit more complicated. Try this:</p>
<pre><code> print ${ZSH_V<TAB>
</code></pre>
<p>You will find that you get the complete word `<code>${ZSH_VERSION}</code>', with
the closing brace and (assuming there are no other matching parameters)
a space afterwards. However, often after you have completed a parameter
in this fashion you want to type something immediately after it, such as
a subscript. With <code>AUTO_PARAM_KEYS</code>, if you type something at this point
which seems likely to have to go after the parameter name, it will
immediately be put there without you having to delete the intervening
characters --- try it with `<code>[</code>', for example. Note that this only
happens if the parameter name and any extra bits were added by
completion; if you type everything by hand, typing `<code>[</code>' will not have
this magic effect.</p>
<p><strong><code>COMPLETE_IN_WORD</code></strong></p>
<p>If this is set, completion always takes place at the cursor position in
the word. For example if you typed `<code>Mafile</code>', went back over the
`<code>f</code>', and hit tab, the shell would complete `<code>Makefile</code>', instead of
its usual behaviour of going to the end of the word and trying to find a
completion there, i.e. something matching `<code>Mafile*</code>'. Some sorts of
new completion (such as filename completion) seem to implement this
behaviour regardless of the option setting; some other features (such as
the `<code>_prefix</code>' completer described below) require it, so it's a good
thing to set and get used to, unless you really need to complete only at
the end of the word.</p>
<p><strong><code>ALWAYS_TO_END</code></strong></p>
<p>If this is set, the cursor is always moved to the end of the word after
it is completed, even if completion took place in the middle. This also
happens with menu completion.</p>
<p><span id="l151"></span></p>
<h3 id="625-changing-the-way-completions-are-displayed"><a class="header" href="#625-changing-the-way-completions-are-displayed">6.2.5: Changing the way completions are displayed</a></h3>
<p><strong><code>LIST_TYPES</code></strong></p>
<p>This is like the <code>-F</code> option to <code>ls</code>; files which appear in the
completion listing have a trailing `<code>/</code>' for a directory, `<code>*</code>' for a
regular file executable by the current process, `<code>@</code>' for a link,
`<code>|</code>' for a named pipe, `<code>%</code>' for a character device and `<code>#</code>' for a
block device. This option is on by default.</p>
<p>Note that the identifiers only appear if the completion system knows
that the item is supposed to be a file. This is automatic if the usual
filename completion commands are used. There is also an option <code>-f</code> to
the builtin <code>compadd</code> if you write your own completion function and want
to tell the shell that the values may be existing files to apply
<code>LIST_TYPES</code> to (though no harm is caused if no such files exist).</p>
<p><strong><code>LIST_PACKED</code>, <code>LIST_ROWS_FIRST</code></strong></p>
<p>These affect the arrangement of the completion listing. With
<code>LIST_PACKED</code>, completion lists are made as compact as possible by
varying the widths of the columns, instead of formatting them into a
completely regular grid. With <code>LIST_ROWS_FIRST</code>, the listing order is
changed so that adjacent items appear along rows instead of down
columns, rather like <code>ls</code>'s <code>-x</code> option.</p>
<p>It is possible to alter both these for particular contexts using the
styles <code>list-packed</code> and <code>list-rows-first</code>. The styles in such cases
always override the option; the option setting is used if no
corresponding style is found.</p>
<p>Note also the discussion of completion groups later on: it is possible
to have different types of completion appear in separate lists, which
may then be formatted differently using these tag-sensitive styles.</p>
<p><span id="l152"></span></p>
<h2 id="63-getting-started-with-new-completion"><a class="header" href="#63-getting-started-with-new-completion">6.3: Getting started with new completion</a></h2>
<p>Before I go into any detail about new completion, here's how to set it
up so that you can try it out. As I said above, the basic objects that
do completions are shell functions. These are all autoloaded, so the
shell needs to know where to find them via the <code>$fpath</code> array. If the
shell was installed properly, and nothing in the initialization files
has removed the required bits from <code>$fpath</code>, this should happen
automatically. It's even possible your system sets up completion for you
(Mandrake Linux 6.1 is the first system known to do this out of the
box), in which case type `<code>which compdef</code>' and you should see a
complete shell function --- actually the one which allows you to define
additional completion functions. Then you can skip the next paragraph.</p>
<p>If you want to load completion, try this at the command line:</p>
<pre><code> autoload -U compinit
compinit
</code></pre>
<p>which should work silently. If not, you need to ask your system
administrator what has happened to the completion functions or find them
yourself, and then add all the required directories to your <code>$fpath</code>.
Either they will all be in one big directory, or in a set of
subdirectories with the names <code>AIX</code>, <code>BSD</code>, <code>Base</code>, <code>Debian</code>, <code>Redhat</code>,
<code>Unix</code>, <code>X</code> and <code>Zsh</code>; in the second case, all the directories need to
be in <code>$fpath</code>. When this works, you can add the same lines, including
any modification of <code>$fpath</code> you needed, to your <code>.zshrc</code>.</p>
<p>You can now see if it's actually working. Type `<code>cd </code>', then <code>^D</code>,
and you should be presented with a list of directories only, no regular
files. If you have <code>$cdpath</code> set, you may see directories that don't
appear with <code>ls</code>. As this suggests, the completion system is supplied
with completions for many common (and some quite abstruse) commands.
Indeed, the idea is that for most users completion just works without
intervention most of the time. If you think it should when it doesn't,
it may be a bug or an oversight, and you should report it.</p>
<p>Another example on the theme of `it just works':</p>
<pre><code> tar xzf archive.tar.gz ^D
</code></pre>
<p>will look inside the gzipped tar archive --- assuming the GNU version of
<code>tar</code>, for which the `<code>z</code>' in the first set of arguments reports that
the archive has been compressed with gzip --- and give you a list of
files or directories you can extract. This is done in a very similar way
to normal file completion; although there are differences, you can do
completion down to any directory depth within the archive. (At this
point, you're supposed to be impressed.)</p>
<p>The completion system knows about more than just commands and their
arguments, it also understands some of the shell syntax. For example,
there's an associative array called <code>$_comps</code> which stores the names of
commands as keys and the names of completion functions as the
corresponding values. Try typing:</p>
<pre><code> print ${_comps[
</code></pre>
<p>and then <code>^D</code>. You'll probably get a message asking if you really want
to see all the possible completions, i.e. the keys for <code>$_comps</code>; if you
say `<code>y</code>' you'll see a list. If you insert any of those keys, then
close the braces so you have e.g. `<code>${_comps[mozilla]}</code>' and hit
return, you'll see the completion function which handles that command;
in this case (at the time of writing) it's <code>_webbrowser</code>. This is one
way of finding out what function is handling a particular command. If
there is no entry --- i.e. the `<code>print ${_comps[mycmd]}</code>' gives you a
blank line --- then the command is not handled specially and will simply
use whatever function is defined for the `<code>-default-</code>' context, usually
<code>_default</code>. Usually this will just try to complete file names. You can
customize <code>_default</code>, if you like.</p>
<p>Apart from <code>-default-</code>, some other of those keys for <code>_comps</code> also look
like <code>-this-</code>: they are special contexts, places other than the
arguments of a command. We were using the context called <code>-subscript-</code>;
you'll find that the function in this case is called <code>_subscript</code>. Many
completion functions have names which are simply an underscore followed
by the command or context name, minus any hyphens. If you want a taster
of how a completion function looks, try `<code>which _subscript</code>'; you may
well find there are a lot of other commands in there that you don't know
yet.</p>
<p>It's important to remember that the function found in this way is at the
root of how a completion is performed. No amount of fiddling with
options or styles --- the stuff I'm going to be talking about for the
next few sections --- will change that; if you want to change the basic
completion, you will just have to write your own function.</p>
<p>By the way, you may have old-style completions you want to mix-in --- or
maybe you specifically don't want to mix them in so that you can make
sure everything is working with the new format. By default, the new
completion system will first try to find a specific new-style
completion, and if it can't it will try to find a <code>compctl</code>-defined
completion for the command in question. If all that fails, it will try
the usual new-style default completion, probably just filename
completion. Note that specific new-style completions take precedence,
which is fair enough, since if you've added them you almost certainly
don't want to go back and use the old form. However, if you don't ever
want to try old-style completion, you can put the following incantation
in your <code>.zshrc</code>:</p>
<pre><code> zstyle ':completion:*' use-compctl false
</code></pre>
<p>For now, that's just black magic, but later I will explain the `style'
mechanism in more detail and you will see that this fits in with the
normal way of turning things off in new-style completion.</p>
<p><span id="l153"></span></p>
<h2 id="64-how-the-shell-finds-the-right-completions"><a class="header" href="#64-how-the-shell-finds-the-right-completions">6.4: How the shell finds the right completions</a></h2>
<p><span id="l154"></span></p>
<h3 id="641-contexts"><a class="header" href="#641-contexts">6.4.1: Contexts</a></h3>
<p>The examples above show that the completion system is highly
context-sensitive, so it's important to know how these contexts are
described. This system evolved gradually, but everything I say applies
to all versions of zsh with the major version 4.</p>
<p>state we are at in completion, and is given as a sort of colon-separated
path, starting with the least specific part. There's an easy way of
finding out what context you are in: at the point where you want to
complete something, instead type `<code>^Xh</code>', and it will tell you. In the
case of the <code>$_comps</code> example, you will find,</p>
<pre><code> :completion::complete:-subscript-::
</code></pre>
<p>plus a list of so-called `tags' and completion functions, which I'll
talk about later. The full form is:</p>
<pre><code> :completion:<func>:<completer>:<command>:<argument>:<tag>
</code></pre>
<p>where the elements may be missing if they are not set, but the colons
will always be there to make pattern matching easier. Here's what the
bits of the context mean after the <code>:completion:</code> part, which is common
to the whole completion system.</p>
<ul>
<li><em><strong><func></strong></em><br />
is the name of a function from which completion is called --- this
is blank if it was started from the standard completion system, and
only appears in a few special cases, listed in section six of this
chapter.</li>
<li><em><strong><completer></strong></em><br />
is called `<code>complete</code>' in this case: this refers to the fact that
the completion system can do more than just simple completion; for
example, it can do a more controlled form of expansion (as I
mentioned), spelling correction, and completing words with spelling
mistakes. I'll introduce the other completers later; `<code>complete</code>'
is the simplest one, which just does basic completion.</li>
<li><em><strong><command></strong></em><br />
is the name of a command or other similar context as described
above, here `<code>-subscript-</code>'.</li>
<li><em><strong><argument></strong></em><br />
is most useful when <code><command></code> is the name of a real command; it
describes where in the arguments to that command we are. You'll see
how it works in a moment. Many of the simpler completions don't use
this; only the ones with complicated option and argument
combinations. You just have to find out with <code>^Xh</code> if you need to
know.</li>
<li><em><strong><tag></strong></em><br />
describes the type of a completion, essentially a way of
discriminating between the different things which can be completed
at the same point on the command line.</li>
</ul>
<p>Now look at the context for a more normal command-argument completion,
e.g. after <code>cd</code>; here you'll see the context
`<code>:completion::complete:cd::</code>'. Here the command-name part of the
context is a real command.</p>
<p>For something more complicated, try after `<code>cvs add</code>' (it doesn't
matter for this if you don't have the <code>cvs</code> command). You'll see a long
and repetitive list of tags, for two possible contexts,</p>
<pre><code> :completion::complete:cvs:argument-rest:
:completion::complete:cvs-add:argument-rest:
</code></pre>
<p>The reason you have both is that the `<code>add</code>' is not only an argument to
<code>cvs</code>, as the first context would suggest, it's also a subcommand in its
own right, with its own arguments, and that's what the second context is
for. The first context implies there might be more subcommands after
`<code>add</code>' and its arguments which are completely separate from them ---
though in fact CVS doesn't work that way, so that form won't give you
any completions here.</p>
<p>In both, `<code>argument-rest</code>' shows that completion is looking for another
argument, the `<code>rest</code>' indicating that it is the list of arguments at
the end of the line; if position were important (see `<code>cvs import</code>' for
an example), the context would contain `<code>argument-1</code>', or whatever. The
`<code>cvs-add</code>' shows how subcommands are handled, by separating with a
hyphen instead of a colon, so as not to confuse the different bits of
the context.</p>
<p>Apart from arguments to commands and subcommands, arguments to options
are another frequent possibility; for an example of this, try typing
<code>^Xh</code> after `<code>dvips -o</code>' and you will see the context
`<code>:completion::complete:dvips:option-o-1:</code>'; this shows you are
completing the first argument to <code>dvips</code>'s <code>-o</code> option, (it only takes
one argument) which happens to be the name of a file for output.</p>
<p><span id="l155"></span></p>
<h3 id="642-tags"><a class="header" href="#642-tags">6.4.2: Tags</a></h3>
<p>Now on to the other matter to do with contexts, tags. Let's go back and
look at the output from the <code>^Xh</code> help test after the <code>cd</code> command in
full:</p>
<pre><code> tags in context :completion::complete:cd::
local-directories path-directories (_alternative _cd)
</code></pre>
<p>Unlike the contexts considered so far, which tell you how completion
arrived at the point it did, the tags describe the things it can
complete here. In this case, there are three: <code>directory-stack</code> refers
to entries such as `<code>+1</code>'; the directory stack is the set of
directories defined by using the <code>pushd</code> command, which you can see by
using the <code>dirs</code> command. Next, <code>local-directories</code> refers to
subdirectories of the current working directory, while
<code>path-directories</code> refers to any directories found by searching the
<code>$cdpath</code> array. Each of the possible completions which the system
offers belongs to one of those classes.</p>
<p>In parentheses, you see the names of the functions which were called to
generate the completions; these are what you need to change or replace
if you want to alter the basic completion behaviour. Calling functions
appear on the right and called functions on the left, so that in this
case the function `<code>_cd</code>' was the function first called to handle
arguments for the <code>cd</code> command, fitting the usual convention. Some
standard completion functions have been filtered out of this list --- it
wouldn't help you to know it had been through <code>_main_complete</code> and
<code>_complete</code>, for example.</p>
<p>Maybe it's already obvious that having the system treat different types
of completion in different ways is useful, but here's an example, which
gives you a preview of the `styles' mechanism, discussed later. Styles
are a sort of glorified shell parameter; they are defined with the
<code>zstyle</code> command, using a style name and possible values which may be an
array; you can always define a style as an array, but some styles may
simply use it as a string, joining together the arguments you gave it
with spaces. You can also use the <code>zstyle</code> command, with different
arguments, to retrieve their value, which is what the completion system
itself does; there's no actual overlap with parameters and their values,
so they don't get in the way of normal shell programming.</p>
<p>Where styles differ from parameters is that they can take different
values in different contexts. The first argument to the <code>zstyle</code> command
gives a context; when you define a style, this argument is actually a
pattern which will be matched against the current context to see if the
style applies. The rule for finding out what applies is: exact string
matches are preferred before patterns, and longer patterns are preferred
before shorter patterns. Here's that example:</p>
<pre><code> zstyle ':completion:*:cd:*' tag-order local-directories \
path-directories
</code></pre>
<p>From the discussion of contexts above, the pattern will match any time
an argument to the <code>cd</code> command is being completed. The style being set
is called <code>tag-order</code>, and the values are the two tags valid for
directories in <code>cd</code>.</p>
<p>The <code>tag-order</code> style determines the order in which tags are tried. The
value given above means that first <code>local-directories</code> will be
completed; only if none can be completed will <code>path-directories</code> be
tried. You can enter the command and try this; if you don't have
<code>$cdpath</code> set up you can assign `<code>cdpath=(~)</code>', which will allow `<code>cd foo</code>' to change to a directory `<code>~/foo</code>' and allow completion of
directories accordingly. Go to a directory other than <code>~</code>; completion
for <code>cd</code> will only show subdirectories of where you are, not those of
<code>~</code>, unless you type a string which is the prefix of a directory under
<code>~</code> but not your current directory. For example,</p>
<pre><code> % cdpath=(~)
% ls -F ~
foo/ bar/
% ls -F
rod/ stick/
# Without that tag-order zstyle command, you would get...
% cd ^D
bar/ foo/ rod/ stick/
% zstyle ':completion:*:cd:*' tag-order local-directories \
path-directories
# now you just get the local directories, if there are any...
% cd ^D
rod/ stick/
</code></pre>
<p>There's more you can do with the <code>tag-order</code> style: if you put the tags
into the same word by quoting, for example <code>"local-directories path-directories"</code>, then they would be tried at the same time, which in
this case gives you the effect of the default. In fact, since it's too
much work to know what tags are going to be available for every single
possible completion, the default when there is no appropriate
<code>tag-order</code> is simply to try all the tags available in the context at
once; this was of course what was originally happening for completion
after <code>cd</code>.</p>
<p>Even if there is a <code>tag-order</code> specification, any tags not specified
will usually be tried all together at the end, so you could actually
have missed out <code>path-directories</code> from the end of the original example
and the effect would have been the same. If you don't want that to
happen, you can specify a `<code>-</code>' somewhere in the list of tags, which is
not used as a tag but tells completion that only the tags in the list
should be tried, not any others that may be available. Also, if you
don't want a particular tag to be shown you can include `<code>!tagname</code>' in
the values, and all the others but this will be included. For example,
you may have noticed that when completing in command position you are
offered parameters to set as well as commands etc.:</p>
<pre><code> Completing external command
tex texhash texi2pdf text2sf
texconfig texi2dvi texindex textmode
texdoc texi2dvi4a2ps texlinks texutil
texexec texi2html texshow texview
Completing parameter
TEXINPUTS texinputs
</code></pre>
<p>(I haven't told you how to produce those descriptions, or how to make
the completions for different tags appear separately, but I will --- see
the descriptions of the `<code>format</code>' and `<code>group-name</code>' styles below.)
If you set</p>
<pre><code> zstyle ':completion:*:-command-:*' tag-order '!parameters'
</code></pre>
<p>then the last two lines will disappear from the completion. Of course,
your completion list probably looks completely different from mine
anyway. By the way, one good thing about styles is that it doesn't
matter whether they're defined before or after completion is loaded,
since styles are stored and retrieved by another part of the shell.</p>
<p>To exclude more than one tag name, you need to include the names in the
same word. For example, to exclude both parameters and reserved words
the value would be <code>'!parameters reserved-words'</code>, and <em>not</em>
<code>'!parameters' '!reserved-words'</code>, which would try completion once with
parameters excluded, then again with reserved words excluded.
Furthermore, tags can actually be patterns, or more precisely any word
in one of the arguments to <code>tag-order</code> may contain a pattern, which will
then be tried against all the valid tags to see if it matches. It's
sometimes even useful to use `<code>*</code>' to match all tags, if you are
specifying a special form of one of the tags --- maybe using a label, as
described next --- in the same word. See the manual for all the tag
names understood by the supplied functions.</p>
<p>The <code>tag-order</code> style allows you to give tags `labels', which are a
sort of alias, instructing the completion system to use a tag under a
different name. You arrange this by giving the tag followed by a colon,
followed by the label. The label can also have a hyphen in front, which
means that the original tag name should be put in front when the label
is looked up; this is really just a way of making the names look neater.
The upshot is that by using contexts with the label name in, rather than
the tag name, you can arrange for special behaviour. Furthermore, you
can give an alternative description for the labelled tag; these show up
with the <code>format</code> style which I'll describe below (and which I
personally find very useful). You put the description after another
colon, with any spaces quoted. It would look like this:</p>
<pre><code> zstyle ':completion:*:aliens:*' tag-order \
'frooble:-funny:funny\ frooble' frooble
</code></pre>
<p>which is used when you're completing for the command <code>aliens</code>, which
presumably has completions tagged as `<code>frooble</code>' (if not, you're very
weird). Then completion will first look up styles for that tag under the
name <code>frooble-funny</code>, and if it finds completions using those styles it
will list them with a description (if you are using <code>format</code>) of `funny
frooble'. Otherwise, it will look up the styles for the tag under its
usual name and try completion again. It's presumably obvious that if you
don't have different styles for the two labels of the tag, you get the
same completions each time.</p>
<p>Rather than overload you with information on tags by giving examples of
how to use tag labels now, I'll reserve this for the description of the
<code>ignored-patterns</code> style below, which is one neat use for labels. In
fact, it's the one for which it was invented; there are probably lots of
other ones we haven't thought of yet.</p>
<p>One important note about <code>tag-order</code> which I may not have made as
explicit as I should have: <em>it doesn't change which tags are actually
valid in that completion</em>. Just putting a tag name into the list doesn't
mean that tag name will be used; that's determined entirely by the
completion functions for a particular context. The <code>tag-order</code> style
simply alters the order in which the tags which <em>are</em> valid are
examined. Come back and read this paragraph again when you can't work
out why <code>tag-order</code> isn't doing what you want.</p>
<p>Note that the rule for testing patterns means that you can always
specify a catch-all worst case by `<code>zstyle "*" style ...</code>', which will
always be tried last --- not just in completion, in fact, since other
parts of the shell use the styles mechanism, and without the
`<code>:completion:</code>' at the start of the context this style definition will
be picked up there, too.</p>
<p>Styles like <code>tag-order</code> are the most important case where tags are used
on their own. In other cases, they can be added to the end of the
context; this is useful for styles which can give different results for
different sets of completions, in particular styles that determine how
the list of completions is displayed, or how a completion is to be
inserted into the command line. The tag is the final element, so is not
followed by a colon. A full context then looks something like
`<code>:completion::complete:cd::path-directories</code>'. Later, you'll see some
styles which can usefully be different for different tag contexts.
Remember, however, that the tags part of the context, like other parts,
may be empty if the completion system hasn't figured out what it should
be yet.</p>
<p><span id="l156"></span></p>
<h2 id="65-configuring-completion-using-styles"><a class="header" href="#65-configuring-completion-using-styles">6.5: Configuring completion using styles</a></h2>
<p>You now know how to define a style for a particular context, using</p>
<pre><code> zstyle <context> <style> <value...>
</code></pre>
<p>and some of the cases where it's useful. Before introducing other
styles, here's some more detailed information. I already said that
styles could take an array value, i.e. a set of values at the end of the
<code>zstyle</code> command corresponding to the array elements, and you've already
seen one case (<code>tag-order</code>) where that is useful. Many styles only use
one value, however. There is a particularly common case, where you
simply want to turn a value on or off, i.e. a boolean value. In this
case, you can use any of `<code>true</code>', `<code>yes</code>', `<code>on</code>' or `<code>1</code>' for on
and `<code>false</code>', `<code>no</code>', `<code>off</code>' or `<code>0</code>' for off. You define all
styles the same way; only when they're used is it decided whether they
should be a scalar, an array, or a boolean, nor is the name of a style
checked to see if it is valid, since the shell doesn't know what styles
might later be looked up. The same obviously goes for contexts.</p>
<p>You can list existing styles (not individually, only as a complete list)
using either `<code>zstyle</code>' or `<code>zstyle -L</code>'. In the second case, they are
output as the set of <code>zstyle</code> commands which would regenerate the styles
currently defined. This is also useful with <code>grep</code>, since you can easily
check all possible contexts for a particular style.</p>
<p>The most powerful way of using <code>zstyle</code> is with the option <code>-e</code>. This
says that the words you supply are to be evaluated as if as arguments to
<code>eval</code>. This should set the array <code>$reply</code> to the words to be used. So</p>
<pre><code> zstyle '*' days 'Monday Tuesday'
</code></pre>
<p>and</p>
<pre><code> zstyle -e '*' days 'reply=(Monday Tuesday)'
</code></pre>
<p>are equivalent --- but the intention, of course, is that in the second
case the argument can return a different value each time so that the
style can vary. It will usually be evaluated in the heat of completion,
hence picking up all the editing parameters; so for example</p>
<pre><code> zstyle -e ':completion:*' mystyles 'reply=(${NUMERIC:-0})'
</code></pre>
<p>will make the style return a non-zero integer (possibly indicating
<code>true</code>) if you entered a non-zero prefix argument to the command, as
described in <a href="zshguide04.html#zle">chapter 4</a>. However, the argument can
contain any zsh code whatsoever, not just a simple assignment. Remember
to quote it to prevent it from being turned into something else when the
<code>zstyle</code> command line is run.</p>
<p>Finally, you can delete a context for a style or a list of styles by</p>
<pre><code> zstyle -d [ <context-pattern> [ <style> ] ] ...
</code></pre>
<p>--- note that although the first argument is a pattern, in this case it
is treated exactly, so if you give the pattern `<code>:completion:*:cd:*</code>',
only values given with <em>exactly</em> that pattern will be deleted, not other
values whose context begins with `<code>:completion:</code>' and contains
`<code>:cd:</code>'. The pattern and the style are optional when deleting; if
omitted, all styles for the context, or all styles of any sort, are
deleted. The completion system has its own defaults, but these are
builtin, so anything you specify takes precedence.</p>
<p>By the way, I did mention in passing in <a href="zshguide04.html#zle">chapter 4</a>
that you could use styles in just the same way in ordinary zle widgets
(the ones created with `<code>zle -N</code>'), but you probably forgot about that
straight away. All the instructions about defining styles and using them
in your own functions from this chapter apply to zle functions. The only
difference is that in that case the convention for contexts is that the
context is set to `<code>:zle:</code><em>widget-name</em>' for executing the widget
<em>widget-name</em>.</p>
<p>The rest of this section describes some useful styles. It's up to you to
experiment with contexts if you want the style's values to be different
in different places, or just use `<code>*</code>' if you don't care.</p>
<p><span id="l157"></span></p>
<h3 id="651-specifying-completers-and-their-options"><a class="header" href="#651-specifying-completers-and-their-options">6.5.1: Specifying completers and their options</a></h3>
<p>`Completers' are the behind-the-scenes functions that decide what sort
of completion is being done. You set what completers to use with the
`<code>completer</code>' style, which takes an array of completers to try in
order. For example,</p>
<pre><code> zstyle ':completion:*' completer _complete _correct _approximate
</code></pre>
<p>specifies that first normal completion will be tried (`<code>_complete</code>'),
then spelling correction (`<code>_correct</code>'), and finally approximate
completion (`<code>_approximate</code>'), which is essentially the combined effect
of the previous two, i.e. complete the word typed but allow for spelling
mistakes. All completers set the context, so inside <code>_complete</code> you will
usually find `<code>:completion::complete:...</code>', inside correction
`<code>:completion::correct:..</code>', and so on.</p>
<p>There's a labelling feature for completers, rather like the one for tags
described, but not illustrated in detail, above. You can put a completer
in a list like this:</p>
<pre><code> zstyle ':completion:*' completer ... _complete:comp-label ...
</code></pre>
<p>which calls the completer <code>_complete</code>, but pretends its name is
<code>comp-label</code> when looking things up in styles, so you can try completers
more than once with different features enabled. As with tags, you can
write it like `<code>_complete:-label</code>', and the normal name will be
prepended to get the name `<code>complete-label</code>' --- just a shortcut, it
doesn't introduce anything new. I'll defer an example until you know
what the completers do.</p>
<p>Here is a more detailed description of the existing completers; they are
all functions, so you can simply copy and modify one to make your own
completer.</p>
<p><strong><code>_complete</code></strong></p>
<p>This is the basic completion behaviour, which we've been assuming up to
now. Its major use is simply to check the context --- here meaning
whether we are completing a normal command argument or one of the
special `<code>-context-</code>' places --- and call the appropriate completion
function. It's possible to trick it by setting the parameter
`<code>compcontext</code>' which will be used instead of the one generated
automatically; this can be useful if you write your own completion
commands for special cases. If you do this, you should make the
parameter local to your function.</p>
<p><strong><code>_approximate</code></strong></p>
<p>This does approximate completion: it's actually written as a wrapper for
the <code>_complete</code> completer, so it does all the things that does, but it
also sets up the system to allow completions with misspellings.
Typically, you would want to try to complete without misspellings first,
so this completer usually appears after <code>_complete</code> in the <code>completers</code>
style.</p>
<p>The main means of control is via the <code>max-errors</code> style. You can set
this to the maximum number of errors to allow. An error is defined as
described in the manual for approximate pattern matching: a character
missing such as `<code>rhythm</code>' / `<code>rhytm</code>', an extra character such as
`<code>rhythm</code>' / `<code>rhythms</code>', an incorrect character such as `<code>rhythm</code>' /
`<code>rhxthm</code>', or a pair of characters transposed such as `<code>rhythm</code>'
`<code>rhyhtm</code>' each count as one error. Approximation will first try to
find a match or matches with one error, then two errors, and so on, up
to and including the value of <code>max-errors</code>; the set of matches with the
lowest number of errors is chosen, so that even if you set <code>max-errors</code>
large, matches with a lower number of errors will always be preferred.
