3 @setfilename coreutils.info
4 @settitle @sc{gnu} Coreutils
9 @include constants.texi
11 @c Define new indices.
15 @c Put everything in one index (arbitrarily chosen to be the concept index).
25 * Coreutils: (coreutils). Core GNU (file, text, shell) utilities.
26 * Common options: (coreutils)Common options.
27 * File permissions: (coreutils)File permissions. Access modes.
28 * Date input formats: (coreutils)Date input formats.
31 @c FIXME: the following need documentation
32 @c * [: (coreutils)[ invocation. File/string tests.
33 @c * pinky: (coreutils)pinky invocation. FIXME.
35 @dircategory Individual utilities
37 * arch: (coreutils)arch invocation. Print machine hardware name.
38 * base64: (coreutils)base64 invocation. Base64 encode/decode data.
39 * basename: (coreutils)basename invocation. Strip directory and suffix.
40 * cat: (coreutils)cat invocation. Concatenate and write files.
41 * chcon: (coreutils)chcon invocation. Change SELinux CTX of files.
42 * chgrp: (coreutils)chgrp invocation. Change file groups.
43 * chmod: (coreutils)chmod invocation. Change access permissions.
44 * chown: (coreutils)chown invocation. Change file owners and groups.
45 * chroot: (coreutils)chroot invocation. Specify the root directory.
46 * cksum: (coreutils)cksum invocation. Print POSIX CRC checksum.
47 * comm: (coreutils)comm invocation. Compare sorted files by line.
48 * cp: (coreutils)cp invocation. Copy files.
49 * csplit: (coreutils)csplit invocation. Split by context.
50 * cut: (coreutils)cut invocation. Print selected parts of lines.
51 * date: (coreutils)date invocation. Print/set system date and time.
52 * dd: (coreutils)dd invocation. Copy and convert a file.
53 * df: (coreutils)df invocation. Report file system disk usage.
54 * dir: (coreutils)dir invocation. List directories briefly.
55 * dircolors: (coreutils)dircolors invocation. Color setup for ls.
56 * dirname: (coreutils)dirname invocation. Strip last file name component.
57 * du: (coreutils)du invocation. Report on disk usage.
58 * echo: (coreutils)echo invocation. Print a line of text.
59 * env: (coreutils)env invocation. Modify the environment.
60 * expand: (coreutils)expand invocation. Convert tabs to spaces.
61 * expr: (coreutils)expr invocation. Evaluate expressions.
62 * factor: (coreutils)factor invocation. Print prime factors
63 * false: (coreutils)false invocation. Do nothing, unsuccessfully.
64 * fmt: (coreutils)fmt invocation. Reformat paragraph text.
65 * fold: (coreutils)fold invocation. Wrap long input lines.
66 * groups: (coreutils)groups invocation. Print group names a user is in.
67 * head: (coreutils)head invocation. Output the first part of files.
68 * hostid: (coreutils)hostid invocation. Print numeric host identifier.
69 * hostname: (coreutils)hostname invocation. Print or set system name.
70 * id: (coreutils)id invocation. Print user identity.
71 * install: (coreutils)install invocation. Copy and change attributes.
72 * join: (coreutils)join invocation. Join lines on a common field.
73 * kill: (coreutils)kill invocation. Send a signal to processes.
74 * link: (coreutils)link invocation. Make hard links between files.
75 * ln: (coreutils)ln invocation. Make links between files.
76 * logname: (coreutils)logname invocation. Print current login name.
77 * ls: (coreutils)ls invocation. List directory contents.
78 * md5sum: (coreutils)md5sum invocation. Print or check MD5 digests.
79 * mkdir: (coreutils)mkdir invocation. Create directories.
80 * mkfifo: (coreutils)mkfifo invocation. Create FIFOs (named pipes).
81 * mknod: (coreutils)mknod invocation. Create special files.
82 * mktemp: (coreutils)mktemp invocation. Create temporary files.
83 * mv: (coreutils)mv invocation. Rename files.
84 * nice: (coreutils)nice invocation. Modify niceness.
85 * nl: (coreutils)nl invocation. Number lines and write files.
86 * nohup: (coreutils)nohup invocation. Immunize to hangups.
87 * nproc: (coreutils)nproc invocation. Print the number of processors.
88 * od: (coreutils)od invocation. Dump files in octal, etc.
89 * paste: (coreutils)paste invocation. Merge lines of files.
90 * pathchk: (coreutils)pathchk invocation. Check file name portability.
91 * pr: (coreutils)pr invocation. Paginate or columnate files.
92 * printenv: (coreutils)printenv invocation. Print environment variables.
93 * printf: (coreutils)printf invocation. Format and print data.
94 * ptx: (coreutils)ptx invocation. Produce permuted indexes.
95 * pwd: (coreutils)pwd invocation. Print working directory.
96 * readlink: (coreutils)readlink invocation. Print referent of a symlink.
97 * realpath: (coreutils)readpath invocation. Print resolved file names.
98 * rm: (coreutils)rm invocation. Remove files.
99 * rmdir: (coreutils)rmdir invocation. Remove empty directories.
100 * runcon: (coreutils)runcon invocation. Run in specified SELinux CTX.
101 * seq: (coreutils)seq invocation. Print numeric sequences
102 * sha1sum: (coreutils)sha1sum invocation. Print or check SHA-1 digests.
103 * sha2: (coreutils)sha2 utilities. Print or check SHA-2 digests.
104 * shred: (coreutils)shred invocation. Remove files more securely.
105 * shuf: (coreutils)shuf invocation. Shuffling text files.
106 * sleep: (coreutils)sleep invocation. Delay for a specified time.
107 * sort: (coreutils)sort invocation. Sort text files.
108 * split: (coreutils)split invocation. Split into pieces.
109 * stat: (coreutils)stat invocation. Report file(system) status.
110 * stdbuf: (coreutils)stdbuf invocation. Modify stdio buffering.
111 * stty: (coreutils)stty invocation. Print/change terminal settings.
112 * sum: (coreutils)sum invocation. Print traditional checksum.
113 * sync: (coreutils)sync invocation. Synchronize memory and disk.
114 * tac: (coreutils)tac invocation. Reverse files.
115 * tail: (coreutils)tail invocation. Output the last part of files.
116 * tee: (coreutils)tee invocation. Redirect to multiple files.
117 * test: (coreutils)test invocation. File/string tests.
118 * timeout: (coreutils)timeout invocation. Run with time limit.
119 * touch: (coreutils)touch invocation. Change file timestamps.
120 * tr: (coreutils)tr invocation. Translate characters.
121 * true: (coreutils)true invocation. Do nothing, successfully.
122 * truncate: (coreutils)truncate invocation. Shrink/extend size of a file.
123 * tsort: (coreutils)tsort invocation. Topological sort.
124 * tty: (coreutils)tty invocation. Print terminal name.
125 * uname: (coreutils)uname invocation. Print system information.
126 * unexpand: (coreutils)unexpand invocation. Convert spaces to tabs.
127 * uniq: (coreutils)uniq invocation. Uniquify files.
128 * unlink: (coreutils)unlink invocation. Removal via unlink(2).
129 * uptime: (coreutils)uptime invocation. Print uptime and load.
130 * users: (coreutils)users invocation. Print current user names.
131 * vdir: (coreutils)vdir invocation. List directories verbosely.
132 * wc: (coreutils)wc invocation. Line, word, and byte counts.
133 * who: (coreutils)who invocation. Print who is logged in.
134 * whoami: (coreutils)whoami invocation. Print effective user ID.
135 * yes: (coreutils)yes invocation. Print a string indefinitely.
139 This manual documents version @value{VERSION} of the @sc{gnu} core
140 utilities, including the standard programs for text and file manipulation.
142 Copyright @copyright{} 1994-2012 Free Software Foundation, Inc.
145 Permission is granted to copy, distribute and/or modify this document
146 under the terms of the GNU Free Documentation License, Version 1.3 or
147 any later version published by the Free Software Foundation; with no
148 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
149 Texts. A copy of the license is included in the section entitled ``GNU
150 Free Documentation License''.
155 @title @sc{gnu} @code{Coreutils}
156 @subtitle Core GNU utilities
157 @subtitle for version @value{VERSION}, @value{UPDATED}
158 @author David MacKenzie et al.
161 @vskip 0pt plus 1filll
174 @cindex core utilities
175 @cindex text utilities
176 @cindex shell utilities
177 @cindex file utilities
180 * Introduction:: Caveats, overview, and authors
181 * Common options:: Common options
182 * Output of entire files:: cat tac nl od base64
183 * Formatting file contents:: fmt pr fold
184 * Output of parts of files:: head tail split csplit
185 * Summarizing files:: wc sum cksum md5sum sha1sum sha2
186 * Operating on sorted files:: sort shuf uniq comm ptx tsort
187 * Operating on fields:: cut paste join
188 * Operating on characters:: tr expand unexpand
189 * Directory listing:: ls dir vdir dircolors
190 * Basic operations:: cp dd install mv rm shred
191 * Special file types:: mkdir rmdir unlink mkfifo mknod ln link readlink
192 * Changing file attributes:: chgrp chmod chown touch
193 * Disk usage:: df du stat sync truncate
194 * Printing text:: echo printf yes
195 * Conditions:: false true test expr
197 * File name manipulation:: dirname basename pathchk mktemp realpath
198 * Working context:: pwd stty printenv tty
199 * User information:: id logname whoami groups users who
200 * System context:: date arch nproc uname hostname hostid uptime
201 * SELinux context:: chcon runcon
202 * Modified command invocation:: chroot env nice nohup stdbuf timeout
203 * Process control:: kill
205 * Numeric operations:: factor seq
206 * File permissions:: Access modes
207 * Date input formats:: Specifying date strings
208 * Opening the software toolbox:: The software tools philosophy
209 * GNU Free Documentation License:: Copying and sharing this manual
210 * Concept index:: General index
213 --- The Detailed Node Listing ---
217 * Exit status:: Indicating program success or failure
218 * Backup options:: Backup options
219 * Block size:: Block size
220 * Floating point:: Floating point number representation
221 * Signal specifications:: Specifying signals
222 * Disambiguating names and IDs:: chgrp and chown owner and group syntax
223 * Random sources:: Sources of random data
224 * Target directory:: Target directory
225 * Trailing slashes:: Trailing slashes
226 * Traversing symlinks:: Traversing symlinks to directories
227 * Treating / specially:: Treating / specially
228 * Standards conformance:: Standards conformance
230 Output of entire files
232 * cat invocation:: Concatenate and write files
233 * tac invocation:: Concatenate and write files in reverse
234 * nl invocation:: Number lines and write files
235 * od invocation:: Write files in octal or other formats
236 * base64 invocation:: Transform data into printable data
238 Formatting file contents
240 * fmt invocation:: Reformat paragraph text
241 * pr invocation:: Paginate or columnate files for printing
242 * fold invocation:: Wrap input lines to fit in specified width
244 Output of parts of files
246 * head invocation:: Output the first part of files
247 * tail invocation:: Output the last part of files
248 * split invocation:: Split a file into fixed-size pieces
249 * csplit invocation:: Split a file into context-determined pieces
253 * wc invocation:: Print newline, word, and byte counts
254 * sum invocation:: Print checksum and block counts
255 * cksum invocation:: Print CRC checksum and byte counts
256 * md5sum invocation:: Print or check MD5 digests
257 * sha1sum invocation:: Print or check SHA-1 digests
258 * sha2 utilities:: Print or check SHA-2 digests
260 Operating on sorted files
262 * sort invocation:: Sort text files
263 * shuf invocation:: Shuffle text files
264 * uniq invocation:: Uniquify files
265 * comm invocation:: Compare two sorted files line by line
266 * ptx invocation:: Produce a permuted index of file contents
267 * tsort invocation:: Topological sort
269 @command{ptx}: Produce permuted indexes
271 * General options in ptx:: Options which affect general program behavior
272 * Charset selection in ptx:: Underlying character set considerations
273 * Input processing in ptx:: Input fields, contexts, and keyword selection
274 * Output formatting in ptx:: Types of output format, and sizing the fields
275 * Compatibility in ptx:: The GNU extensions to @command{ptx}
279 * cut invocation:: Print selected parts of lines
280 * paste invocation:: Merge lines of files
281 * join invocation:: Join lines on a common field
283 Operating on characters
285 * tr invocation:: Translate, squeeze, and/or delete characters
286 * expand invocation:: Convert tabs to spaces
287 * unexpand invocation:: Convert spaces to tabs
289 @command{tr}: Translate, squeeze, and/or delete characters
291 * Character sets:: Specifying sets of characters
292 * Translating:: Changing one set of characters to another
293 * Squeezing:: Squeezing repeats and deleting
297 * ls invocation:: List directory contents
298 * dir invocation:: Briefly list directory contents
299 * vdir invocation:: Verbosely list directory contents
300 * dircolors invocation:: Color setup for @command{ls}
302 @command{ls}: List directory contents
304 * Which files are listed:: Which files are listed
305 * What information is listed:: What information is listed
306 * Sorting the output:: Sorting the output
307 * Details about version sort:: More details about version sort
308 * General output formatting:: General output formatting
309 * Formatting the file names:: Formatting the file names
313 * cp invocation:: Copy files and directories
314 * dd invocation:: Convert and copy a file
315 * install invocation:: Copy files and set attributes
316 * mv invocation:: Move (rename) files
317 * rm invocation:: Remove files or directories
318 * shred invocation:: Remove files more securely
322 * link invocation:: Make a hard link via the link syscall
323 * ln invocation:: Make links between files
324 * mkdir invocation:: Make directories
325 * mkfifo invocation:: Make FIFOs (named pipes)
326 * mknod invocation:: Make block or character special files
327 * readlink invocation:: Print value of a symlink or canonical file name
328 * rmdir invocation:: Remove empty directories
329 * unlink invocation:: Remove files via unlink syscall
331 Changing file attributes
333 * chown invocation:: Change file owner and group
334 * chgrp invocation:: Change group ownership
335 * chmod invocation:: Change access permissions
336 * touch invocation:: Change file timestamps
340 * df invocation:: Report file system disk space usage
341 * du invocation:: Estimate file space usage
342 * stat invocation:: Report file or file system status
343 * sync invocation:: Synchronize data on disk with memory
344 * truncate invocation:: Shrink or extend the size of a file
348 * echo invocation:: Print a line of text
349 * printf invocation:: Format and print data
350 * yes invocation:: Print a string until interrupted
354 * false invocation:: Do nothing, unsuccessfully
355 * true invocation:: Do nothing, successfully
356 * test invocation:: Check file types and compare values
357 * expr invocation:: Evaluate expressions
359 @command{test}: Check file types and compare values
361 * File type tests:: File type tests
362 * Access permission tests:: Access permission tests
363 * File characteristic tests:: File characteristic tests
364 * String tests:: String tests
365 * Numeric tests:: Numeric tests
367 @command{expr}: Evaluate expression
369 * String expressions:: + : match substr index length
370 * Numeric expressions:: + - * / %
371 * Relations for expr:: | & < <= = == != >= >
372 * Examples of expr:: Examples of using @command{expr}
376 * tee invocation:: Redirect output to multiple files or processes
378 File name manipulation
380 * basename invocation:: Strip directory and suffix from a file name
381 * dirname invocation:: Strip last file name component
382 * pathchk invocation:: Check file name validity and portability
383 * mktemp invocation:: Create temporary file or directory
384 * realpath invocation:: Print resolved file names
388 * pwd invocation:: Print working directory
389 * stty invocation:: Print or change terminal characteristics
390 * printenv invocation:: Print all or some environment variables
391 * tty invocation:: Print file name of terminal on standard input
393 @command{stty}: Print or change terminal characteristics
395 * Control:: Control settings
396 * Input:: Input settings
397 * Output:: Output settings
398 * Local:: Local settings
399 * Combination:: Combination settings
400 * Characters:: Special characters
401 * Special:: Special settings
405 * id invocation:: Print user identity
406 * logname invocation:: Print current login name
407 * whoami invocation:: Print effective user ID
408 * groups invocation:: Print group names a user is in
409 * users invocation:: Print login names of users currently logged in
410 * who invocation:: Print who is currently logged in
414 * arch invocation:: Print machine hardware name
415 * date invocation:: Print or set system date and time
416 * nproc invocation:: Print the number of processors
417 * uname invocation:: Print system information
418 * hostname invocation:: Print or set system name
419 * hostid invocation:: Print numeric host identifier
420 * uptime invocation:: Print system uptime and load
422 @command{date}: Print or set system date and time
424 * Time conversion specifiers:: %[HIklMNpPrRsSTXzZ]
425 * Date conversion specifiers:: %[aAbBcCdDeFgGhjmuUVwWxyY]
426 * Literal conversion specifiers:: %[%nt]
427 * Padding and other flags:: Pad with zeros, spaces, etc.
428 * Setting the time:: Changing the system clock
429 * Options for date:: Instead of the current time
430 * Date input formats:: Specifying date strings
431 * Examples of date:: Examples
435 * chcon invocation:: Change SELinux context of file
436 * runcon invocation:: Run a command in specified SELinux context
438 Modified command invocation
440 * chroot invocation:: Run a command with a different root directory
441 * env invocation:: Run a command in a modified environment
442 * nice invocation:: Run a command with modified niceness
443 * nohup invocation:: Run a command immune to hangups
444 * stdbuf invocation:: Run a command with modified I/O buffering
445 * timeout invocation:: Run a command with a time limit
449 * kill invocation:: Sending a signal to processes.
453 * sleep invocation:: Delay for a specified time
457 * factor invocation:: Print prime factors
458 * seq invocation:: Print numeric sequences
462 * Mode Structure:: Structure of file mode bits
463 * Symbolic Modes:: Mnemonic representation of file mode bits
464 * Numeric Modes:: File mode bits as octal numbers
465 * Directory Setuid and Setgid:: Set-user-ID and set-group-ID on directories
469 * General date syntax:: Common rules
470 * Calendar date items:: 19 Dec 1994
471 * Time of day items:: 9:20pm
472 * Time zone items:: @sc{est}, @sc{pdt}, @sc{gmt}
473 * Day of week items:: Monday and others
474 * Relative items in date strings:: next tuesday, 2 years ago
475 * Pure numbers in date strings:: 19931219, 1440
476 * Seconds since the Epoch:: @@1078100502
477 * Specifying time zone rules:: TZ="America/New_York", TZ="UTC0"
478 * Authors of parse_datetime:: Bellovin, Eggert, Salz, Berets, et al
480 Opening the software toolbox
482 * Toolbox introduction:: Toolbox introduction
483 * I/O redirection:: I/O redirection
484 * The who command:: The @command{who} command
485 * The cut command:: The @command{cut} command
486 * The sort command:: The @command{sort} command
487 * The uniq command:: The @command{uniq} command
488 * Putting the tools together:: Putting the tools together
492 * GNU Free Documentation License:: Copying and sharing this manual
499 @chapter Introduction
501 This manual is a work in progress: many sections make no attempt to explain
502 basic concepts in a way suitable for novices. Thus, if you are interested,
503 please get involved in improving this manual. The entire @sc{gnu} community
507 The @sc{gnu} utilities documented here are mostly compatible with the
509 @cindex bugs, reporting
510 Please report bugs to @email{bug-coreutils@@gnu.org}. Remember
511 to include the version number, machine architecture, input files, and
512 any other information needed to reproduce the bug: your input, what you
513 expected, what you got, and why it is wrong. Diffs are welcome, but
514 please include a description of the problem as well, since this is
515 sometimes difficult to infer. @xref{Bugs, , , gcc, Using and Porting GNU CC}.
521 @cindex MacKenzie, D.
524 This manual was originally derived from the Unix man pages in the
525 distributions, which were written by David MacKenzie and updated by Jim
526 Meyering. What you are reading now is the authoritative documentation
527 for these utilities; the man pages are no longer being maintained. The
528 original @command{fmt} man page was written by Ross Paterson. Fran@,{c}ois
529 Pinard did the initial conversion to Texinfo format. Karl Berry did the
530 indexing, some reorganization, and editing of the results. Brian
531 Youmans of the Free Software Foundation office staff combined the
532 manuals for textutils, fileutils, and sh-utils to produce the present
533 omnibus manual. Richard Stallman contributed his usual invaluable
534 insights to the overall process.
537 @chapter Common options
541 @itemx @w{@kbd{--backup}[=@var{method}]}
544 @vindex VERSION_CONTROL
545 @cindex backups, making
546 @xref{Backup options}.
547 Make a backup of each file that would otherwise be overwritten or removed.
550 @macro optBackupSuffix
551 @item -S @var{suffix}
552 @itemx --suffix=@var{suffix}
555 Append @var{suffix} to each backup file made with @option{-b}.
556 @xref{Backup options}.
559 @macro optTargetDirectory
560 @item -t @var{directory}
561 @itemx @w{@kbd{--target-directory}=@var{directory}}
563 @opindex --target-directory
564 @cindex target directory
565 @cindex destination directory
566 Specify the destination @var{directory}.
567 @xref{Target directory}.
570 @macro optNoTargetDirectory
572 @itemx --no-target-directory
574 @opindex --no-target-directory
575 @cindex target directory
576 @cindex destination directory
577 Do not treat the last operand specially when it is a directory or a
578 symbolic link to a directory. @xref{Target directory}.
586 @cindex output @sc{nul}-byte-terminated lines
587 Output a zero byte (ASCII @sc{nul}) at the end of each line,
588 rather than a newline. This option enables other programs to parse the
589 output of @command{\cmd\} even when that output would contain data
590 with embedded newlines.
597 Append an SI-style abbreviation to each size, such as @samp{M} for
598 megabytes. Powers of 1000 are used, not 1024; @samp{M} stands for
599 1,000,000 bytes. This option is equivalent to
600 @option{--block-size=si}. Use the @option{-h} or
601 @option{--human-readable} option if
602 you prefer powers of 1024.
605 @macro optHumanReadable
607 @itemx --human-readable
609 @opindex --human-readable
610 @cindex human-readable output
611 Append a size letter to each size, such as @samp{M} for mebibytes.
612 Powers of 1024 are used, not 1000; @samp{M} stands for 1,048,576 bytes.
613 This option is equivalent to @option{--block-size=human-readable}.
614 Use the @option{--si} option if you prefer powers of 1000.
617 @macro optStripTrailingSlashes
618 @item @w{@kbd{--strip-trailing-slashes}}
619 @opindex --strip-trailing-slashes
620 @cindex stripping trailing slashes
621 Remove any trailing slashes from each @var{source} argument.
622 @xref{Trailing slashes}.
625 @macro mayConflictWithShellBuiltIn{cmd}
626 @cindex conflicts with shell built-ins
627 @cindex built-in shell commands, conflicts with
628 Due to shell aliases and built-in @command{\cmd\} functions, using an
629 unadorned @command{\cmd\} interactively or in a script may get you
630 different functionality than that described here. Invoke it via
631 @command{env} (i.e., @code{env \cmd\ @dots{}}) to avoid interference
636 @macro multiplierSuffixes{varName}
637 @var{\varName\} may be, or may be an integer optionally followed by,
638 one of the following multiplicative suffixes:
640 @samp{b} => 512 ("blocks")
641 @samp{KB} => 1000 (KiloBytes)
642 @samp{K} => 1024 (KibiBytes)
643 @samp{MB} => 1000*1000 (MegaBytes)
644 @samp{M} => 1024*1024 (MebiBytes)
645 @samp{GB} => 1000*1000*1000 (GigaBytes)
646 @samp{G} => 1024*1024*1024 (GibiBytes)
648 and so on for @samp{T}, @samp{P}, @samp{E}, @samp{Z}, and @samp{Y}.
651 @c FIXME: same as above, but no ``blocks'' line.
652 @macro multiplierSuffixesNoBlocks{varName}
653 @var{\varName\} may be, or may be an integer optionally followed by,
654 one of the following multiplicative suffixes:
656 @samp{KB} => 1000 (KiloBytes)
657 @samp{K} => 1024 (KibiBytes)
658 @samp{MB} => 1000*1000 (MegaBytes)
659 @samp{M} => 1024*1024 (MebiBytes)
660 @samp{GB} => 1000*1000*1000 (GigaBytes)
661 @samp{G} => 1024*1024*1024 (GibiBytes)
663 and so on for @samp{T}, @samp{P}, @samp{E}, @samp{Z}, and @samp{Y}.
666 @cindex common options
668 Certain options are available in all of these programs. Rather than
669 writing identical descriptions for each of the programs, they are
670 described here. (In fact, every @sc{gnu} program accepts (or should accept)
673 @vindex POSIXLY_CORRECT
674 Normally options and operands can appear in any order, and programs act
675 as if all the options appear before any operands. For example,
676 @samp{sort -r passwd -t :} acts like @samp{sort -r -t : passwd}, since
677 @samp{:} is an option-argument of @option{-t}. However, if the
678 @env{POSIXLY_CORRECT} environment variable is set, options must appear
679 before operands, unless otherwise specified for a particular command.
681 A few programs can usefully have trailing operands with leading
682 @samp{-}. With such a program, options must precede operands even if
683 @env{POSIXLY_CORRECT} is not set, and this fact is noted in the
684 program description. For example, the @command{env} command's options
685 must appear before its operands, since in some cases the operands
686 specify a command that itself contains options.
688 Most programs that accept long options recognize unambiguous
689 abbreviations of those options. For example, @samp{rmdir
690 --ignore-fail-on-non-empty} can be invoked as @samp{rmdir
691 --ignore-fail} or even @samp{rmdir --i}. Ambiguous options, such as
692 @samp{ls --h}, are identified as such.
694 Some of these programs recognize the @option{--help} and @option{--version}
695 options only when one of them is the sole command line argument. For
696 these programs, abbreviations of the long options are not always recognized.
703 Print a usage message listing all available options, then exit successfully.
707 @cindex version number, finding
708 Print the version number, then exit successfully.
712 @cindex option delimiter
713 Delimit the option list. Later arguments, if any, are treated as
714 operands even if they begin with @samp{-}. For example, @samp{sort --
715 -r} reads from the file named @file{-r}.
719 @cindex standard input
720 @cindex standard output
721 A single @samp{-} operand is not really an option, though it looks like one. It
722 stands for standard input, or for standard output if that is clear from
723 the context. For example, @samp{sort -} reads from standard input,
724 and is equivalent to plain @samp{sort}, and @samp{tee -} writes an
725 extra copy of its input to standard output. Unless otherwise
726 specified, @samp{-} can appear as any operand that requires a file
730 * Exit status:: Indicating program success or failure.
731 * Backup options:: -b -S, in some programs.
732 * Block size:: BLOCK_SIZE and --block-size, in some programs.
733 * Floating point:: Floating point number representation.
734 * Signal specifications:: Specifying signals using the --signal option.
735 * Disambiguating names and IDs:: chgrp and chown owner and group syntax
736 * Random sources:: --random-source, in some programs.
737 * Target directory:: Specifying a target directory, in some programs.
738 * Trailing slashes:: --strip-trailing-slashes, in some programs.
739 * Traversing symlinks:: -H, -L, or -P, in some programs.
740 * Treating / specially:: --preserve-root and --no-preserve-root.
741 * Special built-in utilities:: @command{break}, @command{:}, @dots{}
742 * Standards conformance:: Conformance to the POSIX standard.
750 An exit status of zero indicates success,
751 and a nonzero value indicates failure.
754 Nearly every command invocation yields an integral @dfn{exit status}
755 that can be used to change how other commands work.
756 For the vast majority of commands, an exit status of zero indicates
757 success. Failure is indicated by a nonzero value---typically
758 @samp{1}, though it may differ on unusual platforms as POSIX
759 requires only that it be nonzero.
761 However, some of the programs documented here do produce
762 other exit status values and a few associate different
763 meanings with the values @samp{0} and @samp{1}.
