{ "mode": "man", "parameter": "zshmodules", "section": "1", "url": "https://www.chedong.com/phpMan.php/man/zshmodules/1/json", "generated": "2026-07-05T13:04:11Z", "sections": { "NAME": { "content": "zshmodules - zsh loadable modules\n", "subsections": [] }, "DESCRIPTION": { "content": "Some optional parts of zsh are in modules, separate from the core of the shell. Each of\nthese modules may be linked in to the shell at build time, or can be dynamically linked while\nthe shell is running if the installation supports this feature. Modules are linked at run‐\ntime with the zmodload command, see zshbuiltins(1).\n\nThe modules that are bundled with the zsh distribution are:\n", "subsections": [ { "name": "zsh/attr", "content": "Builtins for manipulating extended attributes (xattr).\n" }, { "name": "zsh/cap", "content": "Builtins for manipulating POSIX.1e (POSIX.6) capability (privilege) sets.\n" }, { "name": "zsh/clone", "content": "A builtin that can clone a running shell onto another terminal.\n" }, { "name": "zsh/compctl", "content": "The compctl builtin for controlling completion.\n" }, { "name": "zsh/complete", "content": "The basic completion code.\n" }, { "name": "zsh/complist", "content": "Completion listing extensions.\n" }, { "name": "zsh/computil", "content": "A module with utility builtins needed for the shell function based completion system.\n" }, { "name": "zsh/curses", "content": "curses windowing commands\n" }, { "name": "zsh/datetime", "content": "Some date/time commands and parameters.\n" }, { "name": "zsh/db/gdbm", "content": "Builtins for managing associative array parameters tied to GDBM databases.\n" }, { "name": "zsh/deltochar", "content": "A ZLE function duplicating EMACS' zap-to-char.\n" }, { "name": "zsh/example", "content": "An example of how to write a module.\n" }, { "name": "zsh/files", "content": "Some basic file manipulation commands as builtins.\n" }, { "name": "zsh/langinfo", "content": "Interface to locale information.\n" }, { "name": "zsh/mapfile", "content": "Access to external files via a special associative array.\n" }, { "name": "zsh/mathfunc", "content": "Standard scientific functions for use in mathematical evaluations.\n" }, { "name": "zsh/nearcolor", "content": "Map colours to the nearest colour in the available palette.\n" }, { "name": "zsh/newuser", "content": "Arrange for files for new users to be installed.\n" }, { "name": "zsh/parameter", "content": "Access to internal hash tables via special associative arrays.\n" }, { "name": "zsh/pcre", "content": "Interface to the PCRE library.\n" }, { "name": "zsh/param/private", "content": "Builtins for managing private-scoped parameters in function context.\n" }, { "name": "zsh/regex", "content": "Interface to the POSIX regex library.\n" }, { "name": "zsh/sched", "content": "A builtin that provides a timed execution facility within the shell.\n" }, { "name": "zsh/net/socket", "content": "Manipulation of Unix domain sockets\n" }, { "name": "zsh/stat", "content": "A builtin command interface to the stat system call.\n" }, { "name": "zsh/system", "content": "A builtin interface to various low-level system features.\n" }, { "name": "zsh/net/tcp", "content": "Manipulation of TCP sockets\n" }, { "name": "zsh/termcap", "content": "Interface to the termcap database.\n" }, { "name": "zsh/terminfo", "content": "Interface to the terminfo database.\n" }, { "name": "zsh/zftp", "content": "A builtin FTP client.\n" }, { "name": "zsh/zle", "content": "The Zsh Line Editor, including the bindkey and vared builtins.\n" }, { "name": "zsh/zleparameter", "content": "Access to internals of the Zsh Line Editor via parameters.\n" }, { "name": "zsh/zprof", "content": "A module allowing profiling for shell functions.\n" }, { "name": "zsh/zpty", "content": "A builtin for starting a command in a pseudo-terminal.\n" }, { "name": "zsh/zselect", "content": "Block and return when file descriptors are ready.\n" }, { "name": "zsh/zutil", "content": "Some utility builtins, e.g. the one for supporting configuration via styles.\n" } ] }, "THE ZSH/ATTR MODULE": { "content": "The zsh/attr module is used for manipulating extended attributes. The -h option causes all\ncommands to operate on symbolic links instead of their targets. The builtins in this module\nare:\n\nzgetattr [ -h ] filename attribute [ parameter ]\nGet the extended attribute attribute from the specified filename. If the optional ar‐\ngument parameter is given, the attribute is set on that parameter instead of being\nprinted to stdout.\n\nzsetattr [ -h ] filename attribute value\nSet the extended attribute attribute on the specified filename to value.\n\nzdelattr [ -h ] filename attribute\nRemove the extended attribute attribute from the specified filename.\n\nzlistattr [ -h ] filename [ parameter ]\nList the extended attributes currently set on the specified filename. If the optional\nargument parameter is given, the list of attributes is set on that parameter instead\nof being printed to stdout.\n\nzgetattr and zlistattr allocate memory dynamically. If the attribute or list of attributes\ngrows between the allocation and the call to get them, they return 2. On all other errors, 1\nis returned. This allows the calling function to check for this case and retry.\n", "subsections": [] }, "THE ZSH/CAP MODULE": { "content": "The zsh/cap module is used for manipulating POSIX.1e (POSIX.6) capability sets. If the oper‐\nating system does not support this interface, the builtins defined by this module will do\nnothing. The builtins in this module are:\n\ncap [ capabilities ]\nChange the shell's process capability sets to the specified capabilities, otherwise\ndisplay the shell's current capabilities.\n\ngetcap filename ...\nThis is a built-in implementation of the POSIX standard utility. It displays the ca‐\npability sets on each specified filename.\n\nsetcap capabilities filename ...\nThis is a built-in implementation of the POSIX standard utility. It sets the capabil‐\nity sets on each specified filename to the specified capabilities.\n", "subsections": [] }, "THE ZSH/CLONE MODULE": { "content": "The zsh/clone module makes available one builtin command:\n\nclone tty\nCreates a forked instance of the current shell, attached to the specified tty. In the\nnew shell, the PID, PPID and TTY special parameters are changed appropriately. $! is\nset to zero in the new shell, and to the new shell's PID in the original shell.\n\nThe return status of the builtin is zero in both shells if successful, and non-zero on\nerror.\n\nThe target of clone should be an unused terminal, such as an unused virtual console or\na virtual terminal created by\n\nxterm -e sh -c 'trap : INT QUIT TSTP; tty;\nwhile :; do sleep 100000000; done'\n\nSome words of explanation are warranted about this long xterm command line: when doing\nclone on a pseudo-terminal, some other session (\"session\" meant as a unix session\ngroup, or SID) is already owning the terminal. Hence the cloned zsh cannot acquire the\npseudo-terminal as a controlling tty. That means two things:\n\n• the job control signals will go to the sh-started-by-xterm process group\n(that's why we disable INT QUIT and TSTP with trap; otherwise the while loop\ncould get suspended or killed)\n\n• the cloned shell will have job control disabled, and the job control keys (con‐\ntrol-C, control-\\ and control-Z) will not work.\n\nThis does not apply when cloning to an unused vc.\n\nCloning to a used (and unprepared) terminal will result in two processes reading si‐\nmultaneously from the same terminal, with input bytes going randomly to either\nprocess.\n\nclone is mostly useful as a shell built-in replacement for openvt.\n", "subsections": [] }, "THE ZSH/COMPCTL MODULE": { "content": "The zsh/compctl module makes available two builtin commands. compctl, is the old, deprecated\nway to control completions for ZLE. See zshcompctl(1). The other builtin command, compcall\ncan be used in user-defined completion widgets, see zshcompwid(1).\n", "subsections": [] }, "THE ZSH/COMPLETE MODULE": { "content": "The zsh/complete module makes available several builtin commands which can be used in\nuser-defined completion widgets, see zshcompwid(1).\n", "subsections": [] }, "THE ZSH/COMPLIST MODULE": { "content": "The zsh/complist module offers three extensions to completion listings: the ability to high‐\nlight matches in such a list, the ability to scroll through long lists and a different style\nof menu completion.\n", "subsections": [ { "name": "Colored completion listings", "content": "Whenever one of the parameters ZLSCOLORS or ZLSCOLOURS is set and the zsh/complist module\nis loaded or linked into the shell, completion lists will be colored. Note, however, that\ncomplist will not automatically be loaded if it is not linked in: on systems with dynamic\nloading, `zmodload zsh/complist' is required.\n\nThe parameters ZLSCOLORS and ZLSCOLOURS describe how matches are highlighted. To turn on\nhighlighting an empty value suffices, in which case all the default values given below will\nbe used. The format of the value of these parameters is the same as used by the GNU version\nof the ls command: a colon-separated list of specifications of the form `name=value'. The\nname may be one of the following strings, most of which specify file types for which the\nvalue will be used. The strings and their default values are:\n\nno 0 for normal text (i.e. when displaying something other than a matched file)\n\nfi 0 for regular files\n\ndi 32 for directories\n\nln 36 for symbolic links. If this has the special value target, symbolic links are derefer‐\nenced and the target file used to determine the display format.\n\npi 31 for named pipes (FIFOs)\n\nso 33 for sockets\n" }, { "name": "bd 44;37", "content": "for block devices\n" }, { "name": "cd 44;37", "content": "for character devices\n\nor none\nfor a symlink to nonexistent file (default is the value defined for ln)\n\nmi none\nfor a non-existent file (default is the value defined for fi); this code is currently\nnot used\n" }, { "name": "su 37;41", "content": "for files with setuid bit set\n" }, { "name": "sg 30;43", "content": "for files with setgid bit set\n" }, { "name": "tw 30;42", "content": "for world writable directories with sticky bit set\n" }, { "name": "ow 34;43", "content": "for world writable directories without sticky bit set\n\nsa none\nfor files with an associated suffix alias; this is only tested after specific suf‐\nfixes, as described below\n" }, { "name": "st 37;44", "content": "for directories with sticky bit set but not world writable\n\nex 35 for executable files\n\nlc \\e[ for the left code (see below)\n\nrc m for the right code\n\ntc 0 for the character indicating the file type printed after filenames if the LISTTYPES\noption is set\n\nsp 0 for the spaces printed after matches to align the next column\n\nec none\nfor the end code\n\nApart from these strings, the name may also be an asterisk (`*') followed by any string. The\nvalue given for such a string will be used for all files whose name ends with the string.\nThe name may also be an equals sign (`=') followed by a pattern; the EXTENDEDGLOB option\nwill be turned on for evaluation of the pattern. The value given for this pattern will be\nused for all matches (not just filenames) whose display string are matched by the pattern.\nDefinitions for the form with the leading equal sign take precedence over the values defined\nfor file types, which in turn take precedence over the form with the leading asterisk (file\nextensions).\n\nThe leading-equals form also allows different parts of the displayed strings to be colored\ndifferently. For this, the pattern has to use the `(#b)' globbing flag and pairs of paren‐\ntheses surrounding the parts of the strings that are to be colored differently. In this case\nthe value may consist of more than one color code separated by equal signs. The first code\nwill be used for all parts for which no explicit code is specified and the following codes\nwill be used for the parts matched by the sub-patterns in parentheses. For example, the\nspecification `=(#b)(?)*(?)=0=3=7' will be used for all matches which are at least two char‐\nacters long and will use the code `3' for the first character, `7' for the last character and\n`0' for the rest.\n\nAll three forms of name may be preceded by a pattern in parentheses. If this is given, the\nvalue will be used only for matches in groups whose names are matched by the pattern given in\nthe parentheses. For example, `(g*)m*=43' highlights all matches beginning with `m' in\ngroups whose names begin with `g' using the color code `43'. In case of the `lc', `rc', and\n`ec' codes, the group pattern is ignored.\n\nNote also that all patterns are tried in the order in which they appear in the parameter\nvalue until the first one matches which is then used. Patterns may be matched against com‐\npletions, descriptions (possibly with spaces appended for padding), or lines consisting of a\ncompletion followed by a description. For consistent coloring it may be necessary to use\nmore than one pattern or a pattern with backreferences.\n\nWhen printing a match, the code prints the value of lc, the value for the file-type or the\nlast matching specification with a `*', the value of rc, the string to display for the match\nitself, and then the value of ec if that is defined or the values of lc, no, and rc if ec is\nnot defined.\n\nThe default values are ISO 6429 (ANSI) compliant and can be used on vt100 compatible termi‐\nnals such as xterms. On monochrome terminals the default values will have no visible effect.\nThe colors function from the contribution can be used to get associative arrays containing\nthe codes for ANSI terminals (see the section `Other Functions' in zshcontrib(1)). For exam‐\nple, after loading colors, one could use `$color[red]' to get the code for foreground color\nred and `$color[bg-green]' for the code for background color green.\n\nIf the completion system invoked by compinit is used, these parameters should not be set di‐\nrectly because the system controls them itself. Instead, the list-colors style should be\nused (see the section `Completion System Configuration' in zshcompsys(1)).\n" }, { "name": "Scrolling in completion listings", "content": "To enable scrolling through a completion list, the LISTPROMPT parameter must be set. Its\nvalue will be used as the prompt; if it is the empty string, a default prompt will be used.\nThe value may contain escapes of the form `%x'. It supports the escapes `%B', `%b', `%S',\n`%s', `%U', `%u', `%F', `%f', `%K', `%k' and `%{...%}' used also in shell prompts as well as\nthree pairs of additional sequences: a `%l' or `%L' is replaced by the number of the last\nline shown and the total number of lines in the form `number/total'; a `%m' or `%M' is re‐\nplaced with the number of the last match shown and the total number of matches; and `%p' or\n`%P' is replaced with `Top', `Bottom' or the position of the first line shown in percent of\nthe total number of lines, respectively. In each of these cases the form with the uppercase\nletter will be replaced with a string of fixed width, padded to the right with spaces, while\nthe lowercase form will not be padded.\n\nIf the parameter LISTPROMPT is set, the completion code will not ask if the list should be\nshown. Instead it immediately starts displaying the list, stopping after the first screen‐\nful, showing the prompt at the bottom, waiting for a keypress after temporarily switching to\nthe listscroll keymap. Some of the zle functions have a special meaning while scrolling\nlists:\n" }, { "name": "send-break", "content": "stops listing discarding the key pressed\n\naccept-line, down-history, down-line-or-history\ndown-line-or-search, vi-down-line-or-history\nscrolls forward one line\n\ncomplete-word, menu-complete, expand-or-complete\nexpand-or-complete-prefix, menu-complete-or-expand\nscrolls forward one screenful\n" }, { "name": "accept-search", "content": "stop listing but take no other action\n\nEvery other character stops listing and immediately processes the key as usual. Any key that\nis not bound in the listscroll keymap or that is bound to undefined-key is looked up in the\nkeymap currently selected.\n\nAs for the ZLSCOLORS and ZLSCOLOURS parameters, LISTPROMPT should not be set directly when\nusing the shell function based completion system. Instead, the list-prompt style should be\nused.\n" }, { "name": "Menu selection", "content": "The zsh/complist module also offers an alternative style of selecting matches from a list,\ncalled menu selection, which can be used if the shell is set up to return to the last prompt\nafter showing a completion list (see the ALWAYSLASTPROMPT option in zshoptions(1)).\n\nMenu selection can be invoked directly by the widget menu-select defined by this module.\nThis is a standard ZLE widget that can be bound to a key in the usual way as described in\nzshzle(1).\n\nAlternatively, the parameter MENUSELECT can be set to an integer, which gives the minimum\nnumber of matches that must be present before menu selection is automatically turned on.\nThis second method requires that menu completion be started, either directly from a widget\nsuch as menu-complete, or due to one of the options MENUCOMPLETE or AUTOMENU being set. If\nMENUSELECT is set, but is 0, 1 or empty, menu selection will always be started during an am‐\nbiguous menu completion.\n\nWhen using the completion system based on shell functions, the MENUSELECT parameter should\nnot be used (like the ZLSCOLORS and ZLSCOLOURS parameters described above). Instead, the\nmenu style should be used with the select=... keyword.\n\nAfter menu selection is started, the matches will be listed. If there are more matches than\nfit on the screen, only the first screenful is shown. The matches to insert into the command\nline can be selected from this list. In the list one match is highlighted using the value\nfor ma from the ZLSCOLORS or ZLSCOLOURS parameter. The default value for this is `7' which\nforces the selected match to be highlighted using standout mode on a vt100-compatible termi‐\nnal. If neither ZLSCOLORS nor ZLSCOLOURS is set, the same terminal control sequence as for\nthe `%S' escape in prompts is used.\n\nIf there are more matches than fit on the screen and the parameter MENUPROMPT is set, its\nvalue will be shown below the matches. It supports the same escape sequences as LISTPROMPT,\nbut the number of the match or line shown will be that of the one where the mark is placed.\nIf its value is the empty string, a default prompt will be used.\n\nThe MENUSCROLL parameter can be used to specify how the list is scrolled. If the parameter\nis unset, this is done line by line, if it is set to `0' (zero), the list will scroll half\nthe number of lines of the screen. If the value is positive, it gives the number of lines to\nscroll and if it is negative, the list will be scrolled the number of lines of the screen mi‐\nnus the (absolute) value.