The real problems with setting a large <code>max-errors</code> are that it will be
slower, and is more likely to generate matches completely unlike what
you want --- with typing errors, two or three are probably the most you
need. Otherwise, there's always Mavis Beacon. Hence:</p>
<pre><code> % zstyle ':completion:*' max-errors 2
# just for the sake of example...
% zstyle ':completion:*' completer _approximate
% ls
ashes sackcloth
% echo siccl<TAB>
-> echo sackcloth
% echo zicc<TAB>
<Beep.>
</code></pre>
<p>because `<code>s[i/a]c[k]cloth</code>' is only two errors, while
`<code>[z/s][i/a]c[k]cloth</code>' would be three, so doesn't complete.</p>
<p>There's another way to give a maximum number of errors, using the
numeric prefix specified with <code>ESC-<digit></code> in Emacs mode, directly with
number keys in vi command mode, or with <code>universal-argument</code>. To enable
this, you have to include the string <code>numeric</code> as one of the values for
<code>max-errors</code> --- hence this can actually be an array, e.g.</p>
<pre><code> zstyle ':completion:*:approximate:*' max-errors 2 numeric
</code></pre>
<p>allows up to two errors automatically, but you can specify a higher
maximum by giving a prefix to the completion command. So to continue the
example above, enter the new <code>zstyle</code> and:</p>
<pre><code> % echo zicc<ESC-3><TAB>
-> echo sackcloth
</code></pre>
<p>because we've allowed three errors. You can start to see the problems
with allowing too many errors: if you had the file `<code>zucchini</code>', that
would be only one error away, and would be found and inserted before
`<code>sackcloth</code>' was even considered.</p>
<p>Note that the context is examined straightaway in the completer, so at
this stage it is simply `<code>:completion::approximate:::</code>'; no more
detailed contextual information is available, so it is not possible to
specify different <code>max-errors</code> for different commands or tags.</p>
<p>The final possibility as a value for the style is `<code>not-numeric</code>': that
means if any numeric prefix is given, approximation will not be done at
all. In the last example, completion would have to find a file beginning
`<code>zicc</code>'.</p>
<p>Other minor styles also control approximation. The style <code>original</code>, if
true means the original value is always treated as a possible
completion, even if it doesn't match anything and even if nothing else
matched. Completing the original and the corrections use different tags,
unimaginatively called <code>original</code> and <code>corrections</code>, so you can organise
this with the <code>tag-order</code> style.</p>
<p>Because the completions in this case usually don't match what's already
on the command line, and may well not match each other, menu completion
is entered straight away for you to pick a completion. You can arrange
that this doesn't happen if there is an unambiguous piece at the start
to insert first by setting the boolean style <code>insert-unambiguous</code>.</p>
<p>Those last two styles (<code>original</code> and <code>insert-unambiguous</code>) are looked
up quite early on, when the context for generating corrections is being
set up, so that only the context up to the completer name is available.
The completer name will be followed by a hyphen and the number of errors
currently being accepted. So for trying approximation with one error the
context is `<code>:completion::approximate-1:::</code>'; if that fails and the
system needs to look for completion with two errors, the context will be
`<code>:completion::approximate-2:::</code>', and so on; the same happens with
correction and `<code>correct-1</code>', etc., for the completer described next.</p>
<p><strong><code>_correct</code></strong></p>
<p>This is very similar to <code>_approximate</code>, except that the context is
`<code>:completion::correct:*</code>' (or `<code>:completion::correct-<num>:*</code>' when
generating corrections, as described immediately above) and it won't
perform completion, just spelling correction, so extra characters which
the completer has to add at the end of the word on the line now count as
extra errors instead of completing in the ordinary way: <code>zicc</code> is
woefully far from <code>sackcloth</code>, seven errors, but <code>ziccloth</code> only counts
three again. The <code>_correct</code> completer is controlled in just the same way
as <code>_approximate</code>.</p>
<p>There is a separate command which only does correction and nothing else,
usually bound to `<code>^Xc</code>', so if you are happy using that you don't need
to include <code>_correct</code> in the list of completers. If you do include it,
and you also have <code>_approximate</code>, <code>_correct</code> should come earlier;
<code>_approximate</code> is bound to generate all the matches <code>_correct</code> does, and
probably more. Like other separate completion commands, it has its own
context, here beginning `<code>:completion:correct-word:</code>', so it's easy to
make this command behave differently from the normal completers.</p>
<p>Old-timers will remember that there is another form of spelling
correction built into the shell, called with `<code>ESC-$</code>' or `<code>ESC-s</code>'.
This only corrects filenames and doesn't understand anything about the
new completion mechanism; the only reason for using it is that it may
well be faster. However, if you use the <code>CORRECT</code> or <code>CORRECT_ALL</code> shell
options, you will be using the old filename correction mechanism; it's
not yet possible to alter this.</p>
<p><strong><code>_expand</code></strong></p>
<p>This actually performs expansion, not completion; the difference was
explained at the start of the chapter. If you use it, you should bind
tab to <code>complete-word</code>, not <code>expand-or-complete</code>, since otherwise
expansion will be performed before the completion mechanism is started
up. As expansion should still usually be attempted before completion,
this completer should appear before <code>_complete</code> and its relatives in the
list of values for the <code>completers</code> style.</p>
<p>The reason for using this completer instead of normal expansion is that
you can control which expansions are performed using styles in the
`<code>:completion:*:expand:*</code>' context. Here are the relevant styles:</p>
<ul>
<li><strong><code>glob</code></strong><br />
expands glob expressions, in other words does filename generation
using wildcards.</li>
<li><strong><code>substitute</code></strong><br />
expands expressions including and active `<code>$</code>' or backquotes.</li>
</ul>
<p>But remember that you need</p>
<pre><code> bindkey '^i' complete-word
</code></pre>
<p>when using this completer as otherwise the built-in expansion mechanism
which is run by the normal binding <code>expand-or-complete</code> will take over.</p>
<p>You can also control how expansions are inserted. The tags for adding
expansions are <code>original</code> (presumably self-explanatory),
<code>all-expansions</code>, which refers to adding a single string containing all
the possible expansions (the default, just like the editor function
<code>expand-word</code>), and <code>expansions</code>, which refers to the results added one
by one. By changing the order in which the tags are tried, as described
for the <code>tag-order</code> style above, you can decide how this happens. For
example,</p>
<pre><code> zstyle ':completion:*' completer _expand _complete
zstyle ':completion::expand:*' tag-order expansions
</code></pre>
<p>sets up for performing glob expansion via completion, with the
expansions being presented one by one (usually via menu completion,
since there is no common prefix). Altering <code>expansions</code> to
<code>all-expansions</code> would insert the list, as done by the normal expansion
mechanism, while altering it to `<code>expansions original</code>' would keep the
one-at-a-time entry but also present the original string as a
possibility. You can even have all three, i.e. the entire list as a
single string becomes just one of the set of possibilities.</p>
<p>There is also a <code>sort</code> style, which determines whether the expansions
generated will be sorted in the way completions usually are, or left
just as the shell produced them from the expansion (for example,
expansion of an array parameter would produce the elements in order). If
it is <code>true</code>, they will always be sorted, if <code>false</code> or unset never, and
if it is <code>menu</code> they will be sorted for the <code>expansions</code> tag, but not
for the <code>all-expansions</code> tag which will be a single string of the values
in the original order.</p>
<p>There is a slight problem when you try just to generate <code>glob</code>
expansions, without <code>substitute</code>. In fact, it doesn't take much thought
to see that an expression like `<code>$PWD/*.c</code>' doesn't mean anything if
<code>substitute</code> is inactive; it must be active to make sense of such
expressions. However, this is annoying if there are no matches: you end
up being offered a completion with the expanded <code>$PWD</code>, but `<code>*.c</code>'
still tacked on the end, which isn't what you want. If you use <code>_expand</code>
mainly for globbing, you might therefore want to set the style
<code>subst-globs-only</code> to true: if a completion just expands the parameters,
and globbing does nothing, then the expansion is rejected and the line
left untouched.</p>
<p>The <code>_expand</code> completer will also use the styles</p>
<ul>
<li><strong><code>accept-exact</code></strong><br />
applies to words beginning with a `<code>$</code>' or `<code>~</code>'. Suppose there is
a parameter `<code>$foo</code>' and a parameter `<code>$foobar</code>' and you have
`<code>$foo</code>' on the line. Normally the completion system will perform
completion at this point. However, with <code>accept-exact</code> set,
`<code>$foo</code>' will be expanded since it matches a parameter.</li>
<li><strong><code>add-space</code></strong><br />
means add a space after the expansion, as with a successful
completion --- although directories are given a `<code>/</code>' instead. For
finer control, it can be set to the word <code>file</code>, which means the
space is only added if the expanded word matches a file that already
exists (the idea being that, if it doesn't, you may want to complete
further). Both <code>true</code> and <code>file</code> may be combined with <code>subst</code>, which
prevents the adding of a space after expanding a substitution of the
form `<code>${...}</code>' or `<code>$(...)</code>'.</li>
<li><strong><code>keep_prefix</code></strong><br />
also addresses the question of whether a `<code>~</code>' or `<code>$</code>' should be
expanded. If set, the prefix will be retained, so expanding
`<code>~/f*</code>' to `<code>~/foo</code>' doesn't turn the `<code>~</code>' into `<code>/home/pws</code>'.
The default is the value `<code>changed</code>', which is a half-way house
been <code>false</code> and <code>true</code>: it means that if there was no other change
in the word, i.e. no other possible expansion was found, the `<code>~</code>'
or `<code>$</code>' will be expanded. If the effect of this style is that the
expansion is the same as the unexpanded word, the next completer in
the list after <code>_expand</code> will be tried.</li>
<li><strong><code>suffix</code></strong><br />
is similar to <code>keep_prefix</code>. The `suffix' referred to is something
after an expression beginning `<code>~</code>' or `<code>$</code>' that wouldn't be part
of that expansion. If this style is set, and such a suffix exists,
the expansion is not performed. So, for example, `<code>~pw<TAB></code>' can
be expanded to `<code>~pws</code>', but `<code>~pw/</code>' is not eligible for
expansion; likewise `<code>$fo</code>' and `<code>$fo/</code>'. This style defaults to
<code>true</code> --- so if you want <code>_expand</code> always to expand such
expressions, you will need to set it to <code>false</code> yourself.</li>
</ul>
<p>An easier way of getting the sort of control over expansion which the
<code>_expand</code> completer provides is with the <code>_expand_word</code> function,
usually bound to <code>\C-xe</code>, which does all the things described above
without getting mixed up with the other completers. In this case the
context string starts `<code>:completion:expand-word</code>', so you can have
different styles for this than for the <code>_expand</code> completer.</p>
<p>Setting different priorities for expansion is one good use for completer
labels, for example</p>
<pre><code> zstyle ':completion:*' completer _expand:-glob _expand:-subst
zstyle ':completion:*:expand-glob:*' glob yes
zstyle ':completion:*:expand-subst:*' substitute yes
</code></pre>
<p>is the basic set up to make <code>_expand</code> try glob completions and failing
that do substitutions, presenting the results as an expansion. You would
almost certainly want to add details to help this along.</p>
<p><strong><code>_history</code></strong></p>
<p>This completes words from the shell's history, in other words everything
you typed or had completed or expanded on previous lines. There are
three styles that affect it, <code>sort</code> and <code>remove-all-dups</code>; they are
described for the command widget <code>_history_complete_word</code> below. That
widget essentially performs the work of this completer as a special
keystroke.</p>
<p><strong><code>_prefix</code></strong></p>
<p>Strictly, this completer doesn't do completion itself, and should hence
be in the group below starting with <code>_match</code>. However, it <em>seems</em> to do
completion... Let me explain.</p>
<p>Many shells including zsh have the facility to complete only the word
before the cursor, which zsh completion jargon refers to as the
`prefix'. I explained this above when I talked about
<code>expand-or-complete-prefix</code>; when you use that instead of the normal
completion functions, the word as it's finally completed looks like
`<code><prefix><completion><suffix></code>' where the completion has changed
`<code><prefix></code>' to `<code><prefix><completion></code>', ignoring <code><suffix></code>
throughout.</p>
<p>The <code>_prefix</code> completer lets you do this as part of normal completion.
What happens is that the completers are evaluated as normal, from left
to right, until a completion is found. If <code>_prefix</code> is reached,
completion is then attempted just on the prefix. So if your completers
are `<code>_complete _prefix</code>', the shell will first try completion on the
whole word, prefix and suffix, then just on the prefix. Only the first
`real' completer (<code>_complete</code>, <code>_approximate</code>, <code>_correct</code>, <code>_expand</code>,
<code>_history</code>) is used.</p>
<p>You can try prefix completion more than once simply by including
<code>_prefix</code> more than once in the list of completers; the second time, it
will try the second `real' completer in the list; so if they are
`<code>_complete _prefix _correct _prefix</code>', you will get first ordinary
completion, then the same for the prefix only, then ordinary correction,
then the same for the prefix only. You can move either of the <code>_prefix</code>
completers to the point in the sequence where you want the prefix-only
version to be tried.</p>
<p>The <code>_prefix</code> completer will re-look up the <code>completer</code> style. This
means that you can use a non-default set of completers for use just with
<code>_prefix</code>. Here, as described in the manual, is how to force <code>_prefix</code>
only to be used as a last resort, and only with normal completion:</p>
<pre><code> zstyle ':completion:::::' completer _complete \
<other-completers> _prefix
zstyle ':completion::prefix:::' completer _complete
</code></pre>
<p>The full contexts are shown, just to emphasise the form; as always, you
can use wildcards if you don't care. In a case like this, you can use
<em>only</em> <code>_prefix</code> as the completer, and completion including the suffix
would never be tried; you then have to make sure you have the
<code>completer</code> style for the <code>prefix</code> context, however, or no completion at
all will be done.</p>
<p>The completer labelling trick is again useful here: you can call
<code>_prefix</code> more than once, wherever you choose in your list of
completers, and force it to look up in a different context each time.</p>
<pre><code> zstyle ':completion:*' completer _complete _prefix:-complete \
_approximate _prefix:-approximate
zstyle ':completion:*:prefix-complete:*' completer _complete
zstyle ':completion:*:prefix-approximate:*' completer _approximate
</code></pre>
<p>This tries ordinary completion, then the same for the prefix only, then
approximation, then the same for the prefix only. As mentioned in the
previous paragraph, it is perfectly legitimate to leave out the raw
<code>_complete</code> and <code>_approximate</code> completers and just use the forms with
the <code>_prefix</code> prefix.</p>
<p>One gotcha with the <code>_prefix</code> completer: you have to make sure the
option <code>COMPLETE_IN_WORD</code> is set. That may sound counter-intuitive:
after all, <code>_prefix</code> forces completion <em>not</em> to complete inside a word.
The point is that without that option, completion is only ever tried at
the end of the word, so when you type <code><TAB></code> in the middle of
<code><prefix><suffix></code>, the cursor is moved to after the end of the suffix
before the completion system has a chance to see what's there, and hence
the whole thing is regarded as a prefix, with no suffix.</p>
<p>There's one more style used with <code>_prefix</code>: `<code>add-space</code>'. This makes
<code>_prefix</code> add a real, live space when it completes the prefix, instead
of just pretending there was one there, hence separating the completed
word from the original suffix; otherwise it would simply leave the
resulting word all joined together, as <code>expand-or-complete-prefix</code>
usually does.</p>
<p><strong><code>_ignored</code></strong></p>
<p>Like <code>_prefix</code> this is a bit of a hybrid, mopping up after completions
which have already been generated. It allows you to have completions
which have already been rejected by the style `<code>ignored-patterns</code>'.
I'll describe that below, but it's effect is very simple: for the
context given, the list of patterns you specify are matched against
possible completions, and any that match are removed from the list. The
<code>_ignored</code> completer allows you to retrieve those removed completions
later in your completer list, in case nothing else matched.</p>
<p>This is used by the <code>$fignore</code> mechanism --- a list of suffixes of files
not normally to be completed --- which is actually built on top of
<code>ignored-patterns</code>, so if you use that in the way familiar to current
zsh users, where the ignored matches are shown if there are no unignored
matches, you need the <code>_ignored</code> completer in your completer list.</p>
<p>One slightly annoying feature with <code>_ignored</code> is if there is only a
single possible completion, since it will then be unconditionally
inserted. Hardly a surprise, but it can be annoying if you really don't
want that choice. There is a style <code>single-ignored</code> which you can set to
<code>show</code> --- just show the single ignored match, don't insert it --- or to
<code>menu</code> --- go to menu completion so that TAB cycles you between the
completion which <code>_ignored</code> produced and what you originally typed. The
latter gives a very natural way of handling ignored files; it's sort of
saying `well, I found this but you might not like it, so hit tab again
if you want to go back to what you had before'.</p>
<p>I said this was like <code>_prefix</code>, and indeed you can specify which
completers are called for the <code>_ignored</code> completer in just the same way,
by giving the <code>completer</code> style in the context
`<code>:completion:*:ignored:*</code>'. That means my description has been a
little over-simplified: <code>_ignored</code> doesn't really use the completions
which were ignored before; rather, when it's called it generates a list
of possibilities where the choices matched by <code>ignore-patterns</code> --- or
internally using <code>$fignore</code> --- are not ignored. So it should really be
called `<code>_not_ignored</code>', but it isn't.</p>
<p><strong><code>_match</code></strong></p>
<p>This and the remaining completers are utilities, which affect the main
completers given above when put into the completion list rather than
doing completion themselves.</p>
<p>The <code>_match</code> completer should appear <em>after</em> <code>_complete</code>; it is a more
flexible form of the <code>GLOB_COMPLETE</code> option. In other words, if
<code>_complete</code> didn't succeed, it will try to match the word on the line as
a pattern, not just a fixed string, against the possible completions. To
make it work like normal completion, it usually acts as if a `<code>*</code>' was
inserted at the cursor position, even if the word already contains
wildcards.</p>
<p>You can control the addition of `<code>*</code>' with the `<code>match-original</code>'
style; the normal behaviour occurs if this is unset. If it is set to
`<code>only</code>', the `<code>*</code>' is not inserted, and if it is `<code>true</code>', or
actually any other string, it will try first without the `<code>*</code>', then
with. For example, consider typing `<code>setopt c*ect<TAB></code>' with the
<code>_match</code> completer in use. Normally this will produce two possibilities,
`<code>correct</code>' and `<code>correctall</code>'. After setting the style,</p>
<pre><code> zstyle ':completion::match:*' original only
</code></pre>
<p>no `<code>*</code>' would be inserted at the place where you hit `<code>TAB</code>', so that
`<code>correct</code>' is the only possible match.</p>
<p>The <code>_match</code> completer uses the style <code>insert-unambiguous</code> in just the
same way as does <code>_approximate</code>.</p>
<p><strong><code>_all_matches</code></strong></p>
<p>This has a similar effect to performing expansion instead of completion:
all the possible completions are inserted onto the command line.
However, it uses the results of ordinary contextual completion to
achieve this. The normal way that the completion system achieves this is
by influencing the behaviour of any subsequent completers which are
called --- hence you will need to put <code>_all_matches</code> in the list of
completers before any which you would like to have this behaviour.</p>
<p>You're unlikely to want to do this with every type of completion, so
there are two ways of limiting its effect. First, there is the
<code>avoid-completer</code> style: you can set this to a list of completers which
should <em>not</em> insert all matches, and they will be handled normally.</p>
<p>Then there is the style <code>old-matches</code>. This forces <code>_all_matches</code> to use
an existing list of matches, if it exists, rather than what would be
generated this time round. You can set the style to <code>only</code> instead of
true; in this case <code>_all_matches</code> will never apply to the completions
which would be generated this time round, it will only use whatever list
of completions already exists.</p>
<p>This can be a nuisance if applied to normal completion generation ---
the usual list would never be generated, since <code>_all_matches</code> would just
insert the non-existent list from last time --- so the manual recommends
two other ways of using the completer with this style. First, you can
add a condition to the use of the style:</p>
<pre><code> zstyle -e ':completion:*' old-matches 'reply=(${NUMERIC:-false})'
</code></pre>
<p>This returns false unless there is a non-zero numeric argument; if you
type <code><ESC>1</code> in emacs mode, or just <code>1</code> in vi mode, before completion,
it will insert all the values generated by the immediately preceding
completion.</p>
<p>Otherwise, you can bind <code>_all_matches</code> separately. This is probably the
more useful; copying the manual entry:</p>
<pre><code> zle -C all-matches complete-word _generic
bindkey '^Xa' all-matches
zstyle ':completion:all-matches:*' completer _all_matches
zstyle ':completion:all-matches:*' old-matches only
</code></pre>
<p>Here we generate ourselves a new completion based on the <code>complete-word</code>
widget, called <code>all-matches</code> --- this name is arbitrary but convenient.
We bind that to the keystroke <code>^Xa</code>, and give it two special styles
which normal completion won't see. For the <code>completer</code> we set just
<code>_all_matches</code>, and for <code>old-matches</code> we set <code>only</code>; the effect is that
<code>^Xa</code> will only ever have the effect of inserting all the completions
which were generated by the last completion, whatever that was --- it
does not have to be an ordinary contextual completion, it may be the
result of any completion widget.</p>
<p><strong><code>_list</code></strong></p>
<p>If you have this in the list of completers (at the beginning is as good
as anything), then the first time you try completion, you only get a
list; nothing changes, not even a common prefix is inserted. The second
time, completion continues as normal. This is like typing <code>^D</code>, then
tab, but using just the one key. This differs from the usual <code>AUTO_LIST</code>
behaviour in that is entirely irrespective of whether the completion is
ambiguous; you always get the list the first time, and it always does
completion in the usual way the second time.</p>
<p>The <code>_list</code> completer also uses the <code>condition</code> style, which works a bit
like the styles for the <code>_expand</code> completer: it must be set to one of
the values corresponding to `true' for the <code>_list</code> delaying behaviour
to take effect. You can test for a particular value of <code>$NUMERIC</code> or any
other condition by using the <code>-e</code> option of <code>zstyle</code> when defining the
style.</p>
<p>Finally, the boolean style <code>word</code> is also relevant. If false or unset,
<code>_list</code> examines the whole line when deciding if it has changed, and
hence completion should be delayed until the next keypress. If true, it
just examines the current word. Note that <code>_list</code> has no knowledge of
what happens between those completion calls; looking at the command line
is its only resource.</p>
<p><strong><code>_menu</code></strong></p>
<p>This just implements menu completion in shell code; it should come
before the `real' completion generators in the <code>completers</code> style. It
ignores the <code>MENU_COMPLETION</code> option and other related options and the
normal menu-completion widgets don't work well with it. However, you can
copy it and write your own completers.</p>
<p><strong><code>_oldlist</code></strong></p>
<p>This completer is most useful when you are in the habit of using special
completion functions, i.e. commands other than the standard completion
system. It is able to hang onto an old completion list which would
otherwise be replaced with a newly generated one. There are two aspects
to this.</p>
<p>First, listing. Suppose you try to complete something from the shell
history, using the command bound to `<code>ESC-/</code>'. For example, I typed
`<code>echo ma<ESC-/></code>' and got `<code>max-errors</code>'. At this point you might
want to list the possible completions. Unfortunately, if you type <code>^D</code>,
it will simply list all the usual contextual completions --- for the
<code>echo</code> command, which is not handled specially, these are simply files.
So it doesn't work. By putting the <code>_oldlist</code> completer into the
<code>completers</code> style <em>before</em> <code>_complete</code>, it does work, because the old
list of matches is kept for <code>^D</code> to use.</p>
<p>In this case, you can force old-listing on or off by setting the
<code>old-list</code> style to <code>always</code> or <code>never</code>; usually it shows the listing
for the current set of completions if that isn't already displayed, and
otherwise generates the standard listing. You can even set the value of
<code>old-list</code> to a list of completers which will always have their list
kept in this way.</p>
<p>The other place where <code>_oldlist</code> is useful is in menu completion, where
exactly the same problem occurs: if you generate a menu from a special
command, then try to cycle through by hitting tab, completion will look
for normal contextual matches instead. There's a way round this time ---
use the special command key repeatedly instead of tab. This is rather
tedious with multiple key sequences. Again, <code>_oldlist</code> cures this, and
again you can control the behaviour with a style, <code>old-menu</code>, which
takes a boolean value (it is on by default). As Orwell put it,
oldlisters unbellyfeel menucomp.</p>
<p><strong>Ordering completers</strong></p>
<p>I've given various suggestions about the order in which completers
should come in, which might be confusing. Here, therefore, is a
suggested order; just miss out any completers you don't want to use:</p>
<pre><code> _all_matches _list _oldlist _menu _expand _complete _match
_ignored _correct _approximate _prefix
</code></pre>
<p>Other orders are certainly possible and maybe even useful: for example,
the <code>_all_matches</code> completer applies to all the completers following not
listed in the <code>avoid-completer</code> style, so you might have good reason to
shift it further down the list.</p>
<p>Here's my example of labels for completers, which I mentioned just above
the list of different completers, whereby completers can be looked up
under different names.</p>
<pre><code> zstyle ':completion:*' completer _complete _approximate:-one \
_complete:-extended _approximate:-four
zstyle ':completion:*:approximate-one:*' max-errors 1
zstyle ':completion:*:complete-extended:*' \
matcher 'r:|[.,_-]=* r:|=*'
zstyle ':completion:*:approximate-four:*' max-errors 4
</code></pre>
<p>This tries the following in order.</p>
<ol>
<li>Ordinary, no-frills completion.</li>
<li>Approximation with one error, as given by the second style.</li>
<li>Ordinary completion with extended completion turned on, as given by
the third style. Sorry, this will be a black box until I talk about
the <code>matcher</code> style later on; for now, you'll just have to take my
word for it that this style allows the characters in the square
brackets to have a wildcard in front, so `<code>a-b</code>' can complete to
`<code>able-baker</code>', and so on.</li>
<li>Approximation with up to four errors, as given by the final style.</li>
</ol>
<p>Here's a rather bogus example. You have a directory containing:</p>
<pre><code> foobar fortified-badger frightfully-barbaric
</code></pre>
<p>Actually, it's not bogus at all, since I just created one. First try
`<code>echo foo<TAB></code>'; no surprise, you get `<code>foobar</code>'. Now try completing
with `<code>fo-b<TAB></code>' after the `<code>echo</code>': basic completion fails, it gets
to `_approximate:-one' and finds that it's allowed one error, so
accepts the completion `<code>foobar</code>' again. Now try `<code>fort-ba<TAB></code>'.
This time nothing kicks in until the third completion, which effectively
allows it to match `<code>fort*-ba*<TAB></code>', so you see `<code>fortified-badger</code>'
(no, I've never seen one myself, but they're nocturnal, you know).
Finally, try `<code>fortfully-ba<TAB></code>'; the last entry, which allows up to
four errors, thoughtfully corrects `<code>or</code>' to `<code>righ</code>', and you get
`<code>frightfully-barbaric</code>'. All right, the example is somewhat unhinged,
but I think you can see the features are useful. If it makes you feel
better, it took me four or five attempts to get the styles right for
this.</p>
<p><span id="l158"></span></p>
<h3 id="652-changing-the-format-of-listings-groups-etc"><a class="header" href="#652-changing-the-format-of-listings-groups-etc">6.5.2: Changing the format of listings: groups etc.</a></h3>
<p><strong><code>format</code></strong></p>
<p>You can use this style if you want to find out where the completions in
a completion listing come from. The most basic use is to set it for the
<code>descriptions</code> tag in any completion context. It takes a string value in
which `<code>%d</code>' should appear; this will be replaced by a description of
whatever is being completed. For example, I use:</p>
<pre><code> zstyle ':completion:*:descriptions' format 'Completing %d'
</code></pre>
<p>and if I type <code>cd^D</code>, I see a listing like this (until I define the
<code>group-name</code> style, that is):</p>
<pre><code> Completing external command
Completing builtin command
Completing shell function
cd cddbsubmit cdp cdrecord
cdctrl cdecl cdparanoia cdswap
cdda2wav cdmatch cdparanoia-yaf
cddaslave cdmatch.newer cdplay
cddbslave cdot cdplayer_applet
</code></pre>
<p>The descriptions at the top are related to the tag names --- usually
there's a unique correspondence --- but are in a more readable form; to
get the tag names, you need to use <code>^Xh</code>. You will no doubt see
something different, but the point is that the completions listed are a
mixture of external commands (e.g. <code>cdplay</code>), builtin commands (<code>cd</code>)
and shell functions (<code>cdmatch</code>, which happens to be a leftover from
old-style completion, showing you how often I clean out my function
directory), and it's often quite handy to know what you have.</p>
<p>You can use some prompt escapes in the description, specifically those
that turn on or off standout mode (`<code>%S</code>', `<code>%s</code>'), bold text
(`<code>%B</code>', `<code>%b</code>'), and underlined text (`<code>%U</code>', `<code>%u</code>'), to make the
descriptions stand out from the completion lists.</p>
<p>You can set this for some other tag than <code>descriptions</code> and the format
thus defined will be used only for completions of that tag.</p>
<p><strong><code>group-name</code>, <code>group-order</code></strong></p>
<p>In the <code>format</code> example just above, you may have wondered if it is
possible to make the different types of completion appear separately,
together with the description. You can do this using <em>groups</em>. They are
also related to tags, although as you can define group names via the
<code>group-name</code> style it is possible to give different names for completion
in any context. However, to start off with it is easiest to give the
value of the style an empty string, which means that group names are
just the names of the tags. In other words,</p>
<pre><code> zstyle ':completion:*' group-name ''
</code></pre>
<p>assigns a different group name for each tag. Later, you can fine-tune
this with more specific patterns, if you decide you want various tags to
have the same group name. If no group name is defined, the group used is
called `<code>-default-</code>', so this is what was happening before you issued
the <code>zstyle</code> command above; all matches were in that group.</p>
<p>The reason for groups is this: matches in the same group are shown
together, matches in different groups are shown separately. So the
completion list from the previous example, with both the <code>format</code> and
<code>group-name</code> styles set, becomes:</p>
<pre><code> Completing external command
cdctrl cddbsubmit cdparanoia cdrecord
cdda2wav cdecl cdparanoia-yaf
cddaslave cdot cdplay
cddbslave cdp cdplayer_applet
Completing builtin command
cd
Completing shell function
cdmatch cdmatch.newer cdswap
</code></pre>
<p>which you may find more helpful, or you may find messier, depending on
deep psychological factors outside my control.</p>
<p>If (and only if) you are using <code>group-name</code>, you can also use
<code>group-order</code>. As its name suggests, it determines the order in which
the different completion groups are displayed. It's a little like
<code>tag-order</code>, which I described when tags were first introduced: the
value is just a set of names of groups, in the order you want to see
them. The example from the manual is relevant to the listing I just
showed:</p>
<pre><code> zstyle ':completion:*:-command-' group-order \
builtins functions commands
</code></pre>
<p>--- remember that the `<code>-command-</code>' context is used when the names of
commands, rather than their arguments, are being completed. Not
surprisingly, that listing now becomes:</p>
<pre><code> Completing builtin command
cd
Completing shell function
cdmatch cdmatch.newer cdswap
Completing external command
cdctrl cddbsubmit cdparanoia cdrecord
cdda2wav cdecl cdparanoia-yaf
cddaslave cdot cdplay
cddbslave cdp cdplayer_applet
</code></pre>
<p>and if you investigate the tags available by using <code>^Xh</code>, you'll see
that there are others such as aliases whose order we haven't defined.