764 Here are some of the exceptions:
765 @command{chroot}, @command{env}, @command{expr}, @command{nice},
766 @command{nohup}, @command{printenv}, @command{sort}, @command{stdbuf},
767 @command{test}, @command{timeout}, @command{tty}.
771 @section Backup options
773 @cindex backup options
775 Some @sc{gnu} programs (at least @command{cp}, @command{install},
776 @command{ln}, and @command{mv}) optionally make backups of files
777 before writing new versions.
778 These options control the details of these backups. The options are also
779 briefly mentioned in the descriptions of the particular programs.
784 @itemx @w{@kbd{--backup}[=@var{method}]}
787 @vindex VERSION_CONTROL
788 @cindex backups, making
789 Make a backup of each file that would otherwise be overwritten or removed.
790 Without this option, the original versions are destroyed.
791 Use @var{method} to determine the type of backups to make.
792 When this option is used but @var{method} is not specified,
793 then the value of the @env{VERSION_CONTROL}
794 environment variable is used. And if @env{VERSION_CONTROL} is not set,
795 the default backup type is @samp{existing}.
797 Note that the short form of this option, @option{-b} does not accept any
798 argument. Using @option{-b} is equivalent to using @option{--backup=existing}.
800 @vindex version-control @r{Emacs variable}
801 This option corresponds to the Emacs variable @samp{version-control};
802 the values for @var{method} are the same as those used in Emacs.
803 This option also accepts more descriptive names.
804 The valid @var{method}s are (unique abbreviations are accepted):
809 @opindex none @r{backup method}
814 @opindex numbered @r{backup method}
815 Always make numbered backups.
819 @opindex existing @r{backup method}
820 Make numbered backups of files that already have them, simple backups
825 @opindex simple @r{backup method}
826 Always make simple backups. Please note @samp{never} is not to be
827 confused with @samp{none}.
831 @item -S @var{suffix}
832 @itemx --suffix=@var{suffix}
835 @cindex backup suffix
836 @vindex SIMPLE_BACKUP_SUFFIX
837 Append @var{suffix} to each backup file made with @option{-b}. If this
838 option is not specified, the value of the @env{SIMPLE_BACKUP_SUFFIX}
839 environment variable is used. And if @env{SIMPLE_BACKUP_SUFFIX} is not
840 set, the default is @samp{~}, just as in Emacs.
849 Some @sc{gnu} programs (at least @command{df}, @command{du}, and
850 @command{ls}) display sizes in ``blocks''. You can adjust the block size
851 and method of display to make sizes easier to read. The block size
852 used for display is independent of any file system block size.
853 Fractional block counts are rounded up to the nearest integer.
855 @opindex --block-size=@var{size}
858 @vindex DF_BLOCK_SIZE
859 @vindex DU_BLOCK_SIZE
860 @vindex LS_BLOCK_SIZE
861 @vindex POSIXLY_CORRECT@r{, and block size}
863 The default block size is chosen by examining the following environment
864 variables in turn; the first one that is set determines the block size.
869 This specifies the default block size for the @command{df} command.
870 Similarly, @env{DU_BLOCK_SIZE} specifies the default for @command{du} and
871 @env{LS_BLOCK_SIZE} for @command{ls}.
874 This specifies the default block size for all three commands, if the
875 above command-specific environment variables are not set.
878 This specifies the default block size for all values that are normally
879 printed as blocks, if neither @env{BLOCK_SIZE} nor the above
880 command-specific environment variables are set. Unlike the other
881 environment variables, @env{BLOCKSIZE} does not affect values that are
882 normally printed as byte counts, e.g., the file sizes contained in
885 @item POSIXLY_CORRECT
886 If neither @env{@var{command}_BLOCK_SIZE}, nor @env{BLOCK_SIZE}, nor
887 @env{BLOCKSIZE} is set, but this variable is set, the block size
892 If none of the above environment variables are set, the block size
893 currently defaults to 1024 bytes in most contexts, but this number may
894 change in the future. For @command{ls} file sizes, the block size
897 @cindex human-readable output
900 A block size specification can be a positive integer specifying the number
901 of bytes per block, or it can be @code{human-readable} or @code{si} to
902 select a human-readable format. Integers may be followed by suffixes
903 that are upward compatible with the
904 @uref{http://www.bipm.org/en/si/si_brochure/chapter3/prefixes.html, SI prefixes}
905 for decimal multiples and with the
906 @uref{http://physics.nist.gov/cuu/Units/binary.html, ISO/IEC 80000-13
907 (formerly IEC 60027-2) prefixes} for binary multiples.
909 With human-readable formats, output sizes are followed by a size letter
910 such as @samp{M} for megabytes. @code{BLOCK_SIZE=human-readable} uses
911 powers of 1024; @samp{M} stands for 1,048,576 bytes.
912 @code{BLOCK_SIZE=si} is similar, but uses powers of 1000 and appends
913 @samp{B}; @samp{MB} stands for 1,000,000 bytes.
916 A block size specification preceded by @samp{'} causes output sizes to
917 be displayed with thousands separators. The @env{LC_NUMERIC} locale
918 specifies the thousands separator and grouping. For example, in an
919 American English locale, @samp{--block-size="'1kB"} would cause a size
920 of 1234000 bytes to be displayed as @samp{1,234}. In the default C
921 locale, there is no thousands separator so a leading @samp{'} has no
924 An integer block size can be followed by a suffix to specify a
925 multiple of that size. A bare size letter,
926 or one followed by @samp{iB}, specifies
927 a multiple using powers of 1024. A size letter followed by @samp{B}
928 specifies powers of 1000 instead. For example, @samp{1M} and
929 @samp{1MiB} are equivalent to @samp{1048576}, whereas @samp{1MB} is
930 equivalent to @samp{1000000}.
932 A plain suffix without a preceding integer acts as if @samp{1} were
933 prepended, except that it causes a size indication to be appended to
934 the output. For example, @samp{--block-size="kB"} displays 3000 as
937 The following suffixes are defined. Large sizes like @code{1Y}
938 may be rejected by your computer due to limitations of its arithmetic.
942 @cindex kilobyte, definition of
943 kilobyte: @math{10^3 = 1000}.
947 @cindex kibibyte, definition of
948 kibibyte: @math{2^{10} = 1024}. @samp{K} is special: the SI prefix is
949 @samp{k} and the ISO/IEC 80000-13 prefix is @samp{Ki}, but tradition and
950 POSIX use @samp{k} to mean @samp{KiB}.
952 @cindex megabyte, definition of
953 megabyte: @math{10^6 = 1,000,000}.
956 @cindex mebibyte, definition of
957 mebibyte: @math{2^{20} = 1,048,576}.
959 @cindex gigabyte, definition of
960 gigabyte: @math{10^9 = 1,000,000,000}.
963 @cindex gibibyte, definition of
964 gibibyte: @math{2^{30} = 1,073,741,824}.
966 @cindex terabyte, definition of
967 terabyte: @math{10^{12} = 1,000,000,000,000}.
970 @cindex tebibyte, definition of
971 tebibyte: @math{2^{40} = 1,099,511,627,776}.
973 @cindex petabyte, definition of
974 petabyte: @math{10^{15} = 1,000,000,000,000,000}.
977 @cindex pebibyte, definition of
978 pebibyte: @math{2^{50} = 1,125,899,906,842,624}.
980 @cindex exabyte, definition of
981 exabyte: @math{10^{18} = 1,000,000,000,000,000,000}.
984 @cindex exbibyte, definition of
985 exbibyte: @math{2^{60} = 1,152,921,504,606,846,976}.
987 @cindex zettabyte, definition of
988 zettabyte: @math{10^{21} = 1,000,000,000,000,000,000,000}
991 @math{2^{70} = 1,180,591,620,717,411,303,424}.
993 @cindex yottabyte, definition of
994 yottabyte: @math{10^{24} = 1,000,000,000,000,000,000,000,000}.
997 @math{2^{80} = 1,208,925,819,614,629,174,706,176}.
1002 @opindex --block-size
1003 @opindex --human-readable
1006 Block size defaults can be overridden by an explicit
1007 @option{--block-size=@var{size}} option. The @option{-k}
1008 option is equivalent to @option{--block-size=1K}, which
1009 is the default unless the @env{POSIXLY_CORRECT} environment variable is
1010 set. The @option{-h} or @option{--human-readable} option is equivalent to
1011 @option{--block-size=human-readable}. The @option{--si} option is
1012 equivalent to @option{--block-size=si}.
1014 @node Floating point
1015 @section Floating point numbers
1016 @cindex floating point
1017 @cindex IEEE floating point
1019 Commands that accept or produce floating point numbers employ the
1020 floating point representation of the underlying system, and suffer
1021 from rounding error, overflow, and similar floating-point issues.
1022 Almost all modern systems use IEEE-754 floating point, and it is
1023 typically portable to assume IEEE-754 behavior these days. IEEE-754
1024 has positive and negative infinity, distinguishes positive from
1025 negative zero, and uses special values called NaNs to represent
1026 invalid computations such as dividing zero by itself. For more
1027 information, please see David Goldberg's paper
1028 @uref{http://@/www.validlab.com/@/goldberg/@/paper.pdf, What Every
1029 Computer Scientist Should Know About Floating-Point Arithmetic}.
1032 Commands that accept floating point numbers as options, operands or
1033 input use the standard C functions @code{strtod} and @code{strtold} to
1034 convert from text to floating point numbers. These floating point
1035 numbers therefore can use scientific notation like @code{1.0e-34} and
1036 @code{-10e100}. Modern C implementations also accept hexadecimal
1037 floating point numbers such as @code{-0x.ep-3}, which stands for
1038 @minus{}14/16 times @math{2^-3}, which equals @minus{}0.109375. The
1039 @env{LC_NUMERIC} locale determines the decimal-point character.
1040 @xref{Parsing of Floats,,, libc, The GNU C Library Reference Manual}.
1042 @node Signal specifications
1043 @section Signal specifications
1044 @cindex signals, specifying
1046 A @var{signal} may be a signal name like @samp{HUP}, or a signal
1047 number like @samp{1}, or an exit status of a process terminated by the
1048 signal. A signal name can be given in canonical form or prefixed by
1049 @samp{SIG}@. The case of the letters is ignored. The following signal names
1050 and numbers are supported on all POSIX compliant systems:
1056 2. Terminal interrupt.
1062 9. Kill (cannot be caught or ignored).
1070 Other supported signal names have system-dependent corresponding
1071 numbers. All systems conforming to POSIX 1003.1-2001 also
1072 support the following signals:
1076 Access to an undefined portion of a memory object.
1078 Child process terminated, stopped, or continued.
1080 Continue executing, if stopped.
1082 Erroneous arithmetic operation.
1084 Illegal Instruction.
1086 Write on a pipe with no one to read it.
1088 Invalid memory reference.
1090 Stop executing (cannot be caught or ignored).
1094 Background process attempting read.
1096 Background process attempting write.
1098 High bandwidth data is available at a socket.
1100 User-defined signal 1.
1102 User-defined signal 2.
1106 POSIX 1003.1-2001 systems that support the XSI extension
1107 also support the following signals:
1113 Profiling timer expired.
1117 Trace/breakpoint trap.
1119 Virtual timer expired.
1121 CPU time limit exceeded.
1123 File size limit exceeded.
1127 POSIX 1003.1-2001 systems that support the XRT extension
1128 also support at least eight real-time signals called @samp{RTMIN},
1129 @samp{RTMIN+1}, @dots{}, @samp{RTMAX-1}, @samp{RTMAX}.
1131 @node Disambiguating names and IDs
1132 @section chown and chgrp: Disambiguating user names and IDs
1133 @cindex user names, disambiguating
1134 @cindex user IDs, disambiguating
1135 @cindex group names, disambiguating
1136 @cindex group IDs, disambiguating
1137 @cindex disambiguating group names and IDs
1139 Since the @var{owner} and @var{group} arguments to @command{chown} and
1140 @command{chgrp} may be specified as names or numeric IDs, there is an
1142 What if a user or group @emph{name} is a string of digits?
1143 @footnote{Using a number as a user name is common in some environments.}
1144 Should the command interpret it as a user name or as an ID@?
1145 POSIX requires that @command{chown} and @command{chgrp}
1146 first attempt to resolve the specified string as a name, and
1147 only once that fails, then try to interpret it as an ID@.
1148 This is troublesome when you want to specify a numeric ID, say 42,
1149 and it must work even in a pathological situation where
1150 @samp{42} is a user name that maps to some other user ID, say 1000.
1151 Simply invoking @code{chown 42 F}, will set @file{F}s owner ID to
1152 1000---not what you intended.
1154 GNU @command{chown} and @command{chgrp} provide a way to work around this,
1155 that at the same time may result in a significant performance improvement
1156 by eliminating a database look-up.
1157 Simply precede each numeric user ID and/or group ID with a @samp{+},
1158 in order to force its interpretation as an integer:
1162 chgrp +$numeric_group_id another-file
1166 GNU @command{chown} and @command{chgrp}
1167 skip the name look-up process for each @samp{+}-prefixed string,
1168 because a string containing @samp{+} is never a valid user or group name.
1169 This syntax is accepted on most common Unix systems, but not on Solaris 10.
1171 @node Random sources
1172 @section Sources of random data
1174 @cindex random sources
1176 The @command{shuf}, @command{shred}, and @command{sort} commands
1177 sometimes need random data to do their work. For example, @samp{sort
1178 -R} must choose a hash function at random, and it needs random data to
1179 make this selection.
1181 By default these commands use an internal pseudorandom generator
1182 initialized by a small amount of entropy, but can be directed to use
1183 an external source with the @option{--random-source=@var{file}} option.
1184 An error is reported if @var{file} does not contain enough bytes.
1186 For example, the device file @file{/dev/urandom} could be used as the
1187 source of random data. Typically, this device gathers environmental
1188 noise from device drivers and other sources into an entropy pool, and
1189 uses the pool to generate random bits. If the pool is short of data,
1190 the device reuses the internal pool to produce more bits, using a
1191 cryptographically secure pseudorandom number generator. But be aware
1192 that this device is not designed for bulk random data generation
1193 and is relatively slow.
1195 @file{/dev/urandom} suffices for most practical uses, but applications
1196 requiring high-value or long-term protection of private data may
1197 require an alternate data source like @file{/dev/random} or
1198 @file{/dev/arandom}. The set of available sources depends on your
1201 To reproduce the results of an earlier invocation of a command, you
1202 can save some random data into a file and then use that file as the
1203 random source in earlier and later invocations of the command.
1205 @node Target directory
1206 @section Target directory
1208 @cindex target directory
1210 The @command{cp}, @command{install}, @command{ln}, and @command{mv}
1211 commands normally treat the last operand specially when it is a
1212 directory or a symbolic link to a directory. For example, @samp{cp
1213 source dest} is equivalent to @samp{cp source dest/source} if
1214 @file{dest} is a directory. Sometimes this behavior is not exactly
1215 what is wanted, so these commands support the following options to
1216 allow more fine-grained control:
1221 @itemx --no-target-directory
1222 @opindex --no-target-directory
1223 @cindex target directory
1224 @cindex destination directory
1225 Do not treat the last operand specially when it is a directory or a
1226 symbolic link to a directory. This can help avoid race conditions in
1227 programs that operate in a shared area. For example, when the command
1228 @samp{mv /tmp/source /tmp/dest} succeeds, there is no guarantee that
1229 @file{/tmp/source} was renamed to @file{/tmp/dest}: it could have been
1230 renamed to @file{/tmp/dest/source} instead, if some other process
1231 created @file{/tmp/dest} as a directory. However, if @file{mv
1232 -T /tmp/source /tmp/dest} succeeds, there is no
1233 question that @file{/tmp/source} was renamed to @file{/tmp/dest}.
1235 In the opposite situation, where you want the last operand to be
1236 treated as a directory and want a diagnostic otherwise, you can use
1237 the @option{--target-directory} (@option{-t}) option.
1239 @item -t @var{directory}
1240 @itemx @w{@kbd{--target-directory}=@var{directory}}
1241 @opindex --target-directory
1242 @cindex target directory
1243 @cindex destination directory
1244 Use @var{directory} as the directory component of each destination
1247 The interface for most programs is that after processing options and a
1248 finite (possibly zero) number of fixed-position arguments, the remaining
1249 argument list is either expected to be empty, or is a list of items
1250 (usually files) that will all be handled identically. The @command{xargs}
1251 program is designed to work well with this convention.
1253 The commands in the @command{mv}-family are unusual in that they take
1254 a variable number of arguments with a special case at the @emph{end}
1255 (namely, the target directory). This makes it nontrivial to perform some
1256 operations, e.g., ``move all files from here to ../d/'', because
1257 @code{mv * ../d/} might exhaust the argument space, and @code{ls | xargs ...}
1258 doesn't have a clean way to specify an extra final argument for each
1259 invocation of the subject command. (It can be done by going through a
1260 shell command, but that requires more human labor and brain power than
1263 The @w{@kbd{--target-directory}} (@option{-t}) option allows the @command{cp},
1264 @command{install}, @command{ln}, and @command{mv} programs to be used
1265 conveniently with @command{xargs}. For example, you can move the files
1266 from the current directory to a sibling directory, @code{d} like this:
1269 ls | xargs mv -t ../d --
1272 However, this doesn't move files whose names begin with @samp{.}.
1273 If you use the @sc{gnu} @command{find} program, you can move those
1274 files too, with this command:
1277 find . -mindepth 1 -maxdepth 1 \
1281 But both of the above approaches fail if there are no files in the
1282 current directory, or if any file has a name containing a blank or
1283 some other special characters.
1284 The following example removes those limitations and requires both
1285 @sc{gnu} @command{find} and @sc{gnu} @command{xargs}:
1288 find . -mindepth 1 -maxdepth 1 -print0 \
1289 | xargs --null --no-run-if-empty \
1296 The @option{--target-directory} (@option{-t}) and
1297 @option{--no-target-directory} (@option{-T})
1298 options cannot be combined.
1300 @node Trailing slashes
1301 @section Trailing slashes
1303 @cindex trailing slashes
1305 Some @sc{gnu} programs (at least @command{cp} and @command{mv}) allow you to
1306 remove any trailing slashes from each @var{source} argument before
1307 operating on it. The @w{@kbd{--strip-trailing-slashes}} option enables
1310 This is useful when a @var{source} argument may have a trailing slash and
1311 @c FIXME: mv's behavior in this case is system-dependent
1312 specify a symbolic link to a directory. This scenario is in fact rather
1313 common because some shells can automatically append a trailing slash when
1314 performing file name completion on such symbolic links. Without this
1315 option, @command{mv}, for example, (via the system's rename function) must
1316 interpret a trailing slash as a request to dereference the symbolic link
1317 and so must rename the indirectly referenced @emph{directory} and not
1318 the symbolic link. Although it may seem surprising that such behavior
1319 be the default, it is required by POSIX and is consistent with
1320 other parts of that standard.
1322 @node Traversing symlinks
1323 @section Traversing symlinks
1325 @cindex symbolic link to directory, controlling traversal of
1327 The following options modify how @command{chown} and @command{chgrp}
1328 @c FIXME: note that 'du' has these options, too, but they have slightly
1329 @c different meaning.
1330 traverse a hierarchy when the @option{--recursive} (@option{-R})
1331 option is also specified.
1332 If more than one of the following options is specified, only the final
1334 These options specify whether processing a symbolic link to a directory
1335 entails operating on just the symbolic link or on all files in the
1336 hierarchy rooted at that directory.
1338 These options are independent of @option{--dereference} and
1339 @option{--no-dereference} (@option{-h}), which control whether to modify
1340 a symlink or its referent.
1347 @cindex symbolic link to directory, traverse if on the command line
1348 If @option{--recursive} (@option{-R}) is specified and
1349 a command line argument is a symbolic link to a directory, traverse it.
1356 @cindex symbolic link to directory, traverse each that is encountered
1357 In a recursive traversal, traverse every symbolic link to a directory
1358 that is encountered.
1365 @cindex symbolic link to directory, never traverse
1366 Do not traverse any symbolic links.
1367 This is the default if none of @option{-H}, @option{-L},
1368 or @option{-P} is specified.
1375 @node Treating / specially
1376 @section Treating @file{/} specially
1378 Certain commands can operate destructively on entire hierarchies.
1379 For example, if a user with appropriate privileges mistakenly runs
1380 @samp{rm -rf / tmp/junk}, that may remove
1381 all files on the entire system. Since there are so few
1382 legitimate uses for such a command,
1383 @sc{gnu} @command{rm} normally declines to operate on any directory
1384 that resolves to @file{/}. If you really want to try to remove all
1385 the files on your system, you can use the @option{--no-preserve-root}
1386 option, but the default behavior, specified by the
1387 @option{--preserve-root} option, is safer for most purposes.
1389 The commands @command{chgrp}, @command{chmod} and @command{chown}
1390 can also operate destructively on entire hierarchies, so they too
1391 support these options. Although, unlike @command{rm}, they don't
1392 actually unlink files, these commands are arguably more dangerous
1393 when operating recursively on @file{/}, since they often work much
1394 more quickly, and hence damage more files before an alert user can
1395 interrupt them. Tradition and POSIX require these commands
1396 to operate recursively on @file{/}, so they default to
1397 @option{--no-preserve-root}, but using the @option{--preserve-root}
1398 option makes them safer for most purposes. For convenience you can
1399 specify @option{--preserve-root} in an alias or in a shell function.
1401 Note that the @option{--preserve-root} option also ensures
1402 that @command{chgrp} and @command{chown} do not modify @file{/}
1403 even when dereferencing a symlink pointing to @file{/}.
1405 @node Special built-in utilities
1406 @section Special built-in utilities
1408 Some programs like @command{nice} can invoke other programs; for
1409 example, the command @samp{nice cat file} invokes the program
1410 @command{cat} by executing the command @samp{cat file}. However,
1411 @dfn{special built-in utilities} like @command{exit} cannot be invoked
1412 this way. For example, the command @samp{nice exit} does not have a
1413 well-defined behavior: it may generate an error message instead of
1416 Here is a list of the special built-in utilities that are standardized
1417 by POSIX 1003.1-2004.
1420 @t{.@: : break continue eval exec exit export readonly
1421 return set shift times trap unset}
1424 For example, because @samp{.}, @samp{:}, and @samp{exec} are special,
1425 the commands @samp{nice . foo.sh}, @samp{nice :}, and @samp{nice exec
1426 pwd} do not work as you might expect.
1428 Many shells extend this list. For example, Bash has several extra
1429 special built-in utilities like @command{history}, and
1430 @command{suspend}, and with Bash the command @samp{nice suspend}
1431 generates an error message instead of suspending.
1433 @node Standards conformance
1434 @section Standards conformance
1436 @vindex POSIXLY_CORRECT
1437 In a few cases, the @sc{gnu} utilities' default behavior is
1438 incompatible with the POSIX standard. To suppress these
1439 incompatibilities, define the @env{POSIXLY_CORRECT} environment
1440 variable. Unless you are checking for POSIX conformance, you
1441 probably do not need to define @env{POSIXLY_CORRECT}.
1443 Newer versions of POSIX are occasionally incompatible with older
1444 versions. For example, older versions of POSIX required the
1445 command @samp{sort +1} to sort based on the second and succeeding
1446 fields in each input line, but starting with POSIX 1003.1-2001
1447 the same command is required to sort the file named @file{+1}, and you
1448 must instead use the command @samp{sort -k 2} to get the field-based
1451 @vindex _POSIX2_VERSION
1452 The @sc{gnu} utilities normally conform to the version of POSIX
1453 that is standard for your system. To cause them to conform to a
1454 different version of POSIX, define the @env{_POSIX2_VERSION}
1455 environment variable to a value of the form @var{yyyymm} specifying
1456 the year and month the standard was adopted. Three values are currently
1457 supported for @env{_POSIX2_VERSION}: @samp{199209} stands for
1458 POSIX 1003.2-1992, @samp{200112} stands for POSIX
1459 1003.1-2001, and @samp{200809} stands for POSIX 1003.1-2008.
1460 For example, if you have a newer system but are running software
1461 that assumes an older version of POSIX and uses @samp{sort +1}
1462 or @samp{tail +10}, you can work around any compatibility problems by setting
1463 @samp{_POSIX2_VERSION=199209} in your environment.
1465 @node Output of entire files
1466 @chapter Output of entire files
1468 @cindex output of entire files
1469 @cindex entire files, output of
1471 These commands read and write entire files, possibly transforming them
1475 * cat invocation:: Concatenate and write files.
1476 * tac invocation:: Concatenate and write files in reverse.
1477 * nl invocation:: Number lines and write files.
1478 * od invocation:: Write files in octal or other formats.
1479 * base64 invocation:: Transform data into printable data.
1482 @node cat invocation
1483 @section @command{cat}: Concatenate and write files
1486 @cindex concatenate and write files
1487 @cindex copying files
1489 @command{cat} copies each @var{file} (@samp{-} means standard input), or
1490 standard input if none are given, to standard output. Synopsis:
1493 cat [@var{option}] [@var{file}]@dots{}
1496 The program accepts the following options. Also see @ref{Common options}.
1504 Equivalent to @option{-vET}.
1507 @itemx --number-nonblank
1509 @opindex --number-nonblank
1510 Number all nonempty output lines, starting with 1.
1514 Equivalent to @option{-vE}.
1519 @opindex --show-ends
1520 Display a @samp{$} after the end of each line.
1526 Number all output lines, starting with 1. This option is ignored
1527 if @option{-b} is in effect.
1530 @itemx --squeeze-blank
1532 @opindex --squeeze-blank
1533 @cindex squeezing empty lines
1534 Suppress repeated adjacent empty lines; output just one empty line
1539 Equivalent to @option{-vT}.
1544 @opindex --show-tabs
1545 Display TAB characters as @samp{^I}.
1549 Ignored; for POSIX compatibility.
1552 @itemx --show-nonprinting
1554 @opindex --show-nonprinting
1555 Display control characters except for LFD and TAB using
1556 @samp{^} notation and precede characters that have the high bit set with
1561 On systems like MS-DOS that distinguish between text and binary files,
1562 @command{cat} normally reads and writes in binary mode. However,
1563 @command{cat} reads in text mode if one of the options
1564 @option{-bensAE} is used or if @command{cat} is reading from standard
1565 input and standard input is a terminal. Similarly, @command{cat}
1566 writes in text mode if one of the options @option{-bensAE} is used or
1567 if standard output is a terminal.
1574 # Output f's contents, then standard input, then g's contents.
1577 # Copy standard input to standard output.
1582 @node tac invocation
1583 @section @command{tac}: Concatenate and write files in reverse
1586 @cindex reversing files
1588 @command{tac} copies each @var{file} (@samp{-} means standard input), or
1589 standard input if none are given, to standard output, reversing the
1590 records (lines by default) in each separately. Synopsis:
1593 tac [@var{option}]@dots{} [@var{file}]@dots{}
1596 @dfn{Records} are separated by instances of a string (newline by
1597 default). By default, this separator string is attached to the end of
1598 the record that it follows in the file.
1600 The program accepts the following options. Also see @ref{Common options}.
1608 The separator is attached to the beginning of the record that it
1609 precedes in the file.
1615 Treat the separator string as a regular expression. Users of @command{tac}
1616 on MS-DOS/MS-Windows should note that, since @command{tac} reads files in
1617 binary mode, each line of a text file might end with a CR/LF pair
1618 instead of the Unix-style LF.