\n\nAs for the ZLSCOLORS, ZLSCOLOURS and LISTPROMPT parameters, neither MENUPROMPT nor\nMENUSCROLL should be set directly when using the shell function based completion system. In‐\nstead, the select-prompt and select-scroll styles should be used.\n\nThe completion code sometimes decides not to show all of the matches in the list. These hid‐\nden matches are either matches for which the completion function which added them explicitly\nrequested that they not appear in the list (using the -n option of the compadd builtin com‐\nmand) or they are matches which duplicate a string already in the list (because they differ\nonly in things like prefixes or suffixes that are not displayed). In the list used for menu\nselection, however, even these matches are shown so that it is possible to select them. To\nhighlight such matches the hi and du capabilities in the ZLSCOLORS and ZLSCOLOURS parame‐\nters are supported for hidden matches of the first and second kind, respectively.\n\nSelecting matches is done by moving the mark around using the zle movement functions. When\nnot all matches can be shown on the screen at the same time, the list will scroll up and down\nwhen crossing the top or bottom line. The following zle functions have special meaning dur‐\ning menu selection. Note that the following always perform the same task within the menu se‐\nlection map and cannot be replaced by user defined widgets, nor can the set of functions be\nextended:\n\naccept-line, accept-search\naccept the current match and leave menu selection (but do not cause the command line\nto be accepted)\n" }, { "name": "send-break", "content": "leaves menu selection and restores the previous contents of the command line\n\nredisplay, clear-screen\nexecute their normal function without leaving menu selection\n\naccept-and-hold, accept-and-menu-complete\naccept the currently inserted match and continue selection allowing to select the next\nmatch to insert into the line\n" }, { "name": "accept-and-infer-next-history", "content": "accepts the current match and then tries completion with menu selection again; in the\ncase of files this allows one to select a directory and immediately attempt to com‐\nplete files in it; if there are no matches, a message is shown and one can use undo\nto go back to completion on the previous level, every other key leaves menu selection\n(including the other zle functions which are otherwise special during menu selection)\n\nundo removes matches inserted during the menu selection by one of the three functions be‐\nfore\n\ndown-history, down-line-or-history\nvi-down-line-or-history, down-line-or-search\nmoves the mark one line down\n\nup-history, up-line-or-history\nvi-up-line-or-history, up-line-or-search\nmoves the mark one line up\n\nforward-char, vi-forward-char\nmoves the mark one column right\n\nbackward-char, vi-backward-char\nmoves the mark one column left\n\nforward-word, vi-forward-word\nvi-forward-word-end, emacs-forward-word\nmoves the mark one screenful down\n\nbackward-word, vi-backward-word, emacs-backward-word\nmoves the mark one screenful up\n\nvi-forward-blank-word, vi-forward-blank-word-end\nmoves the mark to the first line of the next group of matches\n" }, { "name": "vi-backward-blank-word", "content": "moves the mark to the last line of the previous group of matches\n" }, { "name": "beginning-of-history", "content": "moves the mark to the first line\n" }, { "name": "end-of-history", "content": "moves the mark to the last line\n\nbeginning-of-buffer-or-history, beginning-of-line\nbeginning-of-line-hist, vi-beginning-of-line\nmoves the mark to the leftmost column\n\nend-of-buffer-or-history, end-of-line\nend-of-line-hist, vi-end-of-line\nmoves the mark to the rightmost column\n\ncomplete-word, menu-complete, expand-or-complete\nexpand-or-complete-prefix, menu-expand-or-complete\nmoves the mark to the next match\n" }, { "name": "reverse-menu-complete", "content": "moves the mark to the previous match\n" }, { "name": "vi-insert", "content": "this toggles between normal and interactive mode; in interactive mode the keys bound\nto self-insert and self-insert-unmeta insert into the command line as in normal edit‐\ning mode but without leaving menu selection; after each character completion is tried\nagain and the list changes to contain only the new matches; the completion widgets\nmake the longest unambiguous string be inserted in the command line and undo and back‐‐\nward-delete-char go back to the previous set of matches\n" }, { "name": "history-incremental-search-forward", "content": "" }, { "name": "history-incremental-search-backward", "content": "this starts incremental searches in the list of completions displayed; in this mode,\naccept-line only leaves incremental search, going back to the normal menu selection\nmode\n\nAll movement functions wrap around at the edges; any other zle function not listed leaves\nmenu selection and executes that function. It is possible to make widgets in the above list\ndo the same by using the form of the widget with a `.' in front. For example, the widget\n`.accept-line' has the effect of leaving menu selection and accepting the entire command\nline.\n\nDuring this selection the widget uses the keymap menuselect. Any key that is not defined in\nthis keymap or that is bound to undefined-key is looked up in the keymap currently selected.\nThis is used to ensure that the most important keys used during selection (namely the cursor\nkeys, return, and TAB) have sensible defaults. However, keys in the menuselect keymap can be\nmodified directly using the bindkey builtin command (see zshmodules(1)). For example, to make\nthe return key leave menu selection without accepting the match currently selected one could\ncall\n\nbindkey -M menuselect '^M' send-break\n\nafter loading the zsh/complist module.\n" } ] }, "THE ZSH/COMPUTIL MODULE": { "content": "The zsh/computil module adds several builtin commands that are used by some of the completion\nfunctions in the completion system based on shell functions (see zshcompsys(1) ). Except for\ncompquote these builtin commands are very specialised and thus not very interesting when\nwriting your own completion functions. In summary, these builtin commands are:\n", "subsections": [ { "name": "comparguments", "content": "This is used by the arguments function to do the argument and command line parsing.\nLike compdescribe it has an option -i to do the parsing and initialize some internal\nstate and various options to access the state information to decide what should be\ncompleted.\n" }, { "name": "compdescribe", "content": "This is used by the describe function to build the displays for the matches and to\nget the strings to add as matches with their options. On the first call one of the\noptions -i or -I should be supplied as the first argument. In the first case, display\nstrings without the descriptions will be generated, in the second case, the string\nused to separate the matches from their descriptions must be given as the second argu‐\nment and the descriptions (if any) will be shown. All other arguments are like the\ndefinition arguments to describe itself.\n\nOnce compdescribe has been called with either the -i or the -I option, it can be re‐\npeatedly called with the -g option and the names of four parameters as its arguments.\nThis will step through the different sets of matches and store the value of comp‐‐\nstate[list] in the first scalar, the options for compadd in the second array, the\nmatches in the third array, and the strings to be displayed in the completion listing\nin the fourth array. The arrays may then be directly given to compadd to register the\nmatches with the completion code.\n" }, { "name": "compfiles", "content": "Used by the pathfiles function to optimize complex recursive filename generation\n(globbing). It does three things. With the -p and -P options it builds the glob pat‐\nterns to use, including the paths already handled and trying to optimize the patterns\nwith respect to the prefix and suffix from the line and the match specification cur‐\nrently used. The -i option does the directory tests for the ignore-parents style and\nthe -r option tests if a component for some of the matches are equal to the string on\nthe line and removes all other matches if that is true.\n" }, { "name": "compgroups", "content": "Used by the tags function to implement the internals of the group-order style. This\nonly takes its arguments as names of completion groups and creates the groups for it\n(all six types: sorted and unsorted, both without removing duplicates, with removing\nall duplicates and with removing consecutive duplicates).\n\ncompquote [ -p ] names ...\nThere may be reasons to write completion functions that have to add the matches using\nthe -Q option to compadd and perform quoting themselves. Instead of interpreting the\nfirst character of the allquotes key of the compstate special association and using\nthe q flag for parameter expansions, one can use this builtin command. The arguments\nare the names of scalar or array parameters and the values of these parameters are\nquoted as needed for the innermost quoting level. If the -p option is given, quoting\nis done as if there is some prefix before the values of the parameters, so that a\nleading equal sign will not be quoted.\n\nThe return status is non-zero in case of an error and zero otherwise.\n" }, { "name": "comptags", "content": "" }, { "name": "comptry", "content": "These implement the internals of the tags mechanism.\n" }, { "name": "compvalues", "content": "Like comparguments, but for the values function.\n" } ] }, "THE ZSH/CURSES MODULE": { "content": "The zsh/curses module makes available one builtin command and various parameters.\n", "subsections": [ { "name": "Builtin", "content": "" }, { "name": "zcurses init", "content": "" }, { "name": "zcurses end", "content": "zcurses addwin targetwin nlines ncols beginy beginx [ parentwin ]\nzcurses delwin targetwin\nzcurses refresh [ targetwin ... ]\nzcurses touch targetwin ...\nzcurses move targetwin newy newx\nzcurses clear targetwin [ redraw | eol | bot ]\nzcurses position targetwin array\nzcurses char targetwin character\nzcurses string targetwin string\nzcurses border targetwin border\nzcurses attr targetwin [ [+|-]attribute | fgcol/bgcol ] [...]\nzcurses bg targetwin [ [+|-]attribute | fgcol/bgcol | @char ] [...]\nzcurses scroll targetwin [ on | off | [+|-]lines ]\nzcurses input targetwin [ param [ kparam [ mparam ] ] ]\nzcurses mouse [ delay num | [+|-]motion ]\nzcurses timeout targetwin intval\nzcurses querychar targetwin [ param ]\nzcurses resize height width [ endwin | nosave | endwinnosave ]\nManipulate curses windows. All uses of this command should be bracketed by `zcurses\ninit' to initialise use of curses, and `zcurses end' to end it; omitting `zcurses end'\ncan cause the terminal to be in an unwanted state.\n\nThe subcommand addwin creates a window with nlines lines and ncols columns. Its upper\nleft corner will be placed at row beginy and column beginx of the screen. targetwin\nis a string and refers to the name of a window that is not currently assigned. Note\nin particular the curses convention that vertical values appear before horizontal val‐\nues.\n\nIf addwin is given an existing window as the final argument, the new window is created\nas a subwindow of parentwin. This differs from an ordinary new window in that the\nmemory of the window contents is shared with the parent's memory. Subwindows must be\ndeleted before their parent. Note that the coordinates of subwindows are relative to\nthe screen, not the parent, as with other windows.\n\nUse the subcommand delwin to delete a window created with addwin. Note that end does\nnot implicitly delete windows, and that delwin does not erase the screen image of the\nwindow.\n\nThe window corresponding to the full visible screen is called stdscr; it always exists\nafter `zcurses init' and cannot be delete with delwin.\n\nThe subcommand refresh will refresh window targetwin; this is necessary to make any\npending changes (such as characters you have prepared for output with char) visible on\nthe screen. refresh without an argument causes the screen to be cleared and redrawn.\nIf multiple windows are given, the screen is updated once at the end.\n\nThe subcommand touch marks the targetwins listed as changed. This is necessary before\nrefreshing windows if a window that was in front of another window (which may be std‐‐\nscr) is deleted.\n\nThe subcommand move moves the cursor position in targetwin to new coordinates newy\nand newx. Note that the subcommand string (but not the subcommand char) advances the\ncursor position over the characters added.\n\nThe subcommand clear erases the contents of targetwin. One (and no more than one) of\nthree options may be specified. With the option redraw, in addition the next refresh\nof targetwin will cause the screen to be cleared and repainted. With the option eol,\ntargetwin is only cleared to the end of the current cursor line. With the option bot,\ntargetwin is cleared to the end of the window, i.e everything to the right and below\nthe cursor is cleared.\n\nThe subcommand position writes various positions associated with targetwin into the\narray named array. These are, in order:\n- The y and x coordinates of the cursor relative to the top left of targetwin\n- The y and x coordinates of the top left of targetwin on the screen\n- The size of targetwin in y and x dimensions.\n\nOutputting characters and strings are achieved by char and string respectively.\n\nTo draw a border around window targetwin, use border. Note that the border is not\nsubsequently handled specially: in other words, the border is simply a set of charac‐\nters output at the edge of the window. Hence it can be overwritten, can scroll off\nthe window, etc.\n\nThe subcommand attr will set targetwin's attributes or foreground/background color\npair for any successive character output. Each attribute given on the line may be\nprepended by a + to set or a - to unset that attribute; + is assumed if absent. The\nattributes supported are blink, bold, dim, reverse, standout, and underline.\n\nEach fgcol/bgcol attribute (to be read as `fgcol on bgcol') sets the foreground\nand background color for character output. The color default is sometimes available\n(in particular if the library is ncurses), specifying the foreground or background\ncolor with which the terminal started. The color pair default/default is always\navailable. To use more than the 8 named colors (red, green, etc.) construct the\nfgcol/bgcol pairs where fgcol and bgcol are decimal integers, e.g 128/200. The\nmaximum color value is 254 if the terminal supports 256 colors.\n\nbg overrides the color and other attributes of all characters in the window. Its\nusual use is to set the background initially, but it will overwrite the attributes of\nany characters at the time when it is called. In addition to the arguments allowed\nwith attr, an argument @char specifies a character to be shown in otherwise blank ar‐\neas of the window. Owing to limitations of curses this cannot be a multibyte charac‐\nter (use of ASCII characters only is recommended). As the specified set of attributes\noverride the existing background, turning attributes off in the arguments is not use‐\nful, though this does not cause an error.\n\nThe subcommand scroll can be used with on or off to enabled or disable scrolling of a\nwindow when the cursor would otherwise move below the window due to typing or output.\nIt can also be used with a positive or negative integer to scroll the window up or\ndown the given number of lines without changing the current cursor position (which\ntherefore appears to move in the opposite direction relative to the window). In the\nsecond case, if scrolling is off it is temporarily turned on to allow the window to be\nscrolled.\n\nThe subcommand input reads a single character from the window without echoing it back.\nIf param is supplied the character is assigned to the parameter param, else it is as‐\nsigned to the parameter REPLY.\n\nIf both param and kparam are supplied, the key is read in `keypad' mode. In this mode\nspecial keys such as function keys and arrow keys return the name of the key in the\nparameter kparam. The key names are the macros defined in the curses.h or ncurses.h\nwith the prefix `KEY' removed; see also the description of the parameter zcurseskey‐‐\ncodes below. Other keys cause a value to be set in param as before. On a successful\nreturn only one of param or kparam contains a non-empty string; the other is set to an\nempty string.\n\nIf mparam is also supplied, input attempts to handle mouse input. This is only avail‐\nable with the ncurses library; mouse handling can be detected by checking for the exit\nstatus of `zcurses mouse' with no arguments. If a mouse button is clicked (or double-\nor triple-clicked, or pressed or released with a configurable delay from being\nclicked) then kparam is set to the string MOUSE, and mparam is set to an array con‐\nsisting of the following elements:\n- An identifier to discriminate different input devices; this is only rarely use‐\nful.\n- The x, y and z coordinates of the mouse click relative to the full screen, as\nthree elements in that order (i.e. the y coordinate is, unusually, after the x\ncoordinate). The z coordinate is only available for a few unusual input de‐\nvices and is otherwise set to zero.\n- Any events that occurred as separate items; usually there will be just one. An\nevent consists of PRESSED, RELEASED, CLICKED, DOUBLECLICKED or TRIPLECLICKED\nfollowed immediately (in the same element) by the number of the button.\n- If the shift key was pressed, the string SHIFT.\n- If the control key was pressed, the string CTRL.\n- If the alt key was pressed, the string ALT.\n\nNot all mouse events may be passed through to the terminal window; most terminal emu‐\nlators handle some mouse events themselves. Note that the ncurses manual implies that\nusing input both with and without mouse handling may cause the mouse cursor to appear\nand disappear.\n\nThe subcommand mouse can be used to configure the use of the mouse. There is no win‐\ndow argument; mouse options are global. `zcurses mouse' with no arguments returns\nstatus 0 if mouse handling is possible, else status 1. Otherwise, the possible argu‐\nments (which may be combined on the same command line) are as follows. delay num sets\nthe maximum delay in milliseconds between press and release events to be considered as\na click; the value 0 disables click resolution, and the default is one sixth of a sec‐\nond. motion proceeded by an optional `+' (the default) or - turns on or off reporting\nof mouse motion in addition to clicks, presses and releases, which are always re‐\nported. However, it appears reports for mouse motion are not currently implemented.\n\nThe subcommand timeout specifies a timeout value for input from targetwin. If intval\nis negative, `zcurses input' waits indefinitely for a character to be typed; this is\nthe default. If intval is zero, `zcurses input' returns immediately; if there is ty‐\npeahead it is returned, else no input is done and status 1 is returned. If intval is\npositive, `zcurses input' waits intval milliseconds for input and if there is none at\nthe end of that period returns status 1.