These appear after the ones for which you have defined the order and in
some order decided by the function which generated the matches.</p>
<p><strong><code>tag-order</code></strong></p>
<p>As I already said, I've already described this, but it's here again for
completeness.</p>
<p><strong><code>verbose</code>, <code>auto-description</code></strong></p>
<p>These are relatives of <code>format</code> as they add helpful messages to the
listing. If <code>verbose</code> is true, the function generating the matches may,
at its discretion, decide to show more information about them. The most
common case is when describing options; the standard function
<code>_describe</code> that handles descriptions for a whole lot of options tests
the <code>verbose</code> style and will print information about the options it is
completing.</p>
<p>You can also set the string style <code>auto-description</code>; it too is useful
for options, in the case that they don't have a special description, but
they do have a single following argument, which completion already knows
about. Then the description of the argument for verbose printing will be
available as `<code>%d</code>' in <code>auto-describe</code>, so that something like the
manual recommendation `<code>specify: %d</code>' will document the option itself.
So if a command takes `<code>-o <output-file></code>' and the argument has the
description `<code>output file</code>', the `<code>-o</code>', when it appears as a possible
completion, will have the description `<code>specify: output file</code>' if it
does not have its own description. In fact, most options recognized by
the standard completion functions already have their own descriptions
supplied, and this is more subtlety than most people will probably need.</p>
<p><strong><code>list-colors</code></strong></p>
<p>This is used to display lists of matches for files in different colours
depending on the file type. It is based on the syntax of the
<code>$LS_COLORS</code> environment variable, used by the GNU version of <code>ls</code>. You
will need a terminal which is capable of displaying colour such as a
colour xterm, and should make sure the <code>zsh/complist</code> library is loaded,
(it should be automatically if you are using menu selection set up with
the <code>menu</code> style, or if you use this style). But you can make sure
explicitly:</p>
<pre><code> zmodload -i zsh/complist
</code></pre>
<p>The <code>-i</code> keeps it quiet if the module was already loaded. To install a
standard set of default colours, you can use:</p>
<pre><code> zstyle ':completion:*' list-colors ''
</code></pre>
<p>--- note the use of the `<code>default</code>' tag --- since a null string sets
the value to the default.</p>
<p>If that's not good enough for you, here are some more detailed
instructions. The parameter <code>$ZLS_COLORS</code> is the lowest-level part of
the system used by <code>zsh/complist</code>. There is a simple builtin default,
while having the style set to the empty string is equivalent to:</p>
<pre><code> ZLS_COLORS="no=00:fi=00:di=01;34:ln=01;36:\
pi=40;33:so=01;35:bd=40;33;01:cd=40;33;01:\
ex=01;32:lc=\e[:rm=m:tc=00:sp=00:ma=07:hi=00:du=00
</code></pre>
<p>It has essentially the same format as <code>$LS_COLORS</code>, and indeed you can
get a more useful set of values by using the <code>dircolors</code> command which
comes with <code>ls</code>:</p>
<pre><code> ZLS_COLORS="no=00:fi=00:di=01;34:ln=01;36:\
pi=40;33:so=01;35:do=01;35:bd=40;33;01:cd=40;33;01:\
or=40;31;01:ex=01;32:*.tar=01;31:*.tgz=01;31:\
*.arj=01;31:*.taz=01;31:*.lzh=01;31:*.zip=01;31:\
*.z=01;31:*.Z=01;31:*.gz=01;31:*.deb=01;31:\
*.jpg=01;35:*.gif=01;35:*.bmp=01;35:*.ppm=01;35:\
*.tga=01;35:*.xbm=01;35:*.xpm=01;35:*.tif=01;35:\
*.mpg=01;37:*.avi=01;37:*.gl=01;37:*.dl=01;37:"
</code></pre>
<p>You should see the manual for the <code>zsh/complist</code> module for details, but
note in particular the addition of the type `<code>ma</code>', which specifies how
the current match in menu selection is displayed. The default for that
is to use standout mode --- the same effect as the sequence <code>%S</code> in a
prompt, which you can display with `<code>print -P %Sfoo</code>'.</p>
<p>However, you need to define the style directly, since the completion
always uses that to set <code>$ZLS_COLORS</code>; otherwise it doesn't know whether
the value it has found has come from the user or is a previous value
taken from some style. That takes this format:</p>
<pre><code> zstyle ':completion:*' list-colors "no=00" "fi=00" ...
</code></pre>
<p>You can use an already defined <code>$LS_COLORS</code>:</p>
<pre><code> zstyle ':completion:*' list-colors ${(s.:.)LS_COLORS}
</code></pre>
<p>(which splits the parameter to an array on colons) as <code>$LS_COLORS</code> is
still useful for <code>ls</code>, even though it's not worth setting <code>$ZLS_COLORS</code>
directly. This should mean GNU ls and zsh produce similar-looking lists.</p>
<p>There are some special effects allowed. You can use patterns to tell how
filenames are matched: that's part of the default behaviour, in fact,
for example '*.tar=01;31' forces tar files to be coloured red. In that
case, you are limited to `<code>*</code>' followed by a string. However, there's a
way of specifying colouring for any match, not just files, and for any
pattern: use <code>=<pat>=<col></code>. Here are two ways of getting jobs coloured
red in process listings for the `<code>kill</code>' command.</p>
<pre><code> zstyle ':completion:*:*:kill:*' list-colors '=%*=01;31'
</code></pre>
<p>This uses the method just described; jobs begin with `<code>%</code>'.</p>
<pre><code> zstyle ':completion:*:*:kill:*:jobs' list-colors 'no=01;31'
</code></pre>
<p>This uses the tag, rather than the pattern, to match the jobs lines. It
has various advantages. Because you are using the tag, it's much easier
to alter this for all commands using jobs, not just kill --- just miss
out `<code>kill</code>' from the string. That wasn't practical with the other
method because it would have matched too many other things you didn't
want. You're not dependent on using a particular pattern, either. And
finally, if you try it with a `<code>format</code>' description you'll see that
that gets the colour, too, since it matched the correct tag. Note the
use of the `<code>no</code>' to specify that this is to apply for a normal match;
the other two-letter codes for file types aren't useful here.</p>
<p>However, there is one even more special effect you can use with the
general pattern form. By turning on `backreferences' with `<code>(#b)</code>'
inside the pattern, parentheses are active and the bits they match can
be coloured separately. You do this by extending the list of colours,
each code preceded by an `<code>=</code>' sign, and the extra elements will be
used to colour what the parenthesis matched. Here's another example for
`<code>kill</code>', which turns the process number red, but leaves the rest
alone.</p>
<pre><code> zstyle ':completion:*:*:kill:*:processes' list-colors \
'=(#b) #([0-9]#)*=0=01;31'
</code></pre>
<p>The hieroglyphics are extended globbing patterns. You should note that
the <code>EXTENDED_GLOB</code> option is always on inside styles --- it's required
for the `<code>#b</code>' to take effect. In particular, `<code>#</code>' means `zero or
more repetitions of the previous bit of the pattern' with extended glob
patterns; see the globbing manual page for full details.</p>
<p><strong><code>ignored-patterns</code></strong></p>
<p>Many shells, including zsh, have a parameter <code>$fignore</code>, which gives a
list of suffixes; filenames ending in any of these are not to be used in
completion. A typical value is:</p>
<pre><code> fignore=(.o \~ .dvi)
</code></pre>
<p>so that normal file completion will not produce object files, EMACS
backup files, or TeX DVI files.</p>
<p>The <code>ignored-patterns</code> style is an extension of this. It takes an array
value, like <code>fignore</code>, but with various differences. Firstly, these
values are patterns which should match the <em>whole</em> value to be
completed, including prefixes (such as the directory part of a filename)
as well as suffixes. Secondly, they apply to <em>all</em> completions, not just
files, since you can use the style mechanism to tune it to apply
wherever you want, down to particular tags.</p>
<p>Hence you can replace the use of <code>$fignore</code> above with the following:</p>
<pre><code> zstyle ':completion:*:files' ignored-patterns '*?.o' '*?~' '*?.dvi'
</code></pre>
<p>for completion contexts where the tag `<code>files</code>' is in use. The extra
`<code>?</code>'s are because <code>$fignore</code> was careful only to apply to real
suffixes, i.e. strings which had something in front of them, and the
`<code>?</code>' forces there to be at least one character present.</p>
<p>Actually, this isn't quite the same as <code>$fignore</code>, since there are other
file tags than <code>files</code>; apart from those for directories, which you've
already met, there are <code>globbed-files</code> and <code>all-files</code>. The former is
for cases where a pattern is specified by the completion function, for
example `<code>*.dvi</code>' for files following the command name <code>dvips</code>. These
don't use this style, because the pattern was already sufficiently
specified. This follows the behaviour for <code>$fignore</code> in the old
completion system. Another slight difference, as I said above when
discussing the <code>_ignored</code> completer, is that you get to choose whether
you want to see those ignored files if the normal completions fail, by
having <code>_ignored</code> in the completer list or not.</p>
<p>The other tag, <code>all-files</code>, applies when a <code>globbed-files</code> tag failed,
and says any old file is good enough in that case; you can arrange how
this happens with the <code>tag-order</code> style. In this example,</p>
<pre><code> zstyle ':completion:*:*:dvips:argument*' \
tag-order globbed-files all-files
</code></pre>
<p>is enough to say that you want to see all files if no files were
produced from the pattern, i.e. if there were no `<code>*.dvi</code>' files in the
directory. Finally the point of this ramble: as the <code>all-files</code> tag is
separate from the <code>files</code> tag, in this case you really would see all
files (except for those beginning with a `<code>.</code>', as usual). You might
find this useful, but you can easily make the <code>all-files</code> tag behave the
same way as <code>files</code>:</p>
<pre><code> zstyle ':completion:*:(all-|)files' ignored-patterns ...
</code></pre>
<p>Here's the example of using tag labels I promised earlier; it's simply
taken from the manual. To refresh your memory: tag labels are a way of
saying that tags should be looked up under a different name. Here we'll
do:</p>
<pre><code> zstyle ':completion:*:*:-command-:*' tag-order 'functions:-non-comp'
</code></pre>
<p>This applies in command position, from the special `<code>-command-</code>'
context, the place where functions occur most often, along with other
types of command which have their own tags. This says that when
functions are first looked up, they are to be looked up with the name
`<code>functions-non-comp</code>' --- remember that with a hyphen as the first
character of the label part, the bit after the colon, the <code>functions</code>
tag name itself, the bit before the colon, is to be stuck in front to
give the full label name `<code>functions-non-comp</code>'. We can use it as
follows:</p>
<pre><code> zstyle ':completion:*:functions-non-comp' ignored-patterns '_*'
</code></pre>
<p>In the context of this tag label, we have told completion to ignore any
patterns --- i.e. any function names --- beginning with an underscore.
What happens is this: when we try completion in command position,
<code>tag-order</code> is looked up and finds we want to try functions first, but
under the name <code>functions-non-comp</code>; this completes functions apart from
ones beginning with an underscore (presumably completion functions you
don't want to run interactively). Since <code>tag-order</code> normally tries all
the other tags, unless it was told not to, in this case all the normal
command completions will appear, including functions under their normal
tag name, so this just acts as a sort of filter for the first attempt at
completion. This is typically what tag labels are intended for ---
though maybe you can think up a lot of other uses, since the idea is
quite powerful, being backed up by the style mechanism.</p>
<p>You way wonder why you would want to ignore such functions at this
point. After all, you're only likely to be doing completion when you've
already typed the first character, which either is `<code>_</code>' or it isn't.
It becomes useful with correction and approximation --- particularly
since many completion functions are similar to the names of the commands
for which they handle completion. You don't want to be offered
`<code>_zmodload</code>' as a completion if you really want `<code>zmodload</code>'. The
combination of labels and ignored patterns does this for you.</p>
<p>You can generalise this using another feature: tags can actually be
patterns, which I mentioned but didn't demonstrate. Here's a more
sophisticated version of the previous example, adapted from the manual:</p>
<pre><code> zstyle ':completion:*:*:-command-:*' tag-order \
'functions:-non-comp:non-completion\ functions *' functions
</code></pre>
<p>It's enhanced so that completion tries all other possible tags at the
same time as the labelled <code>functions</code>. However, it only ever tries a tag
once at each step, so the `<code>*</code>' doesn't put back <code>functions</code> as you
might expect --- that's still tried under the label
`<code>functions-non-comp</code>', and the <code>ignored-patterns</code> style we set will
still work. In the final word, we try all possible functions, so that
those beginning with an underscore will be restored.</p>
<p>Use of the `<code>_ignored</code>' completer can allow you to play tricks without
having to label your tags:</p>
<pre><code> zstyle ':completion:*' completer _complete _ignored
zstyle ':completion:*:functions' ignored-patterns '_*'
</code></pre>
<p>Now anywhere the <code>functions</code> tag is valid, functions matching `<code>_*</code>'
aren't shown until completion reaches the `<code>_ignored</code>' in the completer
list. Of course, you should manipulate the completer list the way you
want; this just shows the bare bones.</p>
<p><strong><code>prefix-hidden</code>, <code>prefix-needed</code></strong></p>
<p>You will know that when the shell lists matches for files, the directory
part is removed. The boolean style <code>prefix-hidden</code> extends this idea to
various other types of matches. The prefixes referred to are not just
any old common prefix to matches, but only some places defined in the
completion system: the <code>-</code> prefix to options, the `<code>%</code>' prefix to jobs,
the <code>-</code> or <code>+</code> prefix to directory stack entries are the most commonly
used.</p>
<p>The <code>prefix-needed</code> applies not to listings, but instead to what the
user types on the command line. It says that matches will only be
generated if the user has typed the prefix common to them. It applies on
broadly the same occasions as <code>prefix-hidden</code>.</p>
<p><strong><code>list-packed</code>, <code>list-rows-first</code>, <code>accept-exact</code>, <code>last-prompt</code>,
<code>menu</code></strong></p>
<p>The first two of these have already been introduced, and correspond to
the <code>LIST_PACKED</code> and <code>LIST_ROWS_FIRST</code> options. The <code>accept-exact</code> and
<code>last-prompt</code> styles correspond essentially to the <code>REC_EXACT</code> and
<code>ALWAYS_LAST_PROMPT</code> options in the same way.</p>
<p>The style <code>menu</code> roughly corresponds to the <code>MENU_COMPLETE</code> option, but
there is also the business of deciding whether to use menu selection, as
described above. These two uses don't interfere with each other ---
except that, as I explained, menu completion must be started to use menu
selection --- so a value like `<code>true select=6</code>' is valid; it turns on
menu completion for the context, and also activates menu selection if
there are at least 6 choices.</p>
<p>There are some other, slightly more obscure, choices for <code>menu</code>:</p>
<ul>
<li><strong><code>yes=</code><em>num</em></strong><br />
turn on menu completion only if there are at least <em>num</em> matches;</li>
<li><strong><code>no=</code><em>num</em></strong><br />
turn off menu completion if there are as many as <em>num</em> matches;</li>
<li><strong><code>yes=long</code></strong><br />
turn on menu completion if the list does not fit on the screen, and
completion was attempted;</li>
<li><strong><code>yes=long-list</code></strong><br />
the same, but do it even if listing, not completion, was attempted;</li>
<li><strong><code>select=long</code></strong><br />
like <code>yes=long</code>, but this time turn on menu selection, too;</li>
<li><strong><code>select=long-list</code></strong><br />
like <code>yes=long-list</code>, but turn on menu selection, too.</li>
</ul>
<p>In case your eyes glazed over before the end, here's a full description
of the last one, <code>select=long-list</code>, which is quite useful: if you are
attempting completion or even just listing completions, and the list of
matches would be too long to fit on the screen, then menu selection is
turned on, so that you can use the cursor keys (and other selection
keys) to move up and down the list. Generally, the above possibilities
can be combined, unless the combined effect wouldn't work.</p>
<p>As always, <code>yes</code> and <code>true</code> are equivalent, as are <code>no</code> and <code>false</code>. It
just hurts the eyes of programmers to read something which appears to
assign a value to <code>true</code>.</p>
<p><strong><code>hidden</code></strong></p>
<p>This is a little obscure for most users. Its context should be
restricted to specific tags; any corresponding matches will not be shown
in completion listings, but will be available for inserting into the
command line. If its value is `<code>true</code>', then the description for the
tag may still appear; if the value is `<code>all</code>', even that is suppressed.
If you don't want the completions even to be available for insertion,
use the <code>tag-order</code> style.</p>
<p><span id="l159"></span></p>
<h3 id="653-styles-affecting-particular-completions"><a class="header" href="#653-styles-affecting-particular-completions">6.5.3: Styles affecting particular completions</a></h3>
<p>The styles listed here are for use only with certain completions as
noted. I have not included the styles used by particular completers,
which are described with the completer in question in the subsection
`<strong>Specifying completers and their options</strong>'. I have also not
described styles used only in separate widgets that do completion; the
relevant information is all together in the next section.</p>
<p><strong>Filenames (1): patterns: <code>file-patterns</code></strong></p>
<p>It was explained above for the <code>tag-order</code> style that when a function
uses pattern matching to generate file completions, such as all <code>*.ps</code>
files or all <code>*.gz</code> files, the three tags <code>globbed-files</code>, <code>directories</code>
and <code>all-files</code> are tried, in that order.</p>
<p>The <code>file-patterns</code> style allows you to specify a pattern to override
whatever would be completed, even in what would otherwise be a simple
file completion with no pattern. Since this can easily get out of hand,
the best way of using this style is to make sure that you specify it for
a narrowly enough defined context. In particular, you probably want to
restrict it to completions for a single command and for a particular one
of the tags usually applying to files. As always, you can use <code>^Xh</code> to
find out what the context is. It has a labelling mechanism --- you can
specify a tag with a pattern for use in looking up other styles. Hence
`<code>*.o:object-files</code>' gives a pattern `<code>*.o</code>' and a tag name
`<code>object-files</code>' by which to refer to these.</p>
<p>The patterns you specify are tried in order; you don't need to use
<code>tag-order</code>. In fact <code>file-patterns</code> replicates its behaviour in that
you can put patterns in the same word to say they should be tried
together, before going on to the pattern(s) in the next word. Also, you
can give a description after a second colon in the same way. Indeed,
since <code>file-patterns</code> gets its hands on the tags first, any ordering
defined there can't be overridden by <code>tag-order</code>.</p>
<p>So, for example, after</p>
<pre><code> zstyle ':completion:*:*:foo:*:*' file-patterns \
'*.yo:yodl-files:yodl\ files *(-/):directories'
</code></pre>
<p>the command named `<code>foo</code>' will complete files ending in `<code>.yo</code>', as
well as directories. For once, you don't have to change the completer to
alter what's completed: `<code>foo</code>' isn't specially handled, so it causes
default completion, and that means completing files, so that
<code>file-patterns</code> is active anyway.</p>
<p>Here's a slightly enhanced example; it shows how <code>file-patterns</code> can be
used instead of <code>tag-order</code> to offer the tags in the order you want.</p>
<pre><code> zstyle ':completion:*:*:foo:*:*' file-patterns \
'*.yo:yodl-files:yodl\ files' '*(-/):directories:directories' \
'^*.yo(-^/):other-files:other\ files'
</code></pre>
<p>Completion will first try to show you only `<code>.yo</code>' files, if there are
any; otherwise it will show you directories, if there are any; otherwise
it will show you any other files: `<code>^*.yo(-^/)</code>' is an extended glob to
match any file which doesn't end in `<code>.yo</code>' and which isn't a directory
and doesn't link to a directory. As always, you can cycle through the
sets of possibilities using the `<code>_next_tag</code>' completion command.</p>
<p>Note that <code>file-patterns</code> is an exception to the general rule that
styles don't determine <em>which</em> tags are called only <em>where</em> they're
called, or what their behaviour is: this time, you actually get to
specify the set of tags which will be used. This means it doesn't use
the the standard file tags (unless you use those names yourself, of
course), just `<code>files</code>' if you don't specify one. Hence it's good style
to add the tags, following colons, although it'll work without.</p>
<p>Another thing to watch out for is that if there is already a completion
which handles a file type --- for example, if we had tried to alter the
effect of file completion for the `<code>yodl</code>' command instead of the
fictitious `<code>foo</code>' --- the results may well not be quite what you want.</p>
<p>Another feature is that `<code>%p</code>' in the pattern inserts the pattern which
would usually be used. That means that the following is essentially the
same as what file completion normally does:</p>
<pre><code> zstyle ':completion:*' file-patterns '%p:globbed-files' \
'*(-/):directories' '*:all-files'
</code></pre>
<p>You can turn completion for a command that usually doesn't use a pattern
into one that does. Another example taken from the manual:</p>
<pre><code> zstyle ':completion:*:*:rm:*:globbed-files' file-patterns \
'*.o:object-files' '%p:all-files'
</code></pre>
<p>So if there are any <code>*.o</code> files around, completion for <code>rm</code> will just
complete those, even if arguments to <code>rm</code> are otherwise found by default
file completion (which they usually are). The <code>%p</code> will use whatever
file completion normally would have; probably any file at all. You can
change this, if you like; there may be files you don't ever want
automatically completed after <code>rm</code>.</p>
<p>Remember that using explicit patterns overrides the effect of
<code>$fignore</code>; this is obviously useful with <code>rm</code>, since the files you want
to delete are often those you usually don't want to complete.</p>
<p><strong>Filenames (2): paths: <code>ambiguous</code>, <code>expand</code>, <code>file-sort</code>,
<code>special-dirs</code>, <code>ignore-parents</code>, <code>list-suffixes</code>, <code>squeeze-slashes</code></strong></p>
<p>Filename completion is powerful enough to complete all parts of a path
at once, for example `<code>/h/p/z</code>' will complete to `<code>/home/pws/zsh</code>'.
This can cause problems when the match is ambiguous; since several
components of the path may well be ambiguous, how much should the
completion system complete, and where should it leave the cursor? This
facility is associated with all these styles affecting filenames.</p>
<p>With ordinary completion, the usual answer is that the completion is
halted as soon as a path component matches more than one possibility,
and the cursor is moved to that point, with the remainder of the string
left unaltered. With menu completion, you can simply cycle through the
possibilities with the cursor moved to the end as usual. If you set the
style <code>ambiguous</code>, then the system will leave the cursor at the point of
the first ambiguity even if menu completion is in use. Note that this is
always used with the `<code>paths</code>' tag, i.e. the context ends in
`<code>...:paths</code>'.</p>
<p>The style <code>expand</code> is similar and is also applied with the `<code>paths</code>'
tag. It can include either or both of the strings <code>prefix</code> and <code>suffix</code>.
Be careful when setting both --- they have to be separate words, for
example</p>
<pre><code> zstyle ':completion:*' expand prefix suffix
</code></pre>
<p>Don't put quotes around `<code>prefix suffix</code>' as it won't work.</p>
<p>With <code>prefix</code>, <code>expand</code> tells the completion system always to expand
unambiguous prefixes in a path (such as `<code>/u/i</code>' to `<code>/usr/in</code>', which
matches both <code>/usr/include</code> and <code>/usr/info</code>) --- even if the remainder
of the string on the command line doesn't match any file. So this
expansion will now happen even if you try this on
`<code>/u/i/ALoadOfOldCodswallop</code>', which it otherwise wouldn't.</p>
<p>Including <code>suffix</code> in the value of <code>expand</code> extends path completion in
another way: it allows extra unambiguous parts to be added even after
the first ambiguous one. So if `<code>/home/p/.pr</code>' would match
`<code>/home/pws/.procmailrc</code>' or `<code>/home/patricia/.procmailrc</code>', and
nothing else, the last word would be expanded. Set up like this, you
will always get the longest unambiguous match for all parts of the path.</p>
<p>In older versions of the completion system, <code>suffix</code> wasn't used if you
had menu completion active by default, although it was if menu
completion was only started by the <code>AUTO_MENU</code> option. However, in
recent versions, the setting is always respected. This means that
setting the <code>expand</code> style to include the value <code>suffix</code> allows menu
completion to cycle through all possible completions, as if there were a
`<code>*</code>' after each part of the path, so `<code>/u/i/k</code>' will offer all
matches for `<code>/u*/i*/k*</code>'.</p>
<p>The <code>file-sort</code> style allows files to be sorted in a way other than by
alphabetical order: sorting applies both to the list of files, and to
the order in which menu completion presents them. The value should
include one of the following: `<code>size</code>', `<code>links</code>', `<code>modification</code>'
(same as `<code>time</code>', `<code>date</code>'), `<code>access</code>', `<code>inode</code>' (same as
`<code>change</code>'). These pick the obvious properties for sorting: file size,
number of hard links, modification time, access time, inode change time.
You can also add the string `<code>reverse</code>' to the value, which reverses
the order. In this case the tag is always `<code>files</code>'.</p>
<p>The <code>special-dirs</code> style controls completion of the special directories
`<code>.</code>' and `<code>..</code>'. Given that you usually need to type an initial dot
to complete anything at all beginning with one, the idea of
`completing' `<code>.</code>' is a little odd; it simply means that the directory
is accepted when the completion is started on it. You can set the style
to <code>true</code> to allow completion to both of the two, or to `<code>..</code>' to
complete `<code>..</code>' but not `<code>.</code>'. Like <code>ambiguous</code>, this is used with the
tag set to `<code>paths</code>'.</p>
<p>The style <code>ignore-parents</code> is used with the <code>files</code> tag, since it
applies to paths, but not necessarily completion of multiple path names
at once; it can be used when completing just the last element. There are
two main uses, which can be combined. The first case is to include the
string `<code>parent</code>' in the style. This means that when you complete after
(say) <code>foo/../</code>, the string <code>foo</code> won't appear as a choice, since it
already appeared in the string. Secondly, you can include `<code>pwd</code>' in
the value; this means don't complete the current working directory after
`<code>../</code>' --- you can see the sense in that: if you wanted to complete
there, you wouldn't have typed the `<code>..</code>' to get out if it.</p>
<p>Actually, the function performs both those tests on the directories in
question even if the string `<code>..</code>' itself hasn't been typed. That might
be more confusing, and you can make sure that the tests for <code>parent</code> and
<code>pwd</code> are only made when you typed the `<code>..</code>' by including a `<code>..</code>' in
the style's value. Finally, you can include the string `<code>directory</code>' in
the values: that means the tests will only be performed when directories
are being completed, while if some other sort of file, or any file, can
be completed, the special behaviour doesn't occur. You may have to read
that through a couple of times before deciding if you need it or not.</p>
<p>Next, there is <code>list-suffixes</code>. It applies when expanding out earlier
parts of the filename path, not just the last part. In this case, it is
possible that early parts of the path were ambiguous. Normally
completion stops at the point where it finds the ambiguity, and leaves
the rest of the path alone. When <code>list-suffixes</code> is set, it will list
all the possible values of all ambiguous components from the point of
ambiguity onward.</p>
<p>Lastly, there is the style <code>squeeze-slashes</code>. This is rather simpler.