1620 @item -s @var{separator}
1621 @itemx --separator=@var{separator}
1623 @opindex --separator
1624 Use @var{separator} as the record separator, instead of newline.
1632 @section @command{nl}: Number lines and write files
1635 @cindex numbering lines
1636 @cindex line numbering
1638 @command{nl} writes each @var{file} (@samp{-} means standard input), or
1639 standard input if none are given, to standard output, with line numbers
1640 added to some or all of the lines. Synopsis:
1643 nl [@var{option}]@dots{} [@var{file}]@dots{}
1646 @cindex logical pages, numbering on
1647 @command{nl} decomposes its input into (logical) pages; by default, the
1648 line number is reset to 1 at the top of each logical page. @command{nl}
1649 treats all of the input files as a single document; it does not reset
1650 line numbers or logical pages between files.
1652 @cindex headers, numbering
1653 @cindex body, numbering
1654 @cindex footers, numbering
1655 A logical page consists of three sections: header, body, and footer.
1656 Any of the sections can be empty. Each can be numbered in a different
1657 style from the others.
1659 The beginnings of the sections of logical pages are indicated in the
1660 input file by a line containing exactly one of these delimiter strings:
1671 The two characters from which these strings are made can be changed from
1672 @samp{\} and @samp{:} via options (see below), but the pattern and
1673 length of each string cannot be changed.
1675 A section delimiter is replaced by an empty line on output. Any text
1676 that comes before the first section delimiter string in the input file
1677 is considered to be part of a body section, so @command{nl} treats a
1678 file that contains no section delimiters as a single body section.
1680 The program accepts the following options. Also see @ref{Common options}.
1684 @item -b @var{style}
1685 @itemx --body-numbering=@var{style}
1687 @opindex --body-numbering
1688 Select the numbering style for lines in the body section of each
1689 logical page. When a line is not numbered, the current line number
1690 is not incremented, but the line number separator character is still
1691 prepended to the line. The styles are:
1697 number only nonempty lines (default for body),
1699 do not number lines (default for header and footer),
1701 number only lines that contain a match for the basic regular
1702 expression @var{bre}.
1703 @xref{Regular Expressions, , Regular Expressions, grep, The GNU Grep Manual}.
1707 @itemx --section-delimiter=@var{cd}
1709 @opindex --section-delimiter
1710 @cindex section delimiters of pages
1711 Set the section delimiter characters to @var{cd}; default is
1712 @samp{\:}. If only @var{c} is given, the second remains @samp{:}.
1713 (Remember to protect @samp{\} or other metacharacters from shell
1714 expansion with quotes or extra backslashes.)
1716 @item -f @var{style}
1717 @itemx --footer-numbering=@var{style}
1719 @opindex --footer-numbering
1720 Analogous to @option{--body-numbering}.
1722 @item -h @var{style}
1723 @itemx --header-numbering=@var{style}
1725 @opindex --header-numbering
1726 Analogous to @option{--body-numbering}.
1728 @item -i @var{number}
1729 @itemx --line-increment=@var{number}
1731 @opindex --line-increment
1732 Increment line numbers by @var{number} (default 1).
1734 @item -l @var{number}
1735 @itemx --join-blank-lines=@var{number}
1737 @opindex --join-blank-lines
1738 @cindex empty lines, numbering
1739 @cindex blank lines, numbering
1740 Consider @var{number} (default 1) consecutive empty lines to be one
1741 logical line for numbering, and only number the last one. Where fewer
1742 than @var{number} consecutive empty lines occur, do not number them.
1743 An empty line is one that contains no characters, not even spaces
1746 @item -n @var{format}
1747 @itemx --number-format=@var{format}
1749 @opindex --number-format
1750 Select the line numbering format (default is @code{rn}):
1754 @opindex ln @r{format for @command{nl}}
1755 left justified, no leading zeros;
1757 @opindex rn @r{format for @command{nl}}
1758 right justified, no leading zeros;
1760 @opindex rz @r{format for @command{nl}}
1761 right justified, leading zeros.
1765 @itemx --no-renumber
1767 @opindex --no-renumber
1768 Do not reset the line number at the start of a logical page.
1770 @item -s @var{string}
1771 @itemx --number-separator=@var{string}
1773 @opindex --number-separator
1774 Separate the line number from the text line in the output with
1775 @var{string} (default is the TAB character).
1777 @item -v @var{number}
1778 @itemx --starting-line-number=@var{number}
1780 @opindex --starting-line-number
1781 Set the initial line number on each logical page to @var{number} (default 1).
1783 @item -w @var{number}
1784 @itemx --number-width=@var{number}
1786 @opindex --number-width
1787 Use @var{number} characters for line numbers (default 6).
1795 @section @command{od}: Write files in octal or other formats
1798 @cindex octal dump of files
1799 @cindex hex dump of files
1800 @cindex ASCII dump of files
1801 @cindex file contents, dumping unambiguously
1803 @command{od} writes an unambiguous representation of each @var{file}
1804 (@samp{-} means standard input), or standard input if none are given.
1808 od [@var{option}]@dots{} [@var{file}]@dots{}
1809 od [-abcdfilosx]@dots{} [@var{file}] [[+]@var{offset}[.][b]]
1810 od [@var{option}]@dots{} --traditional [@var{file}]@c
1811 [[+]@var{offset}[.][b] [[+]@var{label}[.][b]]]
1814 Each line of output consists of the offset in the input, followed by
1815 groups of data from the file. By default, @command{od} prints the offset in
1816 octal, and each group of file data is a C @code{short int}'s worth of input
1817 printed as a single octal number.
1819 If @var{offset} is given, it specifies how many input bytes to skip
1820 before formatting and writing. By default, it is interpreted as an
1821 octal number, but the optional trailing decimal point causes it to be
1822 interpreted as decimal. If no decimal is specified and the offset
1823 begins with @samp{0x} or @samp{0X} it is interpreted as a hexadecimal
1824 number. If there is a trailing @samp{b}, the number of bytes skipped
1825 will be @var{offset} multiplied by 512.
1827 If a command is of both the first and second forms, the second form is
1828 assumed if the last operand begins with @samp{+} or (if there are two
1829 operands) a digit. For example, in @samp{od foo 10} and @samp{od +10}
1830 the @samp{10} is an offset, whereas in @samp{od 10} the @samp{10} is a
1833 The program accepts the following options. Also see @ref{Common options}.
1837 @item -A @var{radix}
1838 @itemx --address-radix=@var{radix}
1840 @opindex --address-radix
1841 @cindex radix for file offsets
1842 @cindex file offset radix
1843 Select the base in which file offsets are printed. @var{radix} can
1844 be one of the following:
1854 none (do not print offsets).
1857 The default is octal.
1859 @item -j @var{bytes}
1860 @itemx --skip-bytes=@var{bytes}
1862 @opindex --skip-bytes
1863 Skip @var{bytes} input bytes before formatting and writing. If
1864 @var{bytes} begins with @samp{0x} or @samp{0X}, it is interpreted in
1865 hexadecimal; otherwise, if it begins with @samp{0}, in octal; otherwise,
1867 @multiplierSuffixes{bytes}
1869 @item -N @var{bytes}
1870 @itemx --read-bytes=@var{bytes}
1872 @opindex --read-bytes
1873 Output at most @var{bytes} bytes of the input. Prefixes and suffixes on
1874 @code{bytes} are interpreted as for the @option{-j} option.
1876 @item -S @var{bytes}
1877 @itemx --strings[=@var{bytes}]
1880 @cindex string constants, outputting
1881 Instead of the normal output, output only @dfn{string constants}: at
1882 least @var{bytes} consecutive ASCII graphic characters,
1883 followed by a zero byte (ASCII @sc{nul}).
1884 Prefixes and suffixes on @var{bytes} are interpreted as for the
1887 If @var{n} is omitted with @option{--strings}, the default is 3.
1890 @itemx --format=@var{type}
1893 Select the format in which to output the file data. @var{type} is a
1894 string of one or more of the below type indicator characters. If you
1895 include more than one type indicator character in a single @var{type}
1896 string, or use this option more than once, @command{od} writes one copy
1897 of each output line using each of the data types that you specified,
1898 in the order that you specified.
1900 Adding a trailing ``z'' to any type specification appends a display
1901 of the ASCII character representation of the printable characters
1902 to the output line generated by the type specification.
1906 named character, ignoring high-order bit
1908 ASCII character or backslash escape,
1912 floating point (@pxref{Floating point})
1921 The type @code{a} outputs things like @samp{sp} for space, @samp{nl} for
1922 newline, and @samp{nul} for a zero byte. Only the least significant
1923 seven bits of each byte is used; the high-order bit is ignored.
1924 Type @code{c} outputs
1925 @samp{ }, @samp{\n}, and @code{\0}, respectively.
1928 Except for types @samp{a} and @samp{c}, you can specify the number
1929 of bytes to use in interpreting each number in the given data type
1930 by following the type indicator character with a decimal integer.
1931 Alternately, you can specify the size of one of the C compiler's
1932 built-in data types by following the type indicator character with
1933 one of the following characters. For integers (@samp{d}, @samp{o},
1934 @samp{u}, @samp{x}):
1947 For floating point (@code{f}):
1959 @itemx --output-duplicates
1961 @opindex --output-duplicates
1962 Output consecutive lines that are identical. By default, when two or
1963 more consecutive output lines would be identical, @command{od} outputs only
1964 the first line, and puts just an asterisk on the following line to
1965 indicate the elision.
1968 @itemx --width[=@var{n}]
1971 Dump @code{n} input bytes per output line. This must be a multiple of
1972 the least common multiple of the sizes associated with the specified
1975 If this option is not given at all, the default is 16. If @var{n} is
1976 omitted, the default is 32.
1980 The next several options are shorthands for format specifications.
1981 @sc{gnu} @command{od} accepts any combination of shorthands and format
1982 specification options. These options accumulate.
1988 Output as named characters. Equivalent to @samp{-t a}.
1992 Output as octal bytes. Equivalent to @samp{-t o1}.
1996 Output as ASCII characters or backslash escapes. Equivalent to
2001 Output as unsigned decimal two-byte units. Equivalent to @samp{-t u2}.
2005 Output as floats. Equivalent to @samp{-t fF}.
2009 Output as decimal ints. Equivalent to @samp{-t dI}.
2013 Output as decimal long ints. Equivalent to @samp{-t dL}.
2017 Output as octal two-byte units. Equivalent to @option{-t o2}.
2021 Output as decimal two-byte units. Equivalent to @option{-t d2}.
2025 Output as hexadecimal two-byte units. Equivalent to @samp{-t x2}.
2028 @opindex --traditional
2029 Recognize the non-option label argument that traditional @command{od}
2030 accepted. The following syntax:
2033 od --traditional [@var{file}] [[+]@var{offset}[.][b] [[+]@var{label}[.][b]]]
2037 can be used to specify at most one file and optional arguments
2038 specifying an offset and a pseudo-start address, @var{label}.
2039 The @var{label} argument is interpreted
2040 just like @var{offset}, but it specifies an initial pseudo-address. The
2041 pseudo-addresses are displayed in parentheses following any normal
2048 @node base64 invocation
2049 @section @command{base64}: Transform data into printable data
2052 @cindex base64 encoding
2054 @command{base64} transforms data read from a file, or standard input,
2055 into (or from) base64 encoded form. The base64 encoded form uses
2056 printable ASCII characters to represent binary data.
2060 base64 [@var{option}]@dots{} [@var{file}]
2061 base64 --decode [@var{option}]@dots{} [@var{file}]
2064 The base64 encoding expands data to roughly 133% of the original.
2065 The format conforms to
2066 @uref{ftp://ftp.rfc-editor.org/in-notes/rfc4648.txt, RFC 4648}.
2068 The program accepts the following options. Also see @ref{Common options}.
2073 @itemx --wrap=@var{cols}
2077 @cindex column to wrap data after
2078 During encoding, wrap lines after @var{cols} characters. This must be
2081 The default is to wrap after 76 characters. Use the value 0 to
2082 disable line wrapping altogether.
2088 @cindex Decode base64 data
2089 @cindex Base64 decoding
2090 Change the mode of operation, from the default of encoding data, to
2091 decoding data. Input is expected to be base64 encoded data, and the
2092 output will be the original data.
2095 @itemx --ignore-garbage
2097 @opindex --ignore-garbage
2098 @cindex Ignore garbage in base64 stream
2099 When decoding, newlines are always accepted.
2100 During decoding, ignore unrecognized bytes,
2101 to permit distorted data to be decoded.
2108 @node Formatting file contents
2109 @chapter Formatting file contents
2111 @cindex formatting file contents
2113 These commands reformat the contents of files.
2116 * fmt invocation:: Reformat paragraph text.
2117 * pr invocation:: Paginate or columnate files for printing.
2118 * fold invocation:: Wrap input lines to fit in specified width.
2122 @node fmt invocation
2123 @section @command{fmt}: Reformat paragraph text
2126 @cindex reformatting paragraph text
2127 @cindex paragraphs, reformatting
2128 @cindex text, reformatting
2130 @command{fmt} fills and joins lines to produce output lines of (at most)
2131 a given number of characters (75 by default). Synopsis:
2134 fmt [@var{option}]@dots{} [@var{file}]@dots{}
2137 @command{fmt} reads from the specified @var{file} arguments (or standard
2138 input if none are given), and writes to standard output.
2140 By default, blank lines, spaces between words, and indentation are
2141 preserved in the output; successive input lines with different
2142 indentation are not joined; tabs are expanded on input and introduced on
2145 @cindex line-breaking
2146 @cindex sentences and line-breaking
2147 @cindex Knuth, Donald E.
2148 @cindex Plass, Michael F.
2149 @command{fmt} prefers breaking lines at the end of a sentence, and tries to
2150 avoid line breaks after the first word of a sentence or before the last
2151 word of a sentence. A @dfn{sentence break} is defined as either the end
2152 of a paragraph or a word ending in any of @samp{.?!}, followed by two
2153 spaces or end of line, ignoring any intervening parentheses or quotes.
2154 Like @TeX{}, @command{fmt} reads entire ``paragraphs'' before choosing line
2155 breaks; the algorithm is a variant of that given by Donald E. Knuth
2156 and Michael F. Plass in ``Breaking Paragraphs Into Lines'',
2157 @cite{Software---Practice & Experience} @b{11}, 11 (November 1981),
2160 The program accepts the following options. Also see @ref{Common options}.
2165 @itemx --crown-margin
2167 @opindex --crown-margin
2168 @cindex crown margin
2169 @dfn{Crown margin} mode: preserve the indentation of the first two
2170 lines within a paragraph, and align the left margin of each subsequent
2171 line with that of the second line.
2174 @itemx --tagged-paragraph
2176 @opindex --tagged-paragraph
2177 @cindex tagged paragraphs
2178 @dfn{Tagged paragraph} mode: like crown margin mode, except that if
2179 indentation of the first line of a paragraph is the same as the
2180 indentation of the second, the first line is treated as a one-line
2186 @opindex --split-only
2187 Split lines only. Do not join short lines to form longer ones. This
2188 prevents sample lines of code, and other such ``formatted'' text from
2189 being unduly combined.
2192 @itemx --uniform-spacing
2194 @opindex --uniform-spacing
2195 Uniform spacing. Reduce spacing between words to one space, and spacing
2196 between sentences to two spaces.
2199 @itemx -w @var{width}
2200 @itemx --width=@var{width}
2201 @opindex -@var{width}
2204 Fill output lines up to @var{width} characters (default 75 or @var{goal}
2205 plus 10, if @var{goal} is provided).
2208 @itemx --goal=@var{goal}
2211 @command{fmt} initially tries to make lines @var{goal} characters wide.
2212 By default, this is 7% shorter than @var{width}.
2214 @item -p @var{prefix}
2215 @itemx --prefix=@var{prefix}
2216 Only lines beginning with @var{prefix} (possibly preceded by whitespace)
2217 are subject to formatting. The prefix and any preceding whitespace are
2218 stripped for the formatting and then re-attached to each formatted output
2219 line. One use is to format certain kinds of program comments, while
2220 leaving the code unchanged.
2228 @section @command{pr}: Paginate or columnate files for printing
2231 @cindex printing, preparing files for
2232 @cindex multicolumn output, generating
2233 @cindex merging files in parallel
2235 @command{pr} writes each @var{file} (@samp{-} means standard input), or
2236 standard input if none are given, to standard output, paginating and
2237 optionally outputting in multicolumn format; optionally merges all
2238 @var{file}s, printing all in parallel, one per column. Synopsis:
2241 pr [@var{option}]@dots{} [@var{file}]@dots{}
2245 By default, a 5-line header is printed at each page: two blank lines;
2246 a line with the date, the file name, and the page count; and two more
2247 blank lines. A footer of five blank lines is also printed.
2248 The default @var{page_length} is 66
2249 lines. The default number of text lines is therefore 56.
2250 The text line of the header takes the form
2251 @samp{@var{date} @var{string} @var{page}}, with spaces inserted around
2252 @var{string} so that the line takes up the full @var{page_width}. Here,
2253 @var{date} is the date (see the @option{-D} or @option{--date-format}
2254 option for details), @var{string} is the centered header string, and
2255 @var{page} identifies the page number. The @env{LC_MESSAGES} locale
2256 category affects the spelling of @var{page}; in the default C locale, it
2257 is @samp{Page @var{number}} where @var{number} is the decimal page
2260 Form feeds in the input cause page breaks in the output. Multiple form
2261 feeds produce empty pages.
2263 Columns are of equal width, separated by an optional string (default
2264 is @samp{space}). For multicolumn output, lines will always be truncated to
2265 @var{page_width} (default 72), unless you use the @option{-J} option.
2267 column output no line truncation occurs by default. Use @option{-W} option to
2268 truncate lines in that case.
2270 The following changes were made in version 1.22i and apply to later
2271 versions of @command{pr}:
2272 @c FIXME: this whole section here sounds very awkward to me. I
2273 @c made a few small changes, but really it all needs to be redone. - Brian
2274 @c OK, I fixed another sentence or two, but some of it I just don't understand.
2279 Some small @var{letter options} (@option{-s}, @option{-w}) have been
2280 redefined for better POSIX compliance. The output of some further
2281 cases has been adapted to other Unix systems. These changes are not
2282 compatible with earlier versions of the program.
2285 Some @var{new capital letter} options (@option{-J}, @option{-S}, @option{-W})
2286 have been introduced to turn off unexpected interferences of small letter
2287 options. The @option{-N} option and the second argument @var{last_page}
2288 of @samp{+FIRST_PAGE} offer more flexibility. The detailed handling of
2289 form feeds set in the input files requires the @option{-T} option.
2292 Capital letter options override small letter ones.
2295 Some of the option-arguments (compare @option{-s}, @option{-e},
2296 @option{-i}, @option{-n}) cannot be specified as separate arguments from the
2297 preceding option letter (already stated in the POSIX specification).
2300 The program accepts the following options. Also see @ref{Common options}.
2304 @item +@var{first_page}[:@var{last_page}]
2305 @itemx --pages=@var{first_page}[:@var{last_page}]
2306 @c The two following @opindex lines evoke warnings because they contain ':'
2307 @c The 'info' spec does not permit that. If we use those lines, we end
2308 @c up with truncated index entries that don't work.
2309 @c @opindex +@var{first_page}[:@var{last_page}]
2310 @c @opindex --pages=@var{first_page}[:@var{last_page}]
2311 @opindex +@var{page_range}
2312 @opindex --pages=@var{page_range}
2313 Begin printing with page @var{first_page} and stop with @var{last_page}.
2314 Missing @samp{:@var{last_page}} implies end of file. While estimating
2315 the number of skipped pages each form feed in the input file results
2316 in a new page. Page counting with and without @samp{+@var{first_page}}
2317 is identical. By default, counting starts with the first page of input
2318 file (not first page printed). Line numbering may be altered by @option{-N}
2322 @itemx --columns=@var{column}
2323 @opindex -@var{column}
2325 @cindex down columns
2326 With each single @var{file}, produce @var{column} columns of output
2327 (default is 1) and print columns down, unless @option{-a} is used. The
2328 column width is automatically decreased as @var{column} increases; unless
2329 you use the @option{-W/-w} option to increase @var{page_width} as well.
2330 This option might well cause some lines to be truncated. The number of
2331 lines in the columns on each page are balanced. The options @option{-e}
2332 and @option{-i} are on for multiple text-column output. Together with
2333 @option{-J} option column alignment and line truncation is turned off.
2334 Lines of full length are joined in a free field format and @option{-S}
2335 option may set field separators. @option{-@var{column}} may not be used
2336 with @option{-m} option.
2342 @cindex across columns
2343 With each single @var{file}, print columns across rather than down. The
2344 @option{-@var{column}} option must be given with @var{column} greater than one.
2345 If a line is too long to fit in a column, it is truncated.
2348 @itemx --show-control-chars
2350 @opindex --show-control-chars
2351 Print control characters using hat notation (e.g., @samp{^G}); print
2352 other nonprinting characters in octal backslash notation. By default,
2353 nonprinting characters are not changed.
2356 @itemx --double-space
2358 @opindex --double-space
2359 @cindex double spacing
2360 Double space the output.
2362 @item -D @var{format}
2363 @itemx --date-format=@var{format}
2364 @cindex time formats
2365 @cindex formatting times
2366 Format header dates using @var{format}, using the same conventions as
2367 for the command @samp{date +@var{format}}. @xref{date invocation}.
2368 Except for directives, which start with
2369 @samp{%}, characters in @var{format} are printed unchanged. You can use
2370 this option to specify an arbitrary string in place of the header date,
2371 e.g., @option{--date-format="Monday morning"}.
2373 @vindex POSIXLY_CORRECT
2375 The default date format is @samp{%Y-%m-%d %H:%M} (for example,
2376 @samp{2001-12-04 23:59});
2377 but if the @env{POSIXLY_CORRECT} environment variable is set
2378 and the @env{LC_TIME} locale category specifies the POSIX
2379 locale, the default is @samp{%b %e %H:%M %Y} (for example,
2380 @samp{Dec@ @ 4 23:59 2001}.
2383 Time stamps are listed according to the time zone rules specified by
2384 the @env{TZ} environment variable, or by the system default rules if
2385 @env{TZ} is not set. @xref{TZ Variable,, Specifying the Time Zone
2386 with @env{TZ}, libc, The GNU C Library Reference Manual}.
2388 @item -e[@var{in-tabchar}[@var{in-tabwidth}]]
2389 @itemx --expand-tabs[=@var{in-tabchar}[@var{in-tabwidth}]]
2391 @opindex --expand-tabs
2393 Expand @var{tab}s to spaces on input. Optional argument @var{in-tabchar} is
2394 the input tab character (default is the TAB character). Second optional
2395 argument @var{in-tabwidth} is the input tab character's width (default
2403 @opindex --form-feed
2404 Use a form feed instead of newlines to separate output pages. This does
2405 not alter the default page length of 66 lines.
2407 @item -h @var{header}
2408 @itemx --header=@var{header}
2411 Replace the file name in the header with the centered string @var{header}.
2412 When using the shell, @var{header} should be quoted and should be
2413 separated from @option{-h} by a space.
2415 @item -i[@var{out-tabchar}[@var{out-tabwidth}]]
2416 @itemx --output-tabs[=@var{out-tabchar}[@var{out-tabwidth}]]
2418 @opindex --output-tabs
2420 Replace spaces with @var{tab}s on output. Optional argument @var{out-tabchar}
2421 is the output tab character (default is the TAB character). Second optional
2422 argument @var{out-tabwidth} is the output tab character's width (default
2428 @opindex --join-lines
2429 Merge lines of full length. Used together with the column options
2430 @option{-@var{column}}, @option{-a -@var{column}} or @option{-m}. Turns off
2431 @option{-W/-w} line truncation;
2432 no column alignment used; may be used with
2433 @option{--sep-string[=@var{string}]}. @option{-J} has been introduced
2434 (together with @option{-W} and @option{--sep-string})
2435 to disentangle the old (POSIX-compliant) options @option{-w} and
2436 @option{-s} along with the three column options.
2439 @item -l @var{page_length}
2440 @itemx --length=@var{page_length}
2443 Set the page length to @var{page_length} (default 66) lines, including
2444 the lines of the header [and the footer]. If @var{page_length} is less
2445 than or equal to 10, the header and footer are omitted, as if the
2446 @option{-t} option had been given.
2452 Merge and print all @var{file}s in parallel, one in each column. If a
2453 line is too long to fit in a column, it is truncated, unless the @option{-J}
2454 option is used. @option{--sep-string[=@var{string}]} may be used.
2456 some @var{file}s (form feeds set) produce empty columns, still marked
2457 by @var{string}. The result is a continuous line numbering and column
2458 marking throughout the whole merged file. Completely empty merged pages
2459 show no separators or line numbers. The default header becomes
2460 @samp{@var{date} @var{page}} with spaces inserted in the middle; this
2461 may be used with the @option{-h} or @option{--header} option to fill up
2462 the middle blank part.
2464 @item -n[@var{number-separator}[@var{digits}]]
2465 @itemx --number-lines[=@var{number-separator}[@var{digits}]]
2467 @opindex --number-lines
2468 Provide @var{digits} digit line numbering (default for @var{digits} is
2469 5). With multicolumn output the number occupies the first @var{digits}
2470 column positions of each text column or only each line of @option{-m}
2471 output. With single column output the number precedes each line just as
2472 @option{-m} does. Default counting of the line numbers starts with the
2473 first line of the input file (not the first line printed, compare the
2474 @option{--page} option and @option{-N} option).
2475 Optional argument @var{number-separator} is the character appended to
2476 the line number to separate it from the text followed. The default
2477 separator is the TAB character. In a strict sense a TAB is always
2478 printed with single column output only. The TAB width varies
2479 with the TAB position, e.g., with the left @var{margin} specified
2480 by @option{-o} option. With multicolumn output priority is given to
2481 @samp{equal width of output columns} (a POSIX specification).
2482 The TAB width is fixed to the value of the first column and does
2483 not change with different values of left @var{margin}. That means a
2484 fixed number of spaces is always printed in the place of the
2485 @var{number-separator} TAB@. The tabification depends upon the output
2488 @item -N @var{line_number}
2489 @itemx --first-line-number=@var{line_number}
2491 @opindex --first-line-number
2492 Start line counting with the number @var{line_number} at first line of
2493 first page printed (in most cases not the first line of the input file).
2495 @item -o @var{margin}
2496 @itemx --indent=@var{margin}
2499 @cindex indenting lines
2501 Indent each line with a margin @var{margin} spaces wide (default is zero).
2502 The total page width is the size of the margin plus the @var{page_width}
2503 set with the @option{-W/-w} option. A limited overflow may occur with
2504 numbered single column output (compare @option{-n} option).
2507 @itemx --no-file-warnings
2509 @opindex --no-file-warnings
2510 Do not print a warning message when an argument @var{file} cannot be
2511 opened. (The exit status will still be nonzero, however.)
2513 @item -s[@var{char}]
2514 @itemx --separator[=@var{char}]
2516 @opindex --separator
2517 Separate columns by a single character @var{char}. The default for
2518 @var{char} is the TAB character without @option{-w} and @samp{no
2519 character} with @option{-w}. Without @option{-s} the default separator
2520 @samp{space} is set. @option{-s[char]} turns off line truncation of all
2521 three column options (@option{-COLUMN}|@option{-a -COLUMN}|@option{-m}) unless
2522 @option{-w} is set. This is a POSIX-compliant formulation.
2525 @item -S[@var{string}]
2526 @itemx --sep-string[=@var{string}]
2528 @opindex --sep-string
2529 Use @var{string} to separate output columns. The @option{-S} option doesn't
2530 affect the @option{-W/-w} option, unlike the @option{-s} option which does. It
2531 does not affect line truncation or column alignment.
2532 Without @option{-S}, and with @option{-J}, @command{pr} uses the default output
2534 Without @option{-S} or @option{-J}, @command{pr} uses a @samp{space}
2535 (same as @option{-S"@w{ }"}).