\n\nThe subcommand querychar queries the character at the current cursor position. The\nreturn values are stored in the array named param if supplied, else in the array re‐‐\nply. The first value is the character (which may be a multibyte character if the sys‐\ntem supports them); the second is the color pair in the usual fgcol/bgcol notation,\nor 0 if color is not supported. Any attributes other than color that apply to the\ncharacter, as set with the subcommand attr, appear as additional elements.\n\nThe subcommand resize resizes stdscr and all windows to given dimensions (windows that\nstick out from the new dimensions are resized down). The underlying curses extension\n(resizeterm call) can be unavailable. To verify, zeroes can be used for height and\nwidth. If the result of the subcommand is 0, resizeterm is available (2 otherwise).\nTests show that resizing can be normally accomplished by calling zcurses end and\nzcurses refresh. The resize subcommand is provided for versatility. Multiple system\nconfigurations have been checked and zcurses end and zcurses refresh are still needed\nfor correct terminal state after resize. To invoke them with resize, use endwin argu‐\nment. Using nosave argument will cause new terminal state to not be saved internally\nby zcurses. This is also provided for versatility and should normally be not needed.\n" }, { "name": "Parameters", "content": "ZCURSESCOLORS\nReadonly integer. The maximum number of colors the terminal supports. This value is\ninitialised by the curses library and is not available until the first time zcurses\ninit is run.\n\nZCURSESCOLORPAIRS\nReadonly integer. The maximum number of color pairs fgcol/bgcol that may be defined\nin `zcurses attr' commands; note this limit applies to all color pairs that have been\nused whether or not they are currently active. This value is initialised by the\ncurses library and is not available until the first time zcurses init is run.\n\nzcursesattrs\nReadonly array. The attributes supported by zsh/curses; available as soon as the mod‐\nule is loaded.\n\nzcursescolors\nReadonly array. The colors supported by zsh/curses; available as soon as the module\nis loaded.\n\nzcurseskeycodes\nReadonly array. The values that may be returned in the second parameter supplied to\n`zcurses input' in the order in which they are defined internally by curses. Not all\nfunction keys are listed, only F0; curses reserves space for F0 up to F63.\n\nzcurseswindows\nReadonly array. The current list of windows, i.e. all windows that have been created\nwith `zcurses addwin' and not removed with `zcurses delwin'.\n" } ] }, "THE ZSH/DATETIME MODULE": { "content": "The zsh/datetime module makes available one builtin command:\n\nstrftime [ -s scalar ] format [ epochtime [ nanoseconds ] ]\nstrftime -r [ -q ] [ -s scalar ] format timestring\nOutput the date in the format specified. With no epochtime, the current system\ndate/time is used; optionally, epochtime may be used to specify the number of seconds\nsince the epoch, and nanoseconds may additionally be used to specify the number of\nnanoseconds past the second (otherwise that number is assumed to be 0). See strf‐\ntime(3) for details. The zsh extensions described in the section EXPANSION OF PROMPT\nSEQUENCES in zshmisc(1) are also available.\n\n-q Run quietly; suppress printing of all error messages described below. Errors\nfor invalid epochtime values are always printed.\n\n-r With the option -r (reverse), use format to parse the input string timestring\nand output the number of seconds since the epoch at which the time occurred.\nThe parsing is implemented by the system function strptime; see strptime(3).\nThis means that zsh format extensions are not available, but for reverse lookup\nthey are not required.\n\nIn most implementations of strftime any timezone in the timestring is ignored\nand the local timezone declared by the TZ environment variable is used; other\nparameters are set to zero if not present.\n\nIf timestring does not match format the command returns status 1 and prints an\nerror message. If timestring matches format but not all characters in\ntimestring were used, the conversion succeeds but also prints an error message.\n\nIf either of the system functions strptime or mktime is not available, status 2\nis returned and an error message is printed.\n\n-s scalar\nAssign the date string (or epoch time in seconds if -r is given) to scalar in‐\nstead of printing it.\n\nNote that depending on the system's declared integral time type, strftime may produce\nincorrect results for epoch times greater than 2147483647 which corresponds to\n2038-01-19 03:14:07 +0000.\n\nThe zsh/datetime module makes available several parameters; all are readonly:\n\nEPOCHREALTIME\nA floating point value representing the number of seconds since the epoch. The no‐\ntional accuracy is to nanoseconds if the clockgettime call is available and to mi‐\ncroseconds otherwise, but in practice the range of double precision floating point and\nshell scheduling latencies may be significant effects.\n\nEPOCHSECONDS\nAn integer value representing the number of seconds since the epoch.\n", "subsections": [ { "name": "epochtime", "content": "An array value containing the number of seconds since the epoch in the first element\nand the remainder of the time since the epoch in nanoseconds in the second element.\nTo ensure the two elements are consistent the array should be copied or otherwise ref‐\nerenced as a single substitution before the values are used. The following idiom may\nbe used:\n\nfor secs nsecs in $epochtime; do\n...\ndone\n" } ] }, "THE ZSH/DB/GDBM MODULE": { "content": "The zsh/db/gdbm module is used to create \"tied\" associative arrays that interface to database\nfiles. If the GDBM interface is not available, the builtins defined by this module will re‐\nport an error. This module is also intended as a prototype for creating additional database\ninterfaces, so the ztie builtin may move to a more generic module in the future.\n\nThe builtins in this module are:\n\nztie -d db/gdbm -f filename [ -r ] arrayname\nOpen the GDBM database identified by filename and, if successful, create the associa‐\ntive array arrayname linked to the file. To create a local tied array, the parameter\nmust first be declared, so commands similar to the following would be executed inside\na function scope:\n\nlocal -A sampledb\nztie -d db/gdbm -f sample.gdbm sampledb\n\nThe -r option opens the database file for reading only, creating a parameter with the\nreadonly attribute. Without this option, using `ztie' on a file for which the user\ndoes not have write permission is an error. If writable, the database is opened syn‐\nchronously so fields changed in arrayname are immediately written to filename.\n\nChanges to the file modes filename after it has been opened do not alter the state of\narrayname, but `typeset -r arrayname' works as expected.\n\nzuntie [ -u ] arrayname ...\nClose the GDBM database associated with each arrayname and then unset the parameter.\nThe -u option forces an unset of parameters made readonly with `ztie -r'.\n\nThis happens automatically if the parameter is explicitly unset or its local scope\n(function) ends. Note that a readonly parameter may not be explicitly unset, so the\nonly way to unset a global parameter created with `ztie -r' is to use `zuntie -u'.\n\nzgdbmpath parametername\nPut path to database file assigned to parametername into REPLY scalar.\n\nzgdbmtied\nArray holding names of all tied parameters.\n\nThe fields of an associative array tied to GDBM are neither cached nor otherwise stored in\nmemory, they are read from or written to the database on each reference. Thus, for example,\nthe values in a readonly array may be changed by a second writer of the same database file.\n", "subsections": [] }, "THE ZSH/DELTOCHAR MODULE": { "content": "The zsh/deltochar module makes available two ZLE functions:\n", "subsections": [ { "name": "delete-to-char", "content": "Read a character from the keyboard, and delete from the cursor position up to and in‐\ncluding the next (or, with repeat count n, the nth) instance of that character. Nega‐\ntive repeat counts mean delete backwards.\n" }, { "name": "zap-to-char", "content": "This behaves like delete-to-char, except that the final occurrence of the character\nitself is not deleted.\n" } ] }, "THE ZSH/EXAMPLE MODULE": { "content": "The zsh/example module makes available one builtin command:\n\nexample [ -flags ] [ args ... ]\nDisplays the flags and arguments it is invoked with.\n\nThe purpose of the module is to serve as an example of how to write a module.\n", "subsections": [] }, "THE ZSH/FILES MODULE": { "content": "The zsh/files module makes available some common commands for file manipulation as builtins;\nthese commands are probably not needed for many normal situations but can be useful in emer‐\ngency recovery situations with constrained resources. The commands do not implement all fea‐\ntures now required by relevant standards committees.\n\nFor all commands, a variant beginning zf is also available and loaded automatically. Using\nthe features capability of zmodload will let you load only those names you want. Note that\nit's possible to load only the builtins with zsh-specific names using the following command:\n\nzmodload -m -F zsh/files b:zf\\*\n\nThe commands loaded by default are:\n\nchgrp [ -hRs ] group filename ...\nChanges group of files specified. This is equivalent to chown with a user-spec argu‐\nment of `:group'.\n\nchmod [ -Rs ] mode filename ...\nChanges mode of files specified.\n\nThe specified mode must be in octal.\n\nThe -R option causes chmod to recursively descend into directories, changing the mode\nof all files in the directory after changing the mode of the directory itself.\n\nThe -s option is a zsh extension to chmod functionality. It enables paranoid behav‐\niour, intended to avoid security problems involving a chmod being tricked into affect‐\ning files other than the ones intended. It will refuse to follow symbolic links, so\nthat (for example) ``chmod 600 /tmp/foo/passwd'' can't accidentally chmod /etc/passwd\nif /tmp/foo happens to be a link to /etc. It will also check where it is after leav‐\ning directories, so that a recursive chmod of a deep directory tree can't end up re‐\ncursively chmoding /usr as a result of directories being moved up the tree.\n\nchown [ -hRs ] user-spec filename ...\nChanges ownership and group of files specified.\n\nThe user-spec can be in four forms:\n\nuser change owner to user; do not change group\nuser:: change owner to user; do not change group\nuser: change owner to user; change group to user's primary group\nuser:group\nchange owner to user; change group to group\n:group do not change owner; change group to group\n\nIn each case, the `:' may instead be a `.'. The rule is that if there is a `:' then\nthe separator is `:', otherwise if there is a `.' then the separator is `.', otherwise\nthere is no separator.\n\nEach of user and group may be either a username (or group name, as appropriate) or a\ndecimal user ID (group ID). Interpretation as a name takes precedence, if there is an\nall-numeric username (or group name).\n\nIf the target is a symbolic link, the -h option causes chown to set the ownership of\nthe link instead of its target.\n\nThe -R option causes chown to recursively descend into directories, changing the own‐\nership of all files in the directory after changing the ownership of the directory it‐\nself.\n\nThe -s option is a zsh extension to chown functionality. It enables paranoid behav‐\niour, intended to avoid security problems involving a chown being tricked into affect‐\ning files other than the ones intended. It will refuse to follow symbolic links, so\nthat (for example) ``chown luser /tmp/foo/passwd'' can't accidentally chown\n/etc/passwd if /tmp/foo happens to be a link to /etc. It will also check where it is\nafter leaving directories, so that a recursive chown of a deep directory tree can't\nend up recursively chowning /usr as a result of directories being moved up the tree.\n\nln [ -dfhins ] filename dest\nln [ -dfhins ] filename ... dir\nCreates hard (or, with -s, symbolic) links. In the first form, the specified destina‐\ntion is created, as a link to the specified filename. In the second form, each of the\nfilenames is taken in turn, and linked to a pathname in the specified directory that\nhas the same last pathname component.\n\nNormally, ln will not attempt to create hard links to directories. This check can be\noverridden using the -d option. Typically only the super-user can actually succeed in\ncreating hard links to directories. This does not apply to symbolic links in any\ncase.\n\nBy default, existing files cannot be replaced by links. The -i option causes the user\nto be queried about replacing existing files. The -f option causes existing files to\nbe silently deleted, without querying. -f takes precedence.\n\nThe -h and -n options are identical and both exist for compatibility; either one indi‐\ncates that if the target is a symlink then it should not be dereferenced. Typically\nthis is used in combination with -sf so that if an existing link points to a directory\nthen it will be removed, instead of followed. If this option is used with multiple\nfilenames and the target is a symbolic link pointing to a directory then the result is\nan error.\n\nmkdir [ -p ] [ -m mode ] dir ...\nCreates directories. With the -p option, non-existing parent directories are first\ncreated if necessary, and there will be no complaint if the directory already exists.\nThe -m option can be used to specify (in octal) a set of file permissions for the cre‐\nated directories, otherwise mode 777 modified by the current umask (see umask(2)) is\nused.\n\nmv [ -fi ] filename dest\nmv [ -fi ] filename ... dir\nMoves files. In the first form, the specified filename is moved to the specified des‐\ntination. In the second form, each of the filenames is taken in turn, and moved to a\npathname in the specified directory that has the same last pathname component.\n\nBy default, the user will be queried before replacing any file that the user cannot\nwrite to, but writable files will be silently removed. The -i option causes the user\nto be queried about replacing any existing files. The -f option causes any existing\nfiles to be silently deleted, without querying. -f takes precedence.\n\nNote that this mv will not move files across devices. Historical versions of mv, when\nactual renaming is impossible, fall back on copying and removing files; if this behav‐\niour is desired, use cp and rm manually. This may change in a future version.\n\nrm [ -dfiRrs ] filename ...\nRemoves files and directories specified.\n\nNormally, rm will not remove directories (except with the -R or -r options). The -d\noption causes rm to try removing directories with unlink (see unlink(2)), the same\nmethod used for files. Typically only the super-user can actually succeed in unlink‐\ning directories in this way. -d takes precedence over -R and -r.\n\nBy default, the user will be queried before removing any file that the user cannot\nwrite to, but writable files will be silently removed. The -i option causes the user\nto be queried about removing any files. The -f option causes files to be silently\ndeleted, without querying, and suppresses all error indications. -f takes precedence.\n\nThe -R and -r options cause rm to recursively descend into directories, deleting all\nfiles in the directory before removing the directory with the rmdir system call (see\nrmdir(2)).\n\nThe -s option is a zsh extension to rm functionality. It enables paranoid behaviour,\nintended to avoid common security problems involving a root-run rm being tricked into\nremoving files other than the ones intended. It will refuse to follow symbolic links,\nso that (for example) ``rm /tmp/foo/passwd'' can't accidentally remove /etc/passwd if\n/tmp/foo happens to be a link to /etc. It will also check where it is after leaving\ndirectories, so that a recursive removal of a deep directory tree can't end up recur‐\nsively removing /usr as a result of directories being moved up the tree.\n\nrmdir dir ...\nRemoves empty directories specified.\n\nsync Calls the system call of the same name (see sync(2)), which flushes dirty buffers to\ndisk. It might return before the I/O has actually been completed.\n", "subsections": [] }, "THE ZSH/LANGINFO MODULE": { "content": "The zsh/langinfo module makes available one parameter:\n", "subsections": [ { "name": "langinfo", "content": "An associative array that maps langinfo elements to their values.\n\nYour implementation may support a number of the following keys:\n\nCODESET, DTFMT, DFMT, TFMT, RADIXCHAR, THOUSEP, YESEXPR, NOEXPR, CRNCYSTR, AB‐‐\nDAY{1..7}, DAY{1..7}, ABMON{1..12}, MON{1..12}, TFMTAMPM, AMSTR, PMSTR, ERA,\nERADFMT, ERADTFMT, ERATFMT, ALTDIGITS\n" } ] }, "THE ZSH/MAPFILE MODULE": { "content": "The zsh/mapfile module provides one special associative array parameter of the same name.\n", "subsections": [ { "name": "mapfile", "content": "This associative array takes as keys the names of files; the resulting value is the\ncontent of the file. The value is treated identically to any other text coming from a\nparameter. The value may also be assigned to, in which case the file in question is\nwritten (whether or not it originally existed); or an element may be unset, which will\ndelete the file in question. For example, `vared mapfile[myfile]' works as expected,\nediting the file `myfile'.\n\nWhen the array is accessed as a whole, the keys are the names of files in the current\ndirectory, and the values are empty (to save a huge overhead in memory). Thus\n${(k)mapfile} has the same effect as the glob operator *(D), since files beginning\nwith a dot are not special. Care must be taken with expressions such as rm ${(k)map‐‐\nfile}, which will delete every file in the current directory without the usual `rm *'\ntest.\n\nThe parameter mapfile may be made read-only; in that case, files referenced may not be\nwritten or deleted.\n\nA file may conveniently be read into an array as one line per element with the form\n`array=(\"${(f@)mapfile[filename]}\")'. The double quotes and the `@' are necessary to\nprevent empty lines from being removed. Note that if the file ends with a newline,\nthe shell will split on the final newline, generating an additional empty field; this\ncan be suppressed by using `array=(\"${(f@)${mapfile[filename]%$'\\n'}}\")'.\n" }, { "name": "Limitations", "content": "Although reading and writing of the file in question is efficiently handled, zsh's internal\nmemory management may be arbitrarily baroque; however, mapfile is usually very much more ef‐\nficient than anything involving a loop. Note in particular that the whole contents of the\nfile will always reside physically in memory when accessed (possibly multiple times, due to\nstandard parameter substitution operations). In particular, this means handling of suffi‐\nciently long files (greater than the machine's swap space, or than the range of the pointer\ntype) will be incorrect.