You probably already know that in a UNIX filename multiple slashes are
treated just like a single slash (with a few minor exceptions on some
systems). However, path completion usually assumes that multiple slashes
mean multiple directories to be completed: `<code>//termc</code>' completes to
`<code>/etc/termcap</code>' because of this rule. If you want to stick with the
ordinary UNIX rule you can set <code>squeeze-slashes</code> to <code>true</code>. Then in this
example only files in the root directory will be completed.</p>
<p><strong>Processes: <code>command</code>, <code>insert-ids</code></strong></p>
<p>Some functions, such as <code>kill</code>, take process IDs (i.e. numbers) as
arguments. These can be completed by using the <code>ps</code> command to generate
the process numbers. The <code>command</code> style allows you to specify which
arguments are to be passed to <code>ps</code> to generate the numbers; it is simply
<code>eval</code>'d to generate the command line. For example, if you are root and
want to have all processes as possible completions, you might use
`<code>-e</code>', for many modern systems, or `<code>ax</code>', for older BSD-like
systems. The completion system tries to find a column which is headed
`<code>PID</code>' or `<code>pid</code>' (or even `<code>Pid</code>', in fact) to use for the process
IDs; if it doesn't find one, it just uses the first column.</p>
<p>The default is not to use any arguments; most variants of <code>ps</code> will then
just show you interactive processes from your current session. To show
all your own processes on a modern system, you can probably use the
value `<code>ps -u$USER</code>' for the style --- remembering to put this in
single quotes. Clearly, you need to make sure the context is narrow
enough to avoid unexpectedly calling odd commands.</p>
<p>You can make the value begin with a hyphen, then the usual command line
will put afterward and the hyphen removed. The suggested use for this is
adding `<code>command</code>' or `<code>builtin</code>' to make sure the right version of a
command is called.</p>
<p>The completion system allows you to type the name of a command, for
example `<code>emacs</code>', which will be converted to a PID. Note that this is
different from a job name beginning with `<code>%</code>'; in this case, any
command listed by <code>ps</code>, given the setting of the <code>command</code> style, can be
used. Obviously, command names can be ambiguous, unlike the process IDs
themselves, so the names are usually converted immediately to PIDs; if
the name could refer to more than one process, you get a menu of
possible PIDs.</p>
<p>The style <code>insert-ids</code> allows the completion system to keep using the
names rather than the PIDs. If it is set to <code>single</code>, the name will be
retained until you type enough to identify a particular process. If it
is set to <code>true</code> (or anything else but <code>menu</code>, actually), menu
completion is delayed until you have typed a string longer than the
common prefix of the PIDs. This is intended to be similar to
completion's usual logic --- don't do anything which gets rid of
information supplied by the user --- so is probably more useful in
practice than it sounds.</p>
<p><strong>Job control: <code>numbers</code></strong></p>
<p>Builtin functions that take process IDs usually also take job
specifications, strings beginning with `<code>%</code>' and followed either by a
small number or a string. The style <code>numbers</code> determines how these are
completed. By default, the completion system will try to complete an
unambiguous string from the name of the job. If you set <code>numbers</code> to
true, it will instead complete the job number --- though the listing
will still show the full information --- and if you set it to a number,
it will only use that many words of the job name, and switch to using
numbers if those are not unique. In other words, if you set it to `<code>1</code>'
and you have two jobs `<code>vi foo</code>' and `<code>vi bar</code>', then they will
complete as `<code>%1</code>' and `<code>%2</code>' (or maybe other numbers) since the first
words are the same.</p>
<p>Note also that <code>prefix-needed</code> applies here; if it is set, you need to
type the `<code>%</code>' to complete jobs rather than processes.</p>
<p><strong>System information: users, groups, hosts etc.</strong></p>
<p>There are many occasions where you complete the names of users on the
system, groups on the system (not to be confused with completion
groups), names of other hosts you connect to via the network, and ports,
which are essentially the names of internet services available on
another host such as <code>nntp</code> or <code>smtp</code>.</p>
<p>By default, the completion system will query the usual system files to
find the names of users, groups, hosts and ports, though in the final
case it will only look in the file `<code>/etc/hosts</code>', which often includes
only a very small number of not necessarily very useful hosts. It is
possible to tell the completion system always to use a specified set by
setting the appropriate style --- <code>users</code>, <code>groups</code>, <code>hosts</code>, <code>ports</code>
--- to the set of possibilities you want. This is nearly always useful
with <code>hosts</code>, and on some systems you may find it takes an inordinate
amount of time for the system to query the database for groups and
users, so you may want to specify a subset containing just those you use
most often.</p>
<p>There are also three sets of combinations: <code>hosts-ports</code>,
<code>hosts-ports-users</code> and <code>users-hosts</code>. These are used for commands which
can take both or all three arguments. Currently, the command socket uses
<code>hosts-ports</code>, telnet uses <code>hosts-ports-users</code>, while the style
<code>users-hosts</code> is used by remote login commands such as <code>rsh</code> and <code>ssh</code>,
and anywhere the form `<code>user@host</code>' is valid.</p>
<p>The last is probably the most useful, so I'll illustrate that. By
setting:</p>
<pre><code> zstyle ':completion:*' users-hosts \
pws:foo.bar.uk peters@frond.grub.uk
</code></pre>
<p>you tell <code>rsh</code> and friends the possible user/host combinations. Note
that for the separator you can use either `<code>:</code>', as usual inside the
completion system, or `<code>@</code>', which is more natural in this particular
case. If you type `<code>rsh -l </code>', a username is expected and either
<code>pws</code> or <code>peters</code> will be completed. Suppose you picked <code>pws</code>; then for
the next argument, which should be a host, the system now knows that it
must be <code>foo.bar.uk</code>, since the username for the other host doesn't
match.</p>
<p>If you don't need that much control, completion for all these commands
will survive on just the basic `<code>hosts</code>', `<code>users</code>', etc. styles; it
simply won't be as clever in recognising particular combinations. In
fact, even if you set the combined styles, anything that doesn't match
will be looked up in the corresponding basic style, so you can't lose,
in principle.</p>
<p>The other combined styles work in exactly the same way; just set the
values separated by colons or `<code>@</code>', it doesn't matter which.</p>
<p><strong>URLs for web browsers</strong></p>
<p>Completion for URLs is done by setting a parallel path somewhere on your
local machine. The <code>urls</code> style specifies the top directory for this.
For example, to complete the URL <code>http://zsh.org/</code>, you need to make a
set of subdirectories of the <code>path</code> directory <code>http/zsh.org/</code>. You can
extend this for however many levels of directory you need; as you would
expect, if the last object is a file rather than a directory you should
create it with `<code>touch</code>' rather than `<code>mkdir</code>'. The style will always
use the tag `<code>urls</code>' for this purpose, i.e. the context always matches
`<code>:completion:*:urls</code>'. This is a neat way of using the ordinary filing
system for doing the dirty work of turning URLs into components.
Arguably the system should be able to scan your browser's bookmarks
file, but currently it won't; there is, however, a tool provided with
the shell distribution in <code>Misc/make-zsh-urls</code> which should be able to
help --- ask your system administrators about this if it isn't
installed, I'm sure they'll be delighted to help.</p>
<p>If you only have a few URLs you want to complete, you can use one of two
simpler forms for the <code>urls</code> style. First, if the value of the style
contains more than one word, the values are used directly as the URLs to
be completed, e.g.:</p>
<pre><code> zstyle ':completion:*:urls' urls \
http://www.foo.org/ ftp://ftp.bar.net
</code></pre>
<p>Alternatively, you can set the <code>urls</code> style to the name of a normal
file, which contains the URLs to complete separated by white space or
newlines.</p>
<p>Note that many modern browsers allow you to miss out an initial
`<code>http://</code>', and that lots of pseudo-URLs appear in newspapers and
advertisements without it. The completion system needs it, however.</p>
<p>There is a better way when the web pages actually happen to be hosted on
a system whose directories you can access directly. Set the <code>local</code>
style to an array of three strings: a hostname to be considered local
(you can only give one per context), the directory corresponding to the
root of the files, and the directory where a user places their own web
pages, relative to their home directory. For example, if your home page
is usually retrieved as <code>http://www.footling.com/</code>, and that looks for
the index file (often called <code>index.html</code>) in the directory
<code>/usr/local/www/files</code>, and your own web pages live under `<code>~/www</code>',
then you would set</p>
<pre><code> zstyle ':completion:*:urls' local \
www.footling.com /usr/local/www/files www
</code></pre>
<p>and when you type `<code>lynx http://www.footling.com/</code>', all the rest will
be completed automatically.</p>
<p><strong>The X files</strong></p>
<p>There is another use for the <code>path</code> style with the tag `<code>colors</code>': it
gives the path to a file which contains a list of colour names
understood by the X-windows system, usually in file named `<code>rgb.txt</code>'.
This is used in such contexts as `<code>xsetroot -solid </code>', which
completes the name of a colour to set your root window (wallpaper) to.
It may be that the default value works on your system without your
needing to set this.</p>
<p><span id="l160"></span></p>
<h2 id="66-command-widgets"><a class="header" href="#66-command-widgets">6.6: Command widgets</a></h2>
<p><span id="l161"></span></p>
<h3 id="661-_complete_help"><a class="header" href="#661-_complete_help">6.6.1: <code>_complete_help</code></a></h3>
<p>You've already met this, usually bound to `<code>^Xh</code>' unless you already
had that bound when completion started up (in which case you should pick
your own binding and use `<code>bindkey</code>'), but don't forget it, since it's
by far the easiest way of finding out what context to use for setting
particular styles.</p>
<p><span id="l162"></span></p>
<h3 id="662-_correct_word-_correct_filename-_expand_word"><a class="header" href="#662-_correct_word-_correct_filename-_expand_word">6.6.2: <code>_correct_word</code>, <code>_correct_filename</code>, <code>_expand_word</code></a></h3>
<p>The first and last of these have been mentioned in describing the
related completers: <code>_correct_word</code>, usually bound to <code>^Xc</code>, calls the
<code>_correct</code> completer directly to perform spelling correction on the
current word, and <code>_expand_word</code>, usually bound to <code>^Xe</code>, does the same
with the <code>_expand</code> completer. The contexts being
`<code>:completion:complete-word</code>' and `<code>:completion:expand-word</code>'
respectively, so that they can be distinguished in styles from the
ordinary use of the completer. If you want the same styles to be used in
both contexts, but not others, you should define them for patterns
beginning `<code>:completion:complete(|-word)...</code>'.</p>
<p>The middle one simply corrects filenames, regardless of the completion
context. Unlike the others, it can also be called as an ordinary
function: pass it an argument, and it will print out the possible
corrections. It does this because it bypasses most of the usual
completion system. Probably you won't often need it, but it is usually
bound to `<code>^XC</code>' (note the capital `<code>C</code>').</p>
<p><span id="l163"></span></p>
<h3 id="663-_history_complete_word"><a class="header" href="#663-_history_complete_word">6.6.3: <code>_history_complete_word</code></a></h3>
<p>This is usually bound to `<code><ESC-/></code>' for completing back in the
history, and `<code><ESC-,></code>' for completing forward --- this will
automatically turn on menu completion, temporarily if you don't normally
have that set, to cycle through the matches. It will complete words from
the history list, starting with the most recent. Hence</p>
<pre><code> touch supercalifragilisticexpialidocious
cat sup<ESC-/>
</code></pre>
<p>will save you quite a bit of typing --- although in this particular
case, you can use `<code><ESC-.></code>' to insert the last word of the previous
command.</p>
<p>Various styles are available. You can set the `<code>stop</code>' style which
makes it stop once before cycling past the end (or beginning) of the
history list, telling you that the end was reached.</p>
<p>You can also set the `<code>list</code>' style to force matches to be listed, the
`<code>sort</code>' style to sort matches in alphabetical order instead of by
their age in the history list, and the `<code>remove-all-dups</code>' style, which
ensures that each match only occurs once in the completion list ---
normally consecutive identical matches are removed, but the code does
not bother searching for identical matches elsewhere in the list of
possibilities. Finally, the <code>range</code> style is supported via the
<code>_history</code> completer, which does the work. This style restricts the
number of history words to be searched for matches and is most useful if
your history list is large. Setting it to a number <em>n</em> specifies that
only the last <em>n</em> history words should be searched for possible matches.
Alternatively, it can be a value of the form `<em>max</em><code>:</code><em>slice</em>', in
which case it will search through the last <em>slice</em> history words for
matches, and only if it doesn't find any, the <em>slice</em> words before that;
<em>max</em> gives an overall limit on the maximum number of words to search
through.</p>
<p><span id="l164"></span></p>
<h3 id="664-_most_recent_file"><a class="header" href="#664-_most_recent_file">6.6.4: <code>_most_recent_file</code></a></h3>
<p>This function is normally bound to `<code>^Xm</code>'. It simply completes the
most recently modified file that matches what's on the line already. Any
pattern characters in the existing string are active, so this is a cross
between expansion and completion. You can also give it a numeric prefix
to show the <code>N</code>th most recently modified file that matches the pattern.</p>
<p>By the way, you can actually do the same by setting appropriate styles,
without any new functions. The trick is to persuade the system to use
the normal <code>_files</code> completer with the <code>file-sort</code> style. By restricting
the use of the styles to the context of the widget --- which is simply
the <code>_generic</code> completer described above:</p>
<pre><code> zstyle ':completion:(match-word|most-recent-file):*' \
match-original both
zstyle ':completion:most-recent-file::::' completer \
_menu _files _match
zstyle ':completion:most-recent-file:*' file-sort modification
zstyle ':completion:most-recent-file:*' file-patterns \
'*(.):normal\ files'
zstyle ':completion:most-recent-file:*' hidden true
zstyle ':completion:most-recent-file:*:descriptions' format ''
bindkey '^Xm' most-recent-file
zle -C most-recent-file menu-complete _generic
</code></pre>
<p>It may not be obvious how this works, so here's a blow by blow account
if you are interested. (It works even if you aren't interested,
however.)</p>
<ul>
<li>The `<code>zle -C</code>' defines a widget which does menu completion, and
behaves like ordinary completion (that's what <code>_generic</code> is for)
except that the context uses the name of the widget we define.</li>
<li>When we invoke the widget, the system uses the <code>completer</code> style to
decide what completions to perform. This instructs it: use menu
completion, complete files, use pattern matching if the completion
so far didn't work.</li>
<li>First, <code>_menu</code> comes along; it actually does nothing more than tell
the system to use menu completion.</li>
<li>Then <code>_files</code> generates a list of files. This uses the <code>file-sort</code>
and <code>file-patterns</code> styles defined for the <code>most-recent-file</code>
context. They produce a set of files in modification time order, and
include only regular files (so not directories, symlinks, device
files and so on).</li>
<li>If that failed, the <code>_match</code> style allows the word on the command
line to be treated as a pattern; for example, <code>*.c</code> to complete the
most recent C source file. This uses the <code>match-original</code> style; the
setting tells it that it should try first without adding an extra
`<code>*</code>' for matching (this is what we want for the case where we
already have a complete pattern like <code>*.c</code>), and if that fails, add
a <code>*</code> at the end and try again.</li>
<li>The <code>hidden</code> style means that the matches aren't listed; all that
happens is the first is inserted on the line. The setting for the
<code>format</code> tag similarly simplifies the display in this case by
removing verbose descriptions.</li>
<li>The net result is the first step of a menu completion: insert the
first matched file (the most recently modified) onto the line. This
is exactly what you want. Note, however, that as we are in menu
completion you can keep on hitting <code>^xm</code> and the shell will cycle
through the matches, which here gives you files that are
progressively less recently modified.</li>
</ul>
<p>Omit the <code>file-patterns</code> line if you don't want the match restricted to
regular files (I sometimes need the most recently modified directory,
but often it's irrelevant). The whole version using styles comes from
Oliver Kiddle, who recommends using <code>_generic</code> in this way any time you
want to generate a widget from a specific completion such as <code>_files</code>.
There is a brief section on <code>_generic</code> below.</p>
<p><span id="l165"></span></p>
<h3 id="665-_next_tags"><a class="header" href="#665-_next_tags">6.6.5: <code>_next_tags</code></a></h3>
<p>This is a very neat way of getting round the order of tags just with a
key sequence. An example is the best way of showing it; it's bound by
default to the key sequence `<code>^Xn</code>'.</p>
<pre><code> % tex ^D
Completing TeX or LaTeX file
bar.tex foo.tex guff.tex
</code></pre>
<p>Our file is not in that directory, but by default we don't get to see
the directory if there was a file that matched the pattern --- here
`<code>*.tex</code>'. (This will actually change in 4.1, since most people don't
know about <code>_next_tags</code> but do know about directories, but you can still
cycle through the different sets of tags.) You can set the <code>tag-order</code>
style to alter whether they appear at the same time, but <code>_next_tags</code>
lets you do this very simply. Just hit <code>^Xn</code>. You're now looking at</p>
<pre><code> Completing TeX or LaTeX file
dir1/ dir2/ dir3/
</code></pre>
<p>and if you carry on hitting <code>^Xn</code> you will get to all files, and then
you will be taken back to the <code>.tex</code> files again. (Where our file
actually is, is left as an exercise for the reader.)</p>
<p>Of course this works with any set of tags whatsover; it simply has the
effect of cycling you around the tag order.</p>
<p><span id="l166"></span></p>
<h3 id="666-_bash_completions"><a class="header" href="#666-_bash_completions">6.6.6: <code>_bash_completions</code></a></h3>
<p>This function provides compatibility with a set of completion bindings
in bash, in which escape followed by one of the following characters
causes a certain type of (non-contextual) completion: `<code>!</code>', command
names; `<code>$</code>', environment variables; `<code>@</code>', host names; `<code>/</code>',
filenames, and `<code>~</code>' user names. `<code>^X</code>' followed by the same
characters causes the possible completion to be listed. This function
decides by examining its own binding which of those it should be doing,
then calls the appropriate completion function. If you want to use it
for all those possible bindings, you need to issue the right statements
in your <code>.zshrc</code>, since only the bindings with `<code>~</code>' are set up by
default to avoid clashes. This will do it:</p>
<pre><code> for key in '!' '$' '@' '/'; do
bindkey "\e$key" _bash_complete-word
bindkey "^X$key" _bash_list-choices
done
</code></pre>
<p>Unlike most widgets, which are tied to functions of the same name to
minimize confusion, the function <code>_bash_completions</code> is actually called
under the names of the two different widgets shown in that code so as to
be able to implement both completion and listing behaviour.</p>
<p><span id="l167"></span></p>
<h3 id="667-_read_comp"><a class="header" href="#667-_read_comp">6.6.7: <code>_read_comp</code></a></h3>
<p>This function, usually bound to `<code>^X^R</code>', does on-the-fly completion.
When you call it, it prompts for you to enter a type of completion;
usually this will be the name of a completion function with the required
arguments. Thus it's not much use unless you already have some fairly
in-depth knowledge of how the system is set up. For example, try it,
then enter `<code>_files -/</code>', which generates directories. There is a
rudimentary completion for the function names built into it.</p>
<p>The next time you start it up, it will produce the same type of
completion. You need to give it a numeric prefix to tell it to prompt
for a different sort.</p>
<p><span id="l168"></span></p>
<h3 id="668-_generic"><a class="header" href="#668-_generic">6.6.8: <code>_generic</code></a></h3>
<p>Rather than being directly bound, like the others, this widget gives you
a way of creating your own special completions. You define it as a
widget and bind it as if it were any completion function:</p>
<pre><code> zle -C foo complete-word _generic
bindkey '<keys>' foo
</code></pre>
<p>Now the keys bound will perform ordinary contextual completion, but any
styles will be looked up with the command context `<code>foo</code>'. So you can
give it its own set of completers:</p>
<pre><code> zstyle ':completion:foo:*' completer _expand
</code></pre>
<p>and, indeed, give it special values for any style you like. To put it
another way, you've now got a complete, separate copy of the completion
system where the only difference is the extra word in the context.</p>
<p>Good example of the use of this function were given above in the
descriptions of <code>_all_matches</code> and <code>_most_recent_file</code>.</p>
<p><span id="l169"></span></p>
<h3 id="669-predict-on-incremental-complete-word"><a class="header" href="#669-predict-on-incremental-complete-word">6.6.9: <code>predict-on</code>, <code>incremental-complete-word</code></a></h3>
<p>These are not really complete commands at all in the strict sense, they
are normal editing commands which happen to have the effect of
completion. This means that they are not part of the completion system,
and though they are installed with other shell functions they will not
automatically be loaded. You will therefore need an explicit `<code>autoload -U predict-on</code>', etc. --- remember that the `<code>-U</code>' prevents the
functions from expanding any of your own aliases when they are read in
--- as well as an explicit `<code>bindkey</code>' command to bind each function,
and a `<code>zle -N</code>' statement to tell the line editor that the function is
to be regarded as an editing widget. The <code>predict-on</code> file, when loaded,
actually defines two functions, <code>predict-on</code> and <code>predict-off</code>, both of
which need to be defined and bound for them to work. So to use all of
these,</p>
<pre><code> autoload -U incremental-complete-word predict-on
zle -N incremental-complete-word
zle -N predict-on
zle -N predict-off
bindkey '^Xi' incremental-complete-word
bindkey '^Xp' predict-on
bindkey '^X^P' predict-off
</code></pre>
<p>`Prediction' is a sort of dynamic history completion. With <code>predict-on</code>
in effect, the line editor will try to retrieve a line back in the
history which matches what you type. If it does, it will show the line,
extending past the current cursor position. You can then edit the line;
characters which do not insert anything mostly behave as normal. If you
continue to type, and what you type does not match the line which was
found, the line editor will look further back for another line; if no
line matches, editing is essentially as normal. Often this is flexible
enough that you can leave <code>predict-on</code> in effect, but you can return to
basic editing with <code>predict-off</code>.</p>
<p>Note that, with prediction turned on, deleting characters reverses the
direction of the history search, so that you go back to previous lines,
like an ordinary incremental search; unfortunately the previous line
found could be one you've already half-edited, because they don't
disappear from the list until you finally hit `return' on an edited
line to accept it. There's another problem with moving around the line
and inserting characters somewhere else: history searching will resume
as soon as you try to insert the new characters, which means everything
on the right of the cursor is liable to disappear again. So in that case
you need to turn prediction off explicitly. A final problem: prediction
is bad with multi-line buffers.</p>
<p>If prediction fails with <code>predict-on</code> active, completion is
automatically tried. The context for this looks like
`<code>:completion:predict::::</code>'. Various styles are useful at this point:
`<code>list</code>' could be set to <code>always</code>, which will show a possible
completion even if there is only one, for example. The style `<code>cursor</code>'
may have the values `<code>complete</code>' to move to the end of the word
completed, `<code>key</code>' to move past the rightmost occurrence of the
character just typed, allowing you just to keep typing, or anything else
not to move the cursor which is the default behaviour.</p>
<p>The <code>incremental-complete-word</code> function allows you to see a list of
possible completions as you type them character by character after the
first. The function is quite basic; it is really just an example of
using various line editor facilities, and needs some work to make a
useful system. It will understand <code>DEL</code> to delete the previous
character, return to accept, <code>^G</code> to abort, <code>TAB</code> to complete the word
as normal and <code>^D</code> to list possibilities; otherwise, keys which do not
insert are unlikely to have a useful effect. The completion is done
behind the scenes by the standard function <code>complete-word</code>.</p>
<p><span id="l170"></span></p>
<h2 id="67-matching-control-and-controlling-where-things-are-inserted"><a class="header" href="#67-matching-control-and-controlling-where-things-are-inserted">6.7: Matching control and controlling where things are inserted</a></h2>
<p>The final matter before I delve into the system for writing new
completion functions is matching control; the name refers in this case
to how the matching between characters already typed on the command line
and characters in a trial completion is performed. This can be done in
two ways: by setting the <code>matcher-list</code> style, which applies to all
completions, or by using an argument (<code>-M</code>) to the low-level completion
functions. Mostly we will be concerned with the first. All this is best
illustrated by examples, which are taken from the section `<strong>Matching
Control</strong>' in the <code>zshcompwid</code> manual page; in the printed manual and
the `info' pages this occurs within the section `<code>Completion Widgets</code>'.</p>
<p>The <code>matcher-list</code> style takes an array value. The values will be tried
in order from left to right. For example,</p>
<pre><code> zstyle ':completion:*' matcher-list 'm:{a-z-}={A-Z_}' \
'r:|[-_./]=* r:|=*'
</code></pre>
<p>tries the first specification, which is for case-insensitive completion,
and if no matches are generated tries the second, which does partial
word completion; I'll explain both these specifications in detail as we
go along. You can make it do both forms the second time round simply by
combining the values with a space, i.e. the last word on the command
line becomes <code>'m:{a-z-}={A-Z_} r:|[-_./]=* r:|=*'</code>. It is also perfectly
valid to have a first matcher empty, i.e. <code>'``'</code>; this means that
completion is tried with no matching rule the first time, and will only
go on to subsequent matchers in the list if that fails. This is quite a
good practice as it avoids surprises.</p>
<p><span id="l171"></span></p>
<h3 id="671-case-insensitive-matching"><a class="header" href="#671-case-insensitive-matching">6.7.1: Case-insensitive matching</a></h3>
<p>To perform case-insensitive matching for all completions, you can set:</p>
<pre><code> zstyle ':completion:*' matcher-list 'm:{a-z}={A-Z}'
</code></pre>
<p>The `<code>m:</code>' specifies standard matching, with the `<code>{a-z}</code>' describing
what's on the command line, and the `<code>{A-Z}</code>' what's in the trial
completion. The braces indicate `correspondence classes', which are not
lessons taken by email (that's a joke), but a relative of the more usual
character classes like `<code>[a-z]</code>', which, as you no doubt know, would
match any of the letters between <code>a</code> and <code>z</code>. In this context, with the
braces, the letters are forced to match on the left and right hand side
of the `<code>=</code>', so an `<code>a</code>' on the command line must match an `<code>A</code>' in
the trial completion, a `<code>b</code>' must match a `<code>B</code>', and so on. Since an
<code>a</code> in the command line will always match an `<code>a</code>' in the trial
completion, matcher or no matcher, this means that if you type an `<code>a</code>'
it will match either `<code>a</code>' or `<code>A</code>' --- in other words,
case-insensitively. The same goes for any other lowercase letter you
type. The difference from `<code>m:[a-z]=[A-Z]</code>' is that, because ordinary
character classes are unordered, <em>any</em> lowercase letter would have
matched <em>any</em> uppercase letter, which isn't what you want. The rest of
the shell doesn't know about correspondence classes at all.</p>
<p>Finally, the use of a lowercase `<code>m</code>' at the start means that the
characters actually inserted onto the line are those from the trial
completion --- if you type `<code>make<TAB></code>', the completion process
generates file names, and <code>matcher-list</code> allows what you type to match
the file `<code>Makefile</code>', then you need the latter to be inserted on the
command line. Use of `<code>M:</code>' at the start of the matcher would keep
whatever was on the line to begin with there.</p>
<p>If you want completely case-insensitive matching, so that typing
`<code>MAKE<TAB></code>' would also potentially complete to `<code>Makefile</code>' or
`<code>makefile</code>' (and so on), the extension is fairly obvious:</p>
<pre><code> zstyle ':completion:*' matcher-list 'm:{a-zA-Z}={A-Za-z}'
</code></pre>
<p>because now as well as `<code>a</code>' matching `<code>A</code>', `<code>A</code>' will match `<code>a</code>'
--- and, of course, `<code>a</code>' and `<code>A</code>' each still match themselves.</p>
<p>More detail on the patterns: they do not, in fact, allow all the
possible patterns you can use elsewhere in the shell, since that would
be too complicated to implement with little extra use. Apart from
character classes and correspondence classes, you can use `<code>?</code>' which
has its usual meaning of matching one character, or literal characters,
which match themselves; or the pattern for the trial completion only can
be a single `<code>*</code>'. which matches anything. That's it, however; you
can't do other things with the `<code>*</code>' since it's too difficult for the
system to guess what characters should be covered by it.</p>
<p>For the same reason, the `<code>*</code>' must be in an <em>anchored</em> pattern, the
idea behind which is shown in the next example.</p>
<p><span id="l172"></span></p>
<h3 id="672-matching-option-names"><a class="header" href="#672-matching-option-names">6.7.2: Matching option names</a></h3>
<p>I explained back in <a href="zshguide01.html#intro">chapter 1</a> that zsh didn't
care too much how you specified options: `<code>noglob</code>' and `<code>NOGLOB</code>' and
`<code>No_Glob</code>' and `<code>__NO_GLOB_</code>' are all treated the same way. Also,
this is the negation of the option `<code>glob</code>'. Having learnt how to match
case-insensitively, we have two further challenges: how to ignore a
`<code>_</code>' anywhere in the word, and how to ignore the <code>NO</code> at the beginning
so that we can complete an unnegated option name after it.</p>
<p>Well, here's how. Since you don't want this for all completions, just
for option names, I shall show it as an argument for the `<code>compadd</code>'
command, which gives the system the list of possible completions. The
option names should then appear as the remaining arguments to the
command, and the easiest way of doing that is to have the
<code>zsh/parameter</code> module loaded, which it always is for new completion,
and use the keys of the special associative array <code>$options</code>:</p>
<pre><code> compadd -M 'B:|[nN][oO]= M:_= M:{A-Z}={a-z}' - ${(k)options}
</code></pre>
<p>Here, we're interested in the thing in quotes --- it means exactly the
same here as it would as an element of the matcher list, except that it
only applies to the trial completions given after the `<code>-</code>'. It's in
three bits, separated by spaces; as they're in the same word, all are
applied one after the other regardless of any previous ones having
matched.</p>
<p>Starting from the right, you can see that the last part matches letters
case-insensitively; the capital `<code>M</code>' means that, this time, the
letters on the command line, not those in the trial completion are kept;
this is safe because of the way options are parsed, and reduces
unexpected changes.</p>
<p>Moving left, you can now guess `<code>M:_=</code>': it means that the `<code>_</code>'
matches nothing at all in the trial completion --- in other words, it is
simply ignored. The rule for matching across the `<code>=</code>' is that you move
from left to right, pairing off characters or elements of character
classes as I already described, and when you run out, you treat any
missing characters as, well, missing.</p>
<p>The first part has an `anchor', indicated by what lies between the
`<code>:</code>' and the `<code>|</code>'. The <code>B</code> specifies that the case insensitive match
of `<code>no</code>' must occur at the start of the word on the command line (with
`<code>b</code>' it would be the word in the list of matches), but here it is lax
enough to allow this to happen after the `<code>M:_=</code>' has stripped any
initial underscores away. Hence it matches <code>no</code>, <code>NO</code>, <code>No</code> or <code>nO</code> at
the start of the string, and, just like the `<code>M:_=</code>' part, it ignores
it, since there's nothing on the right. Again, the capital `<code>B</code>' at the
start means keep what's on the command line: that's important in this
case, since if you lost the `<code>no</code>', the meaning would change
completely.</p>
<p>So consider the combined effect when trying to complete <code>NO_GL</code>. The
first specification allows it to match against <code>_GL</code>; the second allows
it to match against <code>GL</code>; the third, against <code>gl</code>; and finally the usual
effect of completion means that any option beginning <code>gl</code> may be
completed. Try `<code>setopt NO_GL^D</code>' and you should see something like:</p>
<pre><code> NO_GLob NO_GLobassign NO_GLobdots
NO_GLobalrcs NO_GLobcomplete NO_GLobsubst
</code></pre>
<p>--- after the bit you've typed, the form of the words reverts to
whatever's in the trial completion, i.e. lowercase letters with no
`<code>_</code>'s.</p>
<p><span id="l173"></span></p>
<h3 id="673-partial-word-completion"><a class="header" href="#673-partial-word-completion">6.7.3: Partial word completion</a></h3>
<p>This example shows the other sort of anchoring, on the right, and also
how to use a `<code>*</code>' in the right hand part of a pattern. Consider:</p>
<pre><code> zstyle ':completion:*' matcher-list 'r:|.=* r:|=*'
</code></pre>
<p>The `<code>r:</code>' specifies a right-anchored match, using the characters from
the trial completion rather than what's already on the command line. As
the anchor is on the right this time, the pattern (between `<code>:</code>' and
`<code>|</code>') is empty, and its anchor (between `<code>|</code>' and `<code>=</code>') is `<code>.</code>'.