2536 If no @samp{@var{string}} argument is specified, @samp{""} is assumed.
2539 @itemx --omit-header
2541 @opindex --omit-header
2542 Do not print the usual header [and footer] on each page, and do not fill
2543 out the bottom of pages (with blank lines or a form feed). No page
2544 structure is produced, but form feeds set in the input files are retained.
2545 The predefined pagination is not changed. @option{-t} or @option{-T} may be
2546 useful together with other options; e.g.: @option{-t -e4}, expand TAB characters
2547 in the input file to 4 spaces but don't make any other changes. Use of
2548 @option{-t} overrides @option{-h}.
2551 @itemx --omit-pagination
2553 @opindex --omit-pagination
2554 Do not print header [and footer]. In addition eliminate all form feeds
2555 set in the input files.
2558 @itemx --show-nonprinting
2560 @opindex --show-nonprinting
2561 Print nonprinting characters in octal backslash notation.
2563 @item -w @var{page_width}
2564 @itemx --width=@var{page_width}
2567 Set page width to @var{page_width} characters for multiple text-column
2568 output only (default for @var{page_width} is 72). @option{-s[CHAR]} turns
2569 off the default page width and any line truncation and column alignment.
2570 Lines of full length are merged, regardless of the column options
2571 set. No @var{page_width} setting is possible with single column output.
2572 A POSIX-compliant formulation.
2574 @item -W @var{page_width}
2575 @itemx --page_width=@var{page_width}
2577 @opindex --page_width
2578 Set the page width to @var{page_width} characters. That's valid with and
2579 without a column option. Text lines are truncated, unless @option{-J}
2580 is used. Together with one of the three column options
2581 (@option{-@var{column}}, @option{-a -@var{column}} or @option{-m}) column
2582 alignment is always used. The separator options @option{-S} or @option{-s}
2583 don't affect the @option{-W} option. Default is 72 characters. Without
2584 @option{-W @var{page_width}} and without any of the column options NO line
2585 truncation is used (defined to keep downward compatibility and to meet
2586 most frequent tasks). That's equivalent to @option{-W 72 -J}@. The header
2587 line is never truncated.
2594 @node fold invocation
2595 @section @command{fold}: Wrap input lines to fit in specified width
2598 @cindex wrapping long input lines
2599 @cindex folding long input lines
2601 @command{fold} writes each @var{file} (@option{-} means standard input), or
2602 standard input if none are given, to standard output, breaking long
2606 fold [@var{option}]@dots{} [@var{file}]@dots{}
2609 By default, @command{fold} breaks lines wider than 80 columns. The output
2610 is split into as many lines as necessary.
2612 @cindex screen columns
2613 @command{fold} counts screen columns by default; thus, a tab may count more
2614 than one column, backspace decreases the column count, and carriage
2615 return sets the column to zero.
2617 The program accepts the following options. Also see @ref{Common options}.
2625 Count bytes rather than columns, so that tabs, backspaces, and carriage
2626 returns are each counted as taking up one column, just like other
2633 Break at word boundaries: the line is broken after the last blank before
2634 the maximum line length. If the line contains no such blanks, the line
2635 is broken at the maximum line length as usual.
2637 @item -w @var{width}
2638 @itemx --width=@var{width}
2641 Use a maximum line length of @var{width} columns instead of 80.
2643 For compatibility @command{fold} supports an obsolete option syntax
2644 @option{-@var{width}}. New scripts should use @option{-w @var{width}}
2652 @node Output of parts of files
2653 @chapter Output of parts of files
2655 @cindex output of parts of files
2656 @cindex parts of files, output of
2658 These commands output pieces of the input.
2661 * head invocation:: Output the first part of files.
2662 * tail invocation:: Output the last part of files.
2663 * split invocation:: Split a file into pieces.
2664 * csplit invocation:: Split a file into context-determined pieces.
2667 @node head invocation
2668 @section @command{head}: Output the first part of files
2671 @cindex initial part of files, outputting
2672 @cindex first part of files, outputting
2674 @command{head} prints the first part (10 lines by default) of each
2675 @var{file}; it reads from standard input if no files are given or
2676 when given a @var{file} of @option{-}. Synopsis:
2679 head [@var{option}]@dots{} [@var{file}]@dots{}
2682 If more than one @var{file} is specified, @command{head} prints a
2683 one-line header consisting of:
2686 ==> @var{file name} <==
2690 before the output for each @var{file}.
2692 The program accepts the following options. Also see @ref{Common options}.
2697 @itemx --bytes=@var{k}
2700 Print the first @var{k} bytes, instead of initial lines.
2701 However, if @var{k} starts with a @samp{-},
2702 print all but the last @var{k} bytes of each file.
2703 @multiplierSuffixes{k}
2706 @itemx --lines=@var{k}
2709 Output the first @var{k} lines.
2710 However, if @var{k} starts with a @samp{-},
2711 print all but the last @var{k} lines of each file.
2712 Size multiplier suffixes are the same as with the @option{-c} option.
2720 Never print file name headers.
2726 Always print file name headers.
2730 For compatibility @command{head} also supports an obsolete option syntax
2731 @option{-@var{count}@var{options}}, which is recognized only if it is
2732 specified first. @var{count} is a decimal number optionally followed
2733 by a size letter (@samp{b}, @samp{k}, @samp{m}) as in @option{-c}, or
2734 @samp{l} to mean count by lines, or other option letters (@samp{cqv}).
2735 Scripts intended for standard hosts should use @option{-c @var{count}}
2736 or @option{-n @var{count}} instead. If your script must also run on
2737 hosts that support only the obsolete syntax, it is usually simpler to
2738 avoid @command{head}, e.g., by using @samp{sed 5q} instead of
2744 @node tail invocation
2745 @section @command{tail}: Output the last part of files
2748 @cindex last part of files, outputting
2750 @command{tail} prints the last part (10 lines by default) of each
2751 @var{file}; it reads from standard input if no files are given or
2752 when given a @var{file} of @samp{-}. Synopsis:
2755 tail [@var{option}]@dots{} [@var{file}]@dots{}
2758 If more than one @var{file} is specified, @command{tail} prints a
2759 one-line header consisting of:
2762 ==> @var{file name} <==
2766 before the output for each @var{file}.
2768 @cindex BSD @command{tail}
2769 @sc{gnu} @command{tail} can output any amount of data (some other versions of
2770 @command{tail} cannot). It also has no @option{-r} option (print in
2771 reverse), since reversing a file is really a different job from printing
2772 the end of a file; BSD @command{tail} (which is the one with @option{-r}) can
2773 only reverse files that are at most as large as its buffer, which is
2774 typically 32 KiB@. A more reliable and versatile way to reverse files is
2775 the @sc{gnu} @command{tac} command.
2777 The program accepts the following options. Also see @ref{Common options}.
2782 @itemx --bytes=@var{k}
2785 Output the last @var{k} bytes, instead of final lines.
2786 However, if @var{k} starts with a @samp{+}, start printing with the
2787 @var{k}th byte from the start of each file, instead of from the end.
2788 @multiplierSuffixes{k}
2791 @itemx --follow[=@var{how}]
2794 @cindex growing files
2795 @vindex name @r{follow option}
2796 @vindex descriptor @r{follow option}
2797 Loop forever trying to read more characters at the end of the file,
2798 presumably because the file is growing.
2799 If more than one file is given, @command{tail} prints a header whenever it
2800 gets output from a different file, to indicate which file that output is
2803 There are two ways to specify how you'd like to track files with this option,
2804 but that difference is noticeable only when a followed file is removed or
2806 If you'd like to continue to track the end of a growing file even after
2807 it has been unlinked, use @option{--follow=descriptor}. This is the default
2808 behavior, but it is not useful if you're tracking a log file that may be
2809 rotated (removed or renamed, then reopened). In that case, use
2810 @option{--follow=name} to track the named file, perhaps by reopening it
2811 periodically to see if it has been removed and recreated by some other program.
2812 Note that the inotify-based implementation handles this case without
2813 the need for any periodic reopening.
2815 No matter which method you use, if the tracked file is determined to have
2816 shrunk, @command{tail} prints a message saying the file has been truncated
2817 and resumes tracking the end of the file from the newly-determined endpoint.
2819 When a file is removed, @command{tail}'s behavior depends on whether it is
2820 following the name or the descriptor. When following by name, tail can
2821 detect that a file has been removed and gives a message to that effect,
2822 and if @option{--retry} has been specified it will continue checking
2823 periodically to see if the file reappears.
2824 When following a descriptor, tail does not detect that the file has
2825 been unlinked or renamed and issues no message; even though the file
2826 may no longer be accessible via its original name, it may still be
2829 The option values @samp{descriptor} and @samp{name} may be specified only
2830 with the long form of the option, not with @option{-f}.
2832 The @option{-f} option is ignored if
2833 no @var{file} operand is specified and standard input is a FIFO or a pipe.
2834 Likewise, the @option{-f} option has no effect for any
2835 operand specified as @samp{-}, when standard input is a FIFO or a pipe.
2837 With kernel inotify support, output is triggered by file changes
2838 and is generally very prompt.
2839 Otherwise, @command{tail} sleeps for one second between checks---
2840 use @option{--sleep-interval=@var{n}} to change that default---which can
2841 make the output appear slightly less responsive or bursty.
2842 When using tail without inotify support, you can make it more responsive
2843 by using a sub-second sleep interval, e.g., via an alias like this:
2846 alias tail='tail -s.1'
2851 This option is the same as @option{--follow=name --retry}. That is, tail
2852 will attempt to reopen a file when it is removed. Should this fail, tail
2853 will keep trying until it becomes accessible again.
2857 This option is useful mainly when following by name (i.e., with
2858 @option{--follow=name}).
2859 Without this option, when tail encounters a file that doesn't
2860 exist or is otherwise inaccessible, it reports that fact and
2861 never checks it again.
2863 @item --sleep-interval=@var{number}
2864 @opindex --sleep-interval
2865 Change the number of seconds to wait between iterations (the default is 1.0).
2866 During one iteration, every specified file is checked to see if it has
2868 Historical implementations of @command{tail} have required that
2869 @var{number} be an integer. However, GNU @command{tail} accepts
2870 an arbitrary floating point number. @xref{Floating point}.
2871 When @command{tail} uses inotify, this polling-related option
2872 is usually ignored. However, if you also specify @option{--pid=@var{p}},
2873 @command{tail} checks whether process @var{p} is alive at least
2874 every @var{number} seconds.
2876 @item --pid=@var{pid}
2878 When following by name or by descriptor, you may specify the process ID,
2879 @var{pid}, of the sole writer of all @var{file} arguments. Then, shortly
2880 after that process terminates, tail will also terminate. This will
2881 work properly only if the writer and the tailing process are running on
2882 the same machine. For example, to save the output of a build in a file
2883 and to watch the file grow, if you invoke @command{make} and @command{tail}
2884 like this then the tail process will stop when your build completes.
2885 Without this option, you would have had to kill the @code{tail -f}
2889 $ make >& makerr & tail --pid=$! -f makerr
2892 If you specify a @var{pid} that is not in use or that does not correspond
2893 to the process that is writing to the tailed files, then @command{tail}
2894 may terminate long before any @var{file}s stop growing or it may not
2895 terminate until long after the real writer has terminated.
2896 Note that @option{--pid} cannot be supported on some systems; @command{tail}
2897 will print a warning if this is the case.
2899 @item --max-unchanged-stats=@var{n}
2900 @opindex --max-unchanged-stats
2901 When tailing a file by name, if there have been @var{n} (default
2902 n=@value{DEFAULT_MAX_N_UNCHANGED_STATS_BETWEEN_OPENS}) consecutive
2903 iterations for which the file has not changed, then
2904 @code{open}/@code{fstat} the file to determine if that file name is
2905 still associated with the same device/inode-number pair as before.
2906 When following a log file that is rotated, this is approximately the
2907 number of seconds between when tail prints the last pre-rotation lines
2908 and when it prints the lines that have accumulated in the new log file.
2909 This option is meaningful only when polling (i.e., without inotify)
2910 and when following by name.
2913 @itemx --lines=@var{k}
2916 Output the last @var{k} lines.
2917 However, if @var{k} starts with a @samp{+}, start printing with the
2918 @var{k}th line from the start of each file, instead of from the end.
2919 Size multiplier suffixes are the same as with the @option{-c} option.
2927 Never print file name headers.
2933 Always print file name headers.
2937 For compatibility @command{tail} also supports an obsolete usage
2938 @samp{tail -[@var{count}][bcl][f] [@var{file}]}, which is recognized
2939 only if it does not conflict with the usage described
2940 above. This obsolete form uses exactly one option and at most one
2941 file. In the option, @var{count} is an optional decimal number optionally
2942 followed by a size letter (@samp{b}, @samp{c}, @samp{l}) to mean count
2943 by 512-byte blocks, bytes, or lines, optionally followed by @samp{f}
2944 which has the same meaning as @option{-f}.
2946 @vindex _POSIX2_VERSION
2947 On older systems, the leading @samp{-} can be replaced by @samp{+} in
2948 the obsolete option syntax with the same meaning as in counts, and
2949 obsolete usage overrides normal usage when the two conflict.
2950 This obsolete behavior can be enabled or disabled with the
2951 @env{_POSIX2_VERSION} environment variable (@pxref{Standards
2954 Scripts intended for use on standard hosts should avoid obsolete
2955 syntax and should use @option{-c @var{count}[b]}, @option{-n
2956 @var{count}}, and/or @option{-f} instead. If your script must also
2957 run on hosts that support only the obsolete syntax, you can often
2958 rewrite it to avoid problematic usages, e.g., by using @samp{sed -n
2959 '$p'} rather than @samp{tail -1}. If that's not possible, the script
2960 can use a test like @samp{if tail -c +1 </dev/null >/dev/null 2>&1;
2961 then @dots{}} to decide which syntax to use.
2963 Even if your script assumes the standard behavior, you should still
2964 beware usages whose behaviors differ depending on the POSIX
2965 version. For example, avoid @samp{tail - main.c}, since it might be
2966 interpreted as either @samp{tail main.c} or as @samp{tail -- -
2967 main.c}; avoid @samp{tail -c 4}, since it might mean either @samp{tail
2968 -c4} or @samp{tail -c 10 4}; and avoid @samp{tail +4}, since it might
2969 mean either @samp{tail ./+4} or @samp{tail -n +4}.
2974 @node split invocation
2975 @section @command{split}: Split a file into pieces.
2978 @cindex splitting a file into pieces
2979 @cindex pieces, splitting a file into
2981 @command{split} creates output files containing consecutive or interleaved
2982 sections of @var{input} (standard input if none is given or @var{input}
2983 is @samp{-}). Synopsis:
2986 split [@var{option}] [@var{input} [@var{prefix}]]
2989 By default, @command{split} puts 1000 lines of @var{input} (or whatever is
2990 left over for the last section), into each output file.
2992 @cindex output file name prefix
2993 The output files' names consist of @var{prefix} (@samp{x} by default)
2994 followed by a group of characters (@samp{aa}, @samp{ab}, @dots{} by
2995 default), such that concatenating the output files in traditional
2996 sorted order by file name produces the original input file (except
2997 @option{-nr/@var{n}}). By default split will initially create files
2998 with two generated suffix characters, and will increase this width by two
2999 when the next most significant position reaches the last character.
3000 (@samp{yz}, @samp{zaaa}, @samp{zaab}, @dots{}). In this way an arbitrary
3001 number of output files are supported, which sort as described above,
3002 even in the presence of an @option{--additional-suffix} option.
3003 If the @option{-a} option is specified and the output file names are
3004 exhausted, @command{split} reports an error without deleting the
3005 output files that it did create.
3007 The program accepts the following options. Also see @ref{Common options}.
3011 @item -l @var{lines}
3012 @itemx --lines=@var{lines}
3015 Put @var{lines} lines of @var{input} into each output file.
3017 For compatibility @command{split} also supports an obsolete
3018 option syntax @option{-@var{lines}}. New scripts should use
3019 @option{-l @var{lines}} instead.
3022 @itemx --bytes=@var{size}
3025 Put @var{size} bytes of @var{input} into each output file.
3026 @multiplierSuffixes{size}
3029 @itemx --line-bytes=@var{size}
3031 @opindex --line-bytes
3032 Put into each output file as many complete lines of @var{input} as
3033 possible without exceeding @var{size} bytes. Individual lines longer than
3034 @var{size} bytes are broken into multiple files.
3035 @var{size} has the same format as for the @option{--bytes} option.
3037 @item --filter=@var{command}
3039 With this option, rather than simply writing to each output file,
3040 write through a pipe to the specified shell @var{command} for each output file.
3041 @var{command} should use the $FILE environment variable, which is set
3042 to a different output file name for each invocation of the command.
3043 For example, imagine that you have a 1TiB compressed file
3044 that, if uncompressed, would be too large to reside on disk,
3045 yet you must split it into individually-compressed pieces
3046 of a more manageable size.
3047 To do that, you might run this command:
3050 xz -dc BIG.xz | split -b200G --filter='xz > $FILE.xz' - big-
3053 Assuming a 10:1 compression ratio, that would create about fifty 20GiB files
3054 with names @file{big-aa.xz}, @file{big-ab.xz}, @file{big-ac.xz}, etc.
3056 @item -n @var{chunks}
3057 @itemx --number=@var{chunks}
3061 Split @var{input} to @var{chunks} output files where @var{chunks} may be:
3064 @var{n} generate @var{n} files based on current size of @var{input}
3065 @var{k}/@var{n} only output @var{k}th of @var{n} to stdout
3066 l/@var{n} generate @var{n} files without splitting lines
3067 l/@var{k}/@var{n} likewise but only output @var{k}th of @var{n} to stdout
3068 r/@var{n} like @samp{l} but use round robin distribution
3069 r/@var{k}/@var{n} likewise but only output @var{k}th of @var{n} to stdout
3072 Any excess bytes remaining after dividing the @var{input}
3073 into @var{n} chunks, are assigned to the last chunk.
3074 Any excess bytes appearing after the initial calculation are discarded
3075 (except when using @samp{r} mode).
3077 All @var{n} files are created even if there are fewer than @var{n} lines,
3078 or the @var{input} is truncated.
3080 For @samp{l} mode, chunks are approximately @var{input} size / @var{n}.
3081 The @var{input} is partitioned into @var{n} equal sized portions, with
3082 the last assigned any excess. If a line @emph{starts} within a partition
3083 it is written completely to the corresponding file. Since lines
3084 are not split even if they overlap a partition, the files written
3085 can be larger or smaller than the partition size, and even empty
3086 if a line is so long as to completely overlap the partition.
3088 For @samp{r} mode, the size of @var{input} is irrelevant,
3089 and so can be a pipe for example.
3091 @item -a @var{length}
3092 @itemx --suffix-length=@var{length}
3094 @opindex --suffix-length
3095 Use suffixes of length @var{length}. If a @var{length} of 0 is specified,
3096 this is the same as if (any previous) @option{-a} was not specified, and
3097 thus enables the default behavior, which starts the suffix length at 2,
3098 and unless @option{-n} or @option{--numeric-suffixes=@var{from}} is
3099 specified, will auto increase the length by 2 as required.
3102 @itemx --numeric-suffixes[=@var{from}]
3104 @opindex --numeric-suffixes
3105 Use digits in suffixes rather than lower-case letters. The numerical
3106 suffix counts from @var{from} if specified, 0 otherwise.
3107 Note specifying a @var{from} value also disables the default
3108 auto suffix length expansion described above, and so you may also
3109 want to specify @option{-a} to allow suffixes beyond @samp{99}.
3111 @item --additional-suffix=@var{suffix}
3112 @opindex --additional-suffix
3113 Append an additional @var{suffix} to output file names. @var{suffix}
3114 must not contain slash.
3117 @itemx --elide-empty-files
3119 @opindex --elide-empty-files
3120 Suppress the generation of zero-length output files. This can happen
3121 with the @option{--number} option if a file is (truncated to be) shorter
3122 than the number requested, or if a line is so long as to completely
3123 span a chunk. The output file sequence numbers, always run consecutively
3124 even when this option is specified.
3129 @opindex --unbuffered
3130 Immediately copy input to output in @option{--number r/...} mode,
3131 which is a much slower mode of operation.
3135 Write a diagnostic just before each output file is opened.
3141 Here are a few examples to illustrate how the
3142 @option{--number} (@option{-n}) option works:
3144 Notice how, by default, one line may be split onto two or more:
3147 $ seq -w 6 10 > k; split -n3 k; head xa?
3160 Use the "l/" modifier to suppress that:
3163 $ seq -w 6 10 > k; split -nl/3 k; head xa?
3176 Use the "r/" modifier to distribute lines in a round-robin fashion:
3179 $ seq -w 6 10 > k; split -nr/3 k; head xa?
3192 You can also extract just the Kth chunk.
3193 This extracts and prints just the 7th "chunk" of 33:
3196 $ seq 100 > k; split -nl/7/33 k
3203 @node csplit invocation
3204 @section @command{csplit}: Split a file into context-determined pieces
3207 @cindex context splitting
3208 @cindex splitting a file into pieces by context
3210 @command{csplit} creates zero or more output files containing sections of
3211 @var{input} (standard input if @var{input} is @samp{-}). Synopsis:
3214 csplit [@var{option}]@dots{} @var{input} @var{pattern}@dots{}
3217 The contents of the output files are determined by the @var{pattern}
3218 arguments, as detailed below. An error occurs if a @var{pattern}
3219 argument refers to a nonexistent line of the input file (e.g., if no
3220 remaining line matches a given regular expression). After every
3221 @var{pattern} has been matched, any remaining input is copied into one
3224 By default, @command{csplit} prints the number of bytes written to each
3225 output file after it has been created.
3227 The types of pattern arguments are:
3232 Create an output file containing the input up to but not including line
3233 @var{n} (a positive integer). If followed by a repeat count, also
3234 create an output file containing the next @var{n} lines of the input
3235 file once for each repeat.
3237 @item /@var{regexp}/[@var{offset}]
3238 Create an output file containing the current line up to (but not
3239 including) the next line of the input file that contains a match for
3240 @var{regexp}. The optional @var{offset} is an integer.
3241 If it is given, the input up to (but not including) the
3242 matching line plus or minus @var{offset} is put into the output file,
3243 and the line after that begins the next section of input.
3245 @item %@var{regexp}%[@var{offset}]
3246 Like the previous type, except that it does not create an output
3247 file, so that section of the input file is effectively ignored.
3249 @item @{@var{repeat-count}@}
3250 Repeat the previous pattern @var{repeat-count} additional
3251 times. The @var{repeat-count} can either be a positive integer or an
3252 asterisk, meaning repeat as many times as necessary until the input is
3257 The output files' names consist of a prefix (@samp{xx} by default)
3258 followed by a suffix. By default, the suffix is an ascending sequence
3259 of two-digit decimal numbers from @samp{00} to @samp{99}. In any case,
3260 concatenating the output files in sorted order by file name produces the
3261 original input file.
3263 By default, if @command{csplit} encounters an error or receives a hangup,
3264 interrupt, quit, or terminate signal, it removes any output files
3265 that it has created so far before it exits.
3267 The program accepts the following options. Also see @ref{Common options}.
3271 @item -f @var{prefix}
3272 @itemx --prefix=@var{prefix}
3275 @cindex output file name prefix
3276 Use @var{prefix} as the output file name prefix.
3278 @item -b @var{suffix}
3279 @itemx --suffix=@var{suffix}
3282 @cindex output file name suffix
3283 Use @var{suffix} as the output file name suffix. When this option is
3284 specified, the suffix string must include exactly one
3285 @code{printf(3)}-style conversion specification, possibly including
3286 format specification flags, a field width, a precision specifications,
3287 or all of these kinds of modifiers. The format letter must convert a
3288 binary unsigned integer argument to readable form. The format letters
3289 @samp{d} and @samp{i} are aliases for @samp{u}, and the
3290 @samp{u}, @samp{o}, @samp{x}, and @samp{X} conversions are allowed. The
3291 entire @var{suffix} is given (with the current output file number) to
3292 @code{sprintf(3)} to form the file name suffixes for each of the
3293 individual output files in turn. If this option is used, the
3294 @option{--digits} option is ignored.
3296 @item -n @var{digits}
3297 @itemx --digits=@var{digits}
3300 Use output file names containing numbers that are @var{digits} digits
3301 long instead of the default 2.
3306 @opindex --keep-files
3307 Do not remove output files when errors are encountered.
3310 @itemx --elide-empty-files
3312 @opindex --elide-empty-files
3313 Suppress the generation of zero-length output files. (In cases where
3314 the section delimiters of the input file are supposed to mark the first
3315 lines of each of the sections, the first output file will generally be a
3316 zero-length file unless you use this option.) The output file sequence
3317 numbers always run consecutively starting from 0, even when this option
3328 Do not print counts of output file sizes.
3334 Here is an example of its usage.
3335 First, create an empty directory for the exercise,
3342 Now, split the sequence of 1..14 on lines that end with 0 or 5:
3345 $ seq 14 | csplit - '/[05]$/' '@{*@}'
3351 Each number printed above is the size of an output
3352 file that csplit has just created.
3353 List the names of those output files:
3360 Use @command{head} to show their contents:
3385 @node Summarizing files
3386 @chapter Summarizing files
3388 @cindex summarizing files
3390 These commands generate just a few numbers representing entire
3394 * wc invocation:: Print newline, word, and byte counts.
3395 * sum invocation:: Print checksum and block counts.
3396 * cksum invocation:: Print CRC checksum and byte counts.
3397 * md5sum invocation:: Print or check MD5 digests.
3398 * sha1sum invocation:: Print or check SHA-1 digests.
3399 * sha2 utilities:: Print or check SHA-2 digests.
3404 @section @command{wc}: Print newline, word, and byte counts
3408 @cindex character count
3412 @command{wc} counts the number of bytes, characters, whitespace-separated
3413 words, and newlines in each given @var{file}, or standard input if none
3414 are given or for a @var{file} of @samp{-}. Synopsis:
3417 wc [@var{option}]@dots{} [@var{file}]@dots{}
3420 @cindex total counts
3421 @command{wc} prints one line of counts for each file, and if the file was
3422 given as an argument, it prints the file name following the counts. If
3423 more than one @var{file} is given, @command{wc} prints a final line
3424 containing the cumulative counts, with the file name @file{total}. The
3425 counts are printed in this order: newlines, words, characters, bytes,
3426 maximum line length.
3427 Each count is printed right-justified in a field with at least one
3428 space between fields so that the numbers and file names normally line
3429 up nicely in columns. The width of the count fields varies depending
3430 on the inputs, so you should not depend on a particular field width.
3431 However, as a GNU extension, if only one count is printed,
3432 it is guaranteed to be printed without leading spaces.
3434 By default, @command{wc} prints three counts: the newline, words, and byte
3435 counts. Options can specify that only certain counts be printed.
3436 Options do not undo others previously given, so
3443 prints both the byte counts and the word counts.
3445 With the @option{--max-line-length} option, @command{wc} prints the length
3446 of the longest line per file, and if there is more than one file it
3447 prints the maximum (not the sum) of those lengths. The line lengths here
3448 are measured in screen columns, according to the current locale and
3449 assuming tab positions in every 8th column.
3451 The program accepts the following options. Also see @ref{Common options}.
3459 Print only the byte counts.
3465 Print only the character counts.
3471 Print only the word counts.
3477 Print only the newline counts.
3480 @itemx --max-line-length
3482 @opindex --max-line-length
3483 Print only the maximum line lengths.
3485 @macro filesZeroFromOption{cmd,withTotalOption,subListOutput}
3486 @item --files0-from=@var{file}
3487 @opindex --files0-from=@var{file}
3488 @c This is commented out to avoid a texi2dvi failure.