\n\nNo errors are printed or flagged for non-existent, unreadable, or unwritable files, as the\nparameter mechanism is too low in the shell execution hierarchy to make this convenient.\n\nIt is unfortunate that the mechanism for loading modules does not yet allow the user to spec‐\nify the name of the shell parameter to be given the special behaviour.\n" } ] }, "THE ZSH/MATHFUNC MODULE": { "content": "The zsh/mathfunc module provides standard mathematical functions for use when evaluating\nmathematical formulae. The syntax agrees with normal C and FORTRAN conventions, for example,\n\n(( f = sin(0.3) ))\n\nassigns the sine of 0.3 to the parameter f.\n\nMost functions take floating point arguments and return a floating point value. However, any\nnecessary conversions from or to integer type will be performed automatically by the shell.\nApart from atan with a second argument and the abs, int and float functions, all functions\nbehave as noted in the manual page for the corresponding C function, except that any argu‐\nments out of range for the function in question will be detected by the shell and an error\nreported.\n\nThe following functions take a single floating point argument: acos, acosh, asin, asinh,\natan, atanh, cbrt, ceil, cos, cosh, erf, erfc, exp, expm1, fabs, floor, gamma, j0, j1,\nlgamma, log, log10, log1p, log2, logb, sin, sinh, sqrt, tan, tanh, y0, y1. The atan function\ncan optionally take a second argument, in which case it behaves like the C function atan2.\nThe ilogb function takes a single floating point argument, but returns an integer.\n\nThe function signgam takes no arguments, and returns an integer, which is the C variable of\nthe same name, as described in gamma(3). Note that it is therefore only useful immediately\nafter a call to gamma or lgamma. Note also that `signgam()' and `signgam' are distinct ex‐\npressions.\n\nThe functions min, max, and sum are defined not in this module but in the zmathfunc autoload‐\nable function, described in the section `Mathematical Functions' in zshcontrib(1).\n\nThe following functions take two floating point arguments: copysign, fmod, hypot, nextafter.\n\nThe following take an integer first argument and a floating point second argument: jn, yn.\n\nThe following take a floating point first argument and an integer second argument: ldexp,\nscalb.\n\nThe function abs does not convert the type of its single argument; it returns the absolute\nvalue of either a floating point number or an integer. The functions float and int convert\ntheir arguments into a floating point or integer value (by truncation) respectively.\n\nNote that the C pow function is available in ordinary math evaluation as the `' operator\nand is not provided here.\n\nThe function rand48 is available if your system's mathematical library has the function\nerand48(3). It returns a pseudo-random floating point number between 0 and 1. It takes a\nsingle string optional argument.\n\nIf the argument is not present, the random number seed is initialised by three calls to the\nrand(3) function --- this produces the same random numbers as the next three values of $RAN‐‐\nDOM.\n\nIf the argument is present, it gives the name of a scalar parameter where the current random\nnumber seed will be stored. On the first call, the value must contain at least twelve hexa‐\ndecimal digits (the remainder of the string is ignored), or the seed will be initialised in\nthe same manner as for a call to rand48 with no argument. Subsequent calls to rand48(param)\nwill then maintain the seed in the parameter param as a string of twelve hexadecimal digits,\nwith no base signifier. The random number sequences for different parameters are completely\nindependent, and are also independent from that used by calls to rand48 with no argument.\n\nFor example, consider\n\nprint $(( rand48(seed) ))\nprint $(( rand48() ))\nprint $(( rand48(seed) ))\n\nAssuming $seed does not exist, it will be initialised by the first call. In the second call,\nthe default seed is initialised; note, however, that because of the properties of rand()\nthere is a correlation between the seeds used for the two initialisations, so for more secure\nuses, you should generate your own 12-byte seed. The third call returns to the same sequence\nof random numbers used in the first call, unaffected by the intervening rand48().\n", "subsections": [] }, "THE ZSH/NEARCOLOR MODULE": { "content": "The zsh/nearcolor module replaces colours specified as hex triplets with the nearest colour\nin the 88 or 256 colour palettes that are widely used by terminal emulators. By default,\n24-bit true colour escape codes are generated when colours are specified using hex triplets.\nThese are not supported by all terminals. The purpose of this module is to make it easier to\ndefine colour preferences in a form that can work across a range of terminal emulators.\n\nAside from the default colour, the ANSI standard for terminal escape codes provides for eight\ncolours. The bright attribute brings this to sixteen. These basic colours are commonly used\nin terminal applications due to being widely supported. Expanded 88 and 256 colour palettes\nare also common and, while the first sixteen colours vary somewhat between terminals and con‐\nfigurations, these add a generally consistent and predictable set of colours.\n\nIn order to use the zsh/nearcolor module, it only needs to be loaded. Thereafter, whenever a\ncolour is specified using a hex triplet, it will be compared against each of the available\ncolours and the closest will be selected. The first sixteen colours are never matched in this\nprocess due to being unpredictable.\n\nIt isn't possible to reliably detect support for true colour in the terminal emulator. It is\ntherefore recommended to be selective in loading the zsh/nearcolor module. For example, the\nfollowing checks the COLORTERM environment variable:\n\n[[ $COLORTERM = *(24bit|truecolor)* ]] || zmodload zsh/nearcolor\n\nNote that some terminals accept the true color escape codes but map them internally to a more\nlimited palette in a similar manner to the zsh/nearcolor module.\n", "subsections": [] }, "THE ZSH/NEWUSER MODULE": { "content": "The zsh/newuser module is loaded at boot if it is available, the RCS option is set, and the\nPRIVILEGED option is not set (all three are true by default). This takes place immediately\nafter commands in the global zshenv file (typically /etc/zsh/zshenv), if any, have been exe‐\ncuted. If the module is not available it is silently ignored by the shell; the module may\nsafely be removed from $MODULEPATH by the administrator if it is not required.\n\nOn loading, the module tests if any of the start-up files .zshenv, .zprofile, .zshrc or .zlo‐‐\ngin exist in the directory given by the environment variable ZDOTDIR, or the user's home di‐\nrectory if that is not set. The test is not performed and the module halts processing if the\nshell was in an emulation mode (i.e. had been invoked as some other shell than zsh).\n\nIf none of the start-up files were found, the module then looks for the file newuser first in\na sitewide directory, usually the parent directory of the site-functions directory, and if\nthat is not found the module searches in a version-specific directory, usually the parent of\nthe functions directory containing version-specific functions. (These directories can be\nconfigured when zsh is built using the --enable-site-scriptdir=dir and --enable-scriptdir=dir\nflags to configure, respectively; the defaults are prefix/share/zsh and pre‐\nfix/share/zsh/$ZSHVERSION where the default prefix is /usr/local.)\n\nIf the file newuser is found, it is then sourced in the same manner as a start-up file. The\nfile is expected to contain code to install start-up files for the user, however any valid\nshell code will be executed.\n\nThe zsh/newuser module is then unconditionally unloaded.\n\nNote that it is possible to achieve exactly the same effect as the zsh/newuser module by\nadding code to /etc/zsh/zshenv. The module exists simply to allow the shell to make arrange‐\nments for new users without the need for intervention by package maintainers and system ad‐\nministrators.\n\nThe script supplied with the module invokes the shell function zsh-newuser-install. This may\nbe invoked directly by the user even if the zsh/newuser module is disabled. Note, however,\nthat if the module is not installed the function will not be installed either. The function\nis documented in the section User Configuration Functions in zshcontrib(1).\n", "subsections": [] }, "THE ZSH/PARAMETER MODULE": { "content": "The zsh/parameter module gives access to some of the internal hash tables used by the shell\nby defining some special parameters.\n", "subsections": [ { "name": "options", "content": "The keys for this associative array are the names of the options that can be set and\nunset using the setopt and unsetopt builtins. The value of each key is either the\nstring on if the option is currently set, or the string off if the option is unset.\nSetting a key to one of these strings is like setting or unsetting the option, respec‐\ntively. Unsetting a key in this array is like setting it to the value off.\n" }, { "name": "commands", "content": "This array gives access to the command hash table. The keys are the names of external\ncommands, the values are the pathnames of the files that would be executed when the\ncommand would be invoked. Setting a key in this array defines a new entry in this ta‐\nble in the same way as with the hash builtin. Unsetting a key as in `unset \"com‐‐\nmands[foo]\"' removes the entry for the given key from the command hash table.\n" }, { "name": "functions", "content": "This associative array maps names of enabled functions to their definitions. Setting a\nkey in it is like defining a function with the name given by the key and the body\ngiven by the value. Unsetting a key removes the definition for the function named by\nthe key.\n\ndisfunctions\nLike functions but for disabled functions.\n\nfunctionssource\nThis readonly associative array maps names of enabled functions to the name of the\nfile containing the source of the function.\n\nFor an autoloaded function that has already been loaded, or marked for autoload with\nan absolute path, or that has had its path resolved with `functions -r', this is the\nfile found for autoloading, resolved to an absolute path.\n\nFor a function defined within the body of a script or sourced file, this is the name\nof that file. In this case, this is the exact path originally used to that file,\nwhich may be a relative path.\n\nFor any other function, including any defined at an interactive prompt or an autoload\nfunction whose path has not yet been resolved, this is the empty string. However, the\nhash element is reported as defined just so long as the function is present: the keys\nto this hash are the same as those to $functions.\n\ndisfunctionssource\nLike functionssource but for disabled functions.\n" }, { "name": "builtins", "content": "This associative array gives information about the builtin commands currently enabled.\nThe keys are the names of the builtin commands and the values are either `undefined'\nfor builtin commands that will automatically be loaded from a module if invoked or\n`defined' for builtin commands that are already loaded.\n\ndisbuiltins\nLike builtins but for disabled builtin commands.\n" }, { "name": "reswords", "content": "This array contains the enabled reserved words.\n\ndisreswords\nLike reswords but for disabled reserved words.\n" }, { "name": "patchars", "content": "This array contains the enabled pattern characters.\n\ndispatchars\nLike patchars but for disabled pattern characters.\n" }, { "name": "aliases", "content": "This maps the names of the regular aliases currently enabled to their expansions.\n\ndisaliases\nLike aliases but for disabled regular aliases.\n" }, { "name": "galiases", "content": "Like aliases, but for global aliases.\n\ndisgaliases\nLike galiases but for disabled global aliases.\n" }, { "name": "saliases", "content": "Like raliases, but for suffix aliases.\n\ndissaliases\nLike saliases but for disabled suffix aliases.\n" }, { "name": "parameters", "content": "The keys in this associative array are the names of the parameters currently defined.\nThe values are strings describing the type of the parameter, in the same format used\nby the t parameter flag, see zshexpn(1) . Setting or unsetting keys in this array is\nnot possible.\n" }, { "name": "modules", "content": "An associative array giving information about modules. The keys are the names of the\nmodules loaded, registered to be autoloaded, or aliased. The value says which state\nthe named module is in and is one of the strings `loaded', `autoloaded', or\n`alias:name', where name is the name the module is aliased to.\n\nSetting or unsetting keys in this array is not possible.\n" }, { "name": "dirstack", "content": "A normal array holding the elements of the directory stack. Note that the output of\nthe dirs builtin command includes one more directory, the current working directory.\n" }, { "name": "history", "content": "This associative array maps history event numbers to the full history lines. Although\nit is presented as an associative array, the array of all values (${history[@]}) is\nguaranteed to be returned in order from most recent to oldest history event, that is,\nby decreasing history event number.\n" }, { "name": "historywords", "content": "A special array containing the words stored in the history. These also appear in most\nto least recent order.\n" }, { "name": "jobdirs", "content": "This associative array maps job numbers to the directories from which the job was\nstarted (which may not be the current directory of the job).\n\nThe keys of the associative arrays are usually valid job numbers, and these are the\nvalues output with, for example, ${(k)jobdirs}. Non-numeric job references may be\nused when looking up a value; for example, ${jobdirs[%+]} refers to the current job.\n" }, { "name": "jobtexts", "content": "This associative array maps job numbers to the texts of the command lines that were\nused to start the jobs.\n\nHandling of the keys of the associative array is as described for jobdirs above.\n" }, { "name": "jobstates", "content": "This associative array gives information about the states of the jobs currently known.\nThe keys are the job numbers and the values are strings of the form\n`job-state:mark:pid=state...'. The job-state gives the state the whole job is cur‐\nrently in, one of `running', `suspended', or `done'. The mark is `+' for the current\njob, `-' for the previous job and empty otherwise. This is followed by one\n`:pid=state' for every process in the job. The pids are, of course, the process IDs\nand the state describes the state of that process.\n\nHandling of the keys of the associative array is as described for jobdirs above.\n" }, { "name": "nameddirs", "content": "This associative array maps the names of named directories to the pathnames they stand\nfor.\n" }, { "name": "userdirs", "content": "This associative array maps user names to the pathnames of their home directories.\n" }, { "name": "usergroups", "content": "This associative array maps names of system groups of which the current user is a mem‐\nber to the corresponding group identifiers. The contents are the same as the groups\noutput by the id command.\n" }, { "name": "funcfiletrace", "content": "This array contains the absolute line numbers and corresponding file names for the\npoint where the current function, sourced file, or (if EVALLINENO is set) eval com‐\nmand was called. The array is of the same length as funcsourcetrace and functrace,\nbut differs from funcsourcetrace in that the line and file are the point of call, not\nthe point of definition, and differs from functrace in that all values are absolute\nline numbers in files, rather than relative to the start of a function, if any.\n" }, { "name": "funcsourcetrace", "content": "This array contains the file names and line numbers of the points where the functions,\nsourced files, and (if EVALLINENO is set) eval commands currently being executed were\ndefined. The line number is the line where the `function name' or `name ()' started.\nIn the case of an autoloaded function the line number is reported as zero. The for‐\nmat of each element is filename:lineno.\n\nFor functions autoloaded from a file in native zsh format, where only the body of the\nfunction occurs in the file, or for files that have been executed by the source or `.'\nbuiltins, the trace information is shown as filename:0, since the entire file is the\ndefinition. The source file name is resolved to an absolute path when the function is\nloaded or the path to it otherwise resolved.\n\nMost users will be interested in the information in the funcfiletrace array instead.\n" }, { "name": "funcstack", "content": "This array contains the names of the functions, sourced files, and (if EVALLINENO is\nset) eval commands. currently being executed. The first element is the name of the\nfunction using the parameter.\n\nThe standard shell array zshevalcontext can be used to determine the type of shell\nconstruct being executed at each depth: note, however, that is in the opposite order,\nwith the most recent item last, and it is more detailed, for example including an en‐\ntry for toplevel, the main shell code being executed either interactively or from a\nscript, which is not present in $funcstack.\n" }, { "name": "functrace", "content": "This array contains the names and line numbers of the callers corresponding to the\nfunctions currently being executed. The format of each element is name:lineno. Call‐\ners are also shown for sourced files; the caller is the point where the source or `.'\ncommand was executed.\n" } ] }, "THE ZSH/PCRE MODULE": { "content": "The zsh/pcre module makes some commands available as builtins:\n\npcrecompile [ -aimxs ] PCRE\nCompiles a perl-compatible regular expression.\n\nOption -a will force the pattern to be anchored. Option -i will compile a case-insen‐\nsitive pattern. Option -m will compile a multi-line pattern; that is, ^ and $ will\nmatch newlines within the pattern. Option -x will compile an extended pattern,\nwherein whitespace and # comments are ignored. Option -s makes the dot metacharacter\nmatch all characters, including those that indicate newline.\n\npcrestudy\nStudies the previously-compiled PCRE which may result in faster matching.\n\npcrematch [ -v var ] [ -a arr ] [ -n offset ] [ -b ] string\nReturns successfully if string matches the previously-compiled PCRE.