So this specifies that nothing --- a zero length string, or a gap
between characters if you want to think of it like that --- when
followed by a `<code>.</code>', matches anything at all in the trial completion.</p>
<p>Consequently, the second part says that nothing anchored on the right by
nothing --- in other words, the right hand end of the command line
string --- matches anything. This is what completion normally does, add
anything at all at the end of the string; we've added this part to the
matcher in case the cursor is in the middle of the word. It means that
the right hand end will always be completed, too.</p>
<p>Let's see that in action. Here are the actual contents of my actual
<code>tmp</code> directory, never mind why:</p>
<pre><code> regframe.rpm t.c testpage.dvi testpage.log testpage.ps
</code></pre>
<p>Now I set the <code>matcher-list</code> style as above and type:</p>
<pre><code> echo t.p<TAB>
</code></pre>
<p>and get</p>
<pre><code> echo testpage.ps
</code></pre>
<p>So, apart from the normal completion at the end (<code>p</code> to <code>ps</code>), the empty
string followed by a <code>.</code> was allowed to match anything, too, and I got
the effect of completing both bits of the word.</p>
<p>You might wonder what happens when there's a file <code>testpage.old.ps</code>
around, i.e. the anchor appears twice in that. With the matcher set as
given above, that won't be completed; the anchor needs to be matched
explicitly, not by a wildcard. If you don't like that, you can change
the `<code>*</code>' after the `<code>=</code>' in the specification to `<code>**</code>'; this form
allows the anchor to occur in the string being matched. You can think of
`<code>*</code>' and `<code>**</code>' as taking the shortest and the longest possible
matches respectively. If you use a lot of `<code>**</code>' specifications in your
matches, things can get very confusing, however.</p>
<p>Other shells have a facility for completing inside words like this,
where it goes by such names as `enhanced' completion, although it is
usually not so flexible. In the case of tcsh, not just `<code>.</code>' but also
`<code>-</code>' and `<code>_</code>' have this effect. You can force this with</p>
<pre><code> zstyle ':completion:*' matcher-list 'r:|[._-]=* r:|=*'
</code></pre>
<p><span id="l174"></span></p>
<h3 id="674-substring-completion"><a class="header" href="#674-substring-completion">6.7.4: Substring completion</a></h3>
<p>I've mentioned `<code>r</code>' and `<code>B</code>', but corresponding to `<code>r</code>' there is
`<code>l</code>', which anchors on the left instead of the right, and
corresponding to `<code>B</code>' there is `<code>E</code>' which matches at the end instead
of the beginning; and, of course, all exist in both upper- and lowercase
forms, meaning `keep what the user typed' and `keep what is in the
list of possible matches', respectively.</p>
<p>Here is an example of using `<code>l:|=*</code>' to match anything at the start of
the word: this is the effect of having an empty anchor, as you saw with
`<code>r</code>' above, but note with `<code>l</code>', the anchor appears, logically
enough, on the left of the `<code>|</code>', in the order they would appear on the
command line. By combining this with the `<code>r</code>' form, you can make the
completion system work when what is on the command line matches only a
substring of a trial completion --- i.e., has anything else on the left
and on the right. Since this can potentially generate a lot of matches,
it might by an idea to try it after any other matcher specifications you
have. So the following tries case-insensitive completion, then
partial-word completion (case-sensitively), then substring completion:</p>
<pre><code> zstyle ':completion:*' matcher-list 'm:{a-z}={A-Z}' \
'r:|[._-]=* r:|=*' 'l:|=* r:|=*'
</code></pre>
<p><span id="l175"></span></p>
<h3 id="675-partial-words-with-capitals"><a class="header" href="#675-partial-words-with-capitals">6.7.5: Partial words with capitals</a></h3>
<p>This section illustrates another feature: if you use `<code>||</code>' when
specifying anchors for `<code>L</code>' or `<code>R</code>' or their lowercase variants, the
pattern part for what appears on the command line, which would usually
be translated into some other pattern, is treated instead as another
anchor on the other side of the pattern --- which isn't matched against
the pattern in the word, it just has to appear. In other words, this
part matches without being `swallowed up' in the process. An example
(again adapted from the manual) will make this clearer.</p>
<pre><code> compadd -M 'r:[^A-Z0-9]||[A-Z0-9]=** r:|=*' \
LikeTHIS LooHoo foo123 bar234
</code></pre>
<p>The four possible completions are on the second line. The second of the
two matcher specifications just allows anything to match on the right,
so if we are inside the word, the remainder may be completed. The first
word is where the action is; it says `A part of the completion which
has on the left something other than an upper case letter or a digit,
and on the right an upper case letter or a digit, may match anything,
including the anchor'. So in particular, this would allow `<code>LH</code>' to
complete to `<code>LooHoo</code>' --- and only that, since `<code>LikeTHIS</code>' has an
uppercase letter to the left of the `<code>H</code>', which is not allowed. In
other words, the chunks of word beginning with uppercase letters and
digits act like the start of substrings. (If you like, remember that
last sentence and the specification, and forget the rest.)</p>
<p><span id="l176"></span></p>
<h3 id="676-final-notes"><a class="header" href="#676-final-notes">6.7.6: Final notes</a></h3>
<p>To put everything together, the possible specifications are
`<code>m:...=...</code>', `<code>l:...|...=...</code>', `<code>r:...|...=...</code>',
`<code>b:...|...=...</code>' and `<code>e:...|...=...</code>', which cause the command line
to be altered to the match found, and their counterparts with an
uppercase letter, which cause what's already on the command line to be
left alone and the remaining characters to be inserted directly from the
completion found. The `<code>...</code>' are patterns, which all use the same
format. They can include literal characters, a `<code>?</code>', and character or
correspondence classes, while the rightmost pattern in each type may
also consist of a `<code>*</code>' on its own. Characters are matched from left to
right; a missing character matches an empty string, `<code>*</code>' matches any
number of characters. Specifications may be joined in a single string,
in which case all parts will be applied together.</p>
<p>When using the <code>matcher-list</code> style, a list of different specifications
can be given; in this case, they will be tried in turn until one of them
generates matches, and the rest will not be used.</p>
<p>There's another style apart from <code>matcher-list</code>, called <code>matcher</code>. This
can be set for a particular context, possibly with specific tags, and
will add the given matcher specifications using exactly the same syntax
as <code>matcher-list</code> for that context, except that here all specifications
are used at once, even if they are given as different elements of an
array. This is possibly useful because <code>matcher-list</code> is only aware of
the completer, not of any more specific part of the context.</p>
<p>Although I won't talk about matching control after this section, there
may be cases where you want to include `<code>compadd -M ...</code>' in a
completion function of your own to help the user. Many of the existing
completion functions provide partial word completion where it seems
useful; for example, completion of zle functions allows <code>i-c-w</code> to be
completed to <code>incremental-complete-word</code> in this way.</p>
<p>Actually, you can configure this to a considerable extent without
altering a function, using styles and labelled tags. From the manual:</p>
<pre><code> zstyle ':completion:*:*:foo:*' tag-order '*' '*:-case'
zstyle ':completion:*-case' matcher 'm:{a-z}={A-Z}'
</code></pre>
<p>In command <code>foo</code>, whatever the tags are, they are to be tried normally
first (the `<code>*</code>' argument to <code>tag-order</code>), then under the same name
with `<code>-case</code>' appended. The second style defines a matcher for any tag
ending in the suffix `<code>-case</code>', which allows lowercase characters to
match uppercase ones. The upshot is that completion of anything at all
for the command <code>foo</code> will be tried first case-sensitively, then
case-insensitively.</p>
<p><span id="l177"></span></p>
<h2 id="68-tutorial"><a class="header" href="#68-tutorial">6.8: Tutorial</a></h2>
<p>Before bamboozling you with everything there is to know about writing
your own completion function, I'll give you an example of something I
wrote myself recently. If you were doing this yourself, you would then
just stick this function somewhere in your function search path, and
next time you started the shell it would start doing its work. However,
the file already exists: it's called <code>_perforce</code> and you should find it
in the function search for versions 4.1.1 and above of zsh. I apologize
if it's not the ideal function to start with, but it is fresh in my
mind, so what I'm saying has some chance of being correct.</p>
<p>This section is subtitled, `How I struggled to write a set of
completions for Perforce'. Perforce is a commercial configuration
management tool (as they now call revision control systems); consult
<a href="http://www.perforce.com/">http://www.perforce.com/</a> for details. It's concepts aren't a million
miles from CVS, the archetypal system of this kind, but it was
sufficiently different that the completion functions needed rewriting
from the ground up. You won't need to know anything about CVS or
Perforce, because at each stage I'll explain what I'm trying to complete
and why. This should give you plenty of meat for writing completions of
your own. After the tutorial, the chapter goes into the individual
details, which will expand on some of the things that appeared briefly
in the tutorial.</p>
<p>What I tend to find the most complicated part of this is making sure the
completion system knows the correct types of completions and their tags
to be completed at once. This probably won't be your first priority when
trying to write completions of your own, but if you do it right, all the
stuff about selecting types and arranging them in groups that I showed
above will just work. In this tutorial we arrange to use enough of the
higher level functions that it will work without too much (apparent)
effort. Of course, working out from scratch which those functions are
isn't always that easy; hence the tutorial.</p>
<p>Needless to say, I will simplify grossly at a lot of points. You can see
the finished product in the zsh 4.1 distribution. It even has a few
comments in.</p>
<p><strong>Basic structure</strong></p>
<p>Like the <code>cvs</code> command and a few other of the more complicated commands
you might use, Perforce is run by a single command, <code>p4</code>, followed by an
argument giving the particular Perforce command, followed by an options
and arguments to that command.</p>
<p>This dictates the basic tasks the completion functions must do:</p>
<ul>
<li>If we are in the first argument, complete the name of the
subcommand.</li>
<li>If we are in a subsequent argument, look up the name of the
subcommand and call the function which handles its arguments.</li>
</ul>
<p>This is more complicated than most commands you will write completions
for. However, one useful feature of the completion system is you can do
completions in a recursive fashion. So once you get to the point where
you are handling arguments for a particular subcommand, you can
completely forget about the first step --- as if the subcommand was the
command on the line.</p>
<p>In addition to the subcommands, there are lots of other types of object
Perforce knows about: files, obviously, plus revisions of files, set of
changes (`changelists') applied at once, numbers of fixes applied to
files (essentially a way of tying changlists to a particular change
request for bugtracking purposes), types of file --- text, binary, etc.,
and several others. We will break down each of these completions into
its own function. That means that any time we need to complete a
particular type of object, wherever it appears (and many of these
objects can appear in lots of different places), we just call the same
function.</p>
<p>Hence there are a large number of different functions:</p>
<ul>
<li>The main dispatcher for the command, called <code>_perforce</code> for clarity
--- the main command it handles is `<code>p4</code>', but the name Perforce is
more familiar.</li>
<li>One function for each subcommand.</li>
<li>One function for each type of object Perforce knows about and we
complete (we don't bother completing dates, for example).</li>
<li>In some cases, in particular files, multiple functions since there
are different types of file --- regular files and directories
completed in the normal way, files completed by asking Perforce
where it has stored them, files opened for some form of change to be
made to them, and so on. Each of these is completed by a different
function.</li>
</ul>
<p>This makes it impractical to put all the functions in separate files
since editing them would be a nightmare. What's more, since we will
always go through the dispatcher <code>_perforce</code>, we don't need to tell the
shell to autoload all the other functions; it can just hook them in from
the main file. The file <code>_perforce</code> therefore has the structure:</p>
<pre><code> #compdef p4
# Main dispatcher
_perforce() {
# ...
}
# Helper functions for the various types of object
_perforce_files() {
# ...
}
# ...
# Dispatchers for the individual subcommands.
_perforce_cmd_help() {
# ...
}
# Code to make sure _perforce is run when we load it
_perforce "$@"
</code></pre>
<p>That last line is probably the least obvious. It's because of the fact
that zsh (unlike other shells) usually treats the file of an autoloaded
function as being the body of the function. Since everything else here
just defines a function, without the last line nothing would happen the
first time it was run; it would define <code>_perforce</code> and all the other
functions, but that was it. The last line makes sure <code>_perforce</code> gets
run with all the arguments passed down. The shell is smart enough to
know that the <code>_perforce</code> function we defined in the file is the one to
keep for future use, not the entire file, so from then on things are
easy; we just have a complete set of ready-defined files.</p>
<p>In fact the various helper functions didn't even need to use the `<code>_</code>'
convention for completion functions, since the completion system didn't
see them directly. However, I've kept it for consistency.</p>
<p>There's one extra trick: apart from <code>_perforce</code> itself, the function
definitions look like this:</p>
<pre><code> (( $+functions[_perforce_cmd_diff] )) ||
_perforce_cmd_diff() {
# body of function
}
</code></pre>
<p>This is to allow the user to override each function separately. The test
uses the <code>$functions</code> special associative array from the <code>zsh/parameter</code>
module, which the completion system loads. If the function is already
defined, because the corresponding element in the <code>$functions</code> parameter
is set, then we skip the definition of the function here, because the
user has already defined it. So if you were to write your own
<code>_perforce_cmd_diff</code> and put it into the function path, it would be
used, as you no doubt intended.</p>
<p><span id="l178"></span></p>
<h3 id="681-the-dispatcher"><a class="header" href="#681-the-dispatcher">6.8.1: The dispatcher</a></h3>
<p>This top level is only necessary for complex commands with multiple
subcommands. There are interesting titbits here, but if you just want to
know how to complete a command with ordinary UNIX-style argument
parsing, skip to the next section.</p>
<p>The main <code>_perforce</code> function has the two purposes described at the top
of the previous subsection. We need to decide whether we are in the
first word after the <code>p4</code> command itself. A simple way of doing that is:</p>
<pre><code> if (( CURRENT > 2 )); then
# Remember the subcommand name
local cmd=${words[2]}
# Set the context for the subcommand.
curcontext="${curcontext%:*:*}:p4-$cmd"
# Narrow the range of words we are looking at to exclude `p4'
(( CURRENT-- ))
shift words
# Run the completion for the subcommand
_perforce_cmd_$cmd
else
local hline
local -a cmdlist
_call_program help-commands p4 help commands | while read -A hline; do
(( ${#hline} < 2 )) && continue
[[ $hline[1] = (#i)perforce ]] && continue
cmdlist=($cmdlist "${hline[1]}:${hline[2,-1]}")
done
_describe -t p4-commands 'Perforce command' cmdlist
fi
</code></pre>
<p>This already looks a bit horrific, but it breaks down quite easily. We
test the <code>$CURRENT</code> parameter, which is a special parameter in the
completion system giving the word on the command line we are on. This is
the syntactic word --- the completion system has already done the hard
job (and that's not an overstatement, I can tell you) of deciding what
makes up a word on the command line, taking into account quoting and
special characters. The array of words is stored, unsurprisingly, in the
array <code>$words</code>. So word 1 will be `<code>p4</code>' and word 2 the subcommand.</p>
<p>Hence if we are past word 2, we look at <code>${words[2]}</code> to get the
subcommand, and use that to decide what to do next. The change to
<code>$curcontext</code> is a bit of cleverness to make it easy for the user to
defined styles for particular subcommands; refresh your mind by looking
at the discussion of styles and contexts above if you need to. For
example, if you are completing after `<code>p4 diff</code>', the context will look
something like
`<code>:completion::complete:p4-diff:argument-1:opened-files</code>' where the
remainder says you are on the first argument and are complete the tag
`<code>opened-files</code>', We'll see down below how we tell the system to use
that tag; the `<code>argument-1</code>' is handled by the <code>_arguments</code> utility
function, which takes away a lot of the load of handling options and
arguments in a standard UNIX format.</p>
<p>Next, we pretend that the `<code>p4</code>' at the start wasn't there by removing
the front of <code>$words</code> and decrementing <code>$CURRENT</code> so as to reflect its
new position in <code>$words</code>. The reason for doing this is that we are going
to use <code>_arguments</code> for handling the subcommand. As is only sensible,
this function looks at the first element of <code>$words</code> to find the command
word, and treats the rest as options or arguments to the command.</p>
<p>We then dispatch the right function for the command simply by
constructing the name of the function on the fly. Of course it's a
little neater to check the function exists first;
<code>$+functions[_perforce_cmd_$cmd]</code> would come to our aid again.</p>
<p>However, if we're still on the second (original) word, we have to
generate a list of functions to complete. We will do this by asking
Perforce's help system to list them, and store the results in the array
<code>$cmdlist</code>. The loop has a couple of checks to remove blank lines and
the title line at the start. The remaining lines have a command and a
description. We take the command, but also tack the description on after
a colon --- we can then show the user the description, too, as a bit of
extra help.</p>
<p>Actually, the Perforce command that generates the list of subcommands is
simply `<code>p4 help command</code>'. (That's really all you need to know; skip
the rest of the paragraph if you just want the basics.) The
`<code>_call_program help-commands</code>' was stuck in front for the name of
configurability. Before executing the command, the system checks in the
current context with the given tag <code>help-commands</code> for the style
<code>command</code>. If it finds a value for that style, it will use that as the
command to execute in the place of the remaining arguments. If the style
it read began with <code>-</code>, then the command it was going to execute ---
i.e. `<code>p4 help commands</code>' is appended to the end of the command read
from the style, so that the user's command can process the original
command if it needs to. This is really extreme sophistication; you will
rarely actually need the <code>command</code> style, but if you are writing a
completion for others to use it's polite to give them a chance to
intercept calls in this way.</p>
<p>The <code>_describe</code> command then does the work for us. The `<code>-t p4-commands</code>' gives the tag we are going to use; the convention is that
tag names are plural, though there's nothing to enforce this. Then we
give an overall description --- this is what appears after
`<code>Completing </code>' in the examples of the <code>format</code> style above; if you
don't have that set, you won't see it. Finally, we give the array name
--- note it is the <em>name</em>, not the substituted value. This is more
efficient because the shell doesn't need to extract the values until the
last minute; until then it can pass around just the single word. The
<code>_description</code> function knows about the `<code>completion:description</code>'
syntax; reread what I said about the <code>verbose</code> style for what the system
does with the descriptions for the completion.</p>
<p>The <code>_describe</code> function is one level above the completion system's
basic builtin command, <code>compadd</code>; it just knows about a single tag, with
a little icing sugar to display verbose descriptions. Later, we'll see
ways of building up alternatives where different types of completion can
be completed at the same point. There are lots of ways of doing this;
some of the more complicated are relegated to the detailed descriptions
that follow the tutorial.</p>
<p><span id="l179"></span></p>
<h3 id="682-subcommand-completion-_arguments"><a class="header" href="#682-subcommand-completion-_arguments">6.8.2: Subcommand completion: <code>_arguments</code></a></h3>
<p>Suppose we are now completing after `<code>p4 diff</code>'. We have altered the
command line so that the function now sees the `<code>diff</code>' as the first
word, as if this were the command. This makes the next step easier; the
<code>_arguments</code> function won't see irrelevant words on the command line,
since it is designed to handle the arguments to a simple command in the
standard form `<code>command [ options ] arguments ...</code>'. Here's the simple
version.</p>
<pre><code>
_perforce_cmd_diff() {
_arguments -s : \
'-f[diff every file]' \
'-t[include non-text files]' \
'(-sd -se -sr)-sa[opened files, different or missing]' \
'(-sa -se -sr)-sd[unopened files, missing]' \
'(-sa -sd -sr)-se[unopened files, different]' \
'(-sa -sd -se)-sr[opened files, same as depot]' \
'-d-[select diff option]:diff option:'\
'((b\:ignore\ blanks c\:context n\:RCS s\:summary'\
'u\:unified w\:ignore\ all\ whitespace))' \
"*::file:_perforce_files"
}
</code></pre>
<p>I've split the argument beginning <code>-d</code> into three lines to fit, but it's
just a single argument. Also, for clarity I've missed out the line with
the `<code>$+functions</code>' test to see if <code>_perforce_cmd_diff</code> was already
defined; I'll forget about that for now.</p>
<p>The function <code>_arguments</code> has been described as having `the syntax from
hell', but with the arguments already laid out in front of you it
doesn't look so bad. The are three types of argument: options to
<code>_arguments</code> itself, arguments saying how to handle options to the
command (i.e. `<code>p4 diff</code>'), and arguments saying how to handle normal
arguments to the command.</p>
<p>The first two are for <code>_arguments</code> itself; `<code>-s</code>' tells it that
single-letter options are allowed, i.e. they can be combined as in
`<code>-ft</code>'. Luckily for our purposes, that doesn't stop us having multiple
word options, too. The colon on its own then says everything else is an
argument relating to the command line being handled.</p>
<p>We then start off with some simple options; as you can probably guess
straight away, the first two say that `<code>p4 diff -f</code>' passes a flag to
say any file can be diff'ed (not just ones open for editing), and that
`<code>p4 diff -t</code>' passes a flag to say that binary files can be diff'ed
(not just text files). Note the use of square brackets for giving a
description; this is handled by the <code>verbose</code> style as I mentioned for
<code>_describe</code>. In fact, the list of possible options and arguments,
suitably rearranged, will end up passing through <code>_describe</code>. The
descriptions in square brackets are optional, as the use of square
brackets might suggest; you could just have `<code>-f</code>' and `<code>-t</code>' (making
it fairly obvious why the `<code>:</code>' to separate off _arguments's own
options is a good idea).</p>
<p>The next step in complexity is that set of functions with the list in
parentheses in front. These give mutually exclusive options. In other
words, if there's already a <code>-sa</code> on the command line, don't complete
any of <code>-sd</code>, <code>-se</code> or <code>-sr</code>, and so on. (Remember that by default you
need to type the first `<code>-</code>' of an option, or the system will go
straight to normal arguments, which we'll come to in a moment.)</p>
<p>Next comes the specification for the option <code>-d</code>. All those colons
indicate that this option has an argument, and the <code>-</code> following
straight after the <code>-d</code> indicates that it has to be in the same word,
i.e. follow the <code>-d</code> without a space. After the first colon comes a
description for the argument. This is what you see when you try to
complete the after <code>-d</code>; compare this with the expression in square
brackets before, which is what you see when you try to complete the <code>-d</code>
itself. Then after the second colon is an expression saying how to
complete that argument.</p>
<p>This final part of the specification for an option with an argument can
take various forms. The simplest is just a single space; this means
there's nothing to complete, but the system is aware the user needs to
type something for that word and can prompt with the description. The
next simplest is a set of words in parentheses: here, we could have had
`<code>(b c n s u w)</code>'. Instead, we've had a variant on that which gives yet
another set of descriptions, namely those for the individual completions
that appear after <code>-d</code>. Note various things: the parentheses are doubled
and the colons and spaces within the completion options are backslashed.
All of these are simply there to make it easy for <code>_arguments</code> to parse
the string. The upshot of this is that in the following context:</p>
<pre><code> p4 diff -d
</code></pre>
<p>a verbose completion using the <code>format</code> style as described above looks
like:</p>
<pre><code> Completing diff option
b # ignore blanks
c # context
n # RCS
s # summary
u # unified
w # ignore all whitespace
</code></pre>
<p>or similar --- I have the <code>list-separator</code> style set to `<code>#</code>', because
it looks like a comment normal shell syntax, but in your case you may
get `<code>-``-</code>' as the separator.</p>
<p>(In case you were wondering why the colons needed to be quoted when it
seemed you'd already got to the last argument: it's possible for options
to have multiple arguments, and you can continue having sets of
<code>:</code><em>description</em><code>:</code><em>action</em> pairs. This means the system needs some way
of distinguishing these colons from ones inside arguments. While I'm
digressing, you may also have noticed that I could have written the
<code>-s</code><em>X</em> as an option with arguments, in which case you can have a bonus
point.)</p>
<p>The final argument starts with a `<code>*</code>', which means it applies to all
remaining arguments to `<code>p4 diff</code>' after the options have been
processed. Most of the rest is similar to the form for options, except
for the doubled colon, which indicates that <code>$CURRENT</code> and <code>$words</code>
should be altered to reflect only the arguments being handled by this
argument specifier --- exactly what we did before calling
<code>_perforce_cmd_diff</code> in the first place, in fact. As we mentioned
before, this makes the next step of processing easier if happens to call
<code>_arguments</code> again. (Actually it doesn't in this case.) The `<code>file</code>'
then describes the arguments and the final part, <code>_perforce_files</code>,
tells the system to call that function to complete a file name.</p>
<p>There are numerous (it sometimes seems, endless) subtleties to
<code>_arguments</code>. I won't try to go into them in the tutorial; see the
description of <code>_arguments</code> below for something more detailed to refer
to, and if you are feeling <em>really</em> brave look at the description in the
<code>zshcompsys</code> manual page. Even better, dig into one of the existing
completion functions --- something handling completion for a UNIX
command is probably good, since these make heavy use of <code>_arguments</code> ---
and see how those work. Despite the complexities, I would definitely
suggest using <code>_arguments</code> wherever possible to take away any need on
your part to do processing of command line arguments.</p>
<p><span id="l180"></span></p>
<h3 id="683-completing-particular-argument-types"><a class="header" href="#683-completing-particular-argument-types">6.8.3: Completing particular argument types</a></h3>
<p>Now we'll look inside the <code>_perforce_files</code> function as an example of
the nitty gritty of completing one particular type of argument, which
might have some quite complicated internal structure. This is true in
Perforce as the filename can have extra information tacked on the end:
`<code>file#</code><em>revision</em>' indicates the revision of a file,
`<code>file@</code><em>change</em>' indicates a change status, and in some cases you can
get `<code>file@</code><em>change1</em><code>,</code><em>change2</em>' to indicate a range of changes
(likewise revisions). Furthermore, <code>file</code> can be specified in different
ways, and the file to be completed may be limited by some kind of
context information. We'll start from simple filenames and gradually add
these possibilities in.</p>
<p><strong>Different types of file, part 1</strong></p>
<p>There are so many possibilities for files that I'm going to split up
<code>_perforce_files</code> into individual functions handling different aspects.
For example, even if we are just handling ordinary files in the way the
completion system normally does, Perforce commands understand a special
file name `<code>...</code>' which means `every subdirectory to any depth'.
(Interestingly, zsh used to have this to mean the same thing, instead of
`<code>**</code>'; it was changed in zsh because as the `<code>.</code>'s are regular
characters there's no easy way of quoting them. You didn't need to know
this.)</p>
<p>I'm going to say we can complete both like this:</p>
<pre><code> _alternative \
"files:file:_path_files" \
"subdirs:subdirectory search:_perforce_subdir_search"
</code></pre>
<p>The function <code>_alternative</code> is a little bit like <code>_arguments</code>, but
thankfully much simpler. It's name gives away its purpose; every
argument specifies one of a set of possible alternatives, all of which
are valid at that point --- so the user is offered anything which
matches out of the choices, unlike <code>_arguments</code>, which has to decide
between the various possibilities. It's a sort of glorified loop around
`<code>_describe</code>', with <code>_arguments</code>'s conventions on the action for
generating completions (up to a point --- <code>_alternative</code> doesn't have
all the whackier ones, though it does have the ones I've been talking
about so far).</p>
<p>Each set of possibilities consists of the name of a tag, a description,
and an argument. The tag isn't present in <code>_arguments</code>. If you use <code>^xh</code>
to tell you about valid tags, you'll see <code>_arguments</code> has its own
generic tag, <code>argument-rest</code>; this isn't usually all that useful, so we
are going to supply more specific ones.</p>
<p>In the first possibility, it's the standard one for files, `<code>files</code>.
The function is the basic low-level one for completing files, too; it's
described below, but you already know a lot about the effect since it's
the completion system's workhorse which you use it all the time without
realising. Actually, it will supply its own tags, but that doesn't
matter since they will silently override what we say.</p>
<p>The second possibility is the new one we're adding. I've therefore
invented a suitable tag `<code>subdirs</code>', a description, `<code>subdirectory search</code>', and the name of the function I'm going to supply to do the
completion. This is quite simple:</p>
<pre><code> _perforce_subdir_search() {
compset -P '*/'
compadd "$@" '...'
}
</code></pre>
<p>The first line tells the completion system to ignore anything up to the
last `<code>/</code>'. That's so we can append a `<code>...</code>' to any directory which
already exists on the command line. The builtin <code>compset</code> does various
low-level transformations of this time. Note that the <code>-P</code> is `greedy'
--- it looks for the longest possible pattern match, which is the usual
default in zsh and other UNIX pattern matchers.</p>
<p>The second line actually adds the `<code>...</code>' as a completion; <code>compadd</code> is
the key builtin for the whole completion system. I've actually passed
some on the arguments which we got to `<code>_perforce_subdir_search</code>' via
`<code>"$@"</code>'. In fact, looking back it seems as if there weren't any!