3489 @c texi2dvi (GNU Texinfo 4.11) 1.104
3490 @c @cindex including files from @command{\cmd\}
3491 Disallow processing files named on the command line, and instead process
3492 those named in file @var{file}; each name being terminated by a zero byte
3494 This is useful \withTotalOption\
3495 when the list of file names is so long that it may exceed a command line
3497 In such cases, running @command{\cmd\} via @command{xargs} is undesirable
3498 because it splits the list into pieces and makes @command{\cmd\} print
3499 \subListOutput\ for each sublist rather than for the entire list.
3500 One way to produce a list of ASCII @sc{nul} terminated file
3501 names is with @sc{gnu}
3502 @command{find}, using its @option{-print0} predicate.
3503 If @var{file} is @samp{-} then the ASCII @sc{nul} terminated
3504 file names are read from standard input.
3506 @filesZeroFromOption{wc,,a total}
3508 For example, to find the length of the longest line in any @file{.c} or
3509 @file{.h} file in the current hierarchy, do this:
3512 find . -name '*.[ch]' -print0 |
3513 wc -L --files0-from=- | tail -n1
3521 @node sum invocation
3522 @section @command{sum}: Print checksum and block counts
3525 @cindex 16-bit checksum
3526 @cindex checksum, 16-bit
3528 @command{sum} computes a 16-bit checksum for each given @var{file}, or
3529 standard input if none are given or for a @var{file} of @samp{-}. Synopsis:
3532 sum [@var{option}]@dots{} [@var{file}]@dots{}
3535 @command{sum} prints the checksum for each @var{file} followed by the
3536 number of blocks in the file (rounded up). If more than one @var{file}
3537 is given, file names are also printed (by default). (With the
3538 @option{--sysv} option, corresponding file names are printed when there is
3539 at least one file argument.)
3541 By default, @sc{gnu} @command{sum} computes checksums using an algorithm
3542 compatible with BSD @command{sum} and prints file sizes in units of
3545 The program accepts the following options. Also see @ref{Common options}.
3551 @cindex BSD @command{sum}
3552 Use the default (BSD compatible) algorithm. This option is included for
3553 compatibility with the System V @command{sum}. Unless @option{-s} was also
3554 given, it has no effect.
3560 @cindex System V @command{sum}
3561 Compute checksums using an algorithm compatible with System V
3562 @command{sum}'s default, and print file sizes in units of 512-byte blocks.
3566 @command{sum} is provided for compatibility; the @command{cksum} program (see
3567 next section) is preferable in new applications.
3572 @node cksum invocation
3573 @section @command{cksum}: Print CRC checksum and byte counts
3576 @cindex cyclic redundancy check
3577 @cindex CRC checksum
3579 @command{cksum} computes a cyclic redundancy check (CRC) checksum for each
3580 given @var{file}, or standard input if none are given or for a
3581 @var{file} of @samp{-}. Synopsis:
3584 cksum [@var{option}]@dots{} [@var{file}]@dots{}
3587 @command{cksum} prints the CRC checksum for each file along with the number
3588 of bytes in the file, and the file name unless no arguments were given.
3590 @command{cksum} is typically used to ensure that files
3591 transferred by unreliable means (e.g., netnews) have not been corrupted,
3592 by comparing the @command{cksum} output for the received files with the
3593 @command{cksum} output for the original files (typically given in the
3596 The CRC algorithm is specified by the POSIX standard. It is not
3597 compatible with the BSD or System V @command{sum} algorithms (see the
3598 previous section); it is more robust.
3600 The only options are @option{--help} and @option{--version}. @xref{Common
3606 @node md5sum invocation
3607 @section @command{md5sum}: Print or check MD5 digests
3611 @cindex 128-bit checksum
3612 @cindex checksum, 128-bit
3613 @cindex fingerprint, 128-bit
3614 @cindex message-digest, 128-bit
3616 @command{md5sum} computes a 128-bit checksum (or @dfn{fingerprint} or
3617 @dfn{message-digest}) for each specified @var{file}.
3619 Note: The MD5 digest is more reliable than a simple CRC (provided by
3620 the @command{cksum} command) for detecting accidental file corruption,
3621 as the chances of accidentally having two files with identical MD5
3622 are vanishingly small. However, it should not be considered secure
3623 against malicious tampering: although finding a file with a given MD5
3624 fingerprint is considered infeasible at the moment, it is known how
3625 to modify certain files, including digital certificates, so that they
3626 appear valid when signed with an MD5 digest.
3627 For more secure hashes, consider using SHA-2. @xref{sha2 utilities}.
3629 If a @var{file} is specified as @samp{-} or if no files are given
3630 @command{md5sum} computes the checksum for the standard input.
3631 @command{md5sum} can also determine whether a file and checksum are
3632 consistent. Synopsis:
3635 md5sum [@var{option}]@dots{} [@var{file}]@dots{}
3638 For each @var{file}, @samp{md5sum} outputs the MD5 checksum, a flag
3639 indicating binary or text input mode, and the file name.
3640 If @var{file} contains a backslash or newline, the
3641 line is started with a backslash, and each problematic character in
3642 the file name is escaped with a backslash, making the output
3643 unambiguous even in the presence of arbitrary file names.
3644 If @var{file} is omitted or specified as @samp{-}, standard input is read.
3646 The program accepts the following options. Also see @ref{Common options}.
3654 @cindex binary input files
3655 Treat each input file as binary, by reading it in binary mode and
3656 outputting a @samp{*} flag. This is the inverse of @option{--text}.
3657 On systems like GNU that do not distinguish between binary
3658 and text files, this option merely flags each input mode as binary:
3659 the MD5 checksum is unaffected. This option is the default on systems
3660 like MS-DOS that distinguish between binary and text files, except
3661 for reading standard input when standard input is a terminal.
3665 Read file names and checksum information (not data) from each
3666 @var{file} (or from stdin if no @var{file} was specified) and report
3667 whether the checksums match the contents of the named files.
3668 The input to this mode of @command{md5sum} is usually the output of
3669 a prior, checksum-generating run of @samp{md5sum}.
3670 Each valid line of input consists of an MD5 checksum, a binary/text
3671 flag, and then a file name.
3672 Binary mode is indicated with @samp{*}, text with @samp{ } (space).
3673 For each such line, @command{md5sum} reads the named file and computes its
3674 MD5 checksum. Then, if the computed message digest does not match the
3675 one on the line with the file name, the file is noted as having
3676 failed the test. Otherwise, the file passes the test.
3677 By default, for each valid line, one line is written to standard
3678 output indicating whether the named file passed the test.
3679 After all checks have been performed, if there were any failures,
3680 a warning is issued to standard error.
3681 Use the @option{--status} option to inhibit that output.
3682 If any listed file cannot be opened or read, if any valid line has
3683 an MD5 checksum inconsistent with the associated file, or if no valid
3684 line is found, @command{md5sum} exits with nonzero status. Otherwise,
3685 it exits successfully.
3689 @cindex verifying MD5 checksums
3690 This option is useful only when verifying checksums.
3691 When verifying checksums, don't generate an 'OK' message per successfully
3692 checked file. Files that fail the verification are reported in the
3693 default one-line-per-file format. If there is any checksum mismatch,
3694 print a warning summarizing the failures to standard error.
3698 @cindex verifying MD5 checksums
3699 This option is useful only when verifying checksums.
3700 When verifying checksums, don't generate the default one-line-per-file
3701 diagnostic and don't output the warning summarizing any failures.
3702 Failures to open or read a file still evoke individual diagnostics to
3704 If all listed files are readable and are consistent with the associated
3705 MD5 checksums, exit successfully. Otherwise exit with a status code
3706 indicating there was a failure.
3711 Output BSD style checksums, which indicate the checksum algorithm used.
3712 As a GNU extension, file names with problematic characters
3713 are escaped as described above, with the same escaping indicator of @samp{\}
3714 at the start of the line, being used.
3715 The @option{--tag} option implies binary mode, and is disallowed with
3716 @option{--text} mode as supporting that would unnecessarily complicate
3717 the output format, while providing little benefit.
3723 @cindex text input files
3724 Treat each input file as text, by reading it in text mode and
3725 outputting a @samp{ } flag. This is the inverse of @option{--binary}.
3726 This option is the default on systems like GNU that do not
3727 distinguish between binary and text files. On other systems, it is
3728 the default for reading standard input when standard input is a
3729 terminal. This mode is never defaulted to if @option{--tag} is used.
3735 @cindex verifying MD5 checksums
3736 When verifying checksums, warn about improperly formatted MD5 checksum lines.
3737 This option is useful only if all but a few lines in the checked input
3742 @cindex verifying MD5 checksums
3743 When verifying checksums,
3744 if one or more input line is invalid,
3745 exit nonzero after all warnings have been issued.
3752 @node sha1sum invocation
3753 @section @command{sha1sum}: Print or check SHA-1 digests
3757 @cindex 160-bit checksum
3758 @cindex checksum, 160-bit
3759 @cindex fingerprint, 160-bit
3760 @cindex message-digest, 160-bit
3762 @command{sha1sum} computes a 160-bit checksum for each specified
3763 @var{file}. The usage and options of this command are precisely the
3764 same as for @command{md5sum}. @xref{md5sum invocation}.
3766 Note: The SHA-1 digest is more secure than MD5, and no collisions of
3767 it are known (different files having the same fingerprint). However,
3768 it is known that they can be produced with considerable, but not
3769 unreasonable, resources. For this reason, it is generally considered
3770 that SHA-1 should be gradually phased out in favor of the more secure
3771 SHA-2 hash algorithms. @xref{sha2 utilities}.
3774 @node sha2 utilities
3775 @section sha2 utilities: Print or check SHA-2 digests
3782 @cindex 224-bit checksum
3783 @cindex 256-bit checksum
3784 @cindex 384-bit checksum
3785 @cindex 512-bit checksum
3786 @cindex checksum, 224-bit
3787 @cindex checksum, 256-bit
3788 @cindex checksum, 384-bit
3789 @cindex checksum, 512-bit
3790 @cindex fingerprint, 224-bit
3791 @cindex fingerprint, 256-bit
3792 @cindex fingerprint, 384-bit
3793 @cindex fingerprint, 512-bit
3794 @cindex message-digest, 224-bit
3795 @cindex message-digest, 256-bit
3796 @cindex message-digest, 384-bit
3797 @cindex message-digest, 512-bit
3799 The commands @command{sha224sum}, @command{sha256sum},
3800 @command{sha384sum} and @command{sha512sum} compute checksums of
3801 various lengths (respectively 224, 256, 384 and 512 bits),
3802 collectively known as the SHA-2 hashes. The usage and options of
3803 these commands are precisely the same as for @command{md5sum}.
3804 @xref{md5sum invocation}.
3806 Note: The SHA384 and SHA512 digests are considerably slower to
3807 compute, especially on 32-bit computers, than SHA224 or SHA256.
3810 @node Operating on sorted files
3811 @chapter Operating on sorted files
3813 @cindex operating on sorted files
3814 @cindex sorted files, operations on
3816 These commands work with (or produce) sorted files.
3819 * sort invocation:: Sort text files.
3820 * shuf invocation:: Shuffle text files.
3821 * uniq invocation:: Uniquify files.
3822 * comm invocation:: Compare two sorted files line by line.
3823 * ptx invocation:: Produce a permuted index of file contents.
3824 * tsort invocation:: Topological sort.
3828 @node sort invocation
3829 @section @command{sort}: Sort text files
3832 @cindex sorting files
3834 @command{sort} sorts, merges, or compares all the lines from the given
3835 files, or standard input if none are given or for a @var{file} of
3836 @samp{-}. By default, @command{sort} writes the results to standard
3840 sort [@var{option}]@dots{} [@var{file}]@dots{}
3843 @command{sort} has three modes of operation: sort (the default), merge,
3844 and check for sortedness. The following options change the operation
3851 @itemx --check=diagnose-first
3854 @cindex checking for sortedness
3855 Check whether the given file is already sorted: if it is not all
3856 sorted, print a diagnostic containing the first out-of-order line and
3857 exit with a status of 1.
3858 Otherwise, exit successfully.
3859 At most one input file can be given.
3862 @itemx --check=quiet
3863 @itemx --check=silent
3866 @cindex checking for sortedness
3867 Exit successfully if the given file is already sorted, and
3868 exit with status 1 otherwise.
3869 At most one input file can be given.
3870 This is like @option{-c}, except it does not print a diagnostic.
3876 @cindex merging sorted files
3877 Merge the given files by sorting them as a group. Each input file must
3878 always be individually sorted. It always works to sort instead of
3879 merge; merging is provided because it is faster, in the case where it
3884 @cindex sort stability
3885 @cindex sort's last-resort comparison
3886 A pair of lines is compared as follows:
3887 @command{sort} compares each pair of fields, in the
3888 order specified on the command line, according to the associated
3889 ordering options, until a difference is found or no fields are left.
3890 If no key fields are specified, @command{sort} uses a default key of
3891 the entire line. Finally, as a last resort when all keys compare
3892 equal, @command{sort} compares entire lines as if no ordering options
3893 other than @option{--reverse} (@option{-r}) were specified. The
3894 @option{--stable} (@option{-s}) option disables this @dfn{last-resort
3895 comparison} so that lines in which all fields compare equal are left
3896 in their original relative order. The @option{--unique}
3897 (@option{-u}) option also disables the last-resort comparison.
3901 Unless otherwise specified, all comparisons use the character collating
3902 sequence specified by the @env{LC_COLLATE} locale.@footnote{If you
3903 use a non-POSIX locale (e.g., by setting @env{LC_ALL}
3904 to @samp{en_US}), then @command{sort} may produce output that is sorted
3905 differently than you're accustomed to. In that case, set the @env{LC_ALL}
3906 environment variable to @samp{C}@. Note that setting only @env{LC_COLLATE}
3907 has two problems. First, it is ineffective if @env{LC_ALL} is also set.
3908 Second, it has undefined behavior if @env{LC_CTYPE} (or @env{LANG}, if
3909 @env{LC_CTYPE} is unset) is set to an incompatible value. For example,
3910 you get undefined behavior if @env{LC_CTYPE} is @code{ja_JP.PCK} but
3911 @env{LC_COLLATE} is @code{en_US.UTF-8}.}
3913 @sc{gnu} @command{sort} (as specified for all @sc{gnu} utilities) has no
3914 limit on input line length or restrictions on bytes allowed within lines.
3915 In addition, if the final byte of an input file is not a newline, @sc{gnu}
3916 @command{sort} silently supplies one. A line's trailing newline is not
3917 part of the line for comparison purposes.
3919 @cindex exit status of @command{sort}
3923 0 if no error occurred
3924 1 if invoked with @option{-c} or @option{-C} and the input is not sorted
3925 2 if an error occurred
3929 If the environment variable @env{TMPDIR} is set, @command{sort} uses its
3930 value as the directory for temporary files instead of @file{/tmp}. The
3931 @option{--temporary-directory} (@option{-T}) option in turn overrides
3932 the environment variable.
3934 The following options affect the ordering of output lines. They may be
3935 specified globally or as part of a specific key field. If no key
3936 fields are specified, global options apply to comparison of entire
3937 lines; otherwise the global options are inherited by key fields that do
3938 not specify any special options of their own. In pre-POSIX
3939 versions of @command{sort}, global options affect only later key fields,
3940 so portable shell scripts should specify global options first.
3945 @itemx --ignore-leading-blanks
3947 @opindex --ignore-leading-blanks
3948 @cindex blanks, ignoring leading
3950 Ignore leading blanks when finding sort keys in each line.
3951 By default a blank is a space or a tab, but the @env{LC_CTYPE} locale
3952 can change this. Note blanks may be ignored by your locale's collating
3953 rules, but without this option they will be significant for character
3954 positions specified in keys with the @option{-k} option.
3957 @itemx --dictionary-order
3959 @opindex --dictionary-order
3960 @cindex dictionary order
3961 @cindex phone directory order
3962 @cindex telephone directory order
3964 Sort in @dfn{phone directory} order: ignore all characters except
3965 letters, digits and blanks when sorting.
3966 By default letters and digits are those of ASCII and a blank
3967 is a space or a tab, but the @env{LC_CTYPE} locale can change this.
3970 @itemx --ignore-case
3972 @opindex --ignore-case
3973 @cindex ignoring case
3974 @cindex case folding
3976 Fold lowercase characters into the equivalent uppercase characters when
3977 comparing so that, for example, @samp{b} and @samp{B} sort as equal.
3978 The @env{LC_CTYPE} locale determines character types.
3979 When used with @option{--unique} those lower case equivalent lines are
3980 thrown away. (There is currently no way to throw away the upper case
3981 equivalent instead. (Any @option{--reverse} given would only affect
3982 the final result, after the throwing away.))
3985 @itemx --general-numeric-sort
3986 @itemx --sort=general-numeric
3988 @opindex --general-numeric-sort
3990 @cindex general numeric sort
3992 Sort numerically, converting a prefix of each line to a long
3993 double-precision floating point number. @xref{Floating point}.
3994 Do not report overflow, underflow, or conversion errors.
3995 Use the following collating sequence:
3999 Lines that do not start with numbers (all considered to be equal).
4001 NaNs (``Not a Number'' values, in IEEE floating point arithmetic)
4002 in a consistent but machine-dependent order.
4006 Finite numbers in ascending numeric order (with @math{-0} and @math{+0} equal).
4011 Use this option only if there is no alternative; it is much slower than
4012 @option{--numeric-sort} (@option{-n}) and it can lose information when
4013 converting to floating point.
4016 @itemx --human-numeric-sort
4017 @itemx --sort=human-numeric
4019 @opindex --human-numeric-sort
4021 @cindex human numeric sort
4023 Sort numerically, first by numeric sign (negative, zero, or positive);
4024 then by SI suffix (either empty, or @samp{k} or @samp{K}, or
4025 one of @samp{MGTPEZY}, in that order; @pxref{Block size}); and finally
4026 by numeric value. For example, @samp{1023M} sorts before @samp{1G}
4027 because @samp{M} (mega) precedes @samp{G} (giga) as an SI
4028 suffix. This option sorts values that are consistently scaled to the
4029 nearest suffix, regardless of whether suffixes denote powers of 1000
4030 or 1024, and it therefore sorts the output of any single invocation of
4031 the @command{df}, @command{du}, or @command{ls} commands that are
4032 invoked with their @option{--human-readable} or @option{--si} options.
4033 The syntax for numbers is the same as for the @option{--numeric-sort}
4034 option; the SI suffix must immediately follow the number.
4037 @itemx --ignore-nonprinting
4039 @opindex --ignore-nonprinting
4040 @cindex nonprinting characters, ignoring
4041 @cindex unprintable characters, ignoring
4043 Ignore nonprinting characters.
4044 The @env{LC_CTYPE} locale determines character types.
4045 This option has no effect if the stronger @option{--dictionary-order}
4046 (@option{-d}) option is also given.
4052 @opindex --month-sort
4054 @cindex months, sorting by
4056 An initial string, consisting of any amount of blanks, followed
4057 by a month name abbreviation, is folded to UPPER case and
4058 compared in the order @samp{JAN} < @samp{FEB} < @dots{} < @samp{DEC}@.
4059 Invalid names compare low to valid names. The @env{LC_TIME} locale
4060 category determines the month spellings.
4061 By default a blank is a space or a tab, but the @env{LC_CTYPE} locale
4065 @itemx --numeric-sort
4066 @itemx --sort=numeric
4068 @opindex --numeric-sort
4070 @cindex numeric sort
4072 Sort numerically. The number begins each line and consists
4073 of optional blanks, an optional @samp{-} sign, and zero or more
4074 digits possibly separated by thousands separators, optionally followed
4075 by a decimal-point character and zero or more digits. An empty
4076 number is treated as @samp{0}. The @env{LC_NUMERIC}
4077 locale specifies the decimal-point character and thousands separator.
4078 By default a blank is a space or a tab, but the @env{LC_CTYPE} locale
4081 Comparison is exact; there is no rounding error.
4083 Neither a leading @samp{+} nor exponential notation is recognized.
4084 To compare such strings numerically, use the
4085 @option{--general-numeric-sort} (@option{-g}) option.
4088 @itemx --version-sort
4090 @opindex --version-sort
4091 @cindex version number sort
4092 Sort by version name and number. It behaves like a standard sort,
4093 except that each sequence of decimal digits is treated numerically
4094 as an index/version number. (@xref{Details about version sort}.)
4100 @cindex reverse sorting
4101 Reverse the result of comparison, so that lines with greater key values
4102 appear earlier in the output instead of later.
4105 @itemx --random-sort
4106 @itemx --sort=random
4108 @opindex --random-sort
4111 Sort by hashing the input keys and then sorting the hash values.
4112 Choose the hash function at random, ensuring that it is free of
4113 collisions so that differing keys have differing hash values. This is
4114 like a random permutation of the inputs (@pxref{shuf invocation}),
4115 except that keys with the same value sort together.
4117 If multiple random sort fields are specified, the same random hash
4118 function is used for all fields. To use different random hash
4119 functions for different fields, you can invoke @command{sort} more
4122 The choice of hash function is affected by the
4123 @option{--random-source} option.
4131 @item --compress-program=@var{prog}
4132 Compress any temporary files with the program @var{prog}.
4134 With no arguments, @var{prog} must compress standard input to standard
4135 output, and when given the @option{-d} option it must decompress
4136 standard input to standard output.
4138 Terminate with an error if @var{prog} exits with nonzero status.
4140 White space and the backslash character should not appear in
4141 @var{prog}; they are reserved for future use.
4143 @filesZeroFromOption{sort,,sorted output}
4145 @item -k @var{pos1}[,@var{pos2}]
4146 @itemx --key=@var{pos1}[,@var{pos2}]
4150 Specify a sort field that consists of the part of the line between
4151 @var{pos1} and @var{pos2} (or the end of the line, if @var{pos2} is
4152 omitted), @emph{inclusive}.
4154 Each @var{pos} has the form @samp{@var{f}[.@var{c}][@var{opts}]},
4155 where @var{f} is the number of the field to use, and @var{c} is the number
4156 of the first character from the beginning of the field. Fields and character
4157 positions are numbered starting with 1; a character position of zero in
4158 @var{pos2} indicates the field's last character. If @samp{.@var{c}} is
4159 omitted from @var{pos1}, it defaults to 1 (the beginning of the field);
4160 if omitted from @var{pos2}, it defaults to 0 (the end of the field).
4161 @var{opts} are ordering options, allowing individual keys to be sorted
4162 according to different rules; see below for details. Keys can span
4165 Example: To sort on the second field, use @option{--key=2,2}
4166 (@option{-k 2,2}). See below for more notes on keys and more examples.
4167 See also the @option{--debug} option to help determine the part
4168 of the line being used in the sort.
4171 Highlight the portion of each line used for sorting.
4172 Also issue warnings about questionable usage to stderr.
4174 @item --batch-size=@var{nmerge}
4175 @opindex --batch-size
4176 @cindex number of inputs to merge, nmerge
4177 Merge at most @var{nmerge} inputs at once.
4179 When @command{sort} has to merge more than @var{nmerge} inputs,
4180 it merges them in groups of @var{nmerge}, saving the result in
4181 a temporary file, which is then used as an input in a subsequent merge.
4183 A large value of @var{nmerge} may improve merge performance and decrease
4184 temporary storage utilization at the expense of increased memory usage
4185 and I/O@. Conversely a small value of @var{nmerge} may reduce memory
4186 requirements and I/O at the expense of temporary storage consumption and
4189 The value of @var{nmerge} must be at least 2. The default value is
4190 currently 16, but this is implementation-dependent and may change in
4193 The value of @var{nmerge} may be bounded by a resource limit for open
4194 file descriptors. The commands @samp{ulimit -n} or @samp{getconf
4195 OPEN_MAX} may display limits for your systems; these limits may be
4196 modified further if your program already has some files open, or if
4197 the operating system has other limits on the number of open files. If
4198 the value of @var{nmerge} exceeds the resource limit, @command{sort}
4199 silently uses a smaller value.
4201 @item -o @var{output-file}
4202 @itemx --output=@var{output-file}
4205 @cindex overwriting of input, allowed
4206 Write output to @var{output-file} instead of standard output.
4207 Normally, @command{sort} reads all input before opening
4208 @var{output-file}, so you can safely sort a file in place by using
4209 commands like @code{sort -o F F} and @code{cat F | sort -o F}@.
4210 However, @command{sort} with @option{--merge} (@option{-m}) can open
4211 the output file before reading all input, so a command like @code{cat
4212 F | sort -m -o F - G} is not safe as @command{sort} might start
4213 writing @file{F} before @command{cat} is done reading it.
4215 @vindex POSIXLY_CORRECT
4216 On newer systems, @option{-o} cannot appear after an input file if
4217 @env{POSIXLY_CORRECT} is set, e.g., @samp{sort F -o F}@. Portable
4218 scripts should specify @option{-o @var{output-file}} before any input
4221 @item --random-source=@var{file}
4222 @opindex --random-source
4223 @cindex random source for sorting
4224 Use @var{file} as a source of random data used to determine which
4225 random hash function to use with the @option{-R} option. @xref{Random
4232 @cindex sort stability
4233 @cindex sort's last-resort comparison
4235 Make @command{sort} stable by disabling its last-resort comparison.
4236 This option has no effect if no fields or global ordering options
4237 other than @option{--reverse} (@option{-r}) are specified.
4240 @itemx --buffer-size=@var{size}
4242 @opindex --buffer-size
4243 @cindex size for main memory sorting
4244 Use a main-memory sort buffer of the given @var{size}. By default,
4245 @var{size} is in units of 1024 bytes. Appending @samp{%} causes
4246 @var{size} to be interpreted as a percentage of physical memory.
4247 Appending @samp{K} multiplies @var{size} by 1024 (the default),
4248 @samp{M} by 1,048,576, @samp{G} by 1,073,741,824, and so on for
4249 @samp{T}, @samp{P}, @samp{E}, @samp{Z}, and @samp{Y}@. Appending
4250 @samp{b} causes @var{size} to be interpreted as a byte count, with no
4253 This option can improve the performance of @command{sort} by causing it
4254 to start with a larger or smaller sort buffer than the default.
4255 However, this option affects only the initial buffer size. The buffer
4256 grows beyond @var{size} if @command{sort} encounters input lines larger
4259 @item -t @var{separator}
4260 @itemx --field-separator=@var{separator}
4262 @opindex --field-separator
4263 @cindex field separator character
4264 Use character @var{separator} as the field separator when finding the
4265 sort keys in each line. By default, fields are separated by the empty
4266 string between a non-blank character and a blank character.
4267 By default a blank is a space or a tab, but the @env{LC_CTYPE} locale
4270 That is, given the input line @w{@samp{ foo bar}}, @command{sort} breaks it
4271 into fields @w{@samp{ foo}} and @w{@samp{ bar}}. The field separator is
4272 not considered to be part of either the field preceding or the field
4273 following, so with @samp{sort @w{-t " "}} the same input line has
4274 three fields: an empty field, @samp{foo}, and @samp{bar}.
4275 However, fields that extend to the end of the line,
4276 as @option{-k 2}, or fields consisting of a range, as @option{-k 2,3},
4277 retain the field separators present between the endpoints of the range.
4279 To specify ASCII @sc{nul} as the field separator,
4280 use the two-character string @samp{\0}, e.g., @samp{sort -t '\0'}.
4282 @item -T @var{tempdir}
4283 @itemx --temporary-directory=@var{tempdir}
4285 @opindex --temporary-directory
4286 @cindex temporary directory
4288 Use directory @var{tempdir} to store temporary files, overriding the
4289 @env{TMPDIR} environment variable. If this option is given more than
4290 once, temporary files are stored in all the directories given. If you
4291 have a large sort or merge that is I/O-bound, you can often improve
4292 performance by using this option to specify directories on different
4293 disks and controllers.