\n\nUpon successful match, if the expression captures substrings within parentheses,\npcrematch will set the array match to those substrings, unless the -a option is\ngiven, in which case it will set the array arr. Similarly, the variable MATCH will be\nset to the entire matched portion of the string, unless the -v option is given, in\nwhich case the variable var will be set. No variables are altered if there is no suc‐\ncessful match. A -n option starts searching for a match from the byte offset position\nin string. If the -b option is given, the variable ZPCREOP will be set to an offset\npair string, representing the byte offset positions of the entire matched portion\nwithin the string. For example, a ZPCREOP set to \"32 45\" indicates that the matched\nportion began on byte offset 32 and ended on byte offset 44. Here, byte offset posi‐\ntion 45 is the position directly after the matched portion. Keep in mind that the\nbyte position isn't necessarily the same as the character position when UTF-8 charac‐\nters are involved. Consequently, the byte offset positions are only to be relied on\nin the context of using them for subsequent searches on string, using an offset posi‐\ntion as an argument to the -n option. This is mostly used to implement the \"find all\nnon-overlapping matches\" functionality.\n\nA simple example of \"find all non-overlapping matches\":\n\nstring=\"The following zip codes: 78884 90210 99513\"\npcrecompile -m \"\\d{5}\"\naccum=()\npcrematch -b -- $string\nwhile [[ $? -eq 0 ]] do\nb=($=ZPCREOP)\naccum+=$MATCH\npcrematch -b -n $b[2] -- $string\ndone\nprint -l $accum\n\nThe zsh/pcre module makes available the following test condition:\n\nexpr -pcre-match pcre\nMatches a string against a perl-compatible regular expression.\n\nFor example,\n\n[[ \"$text\" -pcre-match ^d+$ ]] &&\nprint text variable contains only \"d's\".\n\nIf the REMATCHPCRE option is set, the =~ operator is equivalent to -pcre-match, and\nthe NOCASEMATCH option may be used. Note that NOCASEMATCH never applies to the\npcrematch builtin, instead use the -i switch of pcrecompile.\n", "subsections": [] }, "THE ZSH/PARAM/PRIVATE MODULE": { "content": "The zsh/param/private module is used to create parameters whose scope is limited to the cur‐\nrent function body, and not to other functions called by the current function.\n\nThis module provides a single autoloaded builtin:\n\nprivate [ {+|-}AHUahlprtux ] [ {+|-}EFLRZi [ n ] ] [ name[=value] ... ]\nThe private builtin accepts all the same options and arguments as local (zsh‐\nbuiltins(1)) except for the `-T' option. Tied parameters may not be made private.\n\nIf used at the top level (outside a function scope), private creates a normal parame‐\nter in the same manner as declare or typeset. A warning about this is printed if\nWARNCREATEGLOBAL is set (zshoptions(1)). Used inside a function scope, private cre‐\nates a local parameter similar to one declared with local, except having special prop‐\nerties noted below.\n\nSpecial parameters which expose or manipulate internal shell state, such as ARGC,\nargv, COLUMNS, LINES, UID, EUID, IFS, PROMPT, RANDOM, SECONDS, etc., cannot be made\nprivate unless the `-h' option is used to hide the special meaning of the parameter.\nThis may change in the future.\n\nAs with other typeset equivalents, private is both a builtin and a reserved word, so arrays\nmay be assigned with parenthesized word list name=(value...) syntax. However, the reserved\nword `private' is not available until zsh/param/private is loaded, so care must be taken with\norder of execution and parsing for function definitions which use private. To compensate for\nthis, the module also adds the option `-P' to the `local' builtin to declare private parame‐\nters.\n\nFor example, this construction fails if zsh/param/private has not yet been loaded when\n`baddeclaration' is defined:\nbaddeclaration() {\nzmodload zsh/param/private\nprivate array=( one two three )\n}\n\nThis construction works because local is already a keyword, and the module is loaded before\nthe statement is executed:\ngooddeclaration() {\nzmodload zsh/param/private\nlocal -P array=( one two three )\n}\n\nThe following is usable in scripts but may have trouble with autoload:\nzmodload zsh/param/private\niffydeclaration() {\nprivate array=( one two three )\n}\n\nThe private builtin may always be used with scalar assignments and for declarations without\nassignments.\n\nParameters declared with private have the following properties:\n\n• Within the function body where it is declared, the parameter behaves as a local, ex‐\ncept as noted above for tied or special parameters.\n\n• The type of a parameter declared private cannot be changed in the scope where it was\ndeclared, even if the parameter is unset. Thus an array cannot be assigned to a pri‐\nvate scalar, etc.\n\n• Within any other function called by the declaring function, the private parameter does\nNOT hide other parameters of the same name, so for example a global parameter of the\nsame name is visible and may be assigned or unset. This includes calls to anonymous\nfunctions, although that may also change in the future.\n\n• An exported private remains in the environment of inner scopes but appears unset for\nthe current shell in those scopes. Generally, exporting private parameters should be\navoided.\n\nNote that this differs from the static scope defined by compiled languages derived from C, in\nthat the a new call to the same function creates a new scope, i.e., the parameter is still\nassociated with the call stack rather than with the function definition. It differs from ksh\n`typeset -S' because the syntax used to define the function has no bearing on whether the pa‐\nrameter scope is respected.\n", "subsections": [] }, "THE ZSH/REGEX MODULE": { "content": "The zsh/regex module makes available the following test condition:\n\nexpr -regex-match regex\nMatches a string against a POSIX extended regular expression. On successful match,\nmatched portion of the string will normally be placed in the MATCH variable. If there\nare any capturing parentheses within the regex, then the match array variable will\ncontain those. If the match is not successful, then the variables will not be al‐\ntered.\n\nFor example,\n\n[[ alphabetical -regex-match ^a([^a]+)a([^a]+)a ]] &&\nprint -l $MATCH X $match\n\nIf the option REMATCHPCRE is not set, then the =~ operator will automatically load\nthis module as needed and will invoke the -regex-match operator.\n\nIf BASHREMATCH is set, then the array BASHREMATCH will be set instead of MATCH and\nmatch.\n", "subsections": [] }, "THE ZSH/SCHED MODULE": { "content": "The zsh/sched module makes available one builtin command and one parameter.\n\nsched [-o] [+]hh:mm[:ss] command ...\nsched [-o] [+]seconds command ...\nsched [ -item ]\nMake an entry in the scheduled list of commands to execute. The time may be specified\nin either absolute or relative time, and either as hours, minutes and (optionally)\nseconds separated by a colon, or seconds alone. An absolute number of seconds indi‐\ncates the time since the epoch (1970/01/01 00:00); this is useful in combination with\nthe features in the zsh/datetime module, see the zsh/datetime module entry in zshmod‐\nules(1).\n\nWith no arguments, prints the list of scheduled commands. If the scheduled command\nhas the -o flag set, this is shown at the start of the command.\n\nWith the argument `-item', removes the given item from the list. The numbering of the\nlist is continuous and entries are in time order, so the numbering can change when en‐\ntries are added or deleted.\n\nCommands are executed either immediately before a prompt, or while the shell's line\neditor is waiting for input. In the latter case it is useful to be able to produce\noutput that does not interfere with the line being edited. Providing the option -o\ncauses the shell to clear the command line before the event and redraw it afterwards.\nThis should be used with any scheduled event that produces visible output to the ter‐\nminal; it is not needed, for example, with output that updates a terminal emulator's\ntitle bar.\n\nTo effect changes to the editor buffer when an event executes, use the `zle' command\nwith no arguments to test whether the editor is active, and if it is, then use `zle\nwidget' to access the editor via the named widget.\n\nThe sched builtin is not made available by default when the shell starts in a mode em‐\nulating another shell. It can be made available with the command `zmodload -F\nzsh/sched b:sched'.\n\nzshscheduledevents\nA readonly array corresponding to the events scheduled by the sched builtin. The in‐\ndices of the array correspond to the numbers shown when sched is run with no arguments\n(provided that the KSHARRAYS option is not set). The value of the array consists of\nthe scheduled time in seconds since the epoch (see the section `The zsh/datetime Mod‐\nule' for facilities for using this number), followed by a colon, followed by any op‐\ntions (which may be empty but will be preceded by a `-' otherwise), followed by a\ncolon, followed by the command to be executed.\n\nThe sched builtin should be used for manipulating the events. Note that this will\nhave an immediate effect on the contents of the array, so that indices may become in‐\nvalid.\n", "subsections": [] }, "THE ZSH/NET/SOCKET MODULE": { "content": "The zsh/net/socket module makes available one builtin command:\n\nzsocket [ -altv ] [ -d fd ] [ args ]\nzsocket is implemented as a builtin to allow full use of shell command line editing,\nfile I/O, and job control mechanisms.\n", "subsections": [ { "name": "Outbound Connections", "content": "zsocket [ -v ] [ -d fd ] filename\nOpen a new Unix domain connection to filename. The shell parameter REPLY will be set\nto the file descriptor associated with that connection. Currently, only stream con‐\nnections are supported.\n\nIf -d is specified, its argument will be taken as the target file descriptor for the\nconnection.\n\nIn order to elicit more verbose output, use -v.\n\nFile descriptors can be closed with normal shell syntax when no longer needed, for ex‐\nample:\n\nexec {REPLY}>&-\n" }, { "name": "Inbound Connections", "content": "zsocket -l [ -v ] [ -d fd ] filename\nzsocket -l will open a socket listening on filename. The shell parameter REPLY will\nbe set to the file descriptor associated with that listener. The file descriptor re‐\nmains open in subshells and forked external executables.\n\nIf -d is specified, its argument will be taken as the target file descriptor for the\nconnection.\n\nIn order to elicit more verbose output, use -v.\n\nzsocket -a [ -tv ] [ -d targetfd ] listenfd\nzsocket -a will accept an incoming connection to the socket associated with listenfd.\nThe shell parameter REPLY will be set to the file descriptor associated with the in‐\nbound connection. The file descriptor remains open in subshells and forked external\nexecutables.\n\nIf -d is specified, its argument will be taken as the target file descriptor for the\nconnection.\n\nIf -t is specified, zsocket will return if no incoming connection is pending. Other‐\nwise it will wait for one.\n\nIn order to elicit more verbose output, use -v.\n" } ] }, "THE ZSH/STAT MODULE": { "content": "The zsh/stat module makes available one builtin command under two possible names:\n\nzstat [ -gnNolLtTrs ] [ -f fd ] [ -H hash ] [ -A array ] [ -F fmt ]\n[ +element ] [ file ... ]\nstat ...\nThe command acts as a front end to the stat system call (see stat(2)). The same com‐\nmand is provided with two names; as the name stat is often used by an external command\nit is recommended that only the zstat form of the command is used. This can be ar‐\nranged by loading the module with the command `zmodload -F zsh/stat b:zstat'.\n\nIf the stat call fails, the appropriate system error message printed and status 1 is\nreturned. The fields of struct stat give information about the files provided as ar‐\nguments to the command. In addition to those available from the stat call, an extra\nelement `link' is provided. These elements are:\n\ndevice The number of the device on which the file resides.\n\ninode The unique number of the file on this device (`inode' number).\n\nmode The mode of the file; that is, the file's type and access permissions. With\nthe -s option, this will be returned as a string corresponding to the first\ncolumn in the display of the ls -l command.\n\nnlink The number of hard links to the file.\n\nuid The user ID of the owner of the file. With the -s option, this is displayed as\na user name.\n\ngid The group ID of the file. With the -s option, this is displayed as a group\nname.\n\nrdev The raw device number. This is only useful for special devices.\n\nsize The size of the file in bytes.\n\natime\nmtime\nctime The last access, modification and inode change times of the file, respectively,\nas the number of seconds since midnight GMT on 1st January, 1970. With the -s\noption, these are printed as strings for the local time zone; the format can be\naltered with the -F option, and with the -g option the times are in GMT.\n\nblksize\nThe number of bytes in one allocation block on the device on which the file re‐\nsides.\n\nblock The number of disk blocks used by the file.\n\nlink If the file is a link and the -L option is in effect, this contains the name of\nthe file linked to, otherwise it is empty. Note that if this element is se‐\nlected (``zstat +link'') then the -L option is automatically used.\n\nA particular element may be selected by including its name preceded by a `+' in the\noption list; only one element is allowed. The element may be shortened to any unique\nset of leading characters. Otherwise, all elements will be shown for all files.\n\nOptions:\n\n-A array\nInstead of displaying the results on standard output, assign them to an array,\none struct stat element per array element for each file in order. In this case\nneither the name of the element nor the name of the files appears in array un‐\nless the -t or -n options were given, respectively. If -t is given, the ele‐\nment name appears as a prefix to the appropriate array element; if -n is given,\nthe file name appears as a separate array element preceding all the others.\nOther formatting options are respected.\n\n-H hash\nSimilar to -A, but instead assign the values to hash. The keys are the ele‐\nments listed above. If the -n option is provided then the name of the file is\nincluded in the hash with key name.\n\n-f fd Use the file on file descriptor fd instead of named files; no list of file\nnames is allowed in this case.\n\n-F fmt Supplies a strftime (see strftime(3)) string for the formatting of the time el‐\nements. The format string supports all of the zsh extensions described in the\nsection EXPANSION OF PROMPT SEQUENCES in zshmisc(1). The -s option is implied.\n\n-g Show the time elements in the GMT time zone. The -s option is implied.\n\n-l List the names of the type elements (to standard output or an array as appro‐\npriate) and return immediately; arguments, and options other than -A, are ig‐\nnored.\n\n-L Perform an lstat (see lstat(2)) rather than a stat system call. In this case,\nif the file is a link, information about the link itself rather than the target\nfile is returned. This option is required to make the link element useful.\nIt's important to note that this is the exact opposite from ls(1), etc.\n\n-n Always show the names of files. Usually these are only shown when output is to\nstandard output and there is more than one file in the list.\n\n-N Never show the names of files.\n\n-o If a raw file mode is printed, show it in octal, which is more useful for human\nconsumption than the default of decimal. A leading zero will be printed in\nthis case. Note that this does not affect whether a raw or formatted file mode\nis shown, which is controlled by the -r and -s options, nor whether a mode is\nshown at all.\n\n-r Print raw data (the default format) alongside string data (the -s format); the\nstring data appears in parentheses after the raw data.\n\n-s Print mode, uid, gid and the three time elements as strings instead of numbers.\nIn each case the format is like that of ls -l.\n\n-t Always show the type names for the elements of struct stat. Usually these are\nonly shown when output is to standard output and no individual element has been\nselected.\n\n-T Never show the type names of the struct stat elements.\n", "subsections": [] }, "THE ZSH/SYSTEM MODULE": { "content": "The zsh/system module makes available various builtin commands and parameters.\n", "subsections": [ { "name": "Builtins", "content": "syserror [ -e errvar ] [ -p prefix ] [ errno | errname ]\nThis command prints out the error message associated with errno, a system error num‐\nber, followed by a newline to standard error.\n\nInstead of the error number, a name errname, for example ENOENT, may be used. The set\nof names is the same as the contents of the array errnos, see below.\n\nIf the string prefix is given, it is printed in front of the error message, with no\nintervening space.\n\nIf errvar is supplied, the entire message, without a newline, is assigned to the pa‐\nrameter names errvar and nothing is output.\n\nA return status of 0 indicates the message was successfully printed (although it may\nnot be useful if the error number was out of the system's range), a return status of 1\nindicates an error in the parameters, and a return status of 2 indicates the error\nname was not recognised (no message is printed for this).\n\n\nsysopen [ -arw ] [ -m permissions ] [ -o options ]\n-u fd file\nThis command opens a file. The -r, -w and -a flags indicate whether the file should be\nopened for reading, writing and appending, respectively. The -m option allows the ini‐\ntial permissions to use when creating a file to be specified in octal form. The file\ndescriptor is specified with -u. Either an explicit file descriptor in the range 0 to\n9 can be specified or a variable name can be given to which the file descriptor number\nwill be assigned.\n\nThe -o option allows various system specific options to be specified as a comma-sepa‐\nrated list. The following is a list of possible options. Note that, depending on the\nsystem, some may not be available.\ncloexec\nmark file to be closed when other programs are executed (else the file descrip‐\ntor remains open in subshells and forked external executables)\n\ncreate\ncreat create file if it does not exist\n\nexcl create file, error if it already exists\n\nnoatime\nsuppress updating of the file atime\n\nnofollow\nfail if file is a symbolic link\n\nsync request that writes wait until data has been physically written\n\ntruncate\ntrunc truncate file to size 0\n\nTo close the file, use one of the following:\n\nexec {fd}<&-\nexec {fd}>&-\n\n\nsysread [ -c countvar ] [ -i infd ] [ -o outfd ]\n[ -s bufsize ] [ -t timeout ] [ param ]\nPerform a single system read from file descriptor infd, or zero if that is not given.\nThe result of the read is stored in param or REPLY if that is not given. If countvar\nis given, the number of bytes read is assigned to the parameter named by countvar.\n\nThe maximum number of bytes read is bufsize or 8192 if that is not given, however the\ncommand returns as soon as any number of bytes was successfully read.