However, <code>_alternative</code> actually passed some behind my back --- and it's
a good thing, too, since it's exactly those arguments that give the tag
`<code>subdirs</code>' and the description `<code>subdirectory search</code>'. So that extra
`<code>"$@"</code>' is actually quite important. The buck stops here; there's
nothing below <code>compadd</code>. A function of this simplest only works well
when the handling of tags and contexts has already been done; but we
just saw that <code>_alternative</code> did that, so as long as we always call
<code>_perforce_subdir_search</code> suitably, we're in the clear.</p>
<p><strong>Different types of file, part 2</strong></p>
<p>Furthermore, a Perforce file specification can look like a normal UNIX
file path, or it can look like:</p>
<pre><code> //depot/dirs/moredirs/file
</code></pre>
<p>(don't get confused with paths to network resources, which also use the
doubled slash or backslash on some systems, notably Cygwin). We could
use <code>_alternative</code> to handle this, too, and if I was writing <code>_perforce</code>
again I probably would for simplicity. However, I decided to do it just
by testing for the `<code>//</code>' in <code>_perforce_files</code>. This means that the
structure of <code>p4_files</code> so far looks like:</p>
<pre><code> if [[ $PREFIX = //* ]]; then
# ask Perforce for files that match
local -a altfiles
altfiles=(
'depot-files:file in depot:_perforce_depot_files'
depot-dirs:directory in depot:_perforce_depot_dirs'
)
# add other alternatives, such as the `...' thing
altfiles=($altfiles
"subdirs:subdirectory search:_perforce_subdir_search"
)
_alternative $altfiles
else
_alternative \
"files:file:_path_files" \
"subdirs:subdirectory search:_perforce_subdir_search"
fi
</code></pre>
<p>where we are still to write the functions for the first two alternatives
in the first branch; the `<code>...</code>' is still valid for that branch, so
I've added that as the third alternative. I've used the array
<code>$altfiles</code> because, actually, the structure is more complicated than
I've shown; doing it this way makes it easier to add different sets of
alternatives.</p>
<p>The choice of which branch is made by examining the <code>$PREFIX</code> special
variable, which contains everything (well, everything interesting) that
comes before the cursor position in the word being completed. There is a
counterpart <code>$SUFFIX</code> which we will see in a moment. The `almost
everything' comes because sometimes we definitely don't want to see the
whole <code>$PREFIX</code>. Completing the three dots was such as case --- we
didn't want to see anything up to the last <code>/</code>. What that `<code>compset -P '*/'</code>' actually did was move the matched pattern from the front of
<code>$PREFIX</code> to the end of <code>$IPREFIX</code>, another special parameter which
contains parts of the completion we aren't currently interested in, but
which are still there. This allows us to concentrate on a particular
part of the completion. However you do that --- whether by <code>compset</code> or
directly manipulating <code>$PREFIX</code> and friends --- the completion system
usually restores the parameters when you exit the function where you
altered them. This fits in nicely with what we're doing here with
<code>_alternative</code> --- if we handle adding `<code>...</code>' by ignoring everything
up to the last slash, for example, we don't want the next completion we
try to continue to ignore that; other file completions will want to look
at the directory path.</p>
<p>`Depot' is Perforce's name for what CVS calls a repository --- the
central location where all versions of all files are stored, and from
where they are retrieved when you ask to look at one. I've separated out
`<code>depot-dirs</code>' and `<code>depot-files</code>' for various reasons. First, the
commands to examine files and directories are different, so the
completion function is different. Second, we can offer different tags
for files and directories --- this is what <code>_path_files</code> does for normal
UNIX files. Third, it will later allow us more control --- some commands
only operate on directories. Here's <code>_perforce_depot_files</code>;
<code>_perforce_depot_dirs</code> is extremely similar:</p>
<pre><code> _perforce_depot_files() {
# Normal completion of files in depots
local pfx=${(Q)PREFIX} expl
local -a files
compset -P '*/'
files=(${${${(f)"$(\
_call_program files p4 files \
\"\$pfx\*\$\{\(Q\)SUFFIX\}\" 2>/dev/null)"}%\#*}##*/})
[[ $#files -eq 1 && $files[1] = '' ]] && files=()
compadd "$@" -a files
}
</code></pre>
<p>A little messy (and still not quite the full horror). I've split the key
line in the middle which fetches the list from Perforce to make it fit.
If you ploughed through chapter 5, you'll recognised what's going on
here --- we're reading a list of files, one per line, from the command
`<code>p4 files</code>', and we're stripping off the directory at the front, and
everything from a `<code>#</code>' on at the end. The latter is a revision number;
we're not handling those at this point, though we will later.</p>
<p>Notice the way I remembered <code>$PREFIX</code> before I told the system to ignore
it for the word we're now completing. I remembered it as
`<code>${(Q)PREFIX}</code>' in order to remove any quotes from the name. For
example, if the name on the line so far had a space, <code>$PREFIX</code> (which
comes from what is on the command line without any quotes being
stripped) would have the space quoted somehow, e.g. `<code>name\ with\ space</code>'. We arrange for <code>$pfx</code> to contain `<code>name with space</code>', which is
how Perforce knows the file, using the <code>(Q)</code> parameter flag. We then
pass the argument <code>"$pfx*${(Q)SUFFIX}"</code> to `<code>p4 files</code>'; this generates
matching files internally. The extra layer of backslash-quoting is for
the benefit of <code>_call_program</code>, which re-evaluates its arguments; this
ensures the argument is expanded at the point it gets passed to <code>p4 files</code>. All this goes to show just how difficult getting the quoting
right can be.</p>
<p>Once we've got the list of bare filenames, we check to see if the list
is just one element with no length. That's an artefact of the the
<code>"$(cmd)"</code> syntax; if the output is empty, because its quoted you still
get one zero-length string output, which we don't want.</p>
<p>Finally, we pass the result to <code>compadd</code> as before. Again, tags and the
description have already been handled and we just need to make sure the
appropriate options get passed in with <code>"$@"</code>. This time we use the
`<code>-a</code>' option which tells <code>compadd</code> that any arguments are array name,
not a list of completions. This is more efficient; compadd only needs to
expand the array internally instead of the shell passing a potentially
huge list to the builtin.</p>
<p><strong>Handling extra bits on a completion</strong></p>
<p>`Extra bits' on a completion could be anything; common examples include
an extra value for a comma-separated list (the <code>_values</code> functions is
for this), or some kind of modifier applied to the completion you have
already. We've already seen an example, in fact, since the principle of
handling the directory and basename parts of a file is very similar. The
phrase `extra bits' may already alert you to the fact that we are
heading towards the deeper recesses of completion.</p>
<p>Anyway, here's how we tack a revision or change number onto the end of a
file.</p>
<p>I'll stick with revisions: `<em>filename</em><code>#</code><em>revision</em>', where <em>revision</em>
is a number. For the full sophistication, there are three steps to this.
First, make it easy for the user to add `<code>#</code>' to an existing filename;
second, recognise that a `<code>#</code>' is already there so that revisions need
to be completed; third, find out the actual revisions which can be
completed. As a revision is just a number, you might think completing it
was a bit pointless. However, given the sophistication of zsh's
completion system there's actually one very good reason --- we can
supply a description with the revisions, so that the user is given
information about the revisions and can pick the right one without
running some external command to find out. There was the same sort of
rationale behind the `<code>-d</code>' option to <code>p4 diff</code>; there was just one
letter to type, but zsh was able to generate extra information to
describe the possibilities, so it wasn't just laziness.</p>
<p>First part: make it easy for the user to add the `<code>#</code>'. This actually
depends on a new feature in version 4.1 of zsh; in 4.0 you couldn't play
the trick we need or grabbing the keyboard input after a completion was
finished unless you specified a particular suffix to add to the
completion (such as the `<code>/</code>' after a directory --- this is
historically where this feature came from).</p>
<p>The method is to add an extra argument everywhere we complete a file
name. For example, change the <code>compadd</code> in <code>_perforce_depot_files</code> to:</p>
<pre><code> compadd "$@" -R _perforce_file_suffix -a files
</code></pre>
<p>where the option argument specifies a function:</p>
<pre><code> _perforce_file_suffix() {
[[ $1 = 1 ]] || return
if [[ $LBUFFER[-1] = ' ' ]]; then
if [[ $KEYS = '#' ]]; then
# Suffix removal with an added backslash
LBUFFER="$LBUFFER[1,-2]\\"
elif [[ $KEYS = (*[^[:print:]]*|[[:blank:]\;\&\|@]) ]]; then
# Normal suffix removal
LBUFFER="$LBUFFER[1,-2]"
fi
fi
}
</code></pre>
<p>This has been simplified, too; I've ignored revision ranges in the form
<em>file</em><code>#</code><em>rev1</em><code>,</code><em>rev2</em>. However, I've handled changes (`<code>@</code>'
following a filename) as well as revisions. You'll see this function
looks much more like a zle widget rather than a completion widget ---
which is exactly what it is; it's not called as part of the completion
system at all. After the specified completion, zle reads in the next
keystroke, which is stored in <code>$KEYS</code>, and calls this function as a zle
widget. This means it can manipulate the line buffer; we only need to
look at what is at the left of the cursor, stored in <code>$LBUFFER</code>.</p>
<p>The function is called with the length of the suffix added to the
function. In this case, it's just a space --- we've finished a normal
completion, so the system has automatically added a space to what's on
the command line. We therefore check we've just got one single character
in the suffix, to avoid getting confused.</p>
<p>Next, we look at what's immediately left of the cursor, which is the
last character in <code>$LBUFFER</code>, i.e. <code>$LBUFFER[-1]</code>, to make sure this is
a space.</p>
<p>If everything looks OK, we consider the keys typed and decide whether to
modify the line. You may already have noticed that in some cases zsh
automatically removes that space by itself; for example, if you hit
return --- or any other non-printing character --- or if it's a
character that terminates a command such as `<code>&</code>' or `<code>;</code>'. We emulate
that behaviour --- most of the second test is simply to do that. The
only differences from normal are if the key typed was `<code>@</code>' or `<code>#</code>'.</p>
<p>The `<code>@</code>' is simple --- we just remove the last character, the same as
we do for the other characters. For `<code>#</code>', however, we also add a
backslash to the command line before the `<code>#</code>'. That's because `<code>#</code>'
is a special character with extended globbing, and the completion system
generally runs with extended globbing switched on. Adding the backslash
means the user doesn't have to; it's never harmful.</p>
<p>To show the next effect, suppose we complete a file name:</p>
<pre><code> p4 diff fil<TAB>
</code></pre>
<p>to get:</p>
<pre><code> p4 diff filename _
</code></pre>
<p>where `<code>_</code>' shows the cursor position, and then typed `<code>#</code>'; we would
get:</p>
<pre><code> p4 diff filename\#
</code></pre>
<p>with the cursor right at the end.</p>
<p>So far so good. For the second step, we need to modify <code>_perforce_files</code>
to spot that there is a `<code>#</code>' on the line before the cursor, and to
call the revision code. To do this we add an extra branch at the start
of the `<code>if</code>' in <code>_perforce_files</code> --- at the start, because any `<code>#</code>'
before the cursor forces us to look at revisions, so this takes
precedence over the other choices. When this is added, the code will
look like:</p>
<pre><code> if [[ -prefix *\# ]]; then
_perforce_revisions
elif [[ $PREFIX = //* ]]; then
# as before.
</code></pre>
<p>In fact, that <code>-prefix</code> test is just a fancy way of saying the same
thing as the `<code>[[ $PREFIX = *\# ]]</code>' and if I wasn't so hopelessly
inconsistent I would have written both tests the same.</p>
<p>So now the third step: write <code>_perforce_revisions</code> to complete revisions
numbers with the all-important descriptions.</p>
<pre><code> _perforce_revisions() {
local rline match mbegin mend pfx
local -a rl
pfx=${${(Q)PREFIX}%%\#*}
compset -P '*\#'
# Numerical revision numbers, possibly with text.
if [[ -z $PREFIX || $PREFIX = <-> ]]; then
# always allowed (same as none)
rl=($rl 0)
_call_program filelog p4 filelog \$pfx 2>/dev/null |
while read rline; do
if [[ $rline = (#b)'... #'(<->)*\'(*)\' ]]; then
rl=($l "${match[1]}:${match[2]}")
fi
done
fi
# Non-numerical (special) revision names.
if [[ -z $PREFIX || $PREFIX != <-> ]]; then
rl=($rl 'head:head revision' 'none:empty revision'
'have:current synced revision')
fi
_describe -t revisions 'revision' rl
}
</code></pre>
<p>Thankfully, a lot of the structure of this is already familiar. We
extract the existing prefix before the `<code>#</code>', being careful about
quoting --- this is the filename for which we want a list of revisions.
We ignore everything in the command argument before the `<code>#</code>'. After
generating the completions, we use the <code>_describe</code> function to add them
with the tag `<code>revisions</code>' and the description `<code>revision</code>'.</p>
<p>The main new part is the loop over output from `<code>p4 filelog</code>', which is
the Perforce command that tells us about the revisions of a file. We
extract the revision number and the comment from the line using
backreferences (see previous chapter) and weld them together with a
colon so that <code>_describe</code> will be able to separate the completion from
its description. Then we add a few special non-numerical revisions which
Perforce allows, and pass this list down to <code>_describe</code>. The extra
<code>if</code>'s are a very minor optimization to check if we are completing a
numerical or non-numerical revision.</p>
<p><span id="l181"></span></p>
<h3 id="684-the-rest"><a class="header" href="#684-the-rest">6.8.4: The rest</a></h3>
<p>It's obvious that this tutorial could expand in any number of
directions, but as it's really just to point out some possibilities and
directions, that would would miss the point. So the rest of this chapter
takes the completion system apart and looks at the individual
components. It should at least now be a bit more obvious where each
component fits.</p>
<p><span id="l182"></span></p>
<h2 id="69-writing-new-completion-functions-and-widgets"><a class="header" href="#69-writing-new-completion-functions-and-widgets">6.9: Writing new completion functions and widgets</a></h2>
<p>Now down to the nitty gritty. When I first talked about new completion,
I explained that the functions beginning `<code>_</code>' were the core of the
system. For the remainder of the chapter, I'll explain what goes in them
in more detail than I did in the tutorial. However, I'll try to do it in
such a way that you don't need to know every single detail. The trade
off is that if you just use the simplest way of writing functions, many
of the mechanisms I told you about above, particularly those involving
styles and tags, won't work. For example, much of the code that helps
with smart formatting of completion listings is buried in the function
`<code>_description</code>'; if you don't know how to call that --- which is often
done indirectly --- then your own completions won't appear in the same
format as the pre-defined ones.</p>
<p>The easiest way of getting round that is to take a dual approach: read
the following as far as you need, but also try to find the existing
completion that comes nearest to meeting your needs, then copy that and
change it. For example, here's a function that completes files ending in
<code>.gz</code> (the supplied function which does this has now changed), which are
files compressed by the <code>gzip</code> program, for use by the corresponding
program that does decompression, <code>gunzip</code> --- hence the file and
function are called <code>_gunzip</code>:</p>
<pre><code> #compdef gunzip zcat
local expl
_description files expl 'compressed file'
_files "$expl[@]" -g '*.[gG][zZ]'
</code></pre>
<p>You can probably see straight away that if you want to design your own
completion function for a command which takes, say, files ending in
<code>.exe</code>, you need to change three things: the line at the top, which
gives the names of programmes whose arguments are to be completed here,
the description `<code>compressed file</code>' to some appropriate string, and the
argument following the <code>-g</code> to something like <code>'*.exe'</code> --- any globbing
pattern should work, just remember to quote it, since it shouldn't be
expanded until the inside of the function <code>_files</code>. Once you've
installed that somewhere in your <code>$fpath</code> and restarted the shell,
everything should work, probably following a longer pause than usual as
the completion system has to rescan every completion function when it
finds there is a new one.</p>
<p>What you might miss is that the first argument to <code>_description</code>,
`<code>files</code>', is the all-important mystical tag for the type of
completion. In this case, you would probably want to keep it. Indeed,
the <code>_files</code> function is used for all file completions of any type, and
knows all about the other tags --- <code>globbed-files</code>, <code>directories</code>,
<code>all-files</code> --- so virtually all your work's done for you here.</p>
<p>If you're adding your own functions, you will need your own functions
directory. This was described earlier in this guide, but just to remind
you: all you need to do is create a directory and add it to <code>$fpath</code> in
either <code>.zshenv</code> (which a lot of people use) or <code>.zshrc</code> (which some
sticklers insist on, since it doesn't affect non-interactive shells):</p>
<pre><code> fpath=(~/funcs $fpath)
</code></pre>
<p>It's best to put it before the standard completion directories, since
then you can override a standard completion function simply by copying
it into your own directory; that copy will then be found first and used.
This is a perfectly reasonable thing to do with any completion function
--- although if you find you need to tweak one of the larger standard
functions, that's probably better done with styles, and you should
suggest this to us.</p>
<p><span id="l183"></span></p>
<h3 id="691-loading-completion-functions-compdef"><a class="header" href="#691-loading-completion-functions-compdef">6.9.1: Loading completion functions: <code>compdef</code></a></h3>
<p>The first thing to understand is that top line of <code>_gunzip</code>. The
`<code>#compdef</code>' tag is what tells the system when it checks through all
files beginning with `<code>_</code>' that this is a function implementing a
completion. Files which don't directly implement completions, but are
needed by the system, instead have the single word `<code>#autoload</code>' at
that point. All files are only loaded when needed, using the usual
autoloading system, to keep memory usage down.</p>
<p>You can supply various options to the `<code>#compdef</code>' tag; these are
listed in the `<code>Initialization</code>' section of the <code>zshcompsys(1)</code> manual
page or `<strong>Completion System</strong>' info node. The most useful are <code>-k</code> and
<code>-K</code>, which allow you to define a completion command and binding rather
than a function used in a particular context. There are also <code>-p</code> and
<code>-P</code> which tell the system that what follows is a pattern rather than a
literal command name; any command matching the pattern will use that
completion function, unless you used <code>-P</code> and a normal (non-pattern)
completion function for the name was found first.</p>
<p>For normal <code>#compdef</code> entries, however, what comes next is a list of
command names --- or rather a list of contexts, since the form
`<code>-context-</code>' can be used here. For example, the function <code>_default</code>
has the line `<code>#compdef -default-</code>'. You can give as many words as you
like and that completion will be used for each. Note that contexts in
the colon-separated form can't appear here, just command names or the
special contexts named with hyphens.</p>
<p>The system does its work by using a function <code>compdef</code>; it gets as
arguments more or less what you see, except that the function name is
passed as the first argument. Thus the <code>_gunzip</code> completion is loaded by
`<code>compdef _gunzip gunzip zcat</code>', <code>_default</code> by `<code>compdef _default -default-</code>', and so on. This simply records the name of the function
handling the context in the <code>$_comps</code> associative array which you've
already met. You can make extra commands/contexts be handled by an
existing completion function in this way, too; this is generally more
convenient than copying and modifying the function. Just add `<code>compdef <_function> <command-to-handle></code>' to <code>.zshrc</code> after the call to
<code>compinit</code>.</p>
<p>It's also high time I mentioned an easy way of using the completion
already defined for an existing function: `<code>compdef newcmd=oldcmd</code>'
tells the completion system that the completion arguments for
`<code>newcmd</code>' are to be the same as the ones already defined for
`<code>oldcmd</code>'; it will complain if nothing is known about completing for
<code>oldcmd</code>. This works recursively; you can now define completions in
terms of that for <code>newcmd</code>. If you happen to know the name of the
completion function called, you can use that; the following three lines
are broadly equivalent:</p>
<pre><code> compdef $_comps[typeset] foo
compdef _vars_eq foo
compdef foo=typeset
</code></pre>
<p>since the completion for <code>typeset</code> is stored in <code>$_comps</code> along with all
the others, and this happens to resolve to <code>_vars_eq</code>; but the last
example is easier and safer and the intention more obvious. The manual
refers to <code>typeset</code> here as a `service' for <code>foo</code> (guess what the shell
stores in the associative array element <code>$_services[foo]</code>).</p>
<p>There's actually more to services: when a function is called, the
parameter <code>$service</code> is set. Usually this will just be the name of the
command being completed for, or one of the special contexts like
`<code>-math-</code>'. However, in a case like the last <code>compdef</code> in the list
above, the service will be <code>typeset</code> even though the command name may be
`<code>foo</code>'.</p>
<p>This is also used in `<code>#compdef</code>' lines. The top of `<code>_gzip</code>'
contains:</p>
<pre><code> #compdef gzip gunzip gzcat=gunzip
</code></pre>
<p>which says that the file provides two services, for <code>gzip</code> and <code>gunzip</code>,
and also handles completion for <code>gzcat</code>, but with the service name
<code>gunzip</code>. Only a few of the completion functions actually care what
service they provide (you can check, obviously, by looking to see if
they refer to <code>$service</code>); but you may have uses for this. Note that if
you define services with a <code>compdef</code> command, <em>all</em> the arguments must
be in the <em>foo</em><code>=</code><em>bar</em> form; the mixed form is only useful after a
<code>#compdef</code> inside completion functions.</p>
<p><span id="l184"></span></p>
<h3 id="692-adding-a-set-of-completions-compadd"><a class="header" href="#692-adding-a-set-of-completions-compadd">6.9.2: Adding a set of completions: <code>compadd</code></a></h3>
<p>Once you know how to make a new completion function, there is only one
other basic command you need to know before you can create your own
completions yourself. This is the builtin <code>compadd</code>. It is at the heart
of the completions system; all its arguments, after the options, are
taken as possible completions. This is the list from which the system
selects the possibilities that match what you have already typed. Here's
a very basic example which you can type or paste at the command line:</p>
<pre><code> _foo() { compadd Yan Tan Tethera; }
compdef _foo foo
</code></pre>
<p>Now type `<code>foo </code>' and experiment with completions after it. If only
it were all that simple.</p>
<p>There are a whole list of options to <code>compadd</code>, and you will have to
look in the <code>zshcompwid(1)</code> manual page or the `<strong>Completion Widgets</strong>'
info node for all of them. I've already mentioned <code>-M</code> and (long ago)
<code>-f</code>. Here are other interesting ones. <code>-X <description></code> provides a
description --- this is used by the <code>format</code> style to pass descriptions,
and if you use the normal tags system you shouldn't pass it directly;
I'll explain this later.</p>
<p><code>-P <prefix></code> and <code>-S <suffix></code> allow you to specify bits which are not
treated as part of the completion, but appear on the line none the less.
In fact, they do two different things: if the prefix or suffix is
already there, it is ignored, and if it isn't, it is inserted. There are
also corresponding hidden and ignored prefixes, necessary for the full
power of the completion system, but you will need to read the manual for
the full story. The <code>-q</code> option is useful with <code>-S</code>; it enables
auto-remove behaviour for the suffix you gave, just like <code>/</code> with the
<code>AUTO_REMOVE_SLASH</code> option when completing filenames.</p>
<p><code>-J <group></code> is the way group names are specified, used by the
<code>group-name</code> tag; there is also <code>-V <group></code>, but the group here is not
sorted (and is distinct from any group of the same name passed to <code>-J</code>).
<code>-Q</code> tells the completion code not to quote the words --- this is useful
where you need to have unquoted metacharacters in the final completion.
It is also useful when you are completion something where the result
isn't going to be expanded by the shell.</p>
<p><code>-U</code> tells <code>compadd</code> to use the list of completions even if they don't
match what's on the command line; you will need this if your completion
function modifies the prefix or suffix so that they no longer fit what's
already there. If you use this, you might consider turning on menu
completion (using <code>compstate[insert]=menu</code>), since it might otherwise be
difficult to select the appropriate completion.</p>
<p>Finally, note the <code>-F</code> and <code>-W</code> options which I describe below for
<code>_files</code> actually are options to <code>compadd</code> too.</p>
<p><span id="l185"></span></p>
<h3 id="693-functions-for-generating-filenames-etc"><a class="header" href="#693-functions-for-generating-filenames-etc">6.9.3: Functions for generating filenames, etc.</a></h3>
<p>However, for most types of completion the possibilities will not be a
simple list of things you already know, so that you need to have some
way of generating the required values. In this section, I will describe
some of the existing functions you can call to do the hard work. In the
next section I will show how to retrieve information from some special
parameters made available by the <code>zsh/parameter</code> module.</p>
<p><strong>Files etc.: the function <code>_files</code></strong></p>
<p>You have already seen <code>_files</code> in action. Calling this with no arguments
simply adds all possible files as completions, taking account of the
word on the command line to establish directories and so on.</p>
<p>For more specific use, you can give it various options: `<code>-/</code>' means
complete directories, and, as you saw, `<code>-g "<pattern>"</code>' gives a
filename generation pattern to produce matching files.</p>
<p>A couple of other options, which can be combined with the ones above,
are worthy of mention. If you use `<code>-W <dir></code>', then completion takes
place under directory <code><dir</code>> rather than in the current directory ---
it has no effect if you are using an absolute path. Here, `<code><dir></code>' can
also be a set of directories separated by spaces or, most usefully since
it avoids any problems with quoting, the name of an array variable which
contains the list of possible directories. This is essentially how
completion for <code>cd</code> with the <code>$cdpath</code> array works. So if you have a
program that looks for files with the suffix `<code>.mph</code>', first in the
current directory, then in a standard directory, say,
<code>/usr/local/oomph</code>', you can do this:</p>
<pre><code> local oomph_dirs
oomph_dirs=(. /usr/local/oomph)
_files -W oomph_dirs -g '*.mph'
</code></pre>
<p>--- note there is no `<code>$</code>' before the variable <code>$oomph_dirs</code> here,
since it should only be expanded deep inside <code>_files</code>.</p>
<p>The system that implements <code>$fignore</code> and the <code>ignored-patterns</code> style
can be intercepted, if you need to, with the option `<code>-F "<pat>"</code>';
`<code><pat></code>' is an array of patterns to ignore, in the usual completion
format, in other words the name of a real shell array, or a list of
values inside parentheses. If you make sure all the tags stuff is
handled properly, <code>ignored-patterns</code> will work automatically, however,
and in addition extended globbing allows you to specify patterns with
exclusion directly, so you probably won't use this feature directly
unless you're in one of your superhero moods.</p>
<p>In addition, <code>_files</code> also takes many of the standard completion options
which apply to <code>compadd</code>, for convenience.</p>
<p>Actually, the function <code>_path_files</code> is the real engine room of the
system. The advantage of using <code>_files</code> is that it prepares all the tags
for you, deciding whether you want directories to be completed as well
as the globbed files, and so on. If you have particularly specific needs
you can use <code>_path_files</code> directly, but you won't get the automatic
fallback one <code>directories</code> and <code>all-files</code>. Because it doesn't handle
the tags, <code>_path_files</code> is too lowly to do the usual tricks with label
loops, i.e. pretending `<code>dog:-setter</code>' is a tag `<code>dog-setter</code>' with
the usual completions for `<code>dog</code>'; likewise, it doesn't implement the
<code>file-patterns</code> style. So you need to know what you're doing when you
use it directly.</p>
<p><strong>Parameters and options</strong></p>
<p>These can be completed by calls to the <code>_parameters</code> and <code>_options</code>
functions, respectively. Both set up their own tags, and <code>_options</code> uses
the matching control mechanism described above to allow options to be
given in all the available forms. As with <code>_files</code>, they will also pass
standard <code>compadd</code> options down to that function. Furthermore, they are
all at a high enough level to handle tags with labels: to translate that
into English, you can use them directly without any of the preprocessing
described later on which are necessary to make sure the styles dealing
with tags are respected.</p>
<p>For more detailed control with options, the functions <code>_set_options</code> and
<code>_unset_options</code> behave like <code>_options</code>, but the possible completions
are limited to options which are set or unset, respectively. However,
it's not that simple: the completion system itself alters the options,
and you need to enable some code near the top of <code>_main_complete</code> (it's
clearly marked) to remember the options which were set or unset when
completion started. A straw poll based on a sample of two zsh developers
revealed that in any case many people don't like the completion system
to second guess the options they want to set or unset in this way, so
it's probably better just to stick to <code>_options</code>.</p>
<p><strong>Miscellaneous</strong></p>
<p>There are also many other completion functions adding matches of a
certain type. These can be used in the same way as <code>_parameters</code> and
<code>_options</code>; in other words they do all the work needed for tags
themselves and can be given options for <code>compadd</code> as arguments.
Normally, these functions are named directly after the type of matches
they generate, like <code>_users</code>, <code>_groups</code>, <code>_hosts</code>, <code>_pids</code>, <code>_jobs</code>,
etc.</p>
<p><span id="l186"></span></p>
<h3 id="694-the-zshparameter-module"><a class="header" href="#694-the-zshparameter-module">6.9.4: The <code>zsh/parameter</code> module</a></h3>
<p>The new completion system automatically makes the <code>zsh/parameter</code> module
available for use. This provides an easy way of generating arguments for
<code>compadd</code>. To get the maximum use out of this, you should be familiar
with zsh's rather self-willed syntax for extracting bits out of
associative arrays. Note in particular <code>${(k)assoc}</code>, which expands to a
list of the keys of the associative array <code>$assoc</code>, <code>${(v)assoc}</code>, which
expands to just its values (actually, so does <code>$assoc</code> on its own), and
<code>${(kv)assoc}</code> which produces key/value pairs. For all intents and
purposes, the keys and values, or the pairs of them, are in a random
order, but as the completion system does it's own sorting that shouldn't
be a problem. Mostly, the important parts for completion are in the
keys, i.e. to add all aliases as possible completions, you need
`<code>compadd ${(k)aliases}</code>'.</p>
<p>Here's a list of associative and ordinary arrays provided; for more
information on the values of the associative arrays, which could be
useful in some cases, consult the section <strong>The zsh/parameter Module</strong>
in the <code>zshmodules(1)</code> manual page or the corresponding info node.