4295 @item --parallel=@var{n}
4297 @cindex multithreaded sort
4298 Set the number of sorts run in parallel to @var{n}. By default,
4299 @var{n} is set to the number of available processors, but limited
4300 to 8, as there are diminishing performance gains after that.
4301 Note also that using @var{n} threads increases the memory usage by
4302 a factor of log @var{n}. Also see @ref{nproc invocation}.
4308 @cindex uniquifying output
4310 Normally, output only the first of a sequence of lines that compare
4311 equal. For the @option{--check} (@option{-c} or @option{-C}) option,
4312 check that no pair of consecutive lines compares equal.
4314 This option also disables the default last-resort comparison.
4316 The commands @code{sort -u} and @code{sort | uniq} are equivalent, but
4317 this equivalence does not extend to arbitrary @command{sort} options.
4318 For example, @code{sort -n -u} inspects only the value of the initial
4319 numeric string when checking for uniqueness, whereas @code{sort -n |
4320 uniq} inspects the entire line. @xref{uniq invocation}.
4322 @macro zeroTerminatedOption
4324 @itemx --zero-terminated
4326 @opindex --zero-terminated
4327 @cindex process zero-terminated items
4328 Delimit items with a zero byte rather than a newline (ASCII @sc{lf}).
4329 I.e., treat input as items separated by ASCII @sc{nul}
4330 and terminate output items with ASCII @sc{nul}.
4331 This option can be useful in conjunction with @samp{perl -0} or
4332 @samp{find -print0} and @samp{xargs -0} which do the same in order to
4333 reliably handle arbitrary file names (even those containing blanks
4334 or other special characters).
4336 @zeroTerminatedOption
4340 Historical (BSD and System V) implementations of @command{sort} have
4341 differed in their interpretation of some options, particularly
4342 @option{-b}, @option{-f}, and @option{-n}.
4343 @sc{gnu} sort follows the POSIX
4344 behavior, which is usually (but not always!) like the System V behavior.
4345 According to POSIX, @option{-n} no longer implies @option{-b}. For
4346 consistency, @option{-M} has been changed in the same way. This may
4347 affect the meaning of character positions in field specifications in
4348 obscure cases. The only fix is to add an explicit @option{-b}.
4350 A position in a sort field specified with @option{-k} may have any
4351 of the option letters @samp{MbdfghinRrV} appended to it, in which case no
4352 global ordering options are inherited by that particular field. The
4353 @option{-b} option may be independently attached to either or both of
4354 the start and end positions of a field specification, and if it is
4355 inherited from the global options it will be attached to both.
4356 If input lines can contain leading or adjacent blanks and @option{-t}
4357 is not used, then @option{-k} is typically combined with @option{-b} or
4358 an option that implicitly ignores leading blanks (@samp{Mghn}) as otherwise
4359 the varying numbers of leading blanks in fields can cause confusing results.
4361 If the start position in a sort field specifier falls after the end of
4362 the line or after the end field, the field is empty. If the @option{-b}
4363 option was specified, the @samp{.@var{c}} part of a field specification
4364 is counted from the first nonblank character of the field.
4366 @vindex _POSIX2_VERSION
4367 @vindex POSIXLY_CORRECT
4368 On older systems, @command{sort} supports an obsolete origin-zero
4369 syntax @samp{+@var{pos1} [-@var{pos2}]} for specifying sort keys.
4370 The obsolete sequence @samp{sort +@var{a}.@var{x} -@var{b}.@var{y}}
4371 is equivalent to @samp{sort -k @var{a+1}.@var{x+1},@var{b}} if @var{y}
4372 is @samp{0} or absent, otherwise it is equivalent to @samp{sort -k
4373 @var{a+1}.@var{x+1},@var{b+1}.@var{y}}.
4375 This obsolete behavior can be enabled or disabled with the
4376 @env{_POSIX2_VERSION} environment variable (@pxref{Standards
4377 conformance}); it can also be enabled when @env{POSIXLY_CORRECT} is
4378 not set by using the obsolete syntax with @samp{-@var{pos2}} present.
4380 Scripts intended for use on standard hosts should avoid obsolete
4381 syntax and should use @option{-k} instead. For example, avoid
4382 @samp{sort +2}, since it might be interpreted as either @samp{sort
4383 ./+2} or @samp{sort -k 3}. If your script must also run on hosts that
4384 support only the obsolete syntax, it can use a test like @samp{if sort
4385 -k 1 </dev/null >/dev/null 2>&1; then @dots{}} to decide which syntax
4388 Here are some examples to illustrate various combinations of options.
4393 Sort in descending (reverse) numeric order.
4400 Run no more than 4 sorts concurrently, using a buffer size of 10M.
4403 sort --parallel=4 -S 10M
4407 Sort alphabetically, omitting the first and second fields
4408 and the blanks at the start of the third field.
4409 This uses a single key composed of the characters beginning
4410 at the start of the first nonblank character in field three
4411 and extending to the end of each line.
4418 Sort numerically on the second field and resolve ties by sorting
4419 alphabetically on the third and fourth characters of field five.
4420 Use @samp{:} as the field delimiter.
4423 sort -t : -k 2,2n -k 5.3,5.4
4426 Note that if you had written @option{-k 2n} instead of @option{-k 2,2n}
4427 @command{sort} would have used all characters beginning in the second field
4428 and extending to the end of the line as the primary @emph{numeric}
4429 key. For the large majority of applications, treating keys spanning
4430 more than one field as numeric will not do what you expect.
4432 Also note that the @samp{n} modifier was applied to the field-end
4433 specifier for the first key. It would have been equivalent to
4434 specify @option{-k 2n,2} or @option{-k 2n,2n}. All modifiers except
4435 @samp{b} apply to the associated @emph{field}, regardless of whether
4436 the modifier character is attached to the field-start and/or the
4437 field-end part of the key specifier.
4440 Sort the password file on the fifth field and ignore any
4441 leading blanks. Sort lines with equal values in field five
4442 on the numeric user ID in field three. Fields are separated
4446 sort -t : -k 5b,5 -k 3,3n /etc/passwd
4447 sort -t : -n -k 5b,5 -k 3,3 /etc/passwd
4448 sort -t : -b -k 5,5 -k 3,3n /etc/passwd
4451 These three commands have equivalent effect. The first specifies that
4452 the first key's start position ignores leading blanks and the second
4453 key is sorted numerically. The other two commands rely on global
4454 options being inherited by sort keys that lack modifiers. The inheritance
4455 works in this case because @option{-k 5b,5b} and @option{-k 5b,5} are
4456 equivalent, as the location of a field-end lacking a @samp{.@var{c}}
4457 character position is not affected by whether initial blanks are
4461 Sort a set of log files, primarily by IPv4 address and secondarily by
4462 time stamp. If two lines' primary and secondary keys are identical,
4463 output the lines in the same order that they were input. The log
4464 files contain lines that look like this:
4467 4.150.156.3 - - [01/Apr/2004:06:31:51 +0000] message 1
4468 211.24.3.231 - - [24/Apr/2004:20:17:39 +0000] message 2
4471 Fields are separated by exactly one space. Sort IPv4 addresses
4472 lexicographically, e.g., 212.61.52.2 sorts before 212.129.233.201
4473 because 61 is less than 129.
4476 sort -s -t ' ' -k 4.9n -k 4.5M -k 4.2n -k 4.14,4.21 file*.log |
4477 sort -s -t '.' -k 1,1n -k 2,2n -k 3,3n -k 4,4n
4480 This example cannot be done with a single @command{sort} invocation,
4481 since IPv4 address components are separated by @samp{.} while dates
4482 come just after a space. So it is broken down into two invocations of
4483 @command{sort}: the first sorts by time stamp and the second by IPv4
4484 address. The time stamp is sorted by year, then month, then day, and
4485 finally by hour-minute-second field, using @option{-k} to isolate each
4486 field. Except for hour-minute-second there's no need to specify the
4487 end of each key field, since the @samp{n} and @samp{M} modifiers sort
4488 based on leading prefixes that cannot cross field boundaries. The
4489 IPv4 addresses are sorted lexicographically. The second sort uses
4490 @samp{-s} so that ties in the primary key are broken by the secondary
4491 key; the first sort uses @samp{-s} so that the combination of the two
4495 Generate a tags file in case-insensitive sorted order.
4498 find src -type f -print0 | sort -z -f | xargs -0 etags --append
4501 The use of @option{-print0}, @option{-z}, and @option{-0} in this case means
4502 that file names that contain blanks or other special characters are
4504 by the sort operation.
4506 @c This example is a bit contrived and needs more explanation.
4508 @c Sort records separated by an arbitrary string by using a pipe to convert
4509 @c each record delimiter string to @samp{\0}, then using sort's -z option,
4510 @c and converting each @samp{\0} back to the original record delimiter.
4513 @c printf 'c\n\nb\n\na\n' |
4514 @c perl -0pe 's/\n\n/\n\0/g' |
4516 @c perl -0pe 's/\0/\n/g'
4520 Use the common DSU, Decorate Sort Undecorate idiom to
4521 sort lines according to their length.
4524 awk '@{print length, $0@}' /etc/passwd | sort -n | cut -f2- -d' '
4527 In general this technique can be used to sort data that the @command{sort}
4528 command does not support, or is inefficient at, sorting directly.
4531 Shuffle a list of directories, but preserve the order of files within
4532 each directory. For instance, one could use this to generate a music
4533 playlist in which albums are shuffled but the songs of each album are
4537 ls */* | sort -t / -k 1,1R -k 2,2
4543 @node shuf invocation
4544 @section @command{shuf}: Shuffling text
4547 @cindex shuffling files
4549 @command{shuf} shuffles its input by outputting a random permutation
4550 of its input lines. Each output permutation is equally likely.
4554 shuf [@var{option}]@dots{} [@var{file}]
4555 shuf -e [@var{option}]@dots{} [@var{arg}]@dots{}
4556 shuf -i @var{lo}-@var{hi} [@var{option}]@dots{}
4559 @command{shuf} has three modes of operation that affect where it
4560 obtains its input lines. By default, it reads lines from standard
4561 input. The following options change the operation mode:
4569 @cindex command-line operands to shuffle
4570 Treat each command-line operand as an input line.
4572 @item -i @var{lo}-@var{hi}
4573 @itemx --input-range=@var{lo}-@var{hi}
4575 @opindex --input-range
4576 @cindex input range to shuffle
4577 Act as if input came from a file containing the range of unsigned
4578 decimal integers @var{lo}@dots{}@var{hi}, one per line.
4582 @command{shuf}'s other options can affect its behavior in all
4587 @item -n @var{lines}
4588 @itemx --head-count=@var{count}
4590 @opindex --head-count
4591 @cindex head of output
4592 Output at most @var{count} lines. By default, all input lines are
4595 @item -o @var{output-file}
4596 @itemx --output=@var{output-file}
4599 @cindex overwriting of input, allowed
4600 Write output to @var{output-file} instead of standard output.
4601 @command{shuf} reads all input before opening
4602 @var{output-file}, so you can safely shuffle a file in place by using
4603 commands like @code{shuf -o F <F} and @code{cat F | shuf -o F}.
4605 @item --random-source=@var{file}
4606 @opindex --random-source
4607 @cindex random source for shuffling
4608 Use @var{file} as a source of random data used to determine which
4609 permutation to generate. @xref{Random sources}.
4611 @zeroTerminatedOption
4627 might produce the output
4637 Similarly, the command:
4640 shuf -e clubs hearts diamonds spades
4654 and the command @samp{shuf -i 1-4} might output:
4664 These examples all have four input lines, so @command{shuf} might
4665 produce any of the twenty-four possible permutations of the input. In
4666 general, if there are @var{n} input lines, there are @var{n}! (i.e.,
4667 @var{n} factorial, or @var{n} * (@var{n} - 1) * @dots{} * 1) possible
4668 output permutations.
4673 @node uniq invocation
4674 @section @command{uniq}: Uniquify files
4677 @cindex uniquify files
4679 @command{uniq} writes the unique lines in the given @file{input}, or
4680 standard input if nothing is given or for an @var{input} name of
4684 uniq [@var{option}]@dots{} [@var{input} [@var{output}]]
4687 By default, @command{uniq} prints its input lines, except that
4688 it discards all but the first of adjacent repeated lines, so that
4689 no output lines are repeated. Optionally, it can instead discard
4690 lines that are not repeated, or all repeated lines.
4692 The input need not be sorted, but repeated input lines are detected
4693 only if they are adjacent. If you want to discard non-adjacent
4694 duplicate lines, perhaps you want to use @code{sort -u}.
4695 @xref{sort invocation}.
4698 Comparisons honor the rules specified by the @env{LC_COLLATE}
4701 If no @var{output} file is specified, @command{uniq} writes to standard
4704 The program accepts the following options. Also see @ref{Common options}.
4709 @itemx --skip-fields=@var{n}
4711 @opindex --skip-fields
4712 Skip @var{n} fields on each line before checking for uniqueness. Use
4713 a null string for comparison if a line has fewer than @var{n} fields. Fields
4714 are sequences of non-space non-tab characters that are separated from
4715 each other by at least one space or tab.
4717 For compatibility @command{uniq} supports an obsolete option syntax
4718 @option{-@var{n}}. New scripts should use @option{-f @var{n}} instead.
4721 @itemx --skip-chars=@var{n}
4723 @opindex --skip-chars
4724 Skip @var{n} characters before checking for uniqueness. Use a null string
4725 for comparison if a line has fewer than @var{n} characters. If you use both
4726 the field and character skipping options, fields are skipped over first.
4728 @vindex _POSIX2_VERSION
4729 On older systems, @command{uniq} supports an obsolete option syntax
4731 This obsolete behavior can be enabled or disabled with the
4732 @env{_POSIX2_VERSION} environment variable (@pxref{Standards
4733 conformance}), but portable scripts should avoid commands whose
4734 behavior depends on this variable.
4735 For example, use @samp{uniq ./+10} or @samp{uniq -s 10} rather than
4736 the ambiguous @samp{uniq +10}.
4742 Print the number of times each line occurred along with the line.
4745 @itemx --ignore-case
4747 @opindex --ignore-case
4748 Ignore differences in case when comparing lines.
4754 @cindex repeated lines, outputting
4755 Discard lines that are not repeated. When used by itself, this option
4756 causes @command{uniq} to print the first copy of each repeated line,
4760 @itemx --all-repeated[=@var{delimit-method}]
4762 @opindex --all-repeated
4763 @cindex all repeated lines, outputting
4764 Do not discard the second and subsequent repeated input lines,
4765 but discard lines that are not repeated.
4766 This option is useful mainly in conjunction with other options e.g.,
4767 to ignore case or to compare only selected fields.
4768 The optional @var{delimit-method} tells how to delimit
4769 groups of repeated lines, and must be one of the following:
4774 Do not delimit groups of repeated lines.
4775 This is equivalent to @option{--all-repeated} (@option{-D}).
4778 Output a newline before each group of repeated lines.
4779 With @option{--zero-terminated} (@option{-z}), use a zero
4780 byte (ASCII @sc{nul}) instead of a newline.
4783 Separate groups of repeated lines with a single newline.
4784 With @option{--zero-terminated} (@option{-z}), use a zero
4785 byte (ASCII @sc{nul}) instead of a newline.
4786 This is the same as using @samp{prepend}, except that
4787 no delimiter is inserted before the first group, and hence
4788 may be better suited for output direct to users.
4791 Note that when groups are delimited and the input stream contains
4792 two or more consecutive blank lines, then the output is ambiguous.
4793 To avoid that, filter the input through @samp{tr -s '\n'} to replace
4794 each sequence of consecutive newlines with a single newline.
4796 This is a @sc{gnu} extension.
4797 @c FIXME: give an example showing *how* it's useful
4803 @cindex unique lines, outputting
4804 Discard the first repeated line. When used by itself, this option
4805 causes @command{uniq} to print unique lines, and nothing else.
4808 @itemx --check-chars=@var{n}
4810 @opindex --check-chars
4811 Compare at most @var{n} characters on each line (after skipping any specified
4812 fields and characters). By default the entire rest of the lines are
4815 @zeroTerminatedOption
4822 @node comm invocation
4823 @section @command{comm}: Compare two sorted files line by line
4826 @cindex line-by-line comparison
4827 @cindex comparing sorted files
4829 @command{comm} writes to standard output lines that are common, and lines
4830 that are unique, to two input files; a file name of @samp{-} means
4831 standard input. Synopsis:
4834 comm [@var{option}]@dots{} @var{file1} @var{file2}
4838 Before @command{comm} can be used, the input files must be sorted using the
4839 collating sequence specified by the @env{LC_COLLATE} locale.
4840 If an input file ends in a non-newline
4841 character, a newline is silently appended. The @command{sort} command with
4842 no options always outputs a file that is suitable input to @command{comm}.
4844 @cindex differing lines
4845 @cindex common lines
4846 With no options, @command{comm} produces three-column output. Column one
4847 contains lines unique to @var{file1}, column two contains lines unique
4848 to @var{file2}, and column three contains lines common to both files.
4849 Columns are separated by a single TAB character.
4850 @c FIXME: when there's an option to supply an alternative separator
4851 @c string, append "by default" to the above sentence.
4856 The options @option{-1}, @option{-2}, and @option{-3} suppress printing of
4857 the corresponding columns (and separators). Also see @ref{Common options}.
4859 Unlike some other comparison utilities, @command{comm} has an exit
4860 status that does not depend on the result of the comparison.
4861 Upon normal completion @command{comm} produces an exit code of zero.
4862 If there is an error it exits with nonzero status.
4864 @macro checkOrderOption{cmd}
4865 If the @option{--check-order} option is given, unsorted inputs will
4866 cause a fatal error message. If the option @option{--nocheck-order}
4867 is given, unsorted inputs will never cause an error message. If neither
4868 of these options is given, wrongly sorted inputs are diagnosed
4869 only if an input file is found to contain unpairable
4871 lines, and when both input files are non empty.
4873 @ifclear JOIN_COMMAND
4876 If an input file is diagnosed as being unsorted, the @command{\cmd\}
4877 command will exit with a nonzero status (and the output should not be used).
4879 Forcing @command{\cmd\} to process wrongly sorted input files
4880 containing unpairable lines by specifying @option{--nocheck-order} is
4881 not guaranteed to produce any particular output. The output will
4882 probably not correspond with whatever you hoped it would be.
4884 @checkOrderOption{comm}
4889 Fail with an error message if either input file is wrongly ordered.
4891 @item --nocheck-order
4892 Do not check that both input files are in sorted order.
4896 @item --output-delimiter=@var{str}
4897 Print @var{str} between adjacent output columns,
4898 rather than the default of a single TAB character.
4900 The delimiter @var{str} may not be empty.
4904 @node ptx invocation
4905 @section @command{ptx}: Produce permuted indexes
4909 @command{ptx} reads a text file and essentially produces a permuted index, with
4910 each keyword in its context. The calling sketch is either one of:
4913 ptx [@var{option} @dots{}] [@var{file} @dots{}]
4914 ptx -G [@var{option} @dots{}] [@var{input} [@var{output}]]
4917 The @option{-G} (or its equivalent: @option{--traditional}) option disables
4918 all @sc{gnu} extensions and reverts to traditional mode, thus introducing some
4919 limitations and changing several of the program's default option values.
4920 When @option{-G} is not specified, @sc{gnu} extensions are always enabled.
4921 @sc{gnu} extensions to @command{ptx} are documented wherever appropriate in this
4922 document. @xref{Compatibility in ptx}, for the full list.
4924 Individual options are explained in the following sections.
4926 When @sc{gnu} extensions are enabled, there may be zero, one or several
4927 @var{file}s after the options. If there is no @var{file}, the program
4928 reads the standard input. If there is one or several @var{file}s, they
4929 give the name of input files which are all read in turn, as if all the
4930 input files were concatenated. However, there is a full contextual
4931 break between each file and, when automatic referencing is requested,
4932 file names and line numbers refer to individual text input files. In
4933 all cases, the program outputs the permuted index to the standard
4936 When @sc{gnu} extensions are @emph{not} enabled, that is, when the program
4937 operates in traditional mode, there may be zero, one or two parameters
4938 besides the options. If there are no parameters, the program reads the
4939 standard input and outputs the permuted index to the standard output.
4940 If there is only one parameter, it names the text @var{input} to be read
4941 instead of the standard input. If two parameters are given, they give
4942 respectively the name of the @var{input} file to read and the name of
4943 the @var{output} file to produce. @emph{Be very careful} to note that,
4944 in this case, the contents of file given by the second parameter is
4945 destroyed. This behavior is dictated by System V @command{ptx}
4946 compatibility; @sc{gnu} Standards normally discourage output parameters not
4947 introduced by an option.
4949 Note that for @emph{any} file named as the value of an option or as an
4950 input text file, a single dash @kbd{-} may be used, in which case
4951 standard input is assumed. However, it would not make sense to use this
4952 convention more than once per program invocation.
4955 * General options in ptx:: Options which affect general program behavior.
4956 * Charset selection in ptx:: Underlying character set considerations.
4957 * Input processing in ptx:: Input fields, contexts, and keyword selection.
4958 * Output formatting in ptx:: Types of output format, and sizing the fields.
4959 * Compatibility in ptx::
4963 @node General options in ptx
4964 @subsection General options
4969 @itemx --traditional
4970 As already explained, this option disables all @sc{gnu} extensions to
4971 @command{ptx} and switches to traditional mode.
4974 Print a short help on standard output, then exit without further
4978 Print the program version on standard output, then exit without further
4986 @node Charset selection in ptx
4987 @subsection Charset selection
4989 @c FIXME: People don't necessarily know what an IBM-PC was these days.
4990 As it is set up now, the program assumes that the input file is coded
4991 using 8-bit ISO 8859-1 code, also known as Latin-1 character set,
4992 @emph{unless} it is compiled for MS-DOS, in which case it uses the
4993 character set of the IBM-PC@. (@sc{gnu} @command{ptx} is not known to work on
4994 smaller MS-DOS machines anymore.) Compared to 7-bit ASCII, the set
4995 of characters which are letters is different; this alters the behavior
4996 of regular expression matching. Thus, the default regular expression
4997 for a keyword allows foreign or diacriticized letters. Keyword sorting,
4998 however, is still crude; it obeys the underlying character set ordering
5004 @itemx --ignore-case
5005 Fold lower case letters to upper case for sorting.
5010 @node Input processing in ptx
5011 @subsection Word selection and input processing
5016 @itemx --break-file=@var{file}
5018 This option provides an alternative (to @option{-W}) method of describing
5019 which characters make up words. It introduces the name of a
5020 file which contains a list of characters which can@emph{not} be part of
5021 one word; this file is called the @dfn{Break file}. Any character which
5022 is not part of the Break file is a word constituent. If both options
5023 @option{-b} and @option{-W} are specified, then @option{-W} has precedence and
5024 @option{-b} is ignored.
5026 When @sc{gnu} extensions are enabled, the only way to avoid newline as a
5027 break character is to write all the break characters in the file with no
5028 newline at all, not even at the end of the file. When @sc{gnu} extensions
5029 are disabled, spaces, tabs and newlines are always considered as break
5030 characters even if not included in the Break file.
5033 @itemx --ignore-file=@var{file}
5035 The file associated with this option contains a list of words which will
5036 never be taken as keywords in concordance output. It is called the
5037 @dfn{Ignore file}. The file contains exactly one word in each line; the
5038 end of line separation of words is not subject to the value of the
5042 @itemx --only-file=@var{file}
5044 The file associated with this option contains a list of words which will
5045 be retained in concordance output; any word not mentioned in this file
5046 is ignored. The file is called the @dfn{Only file}. The file contains
5047 exactly one word in each line; the end of line separation of words is
5048 not subject to the value of the @option{-S} option.
5050 There is no default for the Only file. When both an Only file and an
5051 Ignore file are specified, a word is considered a keyword only
5052 if it is listed in the Only file and not in the Ignore file.
5057 On each input line, the leading sequence of non-white space characters will be
5058 taken to be a reference that has the purpose of identifying this input
5059 line in the resulting permuted index.
5060 @xref{Output formatting in ptx},
5061 for more information about reference production.
5062 Using this option changes the default value for option @option{-S}.
5064 Using this option, the program does not try very hard to remove
5065 references from contexts in output, but it succeeds in doing so
5066 @emph{when} the context ends exactly at the newline. If option
5067 @option{-r} is used with @option{-S} default value, or when @sc{gnu} extensions
5068 are disabled, this condition is always met and references are completely
5069 excluded from the output contexts.
5071 @item -S @var{regexp}
5072 @itemx --sentence-regexp=@var{regexp}
5074 This option selects which regular expression will describe the end of a
5075 line or the end of a sentence. In fact, this regular expression is not
5076 the only distinction between end of lines or end of sentences, and input
5077 line boundaries have no special significance outside this option. By
5078 default, when @sc{gnu} extensions are enabled and if @option{-r} option is not
5079 used, end of sentences are used. In this case, this @var{regex} is
5080 imported from @sc{gnu} Emacs:
5083 [.?!][]\"')@}]*\\($\\|\t\\| \\)[ \t\n]*
5086 Whenever @sc{gnu} extensions are disabled or if @option{-r} option is used, end
5087 of lines are used; in this case, the default @var{regexp} is just:
5093 Using an empty @var{regexp} is equivalent to completely disabling end of
5094 line or end of sentence recognition. In this case, the whole file is
5095 considered to be a single big line or sentence. The user might want to
5096 disallow all truncation flag generation as well, through option @option{-F
5097 ""}. @xref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
5100 When the keywords happen to be near the beginning of the input line or
5101 sentence, this often creates an unused area at the beginning of the
5102 output context line; when the keywords happen to be near the end of the
5103 input line or sentence, this often creates an unused area at the end of
5104 the output context line. The program tries to fill those unused areas
5105 by wrapping around context in them; the tail of the input line or
5106 sentence is used to fill the unused area on the left of the output line;
5107 the head of the input line or sentence is used to fill the unused area
5108 on the right of the output line.
5110 As a matter of convenience to the user, many usual backslashed escape
5111 sequences from the C language are recognized and converted to the
5112 corresponding characters by @command{ptx} itself.
5114 @item -W @var{regexp}
5115 @itemx --word-regexp=@var{regexp}
5117 This option selects which regular expression will describe each keyword.
5118 By default, if @sc{gnu} extensions are enabled, a word is a sequence of
5119 letters; the @var{regexp} used is @samp{\w+}. When @sc{gnu} extensions are
5120 disabled, a word is by default anything which ends with a space, a tab
5121 or a newline; the @var{regexp} used is @samp{[^ \t\n]+}.
5123 An empty @var{regexp} is equivalent to not using this option.
5124 @xref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
5127 As a matter of convenience to the user, many usual backslashed escape
5128 sequences, as found in the C language, are recognized and converted to
5129 the corresponding characters by @command{ptx} itself.
5134 @node Output formatting in ptx
5135 @subsection Output formatting
5137 Output format is mainly controlled by the @option{-O} and @option{-T} options
5138 described in the table below. When neither @option{-O} nor @option{-T} are
5139 selected, and if @sc{gnu} extensions are enabled, the program chooses an
5140 output format suitable for a dumb terminal. Each keyword occurrence is
5141 output to the center of one line, surrounded by its left and right
5142 contexts. Each field is properly justified, so the concordance output
5143 can be readily observed. As a special feature, if automatic
5144 references are selected by option @option{-A} and are output before the
5145 left context, that is, if option @option{-R} is @emph{not} selected, then
5146 a colon is added after the reference; this nicely interfaces with @sc{gnu}
5147 Emacs @code{next-error} processing. In this default output format, each
5148 white space character, like newline and tab, is merely changed to
5149 exactly one space, with no special attempt to compress consecutive
5150 spaces. This might change in the future. Except for those white space
5151 characters, every other character of the underlying set of 256
5152 characters is transmitted verbatim.