\n\nIf timeout is given, it specifies a timeout in seconds, which may be zero to poll the\nfile descriptor. This is handled by the poll system call if available, otherwise the\nselect system call if available.\n\nIf outfd is given, an attempt is made to write all the bytes just read to the file de‐\nscriptor outfd. If this fails, because of a system error other than EINTR or because\nof an internal zsh error during an interrupt, the bytes read but not written are\nstored in the parameter named by param if supplied (no default is used in this case),\nand the number of bytes read but not written is stored in the parameter named by\ncountvar if that is supplied. If it was successful, countvar contains the full number\nof bytes transferred, as usual, and param is not set.\n\nThe error EINTR (interrupted system call) is handled internally so that shell inter‐\nrupts are transparent to the caller. Any other error causes a return.\n\nThe possible return statuses are\n0 At least one byte of data was successfully read and, if appropriate, written.\n\n1 There was an error in the parameters to the command. This is the only error\nfor which a message is printed to standard error.\n\n2 There was an error on the read, or on polling the input file descriptor for a\ntimeout. The parameter ERRNO gives the error.\n\n3 Data were successfully read, but there was an error writing them to outfd. The\nparameter ERRNO gives the error.\n\n4 The attempt to read timed out. Note this does not set ERRNO as this is not a\nsystem error.\n\n5 No system error occurred, but zero bytes were read. This usually indicates end\nof file. The parameters are set according to the usual rules; no write to\noutfd is attempted.\n\nsysseek [ -u fd ] [ -w start|end|current ] offset\nThe current file position at which future reads and writes will take place is adjusted\nto the specified byte offset. The offset is evaluated as a math expression. The -u op‐\ntion allows the file descriptor to be specified. By default the offset is specified\nrelative to the start or the file but, with the -w option, it is possible to specify\nthat the offset should be relative to the current position or the end of the file.\n\nsyswrite [ -c countvar ] [ -o outfd ] data\nThe data (a single string of bytes) are written to the file descriptor outfd, or 1 if\nthat is not given, using the write system call. Multiple write operations may be used\nif the first does not write all the data.\n\nIf countvar is given, the number of byte written is stored in the parameter named by\ncountvar; this may not be the full length of data if an error occurred.\n\nThe error EINTR (interrupted system call) is handled internally by retrying; otherwise\nan error causes the command to return. For example, if the file descriptor is set to\nnon-blocking output, an error EAGAIN (on some systems, EWOULDBLOCK) may result in the\ncommand returning early.\n\nThe return status may be 0 for success, 1 for an error in the parameters to the com‐\nmand, or 2 for an error on the write; no error message is printed in the last case,\nbut the parameter ERRNO will reflect the error that occurred.\n\nzsystem flock [ -t timeout ] [ -f var ] [-er] file\nzsystem flock -u fdexpr\nThe builtin zsystem's subcommand flock performs advisory file locking (via the fc‐\nntl(2) system call) over the entire contents of the given file. This form of locking\nrequires the processes accessing the file to cooperate; its most obvious use is be‐\ntween two instances of the shell itself.\n\nIn the first form the named file, which must already exist, is locked by opening a\nfile descriptor to the file and applying a lock to the file descriptor. The lock ter‐\nminates when the shell process that created the lock exits; it is therefore often con‐\nvenient to create file locks within subshells, since the lock is automatically re‐\nleased when the subshell exits. Note that use of the print builtin with the -u option\nwill, as a side effect, release the lock, as will redirection to the file in the shell\nholding the lock. To work around this use a subshell, e.g. `(print message) >> file'.\nStatus 0 is returned if the lock succeeds, else status 1.\n\nIn the second form the file descriptor given by the arithmetic expression fdexpr is\nclosed, releasing a lock. The file descriptor can be queried by using the `-f var'\nform during the lock; on a successful lock, the shell variable var is set to the file\ndescriptor used for locking. The lock will be released if the file descriptor is\nclosed by any other means, for example using `exec {var}>&-'; however, the form de‐\nscribed here performs a safety check that the file descriptor is in use for file lock‐\ning.\n\nBy default the shell waits indefinitely for the lock to succeed. The option -t time‐\nout specifies a timeout for the lock in seconds; currently this must be an integer.\nThe shell will attempt to lock the file once a second during this period. If the at‐\ntempt times out, status 2 is returned.\n\nIf the option -e is given, the file descriptor for the lock is preserved when the\nshell uses exec to start a new process; otherwise it is closed at that point and the\nlock released.\n\nIf the option -r is given, the lock is only for reading, otherwise it is for reading\nand writing. The file descriptor is opened accordingly.\n\nzsystem supports subcommand\nThe builtin zsystem's subcommand supports tests whether a given subcommand is sup‐\nported. It returns status 0 if so, else status 1. It operates silently unless there\nwas a syntax error (i.e. the wrong number of arguments), in which case status 255 is\nreturned. Status 1 can indicate one of two things: subcommand is known but not sup‐\nported by the current operating system, or subcommand is not known (possibly because\nthis is an older version of the shell before it was implemented).\n" }, { "name": "Math Functions", "content": "systell(fd)\nThe systell math function returns the current file position for the file descriptor\npassed as an argument.\n" }, { "name": "Parameters", "content": "errnos A readonly array of the names of errors defined on the system. These are typically\nmacros defined in C by including the system header file errno.h. The index of each\nname (assuming the option KSHARRAYS is unset) corresponds to the error number. Error\nnumbers num before the last known error which have no name are given the name Enum in\nthe array.\n\nNote that aliases for errors are not handled; only the canonical name is used.\n" }, { "name": "sysparams", "content": "A readonly associative array. The keys are:\n\npid Returns the process ID of the current process, even in subshells. Compare $$,\nwhich returns the process ID of the main shell process.\n\nppid Returns the process ID of the parent of the current process, even in subshells.\nCompare $PPID, which returns the process ID of the parent of the main shell\nprocess.\n\nprocsubstpid\nReturns the process ID of the last process started for process substitution,\ni.e. the <(...) and >(...) expansions.\n" } ] }, "THE ZSH/NET/TCP MODULE": { "content": "The zsh/net/tcp module makes available one builtin command:\n\nztcp [ -acflLtv ] [ -d fd ] [ args ]\nztcp is implemented as a builtin to allow full use of shell command line editing, file\nI/O, and job control mechanisms.\n\nIf ztcp is run with no options, it will output the contents of its session table.\n\nIf it is run with only the option -L, it will output the contents of the session table\nin a format suitable for automatic parsing. The option is ignored if given with a\ncommand to open or close a session. The output consists of a set of lines, one per\nsession, each containing the following elements separated by spaces:\n\nFile descriptor\nThe file descriptor in use for the connection. For normal inbound (I) and out‐\nbound (O) connections this may be read and written by the usual shell mecha‐\nnisms. However, it should only be close with `ztcp -c'.\n\nConnection type\nA letter indicating how the session was created:\n\nZ A session created with the zftp command.\n\nL A connection opened for listening with `ztcp -l'.\n\nI An inbound connection accepted with `ztcp -a'.\n\nO An outbound connection created with `ztcp host ...'.\n\nThe local host\nThis is usually set to an all-zero IP address as the address of the localhost\nis irrelevant.\n\nThe local port\nThis is likely to be zero unless the connection is for listening.\n\nThe remote host\nThis is the fully qualified domain name of the peer, if available, else an IP\naddress. It is an all-zero IP address for a session opened for listening.\n\nThe remote port\nThis is zero for a connection opened for listening.\n", "subsections": [ { "name": "Outbound Connections", "content": "ztcp [ -v ] [ -d fd ] host [ port ]\nOpen a new TCP connection to host. If the port is omitted, it will default to port\n23. The connection will be added to the session table and the shell parameter REPLY\nwill be set to the file descriptor associated with that connection.\n\nIf -d is specified, its argument will be taken as the target file descriptor for the\nconnection.\n\nIn order to elicit more verbose output, use -v.\n" }, { "name": "Inbound Connections", "content": "ztcp -l [ -v ] [ -d fd ] port\nztcp -l will open a socket listening on TCP port. The socket will be added to the\nsession table and the shell parameter REPLY will be set to the file descriptor associ‐\nated with that listener.\n\nIf -d is specified, its argument will be taken as the target file descriptor for the\nconnection.\n\nIn order to elicit more verbose output, use -v.\n\nztcp -a [ -tv ] [ -d targetfd ] listenfd\nztcp -a will accept an incoming connection to the port associated with listenfd. The\nconnection will be added to the session table and the shell parameter REPLY will be\nset to the file descriptor associated with the inbound connection.\n\nIf -d is specified, its argument will be taken as the target file descriptor for the\nconnection.\n\nIf -t is specified, ztcp will return if no incoming connection is pending. Otherwise\nit will wait for one.\n\nIn order to elicit more verbose output, use -v.\n" }, { "name": "Closing Connections", "content": "ztcp -cf [ -v ] [ fd ]\nztcp -c [ -v ] [ fd ]\nztcp -c will close the socket associated with fd. The socket will be removed from the\nsession table. If fd is not specified, ztcp will close everything in the session ta‐\nble.\n\nNormally, sockets registered by zftp (see zshmodules(1) ) cannot be closed this way.\nIn order to force such a socket closed, use -f.\n\nIn order to elicit more verbose output, use -v.\n" }, { "name": "Example", "content": "Here is how to create a TCP connection between two instances of zsh. We need to pick an\nunassigned port; here we use the randomly chosen 5123.\n\nOn host1,\nzmodload zsh/net/tcp\nztcp -l 5123\nlistenfd=$REPLY\nztcp -a $listenfd\nfd=$REPLY\nThe second from last command blocks until there is an incoming connection.\n\nNow create a connection from host2 (which may, of course, be the same machine):\nzmodload zsh/net/tcp\nztcp host1 5123\nfd=$REPLY\n\nNow on each host, $fd contains a file descriptor for talking to the other. For example, on\nhost1:\nprint This is a message >&$fd\nand on host2:\nread -r line <&$fd; print -r - $line\nprints `This is a message'.\n\nTo tidy up, on host1:\nztcp -c $listenfd\nztcp -c $fd\nand on host2\nztcp -c $fd\n" } ] }, "THE ZSH/TERMCAP MODULE": { "content": "The zsh/termcap module makes available one builtin command:\n\nechotc cap [ arg ... ]\nOutput the termcap value corresponding to the capability cap, with optional arguments.\n\nThe zsh/termcap module makes available one parameter:\n", "subsections": [ { "name": "termcap", "content": "An associative array that maps termcap capability codes to their values.\n" } ] }, "THE ZSH/TERMINFO MODULE": { "content": "The zsh/terminfo module makes available one builtin command:\n\nechoti cap [ arg ]\nOutput the terminfo value corresponding to the capability cap, instantiated with arg\nif applicable.\n\nThe zsh/terminfo module makes available one parameter:\n", "subsections": [ { "name": "terminfo", "content": "An associative array that maps terminfo capability names to their values.\n" } ] }, "THE ZSH/ZFTP MODULE": { "content": "The zsh/zftp module makes available one builtin command:\n\nzftp subcommand [ args ]\nThe zsh/zftp module is a client for FTP (file transfer protocol). It is implemented\nas a builtin to allow full use of shell command line editing, file I/O, and job con‐\ntrol mechanisms. Often, users will access it via shell functions providing a more\npowerful interface; a set is provided with the zsh distribution and is described in\nzshzftpsys(1). However, the zftp command is entirely usable in its own right.\n\nAll commands consist of the command name zftp followed by the name of a subcommand.\nThese are listed below. The return status of each subcommand is supposed to reflect\nthe success or failure of the remote operation. See a description of the variable\nZFTPVERBOSE for more information on how responses from the server may be printed.\n", "subsections": [ { "name": "Subcommands", "content": "open host[:port] [ user [ password [ account ] ] ]\nOpen a new FTP session to host, which may be the name of a TCP/IP connected host or an\nIP number in the standard dot notation. If the argument is in the form host:port,\nopen a connection to TCP port port instead of the standard FTP port 21. This may be\nthe name of a TCP service or a number: see the description of ZFTPPORT below for\nmore information.\n\nIf IPv6 addresses in colon format are used, the host should be surrounded by quoted\nsquare brackets to distinguish it from the port, for example\n'[fe80::203:baff:fe02:8b56]'. For consistency this is allowed with all forms of host.\n\nRemaining arguments are passed to the login subcommand. Note that if no arguments be‐\nyond host are supplied, open will not automatically call login. If no arguments at\nall are supplied, open will use the parameters set by the params subcommand.\n\nAfter a successful open, the shell variables ZFTPHOST, ZFTPPORT, ZFTPIP and\nZFTPSYSTEM are available; see `Variables' below.\n\nlogin [ name [ password [ account ] ] ]\nuser [ name [ password [ account ] ] ]\nLogin the user name with parameters password and account. Any of the parameters can\nbe omitted, and will be read from standard input if needed (name is always needed).\nIf standard input is a terminal, a prompt for each one will be printed on standard er‐\nror and password will not be echoed. If any of the parameters are not used, a warning\nmessage is printed.\n\nAfter a successful login, the shell variables ZFTPUSER, ZFTPACCOUNT and ZFTPPWD are\navailable; see `Variables' below.\n\nThis command may be re-issued when a user is already logged in, and the server will\nfirst be reinitialized for a new user.\n\nparams [ host [ user [ password [ account ] ] ] ]" }, { "name": "params -", "content": "Store the given parameters for a later open command with no arguments. Only those\ngiven on the command line will be remembered. If no arguments are given, the parame‐\nters currently set are printed, although the password will appear as a line of stars;\nthe return status is one if no parameters were set, zero otherwise.\n\nAny of the parameters may be specified as a `?', which may need to be quoted to pro‐\ntect it from shell expansion. In this case, the appropriate parameter will be read\nfrom stdin as with the login subcommand, including special handling of password. If\nthe `?' is followed by a string, that is used as the prompt for reading the parameter\ninstead of the default message (any necessary punctuation and whitespace should be in‐\ncluded at the end of the prompt). The first letter of the parameter (only) may be\nquoted with a `\\'; hence an argument \"\\\\$word\" guarantees that the string from the\nshell parameter $word will be treated literally, whether or not it begins with a `?'.\n\nIf instead a single `-' is given, the existing parameters, if any, are deleted. In\nthat case, calling open with no arguments will cause an error.\n\nThe list of parameters is not deleted after a close, however it will be deleted if the\nzsh/zftp module is unloaded.\n\nFor example,\n\nzftp params ftp.elsewhere.xx juser '?Password for juser: '\n\nwill store the host ftp.elsewhere.xx and the user juser and then prompt the user for\nthe corresponding password with the given prompt.\n\ntest Test the connection; if the server has reported that it has closed the connection\n(maybe due to a timeout), return status 2; if no connection was open anyway, return\nstatus 1; else return status 0. The test subcommand is silent, apart from messages\nprinted by the $ZFTPVERBOSE mechanism, or error messages if the connection closes.\nThere is no network overhead for this test.\n\nThe test is only supported on systems with either the select(2) or poll(2) system\ncalls; otherwise the message `not supported on this system' is printed instead.\n\nThe test subcommand will automatically be called at the start of any other subcommand\nfor the current session when a connection is open.\n\ncd directory\nChange the remote directory to directory. Also alters the shell variable ZFTPPWD.\n\ncdup Change the remote directory to the one higher in the directory tree. Note that cd ..\nwill also work correctly on non-UNIX systems.\n\ndir [ arg ... ]\nGive a (verbose) listing of the remote directory. The args are passed directly to the\nserver. The command's behaviour is implementation dependent, but a UNIX server will\ntypically interpret args as arguments to the ls command and with no arguments return\nthe result of `ls -l'. The directory is listed to standard output.\n\nls [ arg ... ]\nGive a (short) listing of the remote directory. With no arg, produces a raw list of\nthe files in the directory, one per line. Otherwise, up to vagaries of the server im‐\nplementation, behaves similar to dir.\n\ntype [ type ]\nChange the type for the transfer to type, or print the current type if type is absent.\nThe allowed values are `A' (ASCII), `I' (Image, i.e. binary), or `B' (a synonym for\n`I').\n\nThe FTP default for a transfer is ASCII. However, if zftp finds that the remote host\nis a UNIX machine with 8-bit byes, it will automatically switch to using binary for\nfile transfers upon open. This can subsequently be overridden.\n\nThe transfer type is only passed to the remote host when a data connection is estab‐\nlished; this command involves no network overhead.\n\nascii The same as type A.\n\nbinary The same as type I.\n\nmode [ S | B ]\nSet the mode type to stream (S) or block (B). Stream mode is the default; block mode\nis not widely supported.\n\nremote file ...\nlocal [ file ... ]\nPrint the size and last modification time of the remote or local files. If there is\nmore than one item on the list, the name of the file is printed first. The first num‐\nber is the file size, the second is the last modification time of the file in the for‐\nmat CCYYMMDDhhmmSS consisting of year, month, date, hour, minutes and seconds in GMT.