First, the associative arrays.</p>
<ul>
<li><strong><code>$aliases</code>, <code>$dis_aliases</code>, <code>$galiases</code></strong><br />
The keys of these arrays give ordinary aliases, disabled ordinary
aliases for those where you have done <code>disable -a <alias></code> to turn
them off temporarily, and global aliases as defined with <code>alias -g</code>.</li>
<li><strong><code>$builtins</code>, <code>$dis_builtins</code></strong><br />
The keys give active and disabled shell builtin commands.</li>
<li><strong><code>$commands</code></strong><br />
The keys are all external commands stored in the shells internal
tables; it does this both for the purposes of fast completion, and
to avoid having to search each time a command is executed. It's
possible that a command is missing or incorrectly stored if the
contents of your <code>$path</code> directories has changed since the shell
last updated its tables; the <code>rehash</code> command fixes it.</li>
<li><strong><code>$functions</code>, <code>$dis_functions</code></strong><br />
The keys are active and disabled shell functions.</li>
<li><strong><code>$history</code></strong><br />
Here, the <em>values</em> are complete lines stored in the internal
history. The keys are the numbers of the history line; it's an
associative, rather than an ordinary, array because they don't
necessarily start at line 1. However, see the <code>historywords</code>
ordinary array below.</li>
<li><strong><code>$jobtexts</code>, <code>$jobdirs</code>, <code>$jobstates</code></strong><br />
These give you information about jobs; the keys are the job numbers,
as presented by the <code>jobs</code> command, and the values give you the
other information from jobs: <code>$jobtexts</code> tells you what the job is
executing, <code>$jobdirs</code> its working directory, and <code>$jobstates</code> its
state, where the bit before the colon is the most useful as it
refers to the whole job. The remainder describes the state of
individual processes in the job.</li>
<li><strong><code>$modules</code></strong><br />
The keys give the names of modules which are currently available to
the shell, i.e. loaded or to be autoloaded, essentially the same
principle as with functions.</li>
<li><strong><code>$nameddirs</code></strong><br />
If you have named directories, either explicitly (e.g. assigning
`<code>foo=/mydir</code>' and using `<code>~foo</code>') or via the <code>AUTO_NAME_DIRS</code>
option, the keys of this associative array give the names and the
values the expanded directories.</li>
<li><strong><code>$options</code>, <code>$parameters</code></strong><br />
The keys give shell options and parameters, and are used by the
functions <code>_options</code> and <code>_parameters</code> for completion, so you will
mostly not need to refer to them directly.</li>
<li><strong><code>$userdirs</code></strong><br />
The keys give all the users on the system. The values give the
corresponding home directory, so `<code>${userdirs[juser]}</code>' is
equivalent to having <code>~juser</code> expanded and is thus not all that
interesting, except that by doing it this way you can test whether
the expansion exists without causing an error.</li>
</ul>
<p>Now here are the ordinary arrays, which you would therefore refer to
simply as <code>${reswords}</code> etc.</p>
<ul>
<li><strong><code>$dirstack</code></strong><br />
This contains your directory stack, what you see with `<code>dirs -v</code>'.
Note, however that the current directory, which appears as number 0
with that command, doesn't appear in <code>dirstack</code>. Of course it's easy
to add it to a completion if you want.</li>
<li><strong><code>$funcstack</code></strong><br />
This is the call stack of functions, i.e. all the functions which
are active at the time the array was referenced. <code>^Xh</code> uses this to
display which functions have been called for completion.</li>
<li><strong><code>$historywords</code></strong><br />
Unlike <code>$history</code>, this contains just the individual words of the
shell's command line history, and is therefore likely to be more
useful for completion purposes.</li>
<li><strong><code>$reswords</code>, <code>$dis_reswords</code></strong><br />
The active and disabled reserved words (effectively syntactically
special commands) understood by the shell.</li>
</ul>
<p><strong>Other ways of getting at information</strong></p>
<p>Since the arguments to <code>compadd</code> undergo all the usual shell expansions,
it's easy to get words from other sources for completion, and you can
look in the existing completion functions for many examples. A good
understanding of zsh's parameter and command expansion mechanisms and a
strong stomach will be useful here.</p>
<p>For example, here is the expansion used by the <code>_limits</code> function to
retrieve the names of resource limits from the <code>limit</code> command itself:</p>
<pre><code> print ${${(f)"$(limit)"}%% *}
</code></pre>
<p>which you can test does the right thing. Here's a translation:
<code>"$(limit)"</code> calls the command in a quoted context, which means you get
the output as if it were a single file (just type `<code>limit</code>' to see what
that is). <code>${(f)...}</code> splits this into an array (it is now outside
quotes, so splitting will generate an array) with one element per line.
Finally, <code>${...%% *}</code> removes the trailing end of each array element
from the first piece of whitespace on, so that `<code>cputime unlimited</code>' is
reduced to `<code>cputime</code>', and so on. Type `<code>limit ^D</code>', and you will see
the practical upshot of this.</p>
<p>That's by no means the most complicated example. The nested expansion
facility is used throughout the completion functions, which adds to
brevity but subtracts considerably from readability. It will repay
further study, however.</p>
<p><span id="l187"></span></p>
<h3 id="695-special-completion-parameters-and-compset"><a class="header" href="#695-special-completion-parameters-and-compset">6.9.5: Special completion parameters and <code>compset</code></a></h3>
<p>Up to now, I've assumed that at the start of your completion function
you already know what to complete. In more complicated cases that won't
be the case: different things may need completing in different arguments
of a command, or even some part of a word may need to be handled
differently from another part, or you need to look for a word following
a particular option. I will first describe some of the lower level
facilities which allow you to manipulate this; see the manual page
<code>zshcompwid(1)</code> or the info node <strong>Completion Widgets</strong> for the details
of these. Later, I will show how you can actually skip a lot of this for
ordinary commands with options and arguments by using such functions as
<code>_arguments</code>, where you simply specify what arguments and options the
function takes and what sort of completion they need.</p>
<p>The heart of this is the special parameters made available in completion
for testing what has already been typed. It doesn't matter if there are
parameters of that name outside the completion system; they will be
safely hidden, the special values used, and the original values restored
when completion is over.</p>
<p><code>$words</code> is an array corresponding to the words on the command line ---
where by a `word' I mean as always a single argument to the command,
which may include quoted whitespace. <code>$CURRENT</code> is the index into that
array of the current word. Note that to avoid confusion the ksh-like
array behaviour is explicitly turned off in <code>_main_complete</code>, so the
command itself is <code>$words[1]</code>, and so on.</p>
<p>The word being completed is treated specially. The reason is that you
may only want to complete some of it. An obvious example is a file with
a path: if you are completing at `<code>foo/bar</code>', you don't want to have to
check the entire file system; you want the directory <code>foo</code> to be fixed,
and completion just for files in that. There are actually two parts to
this. First, when completion is entered, <code>$PREFIX</code> and <code>$SUFFIX</code> give
you the part of the current word before the cursor, and the remainder,
respectively. It's done like this to make it possible to write functions
for completing inside a word, not just at the end. The simplest possible
way of completing a file is then to find everything that matches
<code>$PREFIX*$SUFFIX</code>.</p>
<p>But there's more to it than that: you need to separate off the
directory, hence the second part. The parameters <code>$IPREFIX</code> and
<code>$ISUFFIX</code> contain a part of the string which will be ignored for
completion. It's up to you to decide what that is, then to move the bit
you want to be ignored from <code>$PREFIX</code> to <code>$IPREFIX</code> (that's the usual
case) or from <code>$SUFFIX</code> to <code>$ISUFFIX</code>, making sure that the word so far
typed is still given by <code>$IPREFIX$PREFIX$SUFFIX$ISUFFIX</code>. Thus in
completing <code>foo/bar</code>, you would strip <code>foo/</code> from the start of <code>$PREFIX</code>
and tack it onto the end of <code>$IPREFIX</code> --- after recording the fact that
you need to move to directory <code>foo</code>, of course. Then you generate files
in <code>foo</code>, and the completion system will happily accept <code>barrack</code> or
<code>barbarous</code> as completions because it doesn't care about the <code>foo</code> any
more.</p>
<p>Actually, this is already done by the the <code>_files</code> and <code>_path_files</code>
functions for filename completion. Also, you can get some help using the
<code>compset</code> builtin command. In this case, the incantation is</p>
<pre><code> if compset -P "*/"; then
# do whatever you need to with the leading
# string up to / stripped off
else
# no prefix stripped, do whatever's necessary in this case
fi
</code></pre>
<p>In other words, any initial match of the pattern `<code>*/</code>' in <code>$PREFIX</code> is
removed and transferred to the end of <code>$IPREFIX</code>; the command status
tells you whether this was done. Note that it is the longest possible
such match, so if there were multiple slashes, all will be moved into
<code>$IPREFIX</code>. You can control this by putting a number <code><N></code> between the
<code>-P</code> and the pattern, which says to move only up to the <code><N></code>th such
match; here, that would be a pattern with exactly <code><N></code> slashes. Note
that <code>-P</code> stands for prefix, not pattern; there is a corresponding <code>-S</code>
option for the suffix. See the manual for other uses of <code>compset</code>; these
are probably the most frequent.</p>
<p>If you want to make the test made by <code>compset</code>, but without the side
effect of changing the prefixes and suffixes, there are tests like this:</p>
<pre><code> if [[ -prefix */ ]]; then
# same as with `compset -P "*/"', except prefixes were left alone.
fi
</code></pre>
<p>These have the advantage of looking like all the standard tests
understood by the shell.</p>
<p>There are three other parameters special to completion. The <code>$QIPREFIX</code>
and <code>$QISUFFIX</code> are a special prefix and suffix used when you are
dividing up a quoted word --- for example, in `<code>zsh -c "echo hi"</code>', the
word <code>"echo hi"</code> is going to be used as a command line in its own right,
so if you want to do completion there, you need to have it split up. You
can use `<code>compset -q</code>' to split a word in this fashion.</p>
<p>There is also an associative array <code>$compstate</code>, which allows you to
inspect and change the state of many internal aspects of completion,
such as use of menus, context, number of matches, and so on. Again,
consult the manual for more detail. Many of the standard styles work by
altering elements of <code>$compstate</code>.</p>
<p>Finally, in addition to the parameters special to completion, you can
examine (but not alter) any of the parameters which appear in all
editing widgets: <code>$BUFFER</code>, the contents of the current editing line;
<code>$LBUFFER</code>, the part of that before the cursor; <code>$RBUFFER</code>, the rest;
<code>$CURSOR</code>, the index of the cursor into <code>$BUFFER</code> (with the first
character at zero, in this case --- or you can think of the zero as
being the point before the first character, which is where insertion
would take place with the cursor on the first character); <code>$WIDGET</code> and
<code>$LASTWIDGET</code>, the names of the current and last editing or completion
widget; <code>$KEYS</code>, the keys typed to invoke the current widget;
<code>$NUMERIC</code>, any numeric prefix given, unset if there is none, and a few
other probably less useful values. These are described in the
<code>zshzle(1)</code> manual page and the <strong>Zsh Line Editor</strong> info node. In
particular, I already mentioned <code>$NUMERIC</code> as of possible use in various
styles, and it is used by the completers which understand a `<code>numeric</code>'
value in their relevant styles; the <code>$WIDGET</code> and <code>$KEYS</code> parameters are
useful for deciding between different behaviours based on what the
widget is called (as in <code>_history_complete_word</code>), or which keys are
used to invoke it (as in <code>_bash_completions</code>).</p>
<p>Here are a few examples of using special parameters and <code>compset</code>.</p>
<p>One of the shortest standard completions is this, <code>_precommand</code>:</p>
<pre><code> #compdef - nohup nice eval time rusage noglob nocorrect exec
shift words
(( CURRENT-- ))
_normal
</code></pre>
<p>It applies for all the standard commands which do nothing but evaluate
their remaining arguments as a command, with some change of state, e.g.
ignoring a certain signal (<code>nohup</code>) or altering the priority (<code>nice</code>).
All the completion system does here is shift the first word off the end
of the <code>$words</code> array, decrement the index of the current word into
<code>$words</code>, and call <code>_normal</code>. This is the function called when
completion occurs not in one of the special <code>-context-</code>s, in other words
when an argument to an ordinary command is being completed. It will look
at the new command word <code>$words[1]</code>, which was previously the first
argument to <code>nohup</code> or whatever, and start completion again based on
that, or even complete that word itself as a command if necessary. The
net effect is that the first word is ignored completely, as required.</p>
<p>Here's just an edited chunk of the file <code>_user_at_host</code>; as its name
suggests, it completes words of the form <code><user>@<host></code>, and it's used
anywhere the <code>user-hosts</code> style, described above, is appropriate:</p>
<pre><code> if [[ -prefix 1 *@ ]]; then
local user=${PREFIX%%@*}
compset -P 1 '*@'
# complete the host for which we want the user
else
# no @, so complete the user
fi
</code></pre>
<p>We test to see if there is already a `<code><user>@</code>' part. If there is, we
extract the user with an ordinary parameter substitution (so ordinary
even other shells could do it). Then we strip off that from the bit to
be completed with <code>compset</code>; we already know it matches the prefix, so
we don't need to test the return value. Then we just do normal hostname
completion on what remains --- except that the <code>user-hosts</code> style might
be able to give us a clue as to which hosts have such a user. If the
original test failed, then we simply complete what's there as a user.</p>
<p>Finally, here is essentially what the function <code>_most_recent_file</code> uses
to extract the <code>$NUMERIC</code>th (default first) most recently modified file.</p>
<pre><code> local file
file=($~PREFIX*$~SUFFIX(om[${NUMERIC:-1}]N))
(( $#file )) && compadd -U -i "$IPREFIX" -I "$ISUFFIX" -f -Q - $file
</code></pre>
<p>Instead of doing it with mirrors, this uses globbing qualifiers to
extract the required file; <code>om</code> specifies ordering by modification time,
and the expression in square brackets selects the single match we're
after. The <code>N</code> turns on <code>NULL_GLOB</code>, so <code>$file</code> is empty if there are no
matches, and the parameter expansions with `<code>$~</code>' force patterns in
<code>$PREFIX</code> and <code>$SUFFIX</code> to be available for expansion (a little extra
feature I use, although ordinary completion would work without).</p>
<p>Most of the <code>compadd</code> command is bookkeeping to make sure the parts of
the prefix and suffix we've already removed, if there are any, get
passed on, but the reason for that deserves a mention, since normally
this is handled automatically. The difference here is that <code>-U</code> usually
replaces absolutely everything that was in the word before, so if you
need to keep it you have to pass it back to <code>compadd</code>. For example,
suppose you were in a context where you were completing after
`<code>file=...</code> and you had told the completion system that everything up
to `<code>file=</code>' was not to count and not to be shown as part of the
completion. You would want to keep that when the word was put back on
the command line. However, `<code>-U</code>' would delete that too. Hence the
`<code>-i "$IPREFIX"</code>' to make sure it's retained. The same argument goes
for the ignored suffix. However, there's currently no way of getting
<code>_most_recent_file</code> to work on only a part of a string, so this
explanation really only applies when you call it from another completion
function, not directly from the command line.</p>
<p><span id="l188"></span></p>
<h3 id="696-fancier-completion-using-the-tags-and-styles-mechanism"><a class="header" href="#696-fancier-completion-using-the-tags-and-styles-mechanism">6.9.6: Fancier completion: using the tags and styles mechanism</a></h3>
<p>At this point, you should be in a position to construct, although maybe
not in the best possible way, pretty much any completion list you want.
Now I need to explain how you make sure it all fits in with the usual
tags and styles system. You will need to pick appropriate tags for your
completions. Although there is no real restriction, it's probably best
to pick one of the standard tags, some of which are suitably general to
cover just about anything: <code>files</code>, <code>options</code>, <code>values</code>, etc. There is a
list in the completion system manual entry. Remember that the main use
for tags is to choose what happens when more than one tag can be
completed in the same place. Finding such things that can't be separated
using the standard tag names is a good reason for inventing some new
ones; you don't have to do anything special if the tag names are new,
just make sure they're documented for anyone using the completion
function.</p>
<p><strong>How to call functions so that `It Just Works'</strong></p>
<p>The simplest way of making your own completion function recognize tags
is to use the <code>_description</code> function, which is usually called with
three arguments: the name of the tag you're completing for, the name of
a variable which will become an array containing arguments to pass to
<code>compadd</code>, and the full description. Then you have to make sure that
array gets passed down to <code>compadd</code>, or to any of the higher-level
completion functions which will pass the arguments on to <code>compadd</code>. For
example,</p>
<pre><code> local expl
_description files expl 'my special files'
_files "$expl[@]"
</code></pre>
<p>This sets the files tag; <code>_description</code> sets <code>$expl</code> to pass on the
description, and maybe other things such as a group name for the tag, in
the appropriate format; we pass this down to <code>_files</code> which will use it
for calling <code>compadd</code>. Generally, you will call <code>_description</code> for each
time you call <code>compadd</code> or something that in turn calls <code>compadd</code>.</p>
<p>The <code>_description</code> function calls another function <code>_setup</code> to do much
of the setting up of styles for the particular tag. Mostly, <code>_setup</code> is
buried deeply enough that you don't need to worry about it yourself.
Sometimes you can't do completion, and just want to print a message
unconditionally to say so, irrespective of tags etc.; the function
<code>_message</code> does this, taking the message as its sole argument.</p>
<p>There are two levels above that; these implement the tags mechanism in
full. In <code>_description</code>, all that happens is that the user is informed
what tag is coming up; there's no check what preferences the user has
for tags (the first level), nor whether he wants tags to be split up
using the labelling mechanism, e.g. picking out certain sorts of files
using the labelled tag `<code>file:-myfiles</code>' to get the final tag
`<code>file-myfiles</code>' (the second level).</p>
<p>To get this for simple cases you use the function <code>_wanted</code>. Unlike
<code>_description</code>, it's an interface to the function that generates
completion as well as a handler for tags --- that's so it can loop over
the generated tags, checking the labels. The call above would now look
like this:</p>
<pre><code> _wanted files expl 'my special files' _files
</code></pre>
<p>Note that you now don't pass the <code>"$expl[@]"</code>, which hasn't even been
set yet; <code>_wanted</code> will generate the string using the parameter name you
say (here `<code>expl</code>', as usual), and assume that the function generating
the completions can use the result passed down to it. This is true of
pretty much anything you are likely to want to use.</p>
<p>Note also the fact you need to pass `<code>_files</code>', i.e. the function
generating the completion. You can put pretty much any command line
which generates completions here, down to a simple `<code>compadd</code>'
expression. The reason it has to be here is the tag labelling business:
<code>_wanted</code> could check whether the tag you specify, `<code>files</code>', is wanted
by the user and then return control to you, but it wouldn't be able to
split up and loop over labelled tags set in this case for the
<code>file-patterns</code> style and in other case by the <code>tag-order</code> style.</p>
<p>Unless you're really going into the bowels, <code>_wanted</code> is probably the
lowest level you will want to use. I'd suggest you remember that one,
and only go back and look at the other stuff if you need to do something
more complicated.</p>
<p>If your function handles multiple tags, you need to loop over the
different tags to find out which sort the tag order wants next. For
this, you first need to tell the system which tags are coming up, using
the <code>_tags</code> function with a list. Then you need to to test whether each
tag in turn actually needs to be completed, and go on doing this until
you run out of tags which need completions performing; the <code>_tags</code>
function without arguments does this. Finally, you need to use
<code>_requested</code>, which works a bit like <code>_wanted</code> but is made to fit inside
the loop we are using. The end result looks like this:</p>
<pre><code> local expl ret=1
_tags foo bar rod
while _tags; do
_requested foo expl "This is the description for tag foo" \
compadd all foos completions && ret=0
_requested bar expl "This is the description for tag bar" \
compadd all bars completions && ret=0
_requested rod expl "This is the description for tag rod" \
compadd all rods completions && ret=0
(( ret )) || return 0 # leave if matches were generated
done
</code></pre>
<p>If you do include the completion function line as arguments, the loop
over labels for the tag you specify is automatically handled as with
<code>_wanted</code>. It may be a little confusing that both <code>_requested</code> and
<code>_wanted</code> exist: the specific difference is that with <code>_requested</code> you
call the <code>_tags</code> function yourself, whereas <code>_wanted</code> assumes the only
valid tag is its argument and acts accordingly, and can be used only for
simple, `one-shot' completions.</p>
<p>With <code>_requested</code>, unlike <code>_wanted</code>, you can separate out the arguments
to the completion generator itself --- here <code>compadd</code> --- into a
different statement, remembering the <code>"$expl[@]"</code> argument in that case.
You can miss out the second and third arguments for <code>_requested</code> in this
way. This time the loop which generates labels for tags is not
performed, and you have to arrange it yourself, with the usual trade off
of greater complexity for greater flexibility. To do this, there are two
other functions: <code>_all_labels</code> and <code>_next_label</code>. The simpler case is
with <code>_all_labels</code>, which just implements the loop over the labels using
the same arguments as <code>_wanted</code>:</p>
<pre><code> _requested values &&
_all_labels values expl 'values for my special things' \
compadd alpha bravo charlie delta echo foxtrot.
</code></pre>
<p>In case you haven't understood (and it's quite complicated, I'm afraid):
the <code>_requested</code> looks at whether the tag you use has been asked for by
the user. Having found out that it is, the <code>_all_labels</code> function calls
the command <code>compadd</code> which actually adds the completions, but it does
it in such a way as to take account of labelled tags --- you might have
both a plain `<code>values</code>' tag and `<code>values:-special</code>' labelled tag, and
<code>_all_labels</code> is needed to decide which is being used here. This last
example is actually exactly what <code>_requested</code> does when given the
<code>compadd</code> as argument, so it's only really useful when there is some
code between the <code>_requested</code> and the <code>_all_labels</code>, for example to
compute the strings to complete.</p>
<p>The most complicated case you are likely to come across is when inside
the part of the tags loop which handles a particular tag (i.e. the
<code>_requested</code> lines in the example above), you actually want to add more
than one possible sort of completion. Then <code>_all_labels</code> is no longer
enough, because completion needs to sort out the different things which
are being added. This can also happen when there is only one valid tag,
but that has multiple completions so that <code>_wanted</code> isn't any use. In
this case you need to use <code>_next_label</code> inside a loop, which, as its
names suggests, fixes up labels for the current tag and stops when it's
found the right one. Here's a stripped down example which handles
completion of messages from the <code>MH</code> mail handling system; you'll find
it complete inside the function <code>_mh</code>.</p>
<pre><code> _tags sequences
while _tags; do
while _next_label sequences expl sequence; do
compadd "$expl[@]" $(mark $foldnam 2>/dev/null |
awk -F: '{ print $1 }') && ret=0
compadd "$expl[@]" reply next cur prev \
first last all unseen && ret=0
_files "$expl[@]" -W folddir -g '<->' && ret=0
done
(( ret )) || return 0
done
</code></pre>
<p>Here's what's going on. The <code>_tags</code> call works just as it did in the
first example I showed for that, deciding whether the tag in question,
<code>sequences</code>, has been asked for; the tag name comes because MH allows
you to define sets of messages called exactly `sequences'. The first
`<code>while</code>' selects all values from <code>tag-order</code> where the `<code>sequences</code>'
tag appears, with or without a label. The second `<code>while</code>' loop then
sorts out any occurrences of labelled sequences to be presented to the
user at the same time, i.e. given in the same element of the <code>tag-order</code>
value array. The first <code>compadd</code> extracts from the folder (MH's name for
a directory) identified by the function the names of any sequences you
have defined; the second adds a lot of standard sequences --- although
strictly speaking <code>unseen</code> isn't a standard sequence since you can name
it yourself in <code>~/.mh_profile</code>. Finally, the third adds files in the
folder itself whose names are just digits, which is how MH stores
messages. The handling of <code>return</code> makes sure it stops as soon as you
have matches for one particular element of <code>tag-order</code>; if you put it in
the inner loop, you would just have the first of those sets that
happened to be generated, while here, if you specify that all types of
sequence should appear in the same completion list, they are all
correctly collected.</p>
<p>Why, in that last example, is there no call to <code>_requested</code>, now I've
gone to the trouble of explaining what that does? The answer is that
there is only one tag; <code>_tags</code> can decide if we want it at all, and
after that the tag is known, so we don't need <code>_requested</code> to find that
information out for us. It's only needed if there is more than one type
of match --- indeed, that's why we introduced it, so this is not
actually a new complication, although you can be forgiven for thinking
otherwise.</p>
<p>Here's an example of using that code for sequences. You might decide
that you only want to see named sequences unless there aren't any,
otherwise ordinary messages. You could do this by setting your styles as
follows:</p>
<pre><code> zstyle ':completion:*' tag-order sequences:-name sequences:-num
zstyle ':completion:*:sequences-name' ignored-patterns '(|,)<->'
zstyle ':completion:*:sequences-num' ignored-patterns '^<->'
</code></pre>
<p>which tries <code>sequences</code> under the labels <code>sequences-name</code> and
<code>sequences-num</code>; which ignore completions which are all digits, and
those which are not all digits, respectively. The slight twiddle in the
pattern for <code>sequences-name</code> ignores messages marked for deletion as
well, which have a comma stuck in front of the number (this is
configurable, so your version of MH may be different).</p>
<p>All of <code>_description</code>, <code>_wanted</code>, <code>_requested</code>, <code>_all_labels</code> and
<code>_next_label</code> take the options <code>-J</code> and <code>-V</code> to specify sorted or
unsorted listings and menus, and the options <code>-1</code> and <code>-2</code> for removing
consecutive duplicates or all duplicates. These are also options to
<code>compadd</code>; the reason for handling them here is that they can be
different for each tag, and the function called will set <code>expl</code>
appropriately.</p>
<p>If your requirements are simple enough, you can replace that <code>_tags</code>
loop above with a single function, <code>_alternative</code>. This takes a series
of arguments each in the form `<tag>:<description>:<action>',
with the first two in the form you now know, and the third an action.