5154 Output format is further controlled by the following options.
5158 @item -g @var{number}
5159 @itemx --gap-size=@var{number}
5161 Select the size of the minimum white space gap between the fields on the
5164 @item -w @var{number}
5165 @itemx --width=@var{number}
5167 Select the maximum output width of each final line. If references are
5168 used, they are included or excluded from the maximum output width
5169 depending on the value of option @option{-R}@. If this option is not
5170 selected, that is, when references are output before the left context,
5171 the maximum output width takes into account the maximum length of all
5172 references. If this option is selected, that is, when references are
5173 output after the right context, the maximum output width does not take
5174 into account the space taken by references, nor the gap that precedes
5178 @itemx --auto-reference
5180 Select automatic references. Each input line will have an automatic
5181 reference made up of the file name and the line ordinal, with a single
5182 colon between them. However, the file name will be empty when standard
5183 input is being read. If both @option{-A} and @option{-r} are selected, then
5184 the input reference is still read and skipped, but the automatic
5185 reference is used at output time, overriding the input reference.
5188 @itemx --right-side-refs
5190 In the default output format, when option @option{-R} is not used, any
5191 references produced by the effect of options @option{-r} or @option{-A} are
5192 placed to the far right of output lines, after the right context. With
5193 default output format, when the @option{-R} option is specified, references
5194 are rather placed at the beginning of each output line, before the left
5195 context. For any other output format, option @option{-R} is
5196 ignored, with one exception: with @option{-R} the width of references
5197 is @emph{not} taken into account in total output width given by @option{-w}.
5199 This option is automatically selected whenever @sc{gnu} extensions are
5202 @item -F @var{string}
5203 @itemx --flac-truncation=@var{string}
5205 This option will request that any truncation in the output be reported
5206 using the string @var{string}. Most output fields theoretically extend
5207 towards the beginning or the end of the current line, or current
5208 sentence, as selected with option @option{-S}@. But there is a maximum
5209 allowed output line width, changeable through option @option{-w}, which is
5210 further divided into space for various output fields. When a field has
5211 to be truncated because it cannot extend beyond the beginning or the end of
5212 the current line to fit in, then a truncation occurs. By default,
5213 the string used is a single slash, as in @option{-F /}.
5215 @var{string} may have more than one character, as in @option{-F ...}.
5216 Also, in the particular case when @var{string} is empty (@option{-F ""}),
5217 truncation flagging is disabled, and no truncation marks are appended in
5220 As a matter of convenience to the user, many usual backslashed escape
5221 sequences, as found in the C language, are recognized and converted to
5222 the corresponding characters by @command{ptx} itself.
5224 @item -M @var{string}
5225 @itemx --macro-name=@var{string}
5227 Select another @var{string} to be used instead of @samp{xx}, while
5228 generating output suitable for @command{nroff}, @command{troff} or @TeX{}.
5231 @itemx --format=roff
5233 Choose an output format suitable for @command{nroff} or @command{troff}
5234 processing. Each output line will look like:
5237 .xx "@var{tail}" "@var{before}" "@var{keyword_and_after}"@c
5238 "@var{head}" "@var{ref}"
5241 so it will be possible to write a @samp{.xx} roff macro to take care of
5242 the output typesetting. This is the default output format when @sc{gnu}
5243 extensions are disabled. Option @option{-M} can be used to change
5244 @samp{xx} to another macro name.
5246 In this output format, each non-graphical character, like newline and
5247 tab, is merely changed to exactly one space, with no special attempt to
5248 compress consecutive spaces. Each quote character: @kbd{"} is doubled
5249 so it will be correctly processed by @command{nroff} or @command{troff}.
5254 Choose an output format suitable for @TeX{} processing. Each output
5255 line will look like:
5258 \xx @{@var{tail}@}@{@var{before}@}@{@var{keyword}@}@c
5259 @{@var{after}@}@{@var{head}@}@{@var{ref}@}
5263 so it will be possible to write a @code{\xx} definition to take care of
5264 the output typesetting. Note that when references are not being
5265 produced, that is, neither option @option{-A} nor option @option{-r} is
5266 selected, the last parameter of each @code{\xx} call is inhibited.
5267 Option @option{-M} can be used to change @samp{xx} to another macro
5270 In this output format, some special characters, like @kbd{$}, @kbd{%},
5271 @kbd{&}, @kbd{#} and @kbd{_} are automatically protected with a
5272 backslash. Curly brackets @kbd{@{}, @kbd{@}} are protected with a
5273 backslash and a pair of dollar signs (to force mathematical mode). The
5274 backslash itself produces the sequence @code{\backslash@{@}}.
5275 Circumflex and tilde diacritical marks produce the sequence @code{^\@{ @}} and
5276 @code{~\@{ @}} respectively. Other diacriticized characters of the
5277 underlying character set produce an appropriate @TeX{} sequence as far
5278 as possible. The other non-graphical characters, like newline and tab,
5279 and all other characters which are not part of ASCII, are merely
5280 changed to exactly one space, with no special attempt to compress
5281 consecutive spaces. Let me know how to improve this special character
5282 processing for @TeX{}.
5287 @node Compatibility in ptx
5288 @subsection The @sc{gnu} extensions to @command{ptx}
5290 This version of @command{ptx} contains a few features which do not exist in
5291 System V @command{ptx}. These extra features are suppressed by using the
5292 @option{-G} command line option, unless overridden by other command line
5293 options. Some @sc{gnu} extensions cannot be recovered by overriding, so the
5294 simple rule is to avoid @option{-G} if you care about @sc{gnu} extensions.
5295 Here are the differences between this program and System V @command{ptx}.
5300 This program can read many input files at once, it always writes the
5301 resulting concordance on standard output. On the other hand, System V
5302 @command{ptx} reads only one file and sends the result to standard output
5303 or, if a second @var{file} parameter is given on the command, to that
5306 Having output parameters not introduced by options is a dangerous
5307 practice which @sc{gnu} avoids as far as possible. So, for using @command{ptx}
5308 portably between @sc{gnu} and System V, you should always use it with a
5309 single input file, and always expect the result on standard output. You
5310 might also want to automatically configure in a @option{-G} option to
5311 @command{ptx} calls in products using @command{ptx}, if the configurator finds
5312 that the installed @command{ptx} accepts @option{-G}.
5315 The only options available in System V @command{ptx} are options @option{-b},
5316 @option{-f}, @option{-g}, @option{-i}, @option{-o}, @option{-r}, @option{-t} and
5317 @option{-w}. All other options are @sc{gnu} extensions and are not repeated in
5318 this enumeration. Moreover, some options have a slightly different
5319 meaning when @sc{gnu} extensions are enabled, as explained below.
5322 By default, concordance output is not formatted for @command{troff} or
5323 @command{nroff}. It is rather formatted for a dumb terminal. @command{troff}
5324 or @command{nroff} output may still be selected through option @option{-O}.
5327 Unless @option{-R} option is used, the maximum reference width is
5328 subtracted from the total output line width. With @sc{gnu} extensions
5329 disabled, width of references is not taken into account in the output
5330 line width computations.
5333 All 256 bytes, even ASCII @sc{nul} bytes, are always read and
5334 processed from input file with no adverse effect, even if @sc{gnu} extensions
5335 are disabled. However, System V @command{ptx} does not accept 8-bit
5336 characters, a few control characters are rejected, and the tilde
5337 @kbd{~} is also rejected.
5340 Input line length is only limited by available memory, even if @sc{gnu}
5341 extensions are disabled. However, System V @command{ptx} processes only
5342 the first 200 characters in each line.
5345 The break (non-word) characters default to be every character except all
5346 letters of the underlying character set, diacriticized or not. When @sc{gnu}
5347 extensions are disabled, the break characters default to space, tab and
5351 The program makes better use of output line width. If @sc{gnu} extensions
5352 are disabled, the program rather tries to imitate System V @command{ptx},
5353 but still, there are some slight disposition glitches this program does
5354 not completely reproduce.
5357 The user can specify both an Ignore file and an Only file. This is not
5358 allowed with System V @command{ptx}.
5363 @node tsort invocation
5364 @section @command{tsort}: Topological sort
5367 @cindex topological sort
5369 @command{tsort} performs a topological sort on the given @var{file}, or
5370 standard input if no input file is given or for a @var{file} of
5371 @samp{-}. For more details and some history, see @ref{tsort background}.
5375 tsort [@var{option}] [@var{file}]
5378 @command{tsort} reads its input as pairs of strings, separated by blanks,
5379 indicating a partial ordering. The output is a total ordering that
5380 corresponds to the given partial ordering.
5394 will produce the output
5405 Consider a more realistic example.
5406 You have a large set of functions all in one file, and they may all be
5407 declared static except one. Currently that one (say @code{main}) is the
5408 first function defined in the file, and the ones it calls directly follow
5409 it, followed by those they call, etc. Let's say that you are determined
5410 to take advantage of prototypes, so you have to choose between declaring
5411 all of those functions (which means duplicating a lot of information from
5412 the definitions) and rearranging the functions so that as many as possible
5413 are defined before they are used. One way to automate the latter process
5414 is to get a list for each function of the functions it calls directly.
5415 Many programs can generate such lists. They describe a call graph.
5416 Consider the following list, in which a given line indicates that the
5417 function on the left calls the one on the right directly.
5423 tail_file pretty_name
5424 tail_file write_header
5426 tail_forever recheck
5427 tail_forever pretty_name
5428 tail_forever write_header
5429 tail_forever dump_remainder
5432 tail_lines start_lines
5433 tail_lines dump_remainder
5434 tail_lines file_lines
5435 tail_lines pipe_lines
5437 tail_bytes start_bytes
5438 tail_bytes dump_remainder
5439 tail_bytes pipe_bytes
5440 file_lines dump_remainder
5444 then you can use @command{tsort} to produce an ordering of those
5445 functions that satisfies your requirement.
5448 example$ tsort call-graph | tac
5468 @command{tsort} detects any cycles in the input and writes the first cycle
5469 encountered to standard error.
5471 Note that for a given partial ordering, generally there is no unique
5472 total ordering. In the context of the call graph above, the function
5473 @code{parse_options} may be placed anywhere in the list as long as it
5474 precedes @code{main}.
5476 The only options are @option{--help} and @option{--version}. @xref{Common
5482 * tsort background:: Where tsort came from.
5485 @node tsort background
5486 @subsection @command{tsort}: Background
5488 @command{tsort} exists because very early versions of the Unix linker processed
5489 an archive file exactly once, and in order. As @command{ld} read each object
5490 in the archive, it decided whether it was needed in the program based on
5491 whether it defined any symbols which were undefined at that point in
5494 This meant that dependencies within the archive had to be handled
5495 specially. For example, @code{scanf} probably calls @code{read}. That means
5496 that in a single pass through an archive, it was important for @code{scanf.o}
5497 to appear before read.o, because otherwise a program which calls
5498 @code{scanf} but not @code{read} might end up with an unexpected unresolved
5499 reference to @code{read}.
5501 The way to address this problem was to first generate a set of
5502 dependencies of one object file on another. This was done by a shell
5503 script called @command{lorder}. The GNU tools don't provide a version of
5504 lorder, as far as I know, but you can still find it in BSD
5507 Then you ran @command{tsort} over the @command{lorder} output, and you used the
5508 resulting sort to define the order in which you added objects to the archive.
5510 This whole procedure has been obsolete since about 1980, because
5511 Unix archives now contain a symbol table (traditionally built by
5512 @command{ranlib}, now generally built by @command{ar} itself), and the Unix
5513 linker uses the symbol table to effectively make multiple passes over
5516 Anyhow, that's where tsort came from. To solve an old problem with
5517 the way the linker handled archive files, which has since been solved
5521 @node Operating on fields
5522 @chapter Operating on fields
5525 * cut invocation:: Print selected parts of lines.
5526 * paste invocation:: Merge lines of files.
5527 * join invocation:: Join lines on a common field.
5531 @node cut invocation
5532 @section @command{cut}: Print selected parts of lines
5535 @command{cut} writes to standard output selected parts of each line of each
5536 input file, or standard input if no files are given or for a file name of
5540 cut @var{option}@dots{} [@var{file}]@dots{}
5543 In the table which follows, the @var{byte-list}, @var{character-list},
5544 and @var{field-list} are one or more numbers or ranges (two numbers
5545 separated by a dash) separated by commas. Bytes, characters, and
5546 fields are numbered starting at 1. Incomplete ranges may be
5547 given: @option{-@var{m}} means @samp{1-@var{m}}; @samp{@var{n}-} means
5548 @samp{@var{n}} through end of line or last field. The list elements
5549 can be repeated, can overlap, and can be specified in any order; but
5550 the selected input is written in the same order that it is read, and
5551 is written exactly once.
5553 The program accepts the following options. Also see @ref{Common
5558 @item -b @var{byte-list}
5559 @itemx --bytes=@var{byte-list}
5562 Select for printing only the bytes in positions listed in
5563 @var{byte-list}. Tabs and backspaces are treated like any other
5564 character; they take up 1 byte. If an output delimiter is specified,
5565 (see the description of @option{--output-delimiter}), then output that
5566 string between ranges of selected bytes.
5568 @item -c @var{character-list}
5569 @itemx --characters=@var{character-list}
5571 @opindex --characters
5572 Select for printing only the characters in positions listed in
5573 @var{character-list}. The same as @option{-b} for now, but
5574 internationalization will change that. Tabs and backspaces are
5575 treated like any other character; they take up 1 character. If an
5576 output delimiter is specified, (see the description of
5577 @option{--output-delimiter}), then output that string between ranges
5580 @item -f @var{field-list}
5581 @itemx --fields=@var{field-list}
5584 Select for printing only the fields listed in @var{field-list}.
5585 Fields are separated by a TAB character by default. Also print any
5586 line that contains no delimiter character, unless the
5587 @option{--only-delimited} (@option{-s}) option is specified.
5589 Note @command{awk} supports more sophisticated field processing,
5590 and by default will use (and discard) runs of blank characters to
5591 separate fields, and ignore leading and trailing blanks.
5594 awk '{print $2}' # print the second field
5595 awk '{print $NF-1}' # print the penultimate field
5596 awk '{print $2,$1}' # reorder the first two fields
5600 In the unlikely event that @command{awk} is unavailable,
5601 one can use the @command{join} command, to process blank
5602 characters as @command{awk} does above.
5605 join -a1 -o 1.2 - /dev/null # print the second field
5606 join -a1 -o 1.2,1.1 - /dev/null # reorder the first two fields
5610 @item -d @var{input_delim_byte}
5611 @itemx --delimiter=@var{input_delim_byte}
5613 @opindex --delimiter
5614 With @option{-f}, use the first byte of @var{input_delim_byte} as
5615 the input fields separator (default is TAB).
5619 Do not split multi-byte characters (no-op for now).
5622 @itemx --only-delimited
5624 @opindex --only-delimited
5625 For @option{-f}, do not print lines that do not contain the field separator
5626 character. Normally, any line without a field separator is printed verbatim.
5628 @item --output-delimiter=@var{output_delim_string}
5629 @opindex --output-delimiter
5630 With @option{-f}, output fields are separated by @var{output_delim_string}.
5631 The default with @option{-f} is to use the input delimiter.
5632 When using @option{-b} or @option{-c} to select ranges of byte or
5633 character offsets (as opposed to ranges of fields),
5634 output @var{output_delim_string} between non-overlapping
5635 ranges of selected bytes.
5638 @opindex --complement
5639 This option is a GNU extension.
5640 Select for printing the complement of the bytes, characters or fields
5641 selected with the @option{-b}, @option{-c} or @option{-f} options.
5642 In other words, do @emph{not} print the bytes, characters or fields
5643 specified via those options. This option is useful when you have
5644 many fields and want to print all but a few of them.
5651 @node paste invocation
5652 @section @command{paste}: Merge lines of files
5655 @cindex merging files
5657 @command{paste} writes to standard output lines consisting of sequentially
5658 corresponding lines of each given file, separated by a TAB character.
5659 Standard input is used for a file name of @samp{-} or if no input files
5681 paste [@var{option}]@dots{} [@var{file}]@dots{}
5684 The program accepts the following options. Also see @ref{Common options}.
5692 Paste the lines of one file at a time rather than one line from each
5693 file. Using the above example data:
5696 $ paste -s num2 let3
5701 @item -d @var{delim-list}
5702 @itemx --delimiters=@var{delim-list}
5704 @opindex --delimiters
5705 Consecutively use the characters in @var{delim-list} instead of
5706 TAB to separate merged lines. When @var{delim-list} is
5707 exhausted, start again at its beginning. Using the above example data:
5710 $ paste -d '%_' num2 let3 num2
5721 @node join invocation
5722 @section @command{join}: Join lines on a common field
5725 @cindex common field, joining on
5727 @command{join} writes to standard output a line for each pair of input
5728 lines that have identical join fields. Synopsis:
5731 join [@var{option}]@dots{} @var{file1} @var{file2}
5734 Either @var{file1} or @var{file2} (but not both) can be @samp{-},
5735 meaning standard input. @var{file1} and @var{file2} should be
5736 sorted on the join fields.
5739 Normally, the sort order is that of the
5740 collating sequence specified by the @env{LC_COLLATE} locale. Unless
5741 the @option{-t} option is given, the sort comparison ignores blanks at
5742 the start of the join field, as in @code{sort -b}. If the
5743 @option{--ignore-case} option is given, the sort comparison ignores
5744 the case of characters in the join field, as in @code{sort -f}.
5746 The @command{sort} and @command{join} commands should use consistent
5747 locales and options if the output of @command{sort} is fed to
5748 @command{join}. You can use a command like @samp{sort -k 1b,1} to
5749 sort a file on its default join field, but if you select a non-default
5750 locale, join field, separator, or comparison options, then you should
5751 do so consistently between @command{join} and @command{sort}.
5752 If @samp{join -t ''} is specified then the whole line is considered which
5753 matches the default operation of sort.
5755 If the input has no unpairable lines, a GNU extension is
5756 available; the sort order can be any order that considers two fields
5757 to be equal if and only if the sort comparison described above
5758 considers them to be equal. For example:
5776 @checkOrderOption{join}
5781 @item the join field is the first field in each line;
5782 @item fields in the input are separated by one or more blanks, with leading
5783 blanks on the line ignored;
5784 @item fields in the output are separated by a space;
5785 @item each output line consists of the join field, the remaining
5786 fields from @var{file1}, then the remaining fields from @var{file2}.
5789 The program accepts the following options. Also see @ref{Common options}.
5793 @item -a @var{file-number}
5795 Print a line for each unpairable line in file @var{file-number} (either
5796 @samp{1} or @samp{2}), in addition to the normal output.
5799 Fail with an error message if either input file is wrongly ordered.
5801 @item --nocheck-order
5802 Do not check that both input files are in sorted order. This is the default.
5804 @item -e @var{string}
5806 Replace those output fields that are missing in the input with @var{string}.
5807 I.e., missing fields specified with the @option{-12jo} options.
5811 Treat the first line of each input file as a header line. The header lines
5812 will be joined and printed as the first output line. If @option{-o} is used to
5813 specify output format, the header line will be printed according to the
5814 specified format. The header lines will not be checked for ordering even if
5815 @option{--check-order} is specified. Also if the header lines from each file
5816 do not match, the heading fields from the first file will be used.
5819 @itemx --ignore-case
5821 @opindex --ignore-case
5822 Ignore differences in case when comparing keys.
5823 With this option, the lines of the input files must be ordered in the same way.
5824 Use @samp{sort -f} to produce this ordering.
5826 @item -1 @var{field}
5828 Join on field @var{field} (a positive integer) of file 1.
5830 @item -2 @var{field}
5832 Join on field @var{field} (a positive integer) of file 2.
5834 @item -j @var{field}
5835 Equivalent to @option{-1 @var{field} -2 @var{field}}.
5837 @item -o @var{field-list}
5839 If the keyword @samp{auto} is specified, infer the output format from
5840 the first line in each file. This is the same as the default output format
5841 but also ensures the same number of fields are output for each line.
5842 Missing fields are replaced with the @option{-e} option and extra fields
5845 Otherwise, construct each output line according to the format in
5846 @var{field-list}. Each element in @var{field-list} is either the single
5847 character @samp{0} or has the form @var{m.n} where the file number, @var{m},
5848 is @samp{1} or @samp{2} and @var{n} is a positive field number.
5850 A field specification of @samp{0} denotes the join field.
5851 In most cases, the functionality of the @samp{0} field spec
5852 may be reproduced using the explicit @var{m.n} that corresponds
5853 to the join field. However, when printing unpairable lines
5854 (using either of the @option{-a} or @option{-v} options), there is no way
5855 to specify the join field using @var{m.n} in @var{field-list}
5856 if there are unpairable lines in both files.
5857 To give @command{join} that functionality, POSIX invented the @samp{0}
5858 field specification notation.
5860 The elements in @var{field-list}
5861 are separated by commas or blanks.
5862 Blank separators typically need to be quoted for the shell. For
5863 example, the commands @samp{join -o 1.2,2.2} and @samp{join -o '1.2
5864 2.2'} are equivalent.
5866 All output lines---including those printed because of any -a or -v
5867 option---are subject to the specified @var{field-list}.
5870 Use character @var{char} as the input and output field separator.
5871 Treat as significant each occurrence of @var{char} in the input file.
5872 Use @samp{sort -t @var{char}}, without the @option{-b} option of
5873 @samp{sort}, to produce this ordering. If @samp{join -t ''} is specified,
5874 the whole line is considered, matching the default operation of sort.
5875 If @samp{-t '\0'} is specified then the ASCII @sc{nul}
5876 character is used to delimit the fields.
5878 @item -v @var{file-number}
5879 Print a line for each unpairable line in file @var{file-number}
5880 (either @samp{1} or @samp{2}), instead of the normal output.
5887 @node Operating on characters
5888 @chapter Operating on characters
5890 @cindex operating on characters
5892 This commands operate on individual characters.
5895 * tr invocation:: Translate, squeeze, and/or delete characters.
5896 * expand invocation:: Convert tabs to spaces.
5897 * unexpand invocation:: Convert spaces to tabs.
5902 @section @command{tr}: Translate, squeeze, and/or delete characters
5909 tr [@var{option}]@dots{} @var{set1} [@var{set2}]
5912 @command{tr} copies standard input to standard output, performing
5913 one of the following operations:
5917 translate, and optionally squeeze repeated characters in the result,
5919 squeeze repeated characters,
5923 delete characters, then squeeze repeated characters from the result.
5926 The @var{set1} and (if given) @var{set2} arguments define ordered
5927 sets of characters, referred to below as @var{set1} and @var{set2}. These
5928 sets are the characters of the input that @command{tr} operates on.
5929 The @option{--complement} (@option{-c}, @option{-C}) option replaces
5931 complement (all of the characters that are not in @var{set1}).
5933 Currently @command{tr} fully supports only single-byte characters.
5934 Eventually it will support multibyte characters; when it does, the
5935 @option{-C} option will cause it to complement the set of characters,
5936 whereas @option{-c} will cause it to complement the set of values.
5937 This distinction will matter only when some values are not characters,
5938 and this is possible only in locales using multibyte encodings when
5939 the input contains encoding errors.
5941 The program accepts the @option{--help} and @option{--version}
5942 options. @xref{Common options}. Options must precede operands.
5947 * Character sets:: Specifying sets of characters.
5948 * Translating:: Changing one set of characters to another.
5949 * Squeezing:: Squeezing repeats and deleting.
5953 @node Character sets
5954 @subsection Specifying sets of characters
5956 @cindex specifying sets of characters
5958 The format of the @var{set1} and @var{set2} arguments resembles
5959 the format of regular expressions; however, they are not regular
5960 expressions, only lists of characters. Most characters simply
5961 represent themselves in these strings, but the strings can contain
5962 the shorthands listed below, for convenience. Some of them can be
5963 used only in @var{set1} or @var{set2}, as noted below.
5967 @item Backslash escapes
5968 @cindex backslash escapes
5970 The following backslash escape sequences are recognized:
5988 The 8-bit character with the value given by @var{ooo}, which is 1 to 3
5989 octal digits. Note that @samp{\400} is interpreted as the two-byte
5990 sequence, @samp{\040} @samp{0}.
5995 While a backslash followed by a character not listed above is
5996 interpreted as that character, the backslash also effectively
5997 removes any special significance, so it is useful to escape
5998 @samp{[}, @samp{]}, @samp{*}, and @samp{-}.
6003 The notation @samp{@var{m}-@var{n}} expands to all of the characters
6004 from @var{m} through @var{n}, in ascending order. @var{m} should
6005 collate before @var{n}; if it doesn't, an error results. As an example,
6006 @samp{0-9} is the same as @samp{0123456789}.
6008 @sc{gnu} @command{tr} does not support the System V syntax that uses square
6009 brackets to enclose ranges. Translations specified in that format
6010 sometimes work as expected, since the brackets are often transliterated
6011 to themselves. However, they should be avoided because they sometimes
6012 behave unexpectedly. For example, @samp{tr -d '[0-9]'} deletes brackets
6015 Many historically common and even accepted uses of ranges are not
6016 portable. For example, on EBCDIC hosts using the @samp{A-Z}
6017 range will not do what most would expect because @samp{A} through @samp{Z}
6018 are not contiguous as they are in ASCII@.
6019 If you can rely on a POSIX compliant version of @command{tr}, then
6020 the best way to work around this is to use character classes (see below).
6021 Otherwise, it is most portable (and most ugly) to enumerate the members
6024 @item Repeated characters
6025 @cindex repeated characters
6027 The notation @samp{[@var{c}*@var{n}]} in @var{set2} expands to @var{n}
6028 copies of character @var{c}. Thus, @samp{[y*6]} is the same as
6029 @samp{yyyyyy}. The notation @samp{[@var{c}*]} in @var{string2} expands
6030 to as many copies of @var{c} as are needed to make @var{set2} as long as
6031 @var{set1}. If @var{n} begins with @samp{0}, it is interpreted in
6032 octal, otherwise in decimal.
6034 @item Character classes
6035 @cindex character classes
6037 The notation @samp{[:@var{class}:]} expands to all of the characters in
6038 the (predefined) class @var{class}. The characters expand in no
6039 particular order, except for the @code{upper} and @code{lower} classes,
6040 which expand in ascending order. When the @option{--delete} (@option{-d})
6041 and @option{--squeeze-repeats} (@option{-s}) options are both given, any
6042 character class can be used in @var{set2}. Otherwise, only the
6043 character classes @code{lower} and @code{upper} are accepted in
6044 @var{set2}, and then only if the corresponding character class
6045 (@code{upper} and @code{lower}, respectively) is specified in the same
6046 relative position in @var{set1}. Doing this specifies case conversion.
6047 The class names are given below; an error results when an invalid class
6059 Horizontal whitespace.
6068 Printable characters, not including space.
6074 Printable characters, including space.
6077 Punctuation characters.
6080 Horizontal or vertical whitespace.
6089 @item Equivalence classes
6090 @cindex equivalence classes
6092 The syntax @samp{[=@var{c}=]} expands to all of the characters that are
6093 equivalent to @var{c}, in no particular order. Equivalence classes are
6094 a relatively recent invention intended to support non-English alphabets.
6095 But there seems to be no standard way to define them or determine their
6096 contents. Therefore, they are not fully implemented in @sc{gnu} @command{tr};
6097 each character's equivalence class consists only of that character,
6098 which is of no particular use.
6104 @subsection Translating
6106 @cindex translating characters
6108 @command{tr} performs translation when @var{set1} and @var{set2} are
6109 both given and the @option{--delete} (@option{-d}) option is not given.