\nNote that this format, including the length, is guaranteed, so that time strings can\nbe directly compared via the [[ builtin's < and > operators, even if they are too long\nto be represented as integers.\n\nNot all servers support the commands for retrieving this information. In that case,\nthe remote command will print nothing and return status 2, compared with status 1 for\na file not found.\n\nThe local command (but not remote) may be used with no arguments, in which case the\ninformation comes from examining file descriptor zero. This is the same file as seen\nby a put command with no further redirection.\n\nget file ...\nRetrieve all files from the server, concatenating them and sending them to standard\noutput.\n\nput file ...\nFor each file, read a file from standard input and send that to the remote host with\nthe given name.\n\nappend file ...\nAs put, but if the remote file already exists, data is appended to it instead of over‐\nwriting it.\n\ngetat file point\nputat file point\nappendat file point\nVersions of get, put and append which will start the transfer at the given point in\nthe remote file. This is useful for appending to an incomplete local file. However,\nnote that this ability is not universally supported by servers (and is not quite the\nbehaviour specified by the standard).\n\ndelete file ...\nDelete the list of files on the server.\n\nmkdir directory\nCreate a new directory directory on the server.\n\nrmdir directory\nDelete the directory directory on the server.\n\nrename old-name new-name\nRename file old-name to new-name on the server.\n\nsite arg ...\nSend a host-specific command to the server. You will probably only need this if in‐\nstructed by the server to use it.\n\nquote arg ...\nSend the raw FTP command sequence to the server. You should be familiar with the FTP\ncommand set as defined in RFC959 before doing this. Useful commands may include STAT\nand HELP. Note also the mechanism for returning messages as described for the vari‐\nable ZFTPVERBOSE below, in particular that all messages from the control connection\nare sent to standard error.\n" }, { "name": "close", "content": "quit Close the current data connection. This unsets the shell parameters ZFTPHOST,\nZFTPPORT, ZFTPIP, ZFTPSYSTEM, ZFTPUSER, ZFTPACCOUNT, ZFTPPWD, ZFTPTYPE and\nZFTPMODE.\n\nsession [ sessname ]\nAllows multiple FTP sessions to be used at once. The name of the session is an arbi‐\ntrary string of characters; the default session is called `default'. If this command\nis called without an argument, it will list all the current sessions; with an argu‐\nment, it will either switch to the existing session called sessname, or create a new\nsession of that name.\n\nEach session remembers the status of the connection, the set of connection-specific\nshell parameters (the same set as are unset when a connection closes, as given in the\ndescription of close), and any user parameters specified with the params subcommand.\nChanging to a previous session restores those values; changing to a new session ini‐\ntialises them in the same way as if zftp had just been loaded. The name of the cur‐\nrent session is given by the parameter ZFTPSESSION.\n\nrmsession [ sessname ]\nDelete a session; if a name is not given, the current session is deleted. If the cur‐\nrent session is deleted, the earliest existing session becomes the new current ses‐\nsion, otherwise the current session is not changed. If the session being deleted is\nthe only one, a new session called `default' is created and becomes the current ses‐\nsion; note that this is a new session even if the session being deleted is also called\n`default'. It is recommended that sessions not be deleted while background commands\nwhich use zftp are still active.\n" }, { "name": "Parameters", "content": "The following shell parameters are used by zftp. Currently none of them are special.\n\nZFTPTMOUT\nInteger. The time in seconds to wait for a network operation to complete before re‐\nturning an error. If this is not set when the module is loaded, it will be given the\ndefault value 60. A value of zero turns off timeouts. If a timeout occurs on the\ncontrol connection it will be closed. Use a larger value if this occurs too fre‐\nquently.\n\nZFTPIP\nReadonly. The IP address of the current connection in dot notation.\n\nZFTPHOST\nReadonly. The hostname of the current remote server. If the host was opened as an IP\nnumber, ZFTPHOST contains that instead; this saves the overhead for a name lookup, as\nIP numbers are most commonly used when a nameserver is unavailable.\n\nZFTPPORT\nReadonly. The number of the remote TCP port to which the connection is open (even if\nthe port was originally specified as a named service). Usually this is the standard\nFTP port, 21.\n\nIn the unlikely event that your system does not have the appropriate conversion func‐\ntions, this appears in network byte order. If your system is little-endian, the port\nthen consists of two swapped bytes and the standard port will be reported as 5376. In\nthat case, numeric ports passed to zftp open will also need to be in this format.\n\nZFTPSYSTEM\nReadonly. The system type string returned by the server in response to an FTP SYST\nrequest. The most interesting case is a string beginning \"UNIX Type: L8\", which en‐\nsures maximum compatibility with a local UNIX host.\n\nZFTPTYPE\nReadonly. The type to be used for data transfers , either `A' or `I'. Use the type\nsubcommand to change this.\n\nZFTPUSER\nReadonly. The username currently logged in, if any.\n\nZFTPACCOUNT\nReadonly. The account name of the current user, if any. Most servers do not require\nan account name.\n\nZFTPPWD\nReadonly. The current directory on the server.\n\nZFTPCODE\nReadonly. The three digit code of the last FTP reply from the server as a string.\nThis can still be read after the connection is closed, and is not changed when the\ncurrent session changes.\n\nZFTPREPLY\nReadonly. The last line of the last reply sent by the server. This can still be read\nafter the connection is closed, and is not changed when the current session changes.\n\nZFTPSESSION\nReadonly. The name of the current FTP session; see the description of the session\nsubcommand.\n\nZFTPPREFS\nA string of preferences for altering aspects of zftp's behaviour. Each preference is\na single character. The following are defined:\n\nP Passive: attempt to make the remote server initiate data transfers. This is\nslightly more efficient than sendport mode. If the letter S occurs later in\nthe string, zftp will use sendport mode if passive mode is not available.\n\nS Sendport: initiate transfers by the FTP PORT command. If this occurs before\nany P in the string, passive mode will never be attempted.\n\nD Dumb: use only the bare minimum of FTP commands. This prevents the variables\nZFTPSYSTEM and ZFTPPWD from being set, and will mean all connections default\nto ASCII type. It may prevent ZFTPSIZE from being set during a transfer if\nthe server does not send it anyway (many servers do).\n\nIf ZFTPPREFS is not set when zftp is loaded, it will be set to a default of `PS',\ni.e. use passive mode if available, otherwise fall back to sendport mode.\n\nZFTPVERBOSE\nA string of digits between 0 and 5 inclusive, specifying which responses from the\nserver should be printed. All responses go to standard error. If any of the numbers\n1 to 5 appear in the string, raw responses from the server with reply codes beginning\nwith that digit will be printed to standard error. The first digit of the three digit\nreply code is defined by RFC959 to correspond to:\n\n1. A positive preliminary reply.\n\n2. A positive completion reply.\n\n3. A positive intermediate reply.\n\n4. A transient negative completion reply.\n\n5. A permanent negative completion reply.\n\nIt should be noted that, for unknown reasons, the reply `Service not available', which\nforces termination of a connection, is classified as 421, i.e. `transient negative',\nan interesting interpretation of the word `transient'.\n\nThe code 0 is special: it indicates that all but the last line of multiline replies\nread from the server will be printed to standard error in a processed format. By con‐\nvention, servers use this mechanism for sending information for the user to read. The\nappropriate reply code, if it matches the same response, takes priority.\n\nIf ZFTPVERBOSE is not set when zftp is loaded, it will be set to the default value\n450, i.e., messages destined for the user and all errors will be printed. A null\nstring is valid and specifies that no messages should be printed.\n" }, { "name": "Functions", "content": "zftpchpwd\nIf this function is set by the user, it is called every time the directory changes on\nthe server, including when a user is logged in, or when a connection is closed. In\nthe last case, $ZFTPPWD will be unset; otherwise it will reflect the new directory.\n\nzftpprogress\nIf this function is set by the user, it will be called during a get, put or append op‐\neration each time sufficient data has been received from the host. During a get, the\ndata is sent to standard output, so it is vital that this function should write to\nstandard error or directly to the terminal, not to standard output.\n\nWhen it is called with a transfer in progress, the following additional shell parame‐\nters are set:\n\nZFTPFILE\nThe name of the remote file being transferred from or to.\n\nZFTPTRANSFER\nA G for a get operation and a P for a put operation.\n\nZFTPSIZE\nThe total size of the complete file being transferred: the same as the first\nvalue provided by the remote and local subcommands for a particular file. If\nthe server cannot supply this value for a remote file being retrieved, it will\nnot be set. If input is from a pipe the value may be incorrect and correspond\nsimply to a full pipe buffer.\n\nZFTPCOUNT\nThe amount of data so far transferred; a number between zero and $ZFTPSIZE, if\nthat is set. This number is always available.\n\nThe function is initially called with ZFTPTRANSFER set appropriately and ZFTPCOUNT\nset to zero. After the transfer is finished, the function will be called one more\ntime with ZFTPTRANSFER set to GF or PF, in case it wishes to tidy up. It is other‐\nwise never called twice with the same value of ZFTPCOUNT.\n\nSometimes the progress meter may cause disruption. It is up to the user to decide\nwhether the function should be defined and to use unfunction when necessary.\n" }, { "name": "Problems", "content": "A connection may not be opened in the left hand side of a pipe as this occurs in a subshell\nand the file information is not updated in the main shell. In the case of type or mode\nchanges or closing the connection in a subshell, the information is returned but variables\nare not updated until the next call to zftp. Other status changes in subshells will not be\nreflected by changes to the variables (but should be otherwise harmless).\n\nDeleting sessions while a zftp command is active in the background can have unexpected ef‐\nfects, even if it does not use the session being deleted. This is because all shell subpro‐\ncesses share information on the state of all connections, and deleting a session changes the\nordering of that information.\n\nOn some operating systems, the control connection is not valid after a fork(), so that opera‐\ntions in subshells, on the left hand side of a pipeline, or in the background are not possi‐\nble, as they should be. This is presumably a bug in the operating system.\n" } ] }, "THE ZSH/ZLE MODULE": { "content": "The zsh/zle module contains the Zsh Line Editor. See zshzle(1).\n", "subsections": [] }, "THE ZSH/ZLEPARAMETER MODULE": { "content": "The zsh/zleparameter module defines two special parameters that can be used to access inter‐\nnal information of the Zsh Line Editor (see zshzle(1)).\n", "subsections": [ { "name": "keymaps", "content": "This array contains the names of the keymaps currently defined.\n" }, { "name": "widgets", "content": "This associative array contains one entry per widget. The name of the widget is the\nkey and the value gives information about the widget. It is either\nthe string `builtin' for builtin widgets,\na string of the form `user:name' for user-defined widgets,\nwhere name is the name of the shell function implementing the widget,\na string of the form `completion:type:name'\nfor completion widgets,\nor a null value if the widget is not yet fully defined. In the penultimate case,\ntype is the name of the builtin widget the completion widget imitates in its behavior\nand name is the name of the shell function implementing the completion widget.\n" } ] }, "THE ZSH/ZPROF MODULE": { "content": "When loaded, the zsh/zprof causes shell functions to be profiled. The profiling results can\nbe obtained with the zprof builtin command made available by this module. There is no way to\nturn profiling off other than unloading the module.\n\nzprof [ -c ]\nWithout the -c option, zprof lists profiling results to standard output. The format\nis comparable to that of commands like gprof.\n\nAt the top there is a summary listing all functions that were called at least once.\nThis summary is sorted in decreasing order of the amount of time spent in each. The\nlines contain the number of the function in order, which is used in other parts of the\nlist in suffixes of the form `[num]', then the number of calls made to the function.\nThe next three columns list the time in milliseconds spent in the function and its de‐\nscendants, the average time in milliseconds spent in the function and its descendants\nper call and the percentage of time spent in all shell functions used in this function\nand its descendants. The following three columns give the same information, but\ncounting only the time spent in the function itself. The final column shows the name\nof the function.\n\nAfter the summary, detailed information about every function that was invoked is\nlisted, sorted in decreasing order of the amount of time spent in each function and\nits descendants. Each of these entries consists of descriptions for the functions\nthat called the function described, the function itself, and the functions that were\ncalled from it. The description for the function itself has the same format as in the\nsummary (and shows the same information). The other lines don't show the number of\nthe function at the beginning and have their function named indented to make it easier\nto distinguish the line showing the function described in the section from the sur‐\nrounding lines.\n\nThe information shown in this case is almost the same as in the summary, but only\nrefers to the call hierarchy being displayed. For example, for a calling function the\ncolumn showing the total running time lists the time spent in the described function\nand its descendants only for the times when it was called from that particular calling\nfunction. Likewise, for a called function, this columns lists the total time spent in\nthe called function and its descendants only for the times when it was called from the\nfunction described.\n\nAlso in this case, the column showing the number of calls to a function also shows a\nslash and then the total number of invocations made to the called function.\n\nAs long as the zsh/zprof module is loaded, profiling will be done and multiple invoca‐\ntions of the zprof builtin command will show the times and numbers of calls since the\nmodule was loaded. With the -c option, the zprof builtin command will reset its in‐\nternal counters and will not show the listing.\n", "subsections": [] }, "THE ZSH/ZPTY MODULE": { "content": "The zsh/zpty module offers one builtin:\n\nzpty [ -e ] [ -b ] name [ arg ... ]\nThe arguments following name are concatenated with spaces between, then executed as a\ncommand, as if passed to the eval builtin. The command runs under a newly assigned\npseudo-terminal; this is useful for running commands non-interactively which expect an\ninteractive environment. The name is not part of the command, but is used to refer to\nthis command in later calls to zpty.\n\nWith the -e option, the pseudo-terminal is set up so that input characters are echoed.\n\nWith the -b option, input to and output from the pseudo-terminal are made non-block‐\ning.\n\nThe shell parameter REPLY is set to the file descriptor assigned to the master side of\nthe pseudo-terminal. This allows the terminal to be monitored with ZLE descriptor\nhandlers (see zshzle(1)) or manipulated with sysread and syswrite (see THE ZSH/SYSTEM\nMODULE in zshmodules(1)). Warning: Use of sysread and syswrite is not recommended;\nuse zpty -r and zpty -w unless you know exactly what you are doing.\n\nzpty -d [ name ... ]\nThe second form, with the -d option, is used to delete commands previously started, by\nsupplying a list of their names. If no name is given, all commands are deleted.\nDeleting a command causes the HUP signal to be sent to the corresponding process.\n\nzpty -w [ -n ] name [ string ... ]\nThe -w option can be used to send the to command name the given strings as input (sep‐\narated by spaces). If the -n option is not given, a newline is added at the end.\n\nIf no string is provided, the standard input is copied to the pseudo-terminal; this\nmay stop before copying the full input if the pseudo-terminal is non-blocking. The\nexact input is always copied: the -n option is not applied.\n\nNote that the command under the pseudo-terminal sees this input as if it were typed,\nso beware when sending special tty driver characters such as word-erase, line-kill,\nand end-of-file.\n\nzpty -r [ -mt ] name [ param [ pattern ] ]\nThe -r option can be used to read the output of the command name. With only a name\nargument, the output read is copied to the standard output. Unless the pseudo-termi‐\nnal is non-blocking, copying continues until the command under the pseudo-terminal ex‐\nits; when non-blocking, only as much output as is immediately available is copied.\nThe return status is zero if any output is copied.\n\nWhen also given a param argument, at most one line is read and stored in the parameter\nnamed param. Less than a full line may be read if the pseudo-terminal is non-block‐\ning. The return status is zero if at least one character is stored in param.\n\nIf a pattern is given as well, output is read until the whole string read matches the\npattern, even in the non-blocking case. The return status is zero if the string read\nmatches the pattern, or if the command has exited but at least one character could\nstill be read. If the option -m is present, the return status is zero only if the\npattern matches. As of this writing, a maximum of one megabyte of output can be con‐\nsumed this way; if a full megabyte is read without matching the pattern, the return\nstatus is non-zero.\n\nIn all cases, the return status is non-zero if nothing could be read, and is 2 if this\nis because the command has finished.\n\nIf the -r option is combined with the -t option, zpty tests whether output is avail‐\nable before trying to read. If no output is available, zpty immediately returns the\nstatus 1. When used with a pattern, the behaviour on a failed poll is similar to when\nthe command has exited: the return value is zero if at least one character could\nstill be read even if the pattern failed to match.