These are essentially the same as actions for the <code>_arguments</code> function,
described below, except that the form `<code>->state</code>', which says that the
calling function will handle the action itself by using the value of the
parameter <code>$state</code>, is not available. The most common forms of action
here will be a call to another completion function, maybe with arguments
(e.g. `<code>_files -/</code>'), or a simple list in parentheses (e.g. `<code>(see saw margery daw)</code>'). Here, for example, is how the <code>_cd</code> function handles
the two cases of local directories (under the current directory) and
directories reached via the <code>$cdpath</code> parameter:</p>
<pre><code> local tmpcdpath
tmpcdpath=(${(@)cdpath:#.})
_alternative \
'local-directories:local directories:_path_files -/' \
'path-directories:directories in cdpath:
_path_files -W tmpcdpath -/'
</code></pre>
<p>The only tricky bit is that <code>$tmpcdpath</code>: it removes the `<code>.</code>' from
<code>$cdpath</code>, if it's present, so that the current directory is always
searched for with the tag `<code>local-directories</code>', never with
`<code>path-directories</code>'. Actually, you could argue that it should be
treated as being in `<code>path-directories</code>' when it's present; but that
confuses the issue over what `<code>local-directories</code>' really means, and it
is useful to have the distinction.</p>
<p>It's now an easy exercise to replace the example function I gave for
<code>_requested</code> by a call to <code>_alternative</code> with the arguments to <code>compadd</code>
turned into a list in parentheses as the <code><action></code> part of the
arguments to <code>_alternative</code>.</p>
<p><strong>How to look up styles</strong></p>
<p>If your completion function gets really sophisticated, you may want it
to look up styles to decide what its behaviour should be. The same
advice goes as for tags: only invent a new style if the old ones don't
seem to cover the use you want to make, since by using contexts you can
always restrict the scope of the style. However, by the same token don't
try to squeeze too much meaning into one style, which will force the
user to narrow the context --- it's always much easier to set a style
for the general context `<code>:completion:*</code>' than to have to worry about
all the circumstances where you need a particular value.</p>
<p>Retrieving values of styles is no harder than defining them, but you
will need to know about the parameter <code>$curcontext</code>, which is what
stores the middle part of the context, sans `<code>:completion:</code>' and sans
tag. When you need to look something up, you pass this context to
<code>zstyle</code> with `<code>:completion:</code>' stuck in front:</p>
<pre><code> zstyle -b ":completion:${curcontext}:tag" style-name parameter
</code></pre>
<p>If the tag is irrelevant, you can leave it empty, but you still need the
final colon since there should always be six in total. In some cases
where multiple tags apply it's useful to have a <code>:default</code> tag context
as a fall back if none of the actual tags yield styles for that context;
hence you should test the style first for the specific tag, then with
the <code>default</code>.</p>
<p>Style lookups all have the form just shown; the result for looking up
<code>style-name</code> in the given context will be saved in the <code>parameter</code>
(which you should make local, obviously). In addition, <code>zstyle</code> returns
a zero status if the lookup succeeded and non-zero if it failed. The
<code>-t</code> lookup is different from the rest as it only returns a status for a
boolean, i.e. returns status 0 if the value is <code>true</code>, <code>yes</code>, <code>1</code> or
<code>on</code>, and doesn't require a parameter name. There is also a <code>-T</code>, which
is identical except that it returns status 0 if the style doesn't exist,
i.e. the style is taken to default to true.</p>
<p>The other lookup options return the style as a particular type in the
parameter with exit status zero if the lookup succeeded, i.e. a value
was found, and non-zero otherwise; <code>-b</code>, <code>-s</code>, and <code>-a</code> specify boolean
(<code>parameter</code> is either <code>yes</code> or <code>no</code>), scalar (<code>parameter</code> is a scalar),
and array (<code>parameter</code> is an array, which may still be a single word, of
course), You can retrieve an associative array with <code>-a</code> as long as the
parameter has already been declared as one.</p>
<p>There's also a convenience option for matching, <code>-m</code>; instead of a
<code>parameter</code> this takes a <code>pattern</code> as the final argument, and returns
status zero if and only if the <code>pattern</code> matches one of the values
stored in the style for the given context.</p>
<p>Typical usages are thus:</p>
<pre><code> if zstyle -t ":completion:${curcontext}:" foo; then
# do things in a fooish way
else
# do things in an unfooish way
fi
</code></pre>
<p>or to use the value:</p>
<pre><code> local val
if zstyle -s ":completion:${curcontext}:" foo val; then
# use $val to establish how fooish to be
else
# be defaultly fooish
fi
</code></pre>
<p><span id="l189"></span></p>
<h3 id="697-getting-the-work-done-for-you-handling-arguments-etc"><a class="header" href="#697-getting-the-work-done-for-you-handling-arguments-etc">6.9.7: Getting the work done for you: handling arguments etc.</a></h3>
<p>The last piece of unfinished completion business is to explain the
higher level functions which can save you time writing completions for
commands which behave in a standard way, with arguments and options. The
good news is that all the higher functions here handle tags and labels
internally, so you don't need to worry about <code>_tags</code>, <code>_wanted</code>,
<code>_requested</code>, etc. There's one exception: the `state' mechanism to be
described, where a function signals you that you're in a given state
using the parameter <code>$state</code>, expects you to handle tag labels yourself
--- pretty reasonable, as you have requested that the function return
control to you to generate the completions. I've mentioned that here so
that I don't have to gum up the description of the functions in this
section by mentioning it again.</p>
<p><strong>Handling ordinary arguments</strong></p>
<p>The most useful function is <code>_arguments</code>. There are many examples of
this in the completion functions for external commands, since so many
external commands take the standard format of a command with options,
some taking their own arguments, plus command arguments.</p>
<p>The basic usage is to call it with a series of arguments (which I'll
call `specifications') like:</p>
<pre><code> <where I am>:<description>:<what action to take>
</code></pre>
<p>although there are a whole series of more complicated possibilities.</p>
<p>The initial `<code><where I am></code>' part tells the function whether the
specification applies to an argument in a particular position, or to an
option and possibly any arguments for that option. Let's start with
ordinary arguments, since these are simpler. In this case `<code><where I am></code>' will be either a number, giving the number of the argument, or a
`<code>*</code>', saying that this applies to all remaining arguments (or all
arguments, if you haven't used any of the other form). You can simplify
the first form, by just missing out the number; then the function will
assume it applies to the first argument not yet specified. Hence the
standard way of handling arguments is with a series of specifications
just beginning `<code>:</code>' for arguments that need to be handled their own
way, if any, then one beginning `<code>*:</code> for all remaining arguments, if
any.</p>
<p>The message that follows is a description to be passed on down to
<code>_description</code>. You don't specify the tags at this point; that comes
with the action.</p>
<p>The action can have various forms, chosen to be easily distinguishable
from one another.</p>
<ol>
<li>
<p>A list of strings in parentheses, such as `<code>(red blue green)</code>'.
These are the possible completions, passed straight down to
<code>compadd</code>.</p>
</li>
<li>
<p>The same, but with double parentheses; the list in this case
consists of the completion, a backslashed colon, and a description.
So an extended version of the previous action is `<code>((red\:The\ colour\ red blue:The\ colour\ blue))</code>' and so on. You can escape
other colons inside the specifications in this way, too.</p>
</li>
<li>
<p>A completion function to call, with any arguments, such as `<code>_files -/</code>' to complete directories. Usually this does the business with
<code>$expl</code> which should be familiar from the section on basic tag
handling, however you can put an extra space in front of the action
to have it called exactly as is, after word splitting.</p>
</li>
<li>
<p>A word preceded by `<code>-></code>' for example `<code>->state</code>'. This specifies
that <code>_arguments</code> should return and allow the calling function to
process the argument. To signal back to the calling function, the
parameter <code>$state</code> will be set to what follows the `<code>-></code>'. It's up
to the calling function to make <code>$state</code> a local parameter ---
<code>_arguments</code> can't do that, since then it couldn't return a value.</p>
<p>You should also make the parameters <code>$context</code> and <code>$line</code> local;
the former is set to the new part to be added to <code>$curcontext</code>,
which, as you can find out from <code>^Xh</code>, is <code>option-<option>-<arg></code>,
for example <code>option-file-1</code> for the first argument of the
<code>option-file</code> option, or <code>argument-N</code>, for example <code>argument-2</code> for
the second argument of the command.</p>
<p>In simple cases, you will just test the parameter <code>$state</code> after
<code>_arguments</code> has returned to see what to do: the return value is 300
to distinguish it from other returns where <code>_arguments</code> itself
performed the completion.</p>
</li>
<li>
<p>A chunk of code to evaluate, given in braces, which removes the need
for a special function or processing states. Obviously this is best
used for the simplest cases.</p>
</li>
</ol>
<p>These are the main possibilities, but I have not described every
variation. As always, you should see the manual for all the detail.</p>
<p>Here's a concocted example for that `<code>->state</code>' action specifier, in
case it's confusing you. It's for a command that takes arguments
`<code>alpha</code>', `<code>beta</code>' and `<code>gamma</code>', and takes a single option
`<code>-type</code>' which takes one argument, either `<code>normal</code>' or `<code>unusual</code>'.</p>
<pre><code> local context state line
typeset -A opt_args
_arguments '-type[specify type]:type:->type' \
'*:greek letter:->gklet' && return 0
case $state in
(type) compadd normal unusual && return 0
;;
(gklet) compadd alpha beta gamma && return 0
;;
esac
return 1
</code></pre>
<p>In fact the possibilities here are so simple that you don't need to use
<code>$state</code>; you can just use the form with the values in parentheses as
the action passed to `<code>_arguments</code>'. Anyway, if you put this into a
function `<code>_foo</code>', type `<code>compdef _foo foo</code>', and attempt completion
for the fictitious command `<code>foo</code>', you will see <code>_arguments</code> in
action.</p>
<p>I haven't shown the gory tag handling; as it's written, you'll see that
no tag is ever defined for the <code>compadd</code> arguments shown. In this case
you could just use <code>_wanted</code>. What you get for free with arguments,
however, is the context: in the first case, you would have
`<code>:option-type-1</code>' in the argument field (the second last, just before
the tag), and in the second case `<code>:argument-rest:</code>'. Go back to where
I originally described contexts if you've forgotten about these; I
didn't tell you at the time, but it's the <code>_argument</code> function that is
responsible for them. (However, you can supply a `<code>-C</code>' argument to
<code>_wanted</code> to tell that a context.)</p>
<p>A note about the form: that `<code>&& return 0</code>' makes the completion
function return if <code>_arguments</code> was satisfied that it found a completion
on its own. It's useful in more complex cases. Remember that most
completion functions return status zero if and only if matches were
added; this function is written to follow that convention. I already
showed this in the section on tags, but you might have skipped that.</p>
<p>Note all the things you had to make local: <code>$context</code>, <code>$state</code>, <code>$line</code>
and the associative array <code>$opt_args</code>. The last named allows you to
retrieve the values for a particular option; for example
`<code>$opt_args[-o]</code>' contains any value already on the command line for
the option <code>-o</code>. For options that take multiple arguments, these appear
separated by colons, so if the line contains `<code>-P prefix 3</code>',
<code>$opt_args[-P]</code> will contain `<code>prefix:3</code>'.</p>
<p><strong>Handling options</strong></p>
<p>Option handling is broadly similar, with the `<code><where I am></code>' part just
giving the option name --- I already showed one example with `<code>-type</code>'
above. In this case, the option will just be completed to itself, the
first part of the specification, and the rest says how to complete its
arguments. Since options can take any number of arguments, including
zero, the <code>:description:action</code> pair can be repeated, or omitted
entirely. Otherwise, it behaves similarly to the way described for
ordinary command arguments, with all the same possible actions. So a
simple option specification could be</p>
<pre><code> _arguments '-turnmeon'
</code></pre>
<p>for an option with no arguments,</p>
<pre><code> _arguments '-file:input file:_files'
</code></pre>
<p>for an option with one argument, or</p>
<pre><code> _arguments '-iofiles:input file:_files:output file:_files'
</code></pre>
<p>for an option with two arguments, both files but with different
descriptions.</p>
<p>The first part of the specification for an option can be more
complicated, to reflect the fact that options can be used in all sorts
of different ways. You can specify a description for the option itself
--- as I tried to explain, the descriptions in the rest of the
specification are instead for the arguments to the option. To specify an
option description, just put that after the option, before any colons,
in square brackets:</p>
<pre><code> _arguments '-on[turn me on, why not]'
</code></pre>
<p>Next, some options to a command are mutually exclusive. As <code>_arguments</code>
has to read its way along the command line to parse it, it can record
what options have already appeared, and can ensure that an option
incompatible with one there already will not be completed. To do this,
you need to include the excluded option in parentheses before the option
itself:</p>
<pre><code> _arguments '(-off)-on[turn me on, why not]' \
'(-on)-off[turn me off, please]'
</code></pre>
<p>This completes either of the options `<code>-on</code>' or `<code>-off</code>', but if
you've already given one, it won't complete the other on the same
command line. If you need to give multiple excluded options, just list
them separated by spaces, like `<code>(-off -noton)</code>'.</p>
<p>Some options can themselves be repeated; <code>_arguments</code> usually won't do
that (in a sense, they are mutually exclusive with themselves), but you
can allow it to happen by putting a `<code>*</code>' in front of the option
specification:</p>
<pre><code> _arguments '*-o[specify extra options]:option string:->option'
</code></pre>
<p>allows you to complete any number of `<code>-o <option></code>' sets using the
<code>$state</code> mechanism. The <code>*</code> appears after any list of excluded options.</p>
<p>There are also ways of allowing different methods of option handling. If
the option is followed by <code>-</code>, that means the value must be in the same
word as the option, instead of in the next word; if that is allowed, but
the argument could be in the next word instead, the option should be
followed by a `<code>+</code>'. The latter behaviour is very common for commands
which take single letter options. Some commands, particularly many
recent GNU commands, allow you to have the argument in the next word or
in the current word after an `<code>=</code>' sign; you get this by putting an
`<code>=</code>' after the option name. For example,</p>
<pre><code> _arguments '-file=:input file:_files'
</code></pre>
<p>allows you to complete `<code>-file </code><em><filename></em>' or
`<code>-file=</code><em><filename></em>'. With</p>
<pre><code> _arguments '-file=-:input file:_files'
</code></pre>
<p>only the second is possible, i.e. the argument must be after the `<code>=</code>',
not in its own word.</p>
<p>You can handle optional and repeated arguments to options, too. This
illustrates some possibilities:</p>
<pre><code> _arguments '-option:first arg:->first::optional arg:->second'
</code></pre>
<p>The doubled colon indicates that the second argument is optional. In
other words, at that point on the command line <code>_arguments</code> will either
try to complete via the state <code>second</code>, or will try to start another
specification entirely.</p>
<pre><code> _arguments '-option:first arg:->first:*:other args:->other'
</code></pre>
<p>Here, all arguments after the first --- everything else on the command
line --- is taken as an argument to the option, to be completed using
the state <code>other</code>.</p>
<pre><code> _arguments '-option:first arg:->first:*-:other args till -:->other'
</code></pre>
<p>This is similar, but less drastic: there is a pattern after the `<code>*</code>',
here a `<code>-</code>', and when that is encountered, processing of arguments to
`<code>-option</code>' stops. A command using this might be called as follows:</p>
<pre><code> cmdname -option <first> <other1> <other2> .... - <remainder>
</code></pre>
<p>where of course completion for <code><remainder></code> might be handled by other
specifications.</p>
<p>There are yet more possible ways of handling options. I've assumed that
option names can have multiple letters and hence must occur in separate
words. You can specify single-letter options as well, of course, but
many commands allow you to combine these into one word. To tell
<code>_arguments</code> that's OK you should give it the option <code>-s</code>; it needs to
come before any specifications, to avoid getting mixed up with them.
After you specify this, a command argument beginning with a single
`<code>-</code>' will be treated by <code>_arguments</code> as a list of single options, so
`<code>-lt</code>' is treated the same as `<code>-l -t</code>'. However, options beginning
with `<code>-``-</code>' are still treated as single options, so a `<code>-``-prefix</code>'
on the command line is still handled as a single long option by
<code>_arguments</code>.</p>
<p>One nice feature which can save a lot of trouble when using certain
commands, notably those written by the GNU project and hence installed
on most Linux-based systems, which take an option `<code>-``-help</code>' that
prints out a list of all options. This is in a human-readable form, but
<code>_arguments</code> is usually able to extract a list of available options
which use the `<code>-``-...</code>' form, and even in many cases whether they
take an argument, and if so what type that is. It knows because
`<code><command> -``-help</code>' often prints out a message like
`<code>-``-file=FILE</code>' which would tell <code>_arguments</code> (1) that `<code>-``-file</code>'
is a possible option (2) that it takes an argument because of the `<code>=</code>'
(3) that that argument should be a file because of the message `<code>FILE</code>'
at the end.</p>
<p>You specify that the command in question works in this way by using the
(fairly memorable) option `<code>-``-</code>' to `<code>_arguments</code>'. You can then
help it out with completion of option arguments by including a pattern
to be matched in the help test after the `<code>-``-</code>'; the format is
otherwise similar to a normal specification. For example
`<code>*=FILE*:file:_files</code>' says that any option with `<code>=FILE</code>' in it has
the description `<code>file</code>' and uses the standard <code>_files</code> function for
completion, while `<code>*=DIR*:directory:_files -/</code>' does the same for
directories. These two examples are so common that they are assumed by
`<code>_arguments -``-</code>'.</p>
<p>So for example, here is the completion for <code>gdb</code>, the GNU debugger,
which not surprisingly understands the GNU option format:</p>
<pre><code> _arguments -- '*=(CORE|SYM)FILE:core file:_files' \
'*=EXECFILE:executable:_files -g \*\(\*\)' \
'*=TTY:terminal device:compadd /dev/tty\*' && return 0
</code></pre>
<p>If you run `<code>gdb --help</code>', you'll see where these come from:
`<code>--core=COREFILE</code>', `<code>--exec=EXECFILE</code>' and `<code>--tty=TTY</code>' are all
listed as possible option/argument pairs. Doing it this way neatly
allows the argument completions to work whatever the names of the
options --- though of course it's possible for the rest of the pattern
to change, too, and the commands, being written by lots of different
people, are not necessarily completely consistent in the way their help
text is presented.</p>
<p><span id="l190"></span></p>
<h3 id="698-more-completion-utility-functions"><a class="header" href="#698-more-completion-utility-functions">6.9.8: More completion utility functions</a></h3>
<p>This is now just a ragbag of other functions which might prove useful in
your own completion functions, and which haven't been mentioned before,
with some examples; once again, consult the manual for more detail. Note
that many of these functions can take the most useful arguments to
<code>compadd</code> and pass them on, even where I haven't explicitly said so.</p>
<p><strong><code>_call_function</code></strong></p>
<p>This is a simple front end to calling a function which may not be
defined and hanging onto the return status of the function. One good use
for this is to call a possibly non-existent function which might have
been defined by the user, before doing some default stuff the user might
want to skip. That would look like this:</p>
<pre><code> local ret # returned status from called function, if it was called
_call_function ret _hook_function arg1 arg2 && return ret
# if we get here, _hook_function wasn't called,
# so do the default stuff.
</code></pre>
<p>As you can work out, <code>_call_function</code> itself returns status zero if the
function in the second argument got called, and in that case the first
argument is the name of a parameter with the return status from the
function itself. The whole point is that this is safe if
<code>_hook_function</code> doesn't exist.</p>
<p>This function is too low level to know about the tags mechanism; use
<code>_wanted</code> or similar to handle tags properly.</p>
<p><strong><code>_contexts</code></strong></p>
<p>This is another shorthand: the arguments it takes are a set of short
contexts, in other words either names of commands or special contexts
like `<code>-math-</code>'. The completion for each of these contexts is tried in
turn; <code>_contexts</code> simply handles all the boring looking up of functions
and testing the return values. The definition, if you want to look, is
reassuringly simple. It only has one use at the moment: <code>_subscript</code>,
which handles the <code>-subscript-</code> context we met early in the chapter,
calls `<code>_contexts -math-</code>' to try mathematical completion, since
ordinary array subscripts can contain mathematical expressions.</p>
<p>This is also too low level to handle tags. In zsh 4.1, it is made
obsolete by a cleverer mechanism for handling different contexts which
can be used, for example, for handling of arguments to redirections for
particular commands, or keys in a particular associative array. I expect
I'll describe that when 4.1 is finally released.</p>
<p><strong><code>_describe</code></strong></p>
<p>Don't confuse this with <code>_description</code> which was explained above and is
the basic function for adding a description to a set of completions of a
certain type. I mentioned in the description of the <code>verbose</code> style that
this function was responsible for showing, or not showing, the
descriptions for a whole lot of options at once. It allows you to do
that with several different sets of completions that may require
different options to <code>compadd</code>. The general form looks something like
this:</p>
<pre><code> _describe "description of set 1" descs1 compls1 \
<compadd-opts-1> -- \
"description of set 2" ...
</code></pre>
<p>where you can have any number of sets separated by the `<code>-``-</code>'. The
<code>descs1</code> and <code>compls1</code> are arrays of the same length, giving a list of
descriptions and a list of completions, respectively. Alternatively, you
need only give one array name and each element of that will contain a
completion and a description separated by the now-traditional colon. The
`<code><compadd-opts-1></code>' are a set of any old options recognised by
<code>compadd</code>, such as <code>-q</code>, or <code>-S=/</code>, or what have you. I won't give an
example for this, since to find something requiring it would almost need
me to rewrite the completion system from scratch.</p>
<p><strong><code>_combination</code></strong></p>
<p>This is the function at the heart of the completions such as
<code>users-hosts</code> described above, where combinations of elements need to be
completed at the same time. It's easiest to describe with an example;
let's pick the <code>users-hosts</code> example, and I'll assume you remember how
that works from the user's point of view, including the format of the
<code>users-hosts</code> style itself. The completion for the username part is
performed as:</p>
<pre><code> _combination my-accounts users-hosts users
</code></pre>
<p>where <code>my-accounts</code> is the tag to be used for the completion, then comes
the style, and then the part of the style to be extracted.</p>
<p>Now suppose we come back into the completion function again to complete
the host later on the command line, so that the username is already
there. We can find that by searching the command line; suppose we store
what we find in <code>$userarg</code>. Then we can complete the hostname as
follows:</p>
<pre><code> _combination my-accounts users-hosts users=$userarg hosts
</code></pre>
<p>and the magic part, the fact that we can limit the hostnames to be
completed to only those with a user <code>$userarg</code>, is handled by
<code>_combination</code>. This extends to <code>hosts-ports-users</code> and any larger
combined set in the obvious way: the first field not to contain an
`<code>=</code>' is the one being completed. You don't need to supply other fields
if they are not known; in other words, the field to be completed doesn't
need to be the first one in sequence not known, it can be any, just as
long as it matches part of the style given in the second argument, so
you could have omitted the `<code>users=$userarg</code>' in the last example if
you couldn't extract the right username.</p>
<p>There are various bells and whistles: after the field to be completed
you can add any options to be passed down to <code>compadd</code>; you can give
<code>_combination</code> itself the option `<code>-s <sep></code>' to specify a character
other than colon to separate the parts of the style values; if the style
lookup fails, but there is a corresponding function, which would be
called `<code>_users</code>' or `<code>_hosts</code>' in this example, it is called to
generate the matches, and gets the options at the end which are
otherwise destined for <code>compadd</code>.</p>
<p>As you can see, this function is at a high enough level to handle the
tags mechanism itself.</p>
<p><strong><code>_multi_parts</code></strong></p>
<p>This takes two arguments, a separator and a list of matches. The list of
matches is normal, except that each element is likely to contain the
separator. In the most obvious usage, the separator is `<code>/</code>' and the
list of matches is a lot of files with path components. Here's another
reasonable usage:</p>
<pre><code> local groups expl
groups=($(awk -F: '{ print $1 }' ~/.newsrc))
_wanted groups expl 'newsgroup' _multi_parts "$expl[@]" . groups
</code></pre>
<p>The generated array contains names of Usenet newsgroups, i.e. names with
components separated by a `<code>.</code>', and <code>_multi_parts</code> allows you to
complete these piece by piece instead of in one go. This is a good deal
better for use with menu completion, and the list which appears is
smaller too. The <code>_wanted</code> part handles the tags mechanism, which
<code>_multi_parts</code> doesn't.</p>
<p><strong><code>_sep_parts</code></strong></p>
<p>This also completes a word piece by piece, but unlike <code>_multi_parts</code> the
trial completions are also only supplied for each piece. The arguments
are alternating arrays and separators; arrays are in the usual form, in
other words either the name of an array parameter, or a literal array in
parentheses, quoted to protect it from immediate shell expansion. The
separators are simply strings. For example</p>
<pre><code> local expl
array1=(apple banana cucumber)
_wanted breakfast expl 'breakfast' \
_sep_parts array1 + '(bread toast croissant)' @ '(bowl plate saucer)';
</code></pre>
<p>completes strings like `<code>apple+toast@plate</code>', piece by piece. This is
currently not used by the distributed completion code.</p>
<p><strong><code>_values</code></strong></p>
<p>This works a little like <code>_arguments</code>, but is designed for completing
the values of a single argument in a form like
`<code>key=val,flag,key=other</code>', in which you can specify the list
separator, here `<code>,</code>' by using the option <code>-s</code>, e.g. `<code>-s ,</code>'. The
first argument to <code>_values</code> is the overall description of the set of
arguments. The other arguments are very much like those to <code>_arguments</code>
except that, as you would expect from the form given, no pluses or minus
signs are involved and each value can only have one argument, which must
follow an `<code>=</code>'. Virtually everything else is identical, with the
exception that the associative array where the arguments are stored for
each value is called <code>$val_args</code>.</p>
<p>I won't bother giving the instructions for <code>_arguments</code> again; instead,
here is an example based on the values used by the <code>-o</code> option to the
<code>mount</code> command:</p>
<pre><code> local context state line
typeset -A val_args
_values -s , 'file system options' \
'(rw)ro[mount file system read-only]' \
'(ro)rw[mount file system read-write]' \
'uid[set owner of root]:user ID:' \
'gid[set group of root]:group ID:' \
'bs[specify block size]:block size:(512 1024 2048 4192)'
</code></pre>
<p>I've just picked out a few of the umpteen possibilities for
illustration; see the function <code>_mount</code> if you want more. Remember that
the `<code>(rw)</code>' before the `<code>ro</code>' means that the options are mutually
exclusive, and the one in parentheses won't be offered if the other
appears on the command line; the strings in square brackets are
descriptions of the particular options; and if there is a colon after
the name of the value, the value takes an argument whose own description
comes next. The second colon is followed by possible completions for
that argument, using the usual convention for actions in <code>_arguments</code>;
as you'll see from the <code>local</code> statement, the <code>$state</code> mechanism can be
used here. Only the `<code>bs</code>' argument here is given possible completions;
for <code>uid</code> and <code>gid</code> you'll have to type in the number without
completion; <code>ro</code> and <code>rw</code> don't take arguments.</p>
<p>Hence a typical(?) list to be completed by this would be
`<code>rw,uid=123,bs=2048</code>'.</p>
<p>Remember also that you can use a `<code>*</code>' before the option name to say
that it can appear more than once in the value list. The <code>_values</code>
function handles the context and tags in a similar way to <code>_arguments</code>.</p>
<p><strong><code>_regex_arguments</code></strong></p>
<p>This function is for use when the behaviour of a set of command
arguments is so complicated that even <code>_arguments</code> can't help. It allows
you to describe the arguments as a regular expression (i.e. a pattern).
I won't explain it because I haven't yet figured out how it works. If
you think you need to use it, look at the manual entry and then at the
<code>_apt</code> function which is currently its main application.</p>
<p><span id="l191"></span></p>
<h2 id="610-finally"><a class="header" href="#610-finally">6.10: Finally</a></h2>
<p>Completion is big and complex: this means that there are probably lots
of bugs around, and things that I haven't described simply enough or
which may be implemented in too complicated a way. Please send the
<code>zsh-workers</code> mailing list any reports or constructive criticism on the
subject.</p>
<p>Last of all, remember that the new completion system is ideally just
supposed to work without you needing to worry exactly how. That's a bold
hope, but at least much of the time you should be able to get away with
using just the tab key and ordinary characters.</p>
<div id="chapter_begin" style="break-before: page; page-break-before: always;"></div><!-- START doctoc generated TOC please keep comment here to allow auto update -->
<!-- DON'T EDIT THIS SECTION, INSTEAD RE-RUN doctoc TO UPDATE -->
<p><strong>Table of Contents</strong> <em>generated with <a href="https://github.com/thlorenz/doctoc">DocToc</a></em></p>
<ul>
<li><a href="zshguide07.html#chapter-7-modules-and-other-bits-and-pieces-not-written">Chapter 7: Modules and other bits and pieces <em>Not written</em></a>
<ul>
<li><a href="zshguide07.html#71-control-over-modules-zmodload">7.1: Control over modules: <code>zmodload</code></a>
<ul>
<li><a href="zshguide07.html#711-modules-defining-parameters">7.1.1: Modules defining parameters</a></li>
<li><a href="zshguide07.html#712-low-level-system-interaction">7.1.2: Low-level system interaction</a></li>
<li><a href="zshguide07.html#713-zftp">7.1.3: ZFTP</a></li>
</ul>
</li>
<li><a href="zshguide07.html#72-contributed-bits">7.2: Contributed bits</a>
<ul>
<li><a href="zshguide07.html#721-prompt-themes">7.2.1: Prompt themes</a></li>
</ul>
</li>
<li><a href="zshguide07.html#73-whats-new-in-41">7.3: What's new in 4.1</a></li>
</ul>
</li>
</ul>
<!-- END doctoc generated TOC please keep comment here to allow auto update -->
<p><span id="ragbag"></span><span id="l192"></span></p>
<h1 id="chapter-7-modules-and-other-bits-and-pieces-not-written"><a class="header" href="#chapter-7-modules-and-other-bits-and-pieces-not-written">Chapter 7: Modules and other bits and pieces <em>Not written</em></a></h1>
<p><span id="l193"></span></p>
<h2 id="71-control-over-modules-zmodload"><a class="header" href="#71-control-over-modules-zmodload">7.1: Control over modules: <code>zmodload</code></a></h2>
<p><span id="l194"></span></p>
<h3 id="711-modules-defining-parameters"><a class="header" href="#711-modules-defining-parameters">7.1.1: Modules defining parameters</a></h3>
<p><span id="l195"></span></p>
<h3 id="712-low-level-system-interaction"><a class="header" href="#712-low-level-system-interaction">7.1.2: Low-level system interaction</a></h3>
<p><span id="l196"></span></p>
<h3 id="713-zftp"><a class="header" href="#713-zftp">7.1.3: ZFTP</a></h3>
<p><span id="l197"></span></p>
<h2 id="72-contributed-bits"><a class="header" href="#72-contributed-bits">7.2: Contributed bits</a></h2>
<p><span id="l198"></span></p>
<h3 id="721-prompt-themes"><a class="header" href="#721-prompt-themes">7.2.1: Prompt themes</a></h3>
<p><span id="l199"></span></p>
<h2 id="73-whats-new-in-41"><a class="header" href="#73-whats-new-in-41">7.3: What's new in 4.1</a></h2>
</main>
<nav class="nav-wrapper" aria-label="Page navigation">
<!-- Mobile navigation buttons -->
<div style="clear: both"></div>
</nav>
</div>
</div>
<nav class="nav-wide-wrapper" aria-label="Page navigation">
</nav>
</div>
<!-- Livereload script (if served using the cli tool) -->
<script type="text/javascript">
var socket = new WebSocket("ws://localhost:3000/__livereload");
socket.onmessage = function (event) {
if (event.data === "reload") {
socket.close();
location.reload();
}
};
window.onbeforeunload = function() {
socket.close();
}
</script>
<script type="text/javascript">
window.playground_copyable = true;
</script>
<script src="elasticlunr.min.js" type="text/javascript" charset="utf-8"></script>
<script src="mark.min.js" type="text/javascript" charset="utf-8"></script>
<script src="searcher.js" type="text/javascript" charset="utf-8"></script>
<script src="clipboard.min.js" type="text/javascript" charset="utf-8"></script>
<script src="highlight.js" type="text/javascript" charset="utf-8"></script>
<script src="book.js" type="text/javascript" charset="utf-8"></script>
<!-- Custom JS scripts -->
<script type="text/javascript">
window.addEventListener('load', function() {
window.setTimeout(window.print, 100);
});
</script>
</body>
</html>