6110 @command{tr} translates each character of its input that is in @var{set1}
6111 to the corresponding character in @var{set2}. Characters not in
6112 @var{set1} are passed through unchanged. When a character appears more
6113 than once in @var{set1} and the corresponding characters in @var{set2}
6114 are not all the same, only the final one is used. For example, these
6115 two commands are equivalent:
6122 A common use of @command{tr} is to convert lowercase characters to
6123 uppercase. This can be done in many ways. Here are three of them:
6126 tr abcdefghijklmnopqrstuvwxyz ABCDEFGHIJKLMNOPQRSTUVWXYZ
6128 tr '[:lower:]' '[:upper:]'
6132 But note that using ranges like @code{a-z} above is not portable.
6134 When @command{tr} is performing translation, @var{set1} and @var{set2}
6135 typically have the same length. If @var{set1} is shorter than
6136 @var{set2}, the extra characters at the end of @var{set2} are ignored.
6138 On the other hand, making @var{set1} longer than @var{set2} is not
6139 portable; POSIX says that the result is undefined. In this situation,
6140 BSD @command{tr} pads @var{set2} to the length of @var{set1} by repeating
6141 the last character of @var{set2} as many times as necessary. System V
6142 @command{tr} truncates @var{set1} to the length of @var{set2}.
6144 By default, @sc{gnu} @command{tr} handles this case like BSD @command{tr}.
6145 When the @option{--truncate-set1} (@option{-t}) option is given,
6146 @sc{gnu} @command{tr} handles this case like the System V @command{tr}
6147 instead. This option is ignored for operations other than translation.
6149 Acting like System V @command{tr} in this case breaks the relatively common
6153 tr -cs A-Za-z0-9 '\012'
6157 because it converts only zero bytes (the first element in the
6158 complement of @var{set1}), rather than all non-alphanumerics, to
6162 By the way, the above idiom is not portable because it uses ranges, and
6163 it assumes that the octal code for newline is 012.
6164 Assuming a POSIX compliant @command{tr}, here is a better
6168 tr -cs '[:alnum:]' '[\n*]'
6173 @subsection Squeezing repeats and deleting
6175 @cindex squeezing repeat characters
6176 @cindex deleting characters
6178 When given just the @option{--delete} (@option{-d}) option, @command{tr}
6179 removes any input characters that are in @var{set1}.
6181 When given just the @option{--squeeze-repeats} (@option{-s}) option,
6182 @command{tr} replaces each input sequence of a repeated character that
6183 is in @var{set1} with a single occurrence of that character.
6185 When given both @option{--delete} and @option{--squeeze-repeats}, @command{tr}
6186 first performs any deletions using @var{set1}, then squeezes repeats
6187 from any remaining characters using @var{set2}.
6189 The @option{--squeeze-repeats} option may also be used when translating,
6190 in which case @command{tr} first performs translation, then squeezes
6191 repeats from any remaining characters using @var{set2}.
6193 Here are some examples to illustrate various combinations of options:
6198 Remove all zero bytes:
6205 Put all words on lines by themselves. This converts all
6206 non-alphanumeric characters to newlines, then squeezes each string
6207 of repeated newlines into a single newline:
6210 tr -cs '[:alnum:]' '[\n*]'
6214 Convert each sequence of repeated newlines to a single newline:
6221 Find doubled occurrences of words in a document.
6222 @c Separate the following two "the"s, so typo checkers don't complain.
6223 For example, people often write ``the @w{}the'' with the repeated words
6224 separated by a newline. The Bourne shell script below works first
6225 by converting each sequence of punctuation and blank characters to a
6226 single newline. That puts each ``word'' on a line by itself.
6227 Next it maps all uppercase characters to lower case, and finally it
6228 runs @command{uniq} with the @option{-d} option to print out only the words
6234 | tr -s '[:punct:][:blank:]' '[\n*]' \
6235 | tr '[:upper:]' '[:lower:]' \
6240 Deleting a small set of characters is usually straightforward. For example,
6241 to remove all @samp{a}s, @samp{x}s, and @samp{M}s you would do this:
6247 However, when @samp{-} is one of those characters, it can be tricky because
6248 @samp{-} has special meanings. Performing the same task as above but also
6249 removing all @samp{-} characters, we might try @code{tr -d -axM}, but
6250 that would fail because @command{tr} would try to interpret @option{-a} as
6251 a command-line option. Alternatively, we could try putting the hyphen
6252 inside the string, @code{tr -d a-xM}, but that wouldn't work either because
6253 it would make @command{tr} interpret @code{a-x} as the range of characters
6254 @samp{a}@dots{}@samp{x} rather than the three.
6255 One way to solve the problem is to put the hyphen at the end of the list
6262 Or you can use @samp{--} to terminate option processing:
6268 More generally, use the character class notation @code{[=c=]}
6269 with @samp{-} (or any other character) in place of the @samp{c}:
6275 Note how single quotes are used in the above example to protect the
6276 square brackets from interpretation by a shell.
6281 @node expand invocation
6282 @section @command{expand}: Convert tabs to spaces
6285 @cindex tabs to spaces, converting
6286 @cindex converting tabs to spaces
6288 @command{expand} writes the contents of each given @var{file}, or standard
6289 input if none are given or for a @var{file} of @samp{-}, to standard
6290 output, with tab characters converted to the appropriate number of
6294 expand [@var{option}]@dots{} [@var{file}]@dots{}
6297 By default, @command{expand} converts all tabs to spaces. It preserves
6298 backspace characters in the output; they decrement the column count for
6299 tab calculations. The default action is equivalent to @option{-t 8} (set
6300 tabs every 8 columns).
6302 The program accepts the following options. Also see @ref{Common options}.
6306 @item -t @var{tab1}[,@var{tab2}]@dots{}
6307 @itemx --tabs=@var{tab1}[,@var{tab2}]@dots{}
6310 @cindex tab stops, setting
6311 If only one tab stop is given, set the tabs @var{tab1} spaces apart
6312 (default is 8). Otherwise, set the tabs at columns @var{tab1},
6313 @var{tab2}, @dots{} (numbered from 0), and replace any tabs beyond the
6314 last tab stop given with single spaces. Tab stops can be separated by
6315 blanks as well as by commas.
6317 For compatibility, GNU @command{expand} also accepts the obsolete
6318 option syntax, @option{-@var{t1}[,@var{t2}]@dots{}}. New scripts
6319 should use @option{-t @var{t1}[,@var{t2}]@dots{}} instead.
6325 @cindex initial tabs, converting
6326 Only convert initial tabs (those that precede all non-space or non-tab
6327 characters) on each line to spaces.
6334 @node unexpand invocation
6335 @section @command{unexpand}: Convert spaces to tabs
6339 @command{unexpand} writes the contents of each given @var{file}, or
6340 standard input if none are given or for a @var{file} of @samp{-}, to
6341 standard output, converting blanks at the beginning of each line into
6342 as many tab characters as needed. In the default POSIX
6343 locale, a @dfn{blank} is a space or a tab; other locales may specify
6344 additional blank characters. Synopsis:
6347 unexpand [@var{option}]@dots{} [@var{file}]@dots{}
6350 By default, @command{unexpand} converts only initial blanks (those
6351 that precede all non-blank characters) on each line. It
6352 preserves backspace characters in the output; they decrement the column
6353 count for tab calculations. By default, tabs are set at every 8th
6356 The program accepts the following options. Also see @ref{Common options}.
6360 @item -t @var{tab1}[,@var{tab2}]@dots{}
6361 @itemx --tabs=@var{tab1}[,@var{tab2}]@dots{}
6364 If only one tab stop is given, set the tabs @var{tab1} columns apart
6365 instead of the default 8. Otherwise, set the tabs at columns
6366 @var{tab1}, @var{tab2}, @dots{} (numbered from 0), and leave blanks
6367 beyond the tab stops given unchanged. Tab stops can be separated by
6368 blanks as well as by commas. This option implies the @option{-a} option.
6370 For compatibility, GNU @command{unexpand} supports the obsolete option syntax,
6371 @option{-@var{tab1}[,@var{tab2}]@dots{}}, where tab stops must be
6372 separated by commas. (Unlike @option{-t}, this obsolete option does
6373 not imply @option{-a}.) New scripts should use @option{--first-only -t
6374 @var{tab1}[,@var{tab2}]@dots{}} instead.
6380 Also convert all sequences of two or more blanks just before a tab stop,
6381 even if they occur after non-blank characters in a line.
6388 @node Directory listing
6389 @chapter Directory listing
6391 This chapter describes the @command{ls} command and its variants @command{dir}
6392 and @command{vdir}, which list information about files.
6395 * ls invocation:: List directory contents.
6396 * dir invocation:: Briefly ls.
6397 * vdir invocation:: Verbosely ls.
6398 * dircolors invocation:: Color setup for ls, etc.
6403 @section @command{ls}: List directory contents
6406 @cindex directory listing
6408 The @command{ls} program lists information about files (of any type,
6409 including directories). Options and file arguments can be intermixed
6410 arbitrarily, as usual.
6412 For non-option command-line arguments that are directories, by default
6413 @command{ls} lists the contents of directories, not recursively, and
6414 omitting files with names beginning with @samp{.}. For other non-option
6415 arguments, by default @command{ls} lists just the file name. If no
6416 non-option argument is specified, @command{ls} operates on the current
6417 directory, acting as if it had been invoked with a single argument of @samp{.}.
6420 By default, the output is sorted alphabetically, according to the locale
6421 settings in effect.@footnote{If you use a non-POSIX
6422 locale (e.g., by setting @env{LC_ALL} to @samp{en_US}), then @command{ls} may
6423 produce output that is sorted differently than you're accustomed to.
6424 In that case, set the @env{LC_ALL} environment variable to @samp{C}.}
6425 If standard output is
6426 a terminal, the output is in columns (sorted vertically) and control
6427 characters are output as question marks; otherwise, the output is listed
6428 one per line and control characters are output as-is.
6430 Because @command{ls} is such a fundamental program, it has accumulated many
6431 options over the years. They are described in the subsections below;
6432 within each section, options are listed alphabetically (ignoring case).
6433 The division of options into the subsections is not absolute, since some
6434 options affect more than one aspect of @command{ls}'s operation.
6436 @cindex exit status of @command{ls}
6441 1 minor problems (e.g., failure to access a file or directory not
6442 specified as a command line argument. This happens when listing a
6443 directory in which entries are actively being removed or renamed.)
6444 2 serious trouble (e.g., memory exhausted, invalid option, failure
6445 to access a file or directory specified as a command line argument
6446 or a directory loop)
6449 Also see @ref{Common options}.
6452 * Which files are listed::
6453 * What information is listed::
6454 * Sorting the output::
6455 * Details about version sort::
6456 * General output formatting::
6457 * Formatting file timestamps::
6458 * Formatting the file names::
6462 @node Which files are listed
6463 @subsection Which files are listed
6465 These options determine which files @command{ls} lists information for.
6466 By default, @command{ls} lists files and the contents of any
6467 directories on the command line, except that in directories it ignores
6468 files whose names start with @samp{.}.
6476 In directories, do not ignore file names that start with @samp{.}.
6481 @opindex --almost-all
6482 In directories, do not ignore all file names that start with @samp{.};
6483 ignore only @file{.} and @file{..}. The @option{--all} (@option{-a})
6484 option overrides this option.
6487 @itemx --ignore-backups
6489 @opindex --ignore-backups
6490 @cindex backup files, ignoring
6491 In directories, ignore files that end with @samp{~}. This option is
6492 equivalent to @samp{--ignore='*~' --ignore='.*~'}.
6497 @opindex --directory
6498 List just the names of directories, as with other types of files, rather
6499 than listing their contents.
6500 @c The following sentence is the same as the one for -F.
6501 Do not follow symbolic links listed on the
6502 command line unless the @option{--dereference-command-line} (@option{-H}),
6503 @option{--dereference} (@option{-L}), or
6504 @option{--dereference-command-line-symlink-to-dir} options are specified.
6507 @itemx --dereference-command-line
6509 @opindex --dereference-command-line
6510 @cindex symbolic links, dereferencing
6511 If a command line argument specifies a symbolic link, show information
6512 for the file the link references rather than for the link itself.
6514 @item --dereference-command-line-symlink-to-dir
6515 @opindex --dereference-command-line-symlink-to-dir
6516 @cindex symbolic links, dereferencing
6517 Do not dereference symbolic links, with one exception:
6518 if a command line argument specifies a symbolic link that refers to
6519 a directory, show information for that directory rather than for the
6521 This is the default behavior when no other dereferencing-related
6522 option has been specified (@option{--classify} (@option{-F}),
6523 @option{--directory} (@option{-d}),
6525 @option{--dereference} (@option{-L}), or
6526 @option{--dereference-command-line} (@option{-H})).
6528 @item --group-directories-first
6529 @opindex --group-directories-first
6530 Group all the directories before the files and then sort the
6531 directories and the files separately using the selected sort key
6532 (see --sort option).
6533 That is, this option specifies a primary sort key,
6534 and the --sort option specifies a secondary key.
6535 However, any use of @option{--sort=none}
6536 (@option{-U}) disables this option altogether.
6538 @item --hide=PATTERN
6539 @opindex --hide=@var{pattern}
6540 In directories, ignore files whose names match the shell pattern
6541 @var{pattern}, unless the @option{--all} (@option{-a}) or
6542 @option{--almost-all} (@option{-A}) is also given. This
6543 option acts like @option{--ignore=@var{pattern}} except that it has no
6544 effect if @option{--all} (@option{-a}) or @option{--almost-all}
6545 (@option{-A}) is also given.
6547 This option can be useful in shell aliases. For example, if
6548 @command{lx} is an alias for @samp{ls --hide='*~'} and @command{ly} is
6549 an alias for @samp{ls --ignore='*~'}, then the command @samp{lx -A}
6550 lists the file @file{README~} even though @samp{ly -A} would not.
6552 @item -I @var{pattern}
6553 @itemx --ignore=@var{pattern}
6555 @opindex --ignore=@var{pattern}
6556 In directories, ignore files whose names match the shell pattern
6557 (not regular expression) @var{pattern}. As
6558 in the shell, an initial @samp{.} in a file name does not match a
6559 wildcard at the start of @var{pattern}. Sometimes it is useful
6560 to give this option several times. For example,
6563 $ ls --ignore='.??*' --ignore='.[^.]' --ignore='#*'
6566 The first option ignores names of length 3 or more that start with @samp{.},
6567 the second ignores all two-character names that start with @samp{.}
6568 except @samp{..}, and the third ignores names that start with @samp{#}.
6571 @itemx --dereference
6573 @opindex --dereference
6574 @cindex symbolic links, dereferencing
6575 When showing file information for a symbolic link, show information
6576 for the file the link references rather than the link itself.
6577 However, even with this option, @command{ls} still prints the name
6578 of the link itself, not the name of the file that the link points to.
6583 @opindex --recursive
6584 @cindex recursive directory listing
6585 @cindex directory listing, recursive
6586 List the contents of all directories recursively.
6591 @node What information is listed
6592 @subsection What information is listed
6594 These options affect the information that @command{ls} displays. By
6595 default, only file names are shown.
6601 @cindex hurd, author, printing
6602 List each file's author when producing long format directory listings.
6603 In GNU/Hurd, file authors can differ from their owners, but in other
6604 operating systems the two are the same.
6610 @cindex dired Emacs mode support
6611 With the long listing (@option{-l}) format, print an additional line after
6615 //DIRED// @var{beg1} @var{end1} @var{beg2} @var{end2} @dots{}
6619 The @var{begn} and @var{endn} are unsigned integers that record the
6620 byte position of the beginning and end of each file name in the output.
6621 This makes it easy for Emacs to find the names, even when they contain
6622 unusual characters such as space or newline, without fancy searching.
6624 If directories are being listed recursively (@option{-R}), output a similar
6625 line with offsets for each subdirectory name:
6628 //SUBDIRED// @var{beg1} @var{end1} @dots{}
6631 Finally, output a line of the form:
6634 //DIRED-OPTIONS// --quoting-style=@var{word}
6638 where @var{word} is the quoting style (@pxref{Formatting the file names}).
6640 Here is an actual example:
6643 $ mkdir -p a/sub/deeper a/sub2
6645 $ touch a/sub/deeper/file
6646 $ ls -gloRF --dired a
6649 -rw-r--r-- 1 0 Jun 10 12:27 f1
6650 -rw-r--r-- 1 0 Jun 10 12:27 f2
6651 drwxr-xr-x 3 4096 Jun 10 12:27 sub/
6652 drwxr-xr-x 2 4096 Jun 10 12:27 sub2/
6656 drwxr-xr-x 2 4096 Jun 10 12:27 deeper/
6660 -rw-r--r-- 1 0 Jun 10 12:27 file
6664 //DIRED// 48 50 84 86 120 123 158 162 217 223 282 286
6665 //SUBDIRED// 2 3 167 172 228 240 290 296
6666 //DIRED-OPTIONS// --quoting-style=literal
6669 Note that the pairs of offsets on the @samp{//DIRED//} line above delimit
6670 these names: @file{f1}, @file{f2}, @file{sub}, @file{sub2}, @file{deeper},
6672 The offsets on the @samp{//SUBDIRED//} line delimit the following
6673 directory names: @file{a}, @file{a/sub}, @file{a/sub/deeper}, @file{a/sub2}.
6675 Here is an example of how to extract the fifth entry name, @samp{deeper},
6676 corresponding to the pair of offsets, 222 and 228:
6679 $ ls -gloRF --dired a > out
6680 $ dd bs=1 skip=222 count=6 < out 2>/dev/null; echo
6684 Note that although the listing above includes a trailing slash
6685 for the @samp{deeper} entry, the offsets select the name without
6686 the trailing slash. However, if you invoke @command{ls} with @option{--dired}
6687 along with an option like @option{--escape} (aka @option{-b}) and operate
6688 on a file whose name contains special characters, notice that the backslash
6693 $ ls -blog --dired 'a b'
6694 -rw-r--r-- 1 0 Jun 10 12:28 a\ b
6696 //DIRED-OPTIONS// --quoting-style=escape
6699 If you use a quoting style that adds quote marks
6700 (e.g., @option{--quoting-style=c}), then the offsets include the quote marks.
6701 So beware that the user may select the quoting style via the environment
6702 variable @env{QUOTING_STYLE}@. Hence, applications using @option{--dired}
6703 should either specify an explicit @option{--quoting-style=literal} option
6704 (aka @option{-N} or @option{--literal}) on the command line, or else be
6705 prepared to parse the escaped names.
6708 @opindex --full-time
6709 Produce long format directory listings, and list times in full. It is
6710 equivalent to using @option{--format=long} with
6711 @option{--time-style=full-iso} (@pxref{Formatting file timestamps}).
6715 Produce long format directory listings, but don't display owner information.
6721 Inhibit display of group information in a long format directory listing.
6722 (This is the default in some non-@sc{gnu} versions of @command{ls}, so we
6723 provide this option for compatibility.)
6731 @cindex inode number, printing
6732 Print the inode number (also called the file serial number and index
6733 number) of each file to the left of the file name. (This number
6734 uniquely identifies each file within a particular file system.)
6737 @itemx --format=long
6738 @itemx --format=verbose
6741 @opindex long ls @r{format}
6742 @opindex verbose ls @r{format}
6743 In addition to the name of each file, print the file type, file mode bits,
6744 number of hard links, owner name, group name, size, and
6745 timestamp (@pxref{Formatting file timestamps}), normally
6746 the modification time. Print question marks for information that
6747 cannot be determined.
6749 Normally the size is printed as a byte count without punctuation, but
6750 this can be overridden (@pxref{Block size}). For example, @option{-h}
6751 prints an abbreviated, human-readable count, and
6752 @samp{--block-size="'1"} prints a byte count with the thousands
6753 separator of the current locale.
6755 For each directory that is listed, preface the files with a line
6756 @samp{total @var{blocks}}, where @var{blocks} is the total disk allocation
6757 for all files in that directory. The block size currently defaults to 1024
6758 bytes, but this can be overridden (@pxref{Block size}).
6759 The @var{blocks} computed counts each hard link separately;
6760 this is arguably a deficiency.
6762 The file type is one of the following characters:
6764 @c The commented-out entries are ones we're not sure about.
6772 character special file
6774 high performance (``contiguous data'') file
6778 door (Solaris 2.5 and up)
6780 @c semaphore, if this is a distinct file type
6784 @c multiplexed file (7th edition Unix; obsolete)
6786 off-line (``migrated'') file (Cray DMF)
6788 network special file (HP-UX)
6792 port (Solaris 10 and up)
6794 @c message queue, if this is a distinct file type
6798 @c shared memory object, if this is a distinct file type
6800 @c typed memory object, if this is a distinct file type
6802 @c whiteout (4.4BSD; not implemented)
6804 some other file type
6807 @cindex permissions, output by @command{ls}
6808 The file mode bits listed are similar to symbolic mode specifications
6809 (@pxref{Symbolic Modes}). But @command{ls} combines multiple bits into the
6810 third character of each set of permissions as follows:
6814 If the set-user-ID or set-group-ID bit and the corresponding executable bit
6818 If the set-user-ID or set-group-ID bit is set but the corresponding
6819 executable bit is not set.
6822 If the restricted deletion flag or sticky bit, and the
6823 other-executable bit, are both set. The restricted deletion flag is
6824 another name for the sticky bit. @xref{Mode Structure}.
6827 If the restricted deletion flag or sticky bit is set but the
6828 other-executable bit is not set.
6831 If the executable bit is set and none of the above apply.
6837 Following the file mode bits is a single character that specifies
6838 whether an alternate access method such as an access control list
6839 applies to the file. When the character following the file mode bits is a
6840 space, there is no alternate access method. When it is a printing
6841 character, then there is such a method.
6843 GNU @command{ls} uses a @samp{.} character to indicate a file
6844 with an SELinux security context, but no other alternate access method.
6846 A file with any other combination of alternate access methods
6847 is marked with a @samp{+} character.
6850 @itemx --numeric-uid-gid
6852 @opindex --numeric-uid-gid
6853 @cindex numeric uid and gid
6854 @cindex numeric user and group IDs
6855 Produce long format directory listings, but
6856 display numeric user and group IDs instead of the owner and group names.
6860 Produce long format directory listings, but don't display group information.
6861 It is equivalent to using @option{--format=long} with @option{--no-group} .
6867 @cindex disk allocation
6868 @cindex size of files, reporting
6869 Print the disk allocation of each file to the left of the file name.
6870 This is the amount of disk space used by the file, which is usually a
6871 bit more than the file's size, but it can be less if the file has holes.
6873 Normally the disk allocation is printed in units of
6874 1024 bytes, but this can be overridden (@pxref{Block size}).
6876 @cindex NFS mounts from BSD to HP-UX
6877 For files that are NFS-mounted from an HP-UX system to a BSD system,
6878 this option reports sizes that are half the correct values. On HP-UX
6879 systems, it reports sizes that are twice the correct values for files
6880 that are NFS-mounted from BSD systems. This is due to a flaw in HP-UX;
6881 it also affects the HP-UX @command{ls} program.
6890 @cindex security context
6891 Display the SELinux security context or @samp{?} if none is found.
6892 When used with the @option{-l} option, print the security context
6893 to the left of the size column.
6898 @node Sorting the output
6899 @subsection Sorting the output
6901 @cindex sorting @command{ls} output
6902 These options change the order in which @command{ls} sorts the information
6903 it outputs. By default, sorting is done by character code
6904 (e.g., ASCII order).
6910 @itemx --time=status
6913 @opindex ctime@r{, printing or sorting by}
6914 @opindex status time@r{, printing or sorting by}
6915 @opindex use time@r{, printing or sorting files by}
6916 If the long listing format (e.g., @option{-l}, @option{-o}) is being used,
6917 print the status change time (the @samp{ctime} in the inode) instead of
6918 the modification time.
6919 When explicitly sorting by time (@option{--sort=time} or @option{-t})
6920 or when not using a long listing format,
6921 sort according to the status change time.
6925 @cindex unsorted directory listing
6926 @cindex directory order, listing by
6927 Primarily, like @option{-U}---do not sort; list the files in whatever
6928 order they are stored in the directory. But also enable @option{-a} (list
6929 all files) and disable @option{-l}, @option{--color}, and @option{-s} (if they
6930 were specified before the @option{-f}).
6936 @cindex reverse sorting
6937 Reverse whatever the sorting method is---e.g., list files in reverse
6938 alphabetical order, youngest first, smallest first, or whatever.
6944 @opindex size of files@r{, sorting files by}
6945 Sort by file size, largest first.
6951 @opindex modification time@r{, sorting files by}
6952 Sort by modification time (the @samp{mtime} in the inode), newest first.
6956 @itemx --time=access
6960 @opindex use time@r{, printing or sorting files by}
6961 @opindex atime@r{, printing or sorting files by}
6962 @opindex access time@r{, printing or sorting files by}
6963 If the long listing format (e.g., @option{--format=long}) is being used,
6964 print the last access time (the @samp{atime} in the inode).
6965 When explicitly sorting by time (@option{--sort=time} or @option{-t})
6966 or when not using a long listing format, sort according to the access time.
6972 @opindex none@r{, sorting option for @command{ls}}
6973 Do not sort; list the files in whatever order they are
6974 stored in the directory. (Do not do any of the other unrelated things
6975 that @option{-f} does.) This is especially useful when listing very large
6976 directories, since not doing any sorting can be noticeably faster.
6979 @itemx --sort=version
6982 @opindex version@r{, sorting option for @command{ls}}
6983 Sort by version name and number, lowest first. It behaves like a default
6984 sort, except that each sequence of decimal digits is treated numerically
6985 as an index/version number. (@xref{Details about version sort}.)
6988 @itemx --sort=extension
6991 @opindex extension@r{, sorting files by}
6992 Sort directory contents alphabetically by file extension (characters
6993 after the last @samp{.}); files with no extension are sorted first.
6998 @node Details about version sort
6999 @subsection Details about version sort
7001 Version sorting handles the fact that file names frequently include indices or
7002 version numbers. Standard sorting usually does not produce the order that one
7003 expects because comparisons are made on a character-by-character basis.
7004 Version sorting is especially useful when browsing directories that contain
7005 many files with indices/version numbers in their names:
7009 abc.zml-1.gz abc.zml-1.gz
7010 abc.zml-12.gz abc.zml-2.gz
7011 abc.zml-2.gz abc.zml-12.gz
7014 Version-sorted strings are compared such that if @var{ver1} and @var{ver2}
7015 are version numbers and @var{prefix} and @var{suffix} (@var{suffix} matching
7016 the regular expression @samp{(\.[A-Za-z~][A-Za-z0-9~]*)*}) are strings then
7017 @var{ver1} < @var{ver2} implies that the name composed of
7018 ``@var{prefix} @var{ver1} @var{suffix}'' sorts before
7019 ``@var{prefix} @var{ver2} @var{suffix}''.
7021 Note also that leading zeros of numeric parts are ignored:
7025 abc-1.007.tgz abc-1.01a.tgz
7026 abc-1.012b.tgz abc-1.007.tgz
7027 abc-1.01a.tgz abc-1.012b.tgz
7030 This functionality is implemented using gnulib's @code{filevercmp} function,
7031 which has some caveats worth noting.
7034 @item @env{LC_COLLATE} is ignored, which means @samp{ls -v} and @samp{sort -V}
7035 will sort non-numeric prefixes as if the @env{LC_COLLATE} locale category
7036 was set to @samp{C}@.
7037 @item Some suffixes will not be matched by the regular
7038 expression mentioned above. Consequently these examples may
7039 not sort as you expect:
7047 abc-1.2.3.4.x86_64.rpm
7048 abc-1