\n\nzpty -t name\nThe -t option without the -r option can be used to test whether the command name is\nstill running. It returns a zero status if the command is running and a non-zero\nvalue otherwise.\n\nzpty [ -L ]\nThe last form, without any arguments, is used to list the commands currently defined.\nIf the -L option is given, this is done in the form of calls to the zpty builtin.\n", "subsections": [] }, "THE ZSH/ZSELECT MODULE": { "content": "The zsh/zselect module makes available one builtin command:\n\nzselect [ -rwe ] [ -t timeout ] [ -a array ] [ -A assoc ] [ fd ... ]\nThe zselect builtin is a front-end to the `select' system call, which blocks until a\nfile descriptor is ready for reading or writing, or has an error condition, with an\noptional timeout. If this is not available on your system, the command prints an er‐\nror message and returns status 2 (normal errors return status 1). For more informa‐\ntion, see your systems documentation for select(3). Note there is no connection with\nthe shell builtin of the same name.\n\nArguments and options may be intermingled in any order. Non-option arguments are file\ndescriptors, which must be decimal integers. By default, file descriptors are to be\ntested for reading, i.e. zselect will return when data is available to be read from\nthe file descriptor, or more precisely, when a read operation from the file descriptor\nwill not block. After a -r, -w and -e, the given file descriptors are to be tested\nfor reading, writing, or error conditions. These options and an arbitrary list of\nfile descriptors may be given in any order.\n\n(The presence of an `error condition' is not well defined in the documentation for\nmany implementations of the select system call. According to recent versions of the\nPOSIX specification, it is really an exception condition, of which the only standard\nexample is out-of-band data received on a socket. So zsh users are unlikely to find\nthe -e option useful.)\n\nThe option `-t timeout' specifies a timeout in hundredths of a second. This may be\nzero, in which case the file descriptors will simply be polled and zselect will return\nimmediately. It is possible to call zselect with no file descriptors and a non-zero\ntimeout for use as a finer-grained replacement for `sleep'; note, however, the return\nstatus is always 1 for a timeout.\n\nThe option `-a array' indicates that array should be set to indicate the file descrip‐\ntor(s) which are ready. If the option is not given, the array reply will be used for\nthis purpose. The array will contain a string similar to the arguments for zselect.\nFor example,\n\nzselect -t 0 -r 0 -w 1\n\nmight return immediately with status 0 and $reply containing `-r 0 -w 1' to show that\nboth file descriptors are ready for the requested operations.\n\nThe option `-A assoc' indicates that the associative array assoc should be set to in‐\ndicate the file descriptor(s) which are ready. This option overrides the option -a,\nnor will reply be modified. The keys of assoc are the file descriptors, and the cor‐\nresponding values are any of the characters `rwe' to indicate the condition.\n\nThe command returns status 0 if some file descriptors are ready for reading. If the\noperation timed out, or a timeout of 0 was given and no file descriptors were ready,\nor there was an error, it returns status 1 and the array will not be set (nor modified\nin any way). If there was an error in the select operation the appropriate error mes‐\nsage is printed.\n", "subsections": [] }, "THE ZSH/ZUTIL MODULE": { "content": "The zsh/zutil module only adds some builtins:\n\nzstyle [ -L [ metapattern [ style ] ] ]\nzstyle [ -e | - | -- ] pattern style string ...\nzstyle -d [ pattern [ style ... ] ]\nzstyle -g name [ pattern [ style ] ]\nzstyle -{a|b|s} context style name [ sep ]\nzstyle -{T|t} context style [ string ... ]\nzstyle -m context style pattern\nThis builtin command is used to define and lookup styles. Styles are pairs of names\nand values, where the values consist of any number of strings. They are stored to‐\ngether with patterns and lookup is done by giving a string, called the `context',\nwhich is matched against the patterns. The definition stored for the most specific\npattern that matches will be returned.\n\nA pattern is considered to be more specific than another if it contains more compo‐\nnents (substrings separated by colons) or if the patterns for the components are more\nspecific, where simple strings are considered to be more specific than patterns and\ncomplex patterns are considered to be more specific than the pattern `*'. A `*' in\nthe pattern will match zero or more characters in the context; colons are not treated\nspecially in this regard. If two patterns are equally specific, the tie is broken in\nfavour of the pattern that was defined first.\n\nExample\n\nFor example, to define your preferred form of precipitation depending on which city\nyou're in, you might set the following in your zshrc:\n\nzstyle ':weather:europe:*' preferred-precipitation rain\nzstyle ':weather:europe:germany:* preferred-precipitation none\nzstyle ':weather:europe:germany:*:munich' preferred-precipitation snow\n\nThen, the fictional `weather' plugin might run under the hood a command such as\n\nzstyle -s \":weather:${continent}:${country}:${county}:${city}\" preferred-precipitation REPLY\n\nin order to retrieve your preference into the scalar variable $REPLY.\n\nUsage\n\nThe forms that operate on patterns are the following.\n\nzstyle [ -L [ metapattern [ style ] ] ]\nWithout arguments, lists style definitions. Styles are shown in alphabetic or‐\nder and patterns are shown in the order zstyle will test them.\n\nIf the -L option is given, listing is done in the form of calls to zstyle. The\noptional first argument, metapattern, is a pattern which will be matched\nagainst the string supplied as pattern when the style was defined. Note: this\nmeans, for example, `zstyle -L \":completion:*\"' will match any supplied pattern\nbeginning `:completion:', not just \":completion:*\": use ':completion:\\*' to\nmatch that. The optional second argument limits the output to a specific style\n(not a pattern). -L is not compatible with any other options.\n\nzstyle [ - | -- | -e ] pattern style string ...\nDefines the given style for the pattern with the strings as the value. If the\n-e option is given, the strings will be concatenated (separated by spaces) and\nthe resulting string will be evaluated (in the same way as it is done by the\neval builtin command) when the style is looked up. In this case the parameter\n`reply' must be assigned to set the strings returned after the evaluation. Be‐\nfore evaluating the value, reply is unset, and if it is still unset after the\nevaluation, the style is treated as if it were not set.\n\nzstyle -d [ pattern [ style ... ] ]\nDelete style definitions. Without arguments all definitions are deleted, with a\npattern all definitions for that pattern are deleted and if any styles are\ngiven, then only those styles are deleted for the pattern.\n\nzstyle -g name [ pattern [ style ] ]\nRetrieve a style definition. The name is used as the name of an array in which\nthe results are stored. Without any further arguments, all patterns defined are\nreturned. With a pattern the styles defined for that pattern are returned and\nwith both a pattern and a style, the value strings of that combination is re‐\nturned.\n\nThe other forms can be used to look up or test styles for a given context.\n\nzstyle -s context style name [ sep ]\nThe parameter name is set to the value of the style interpreted as a string.\nIf the value contains several strings they are concatenated with spaces (or\nwith the sep string if that is given) between them.\n\nReturn 0 if the style is set, 1 otherwise.\n\nzstyle -b context style name\nThe value is stored in name as a boolean, i.e. as the string `yes' if the value\nhas only one string and that string is equal to one of `yes', `true', `on', or\n`1'. If the value is any other string or has more than one string, the parame‐\nter is set to `no'.\n\nReturn 0 if name is set to `yes', 1 otherwise.\n\nzstyle -a context style name\nThe value is stored in name as an array. If name is declared as an associative\narray, the first, third, etc. strings are used as the keys and the other\nstrings are used as the values.\n\nReturn 0 if the style is set, 1 otherwise.\n\nzstyle -t context style [ string ... ]\nzstyle -T context style [ string ... ]\nTest the value of a style, i.e. the -t option only returns a status (sets $?).\nWithout any string the return status is zero if the style is defined for at\nleast one matching pattern, has only one string in its value, and that is equal\nto one of `true', `yes', `on' or `1'. If any strings are given the status is\nzero if and only if at least one of the strings is equal to at least one of the\nstrings in the value. If the style is defined but doesn't match, the return\nstatus is 1. If the style is not defined, the status is 2.\n\nThe -T option tests the values of the style like -t, but it returns status zero\n(rather than 2) if the style is not defined for any matching pattern.\n\nzstyle -m context style pattern\nMatch a value. Returns status zero if the pattern matches at least one of the\nstrings in the value.\n\nzformat -f param format spec ...\nzformat -a array sep spec ...\nThis builtin provides two different forms of formatting. The first form is selected\nwith the -f option. In this case the format string will be modified by replacing se‐\nquences starting with a percent sign in it with strings from the specs. Each spec\nshould be of the form `char:string' which will cause every appearance of the sequence\n`%char' in format to be replaced by the string. The `%' sequence may also contain op‐\ntional minimum and maximum field width specifications between the `%' and the `char'\nin the form `%min.maxc', i.e. the minimum field width is given first and if the maxi‐\nmum field width is used, it has to be preceded by a dot. Specifying a minimum field\nwidth makes the result be padded with spaces to the right if the string is shorter\nthan the requested width. Padding to the left can be achieved by giving a negative\nminimum field width. If a maximum field width is specified, the string will be trun‐\ncated after that many characters. After all `%' sequences for the given specs have\nbeen processed, the resulting string is stored in the parameter param.\n\nThe %-escapes also understand ternary expressions in the form used by prompts. The %\nis followed by a `(' and then an ordinary format specifier character as described\nabove. There may be a set of digits either before or after the `('; these specify a\ntest number, which defaults to zero. Negative numbers are also allowed. An arbitrary\ndelimiter character follows the format specifier, which is followed by a piece of\n`true' text, the delimiter character again, a piece of `false' text, and a closing\nparenthesis. The complete expression (without the digits) thus looks like\n`%(X.text1.text2)', except that the `.' character is arbitrary. The value given for\nthe format specifier in the char:string expressions is evaluated as a mathematical ex‐\npression, and compared with the test number. If they are the same, text1 is output,\nelse text2 is output. A parenthesis may be escaped in text2 as %). Either of text1\nor text2 may contain nested %-escapes.\n\nFor example:\n\nzformat -f REPLY \"The answer is '%3(c.yes.no)'.\" c:3\n\noutputs \"The answer is 'yes'.\" to REPLY since the value for the format specifier c is\n3, agreeing with the digit argument to the ternary expression.\n\nThe second form, using the -a option, can be used for aligning strings. Here, the\nspecs are of the form `left:right' where `left' and `right' are arbitrary strings.\nThese strings are modified by replacing the colons by the sep string and padding the\nleft strings with spaces to the right so that the sep strings in the result (and hence\nthe right strings after them) are all aligned if the strings are printed below each\nother. All strings without a colon are left unchanged and all strings with an empty\nright string have the trailing colon removed. In both cases the lengths of the\nstrings are not used to determine how the other strings are to be aligned. A colon in\nthe left string can be escaped with a backslash. The resulting strings are stored in\nthe array.\n", "subsections": [ { "name": "zregexparse", "content": "This implements some internals of the regexarguments function.\n\nzparseopts [ -D -E -F -K -M ] [ -a array ] [ -A assoc ] [ - ] spec ...\nThis builtin simplifies the parsing of options in positional parameters, i.e. the set\nof arguments given by $*. Each spec describes one option and must be of the form\n`opt[=array]'. If an option described by opt is found in the positional parameters it\nis copied into the array specified with the -a option; if the optional `=array' is\ngiven, it is instead copied into that array, which should be declared as a normal ar‐\nray and never as an associative array.\n\nNote that it is an error to give any spec without an `=array' unless one of the -a or\n-A options is used.\n\nUnless the -E option is given, parsing stops at the first string that isn't described\nby one of the specs. Even with -E, parsing always stops at a positional parameter\nequal to `-' or `--'. See also -F.\n\nThe opt description must be one of the following. Any of the special characters can\nappear in the option name provided it is preceded by a backslash.\n\nname\nname+ The name is the name of the option without the leading `-'. To specify a\nGNU-style long option, one of the usual two leading `-' must be included in\nname; for example, a `--file' option is represented by a name of `-file'.\n\nIf a `+' appears after name, the option is appended to array each time it is\nfound in the positional parameters; without the `+' only the last occurrence of\nthe option is preserved.\n\nIf one of these forms is used, the option takes no argument, so parsing stops\nif the next positional parameter does not also begin with `-' (unless the -E\noption is used).\n\nname:\nname:-\nname:: If one or two colons are given, the option takes an argument; with one colon,\nthe argument is mandatory and with two colons it is optional. The argument is\nappended to the array after the option itself.\n\nAn optional argument is put into the same array element as the option name\n(note that this makes empty strings as arguments indistinguishable). A manda‐\ntory argument is added as a separate element unless the `:-' form is used, in\nwhich case the argument is put into the same element.\n\nA `+' as described above may appear between the name and the first colon.\n\nIn all cases, option-arguments must appear either immediately following the option in\nthe same positional parameter or in the next one. Even an optional argument may appear\nin the next parameter, unless it begins with a `-'. There is no special handling of\n`=' as with GNU-style argument parsers; given the spec `-foo:', the positional parame‐\nter `--foo=bar' is parsed as `--foo' with an argument of `=bar'.\n\nWhen the names of two options that take no arguments overlap, the longest one wins, so\nthat parsing for the specs `-foo -foobar' (for example) is unambiguous. However, due\nto the aforementioned handling of option-arguments, ambiguities may arise when at\nleast one overlapping spec takes an argument, as in `-foo: -foobar'. In that case, the\nlast matching spec wins.\n\nThe options of zparseopts itself cannot be stacked because, for example, the stack\n`-DEK' is indistinguishable from a spec for the GNU-style long option `--DEK'. The\noptions of zparseopts itself are:\n\n-a array\nAs described above, this names the default array in which to store the recog‐\nnised options.\n\n-A assoc\nIf this is given, the options and their values are also put into an associative\narray with the option names as keys and the arguments (if any) as the values.\n\n-D If this option is given, all options found are removed from the positional pa‐\nrameters of the calling shell or shell function, up to but not including any\nnot described by the specs. If the first such parameter is `-' or `--', it is\nremoved as well. This is similar to using the shift builtin.\n\n-E This changes the parsing rules to not stop at the first string that isn't de‐\nscribed by one of the specs. It can be used to test for or (if used together\nwith -D) extract options and their arguments, ignoring all other options and\narguments that may be in the positional parameters. As indicated above, pars‐\ning still stops at the first `-' or `--' not described by a spec, but it is not\nremoved when used with -D.\n\n-F If this option is given, zparseopts immediately stops at the first option-like\nparameter not described by one of the specs, prints an error message, and re‐\nturns status 1. Removal (-D) and extraction (-E) are not performed, and option\narrays are not updated. This provides basic validation for the given options.\n\nNote that the appearance in the positional parameters of an option without its\nrequired argument always aborts parsing and returns an error as described above\nregardless of whether this option is used.\n\n-K With this option, the arrays specified with the -a option and with the `=array'\nforms are kept unchanged when none of the specs for them is used. Otherwise\nthe entire array is replaced when any of the specs is used. Individual ele‐\nments of associative arrays specified with the -A option are preserved by -K.\nThis allows assignment of default values to arrays before calling zparseopts.\n\n-M This changes the assignment rules to implement a map among equivalent option\nnames. If any spec uses the `=array' form, the string array is interpreted as\nthe name of another spec, which is used to choose where to store the values.\nIf no other spec is found, the values are stored as usual. This changes only\nthe way the values are stored, not the way $* is parsed, so results may be un‐\npredictable if the `name+' specifier is used inconsistently.\n\nFor example,\n\nset -- -a -bx -c y -cz baz -cend\nzparseopts a=foo b:=bar c+:=bar\n\nwill have the effect of\n\nfoo=(-a)\nbar=(-b x -c y -c z)\n\nThe arguments from `baz' on will not be used.\n\nAs an example for the -E option, consider:\n\nset -- -a x -b y -c z arg1 arg2\nzparseopts -E -D b:=bar\n\nwill have the effect of\n\nbar=(-b y)\nset -- -a x -c z arg1 arg2\n\nI.e., the option -b and its arguments are taken from the positional parameters and put\ninto the array bar.\n\nThe -M option can be used like this:\n\nset -- -a -bx -c y -cz baz -cend\nzparseopts -A bar -M a=foo b+: c:=b\n\nto have the effect of\n\nfoo=(-a)\nbar=(-a '' -b xyz)\n\n\n\nzsh 5.8.1 February 12, 2022 ZSHMODULES(1)" } ] } }, "summary": "zshmodules - zsh loadable modules", "flags": [], "examples": [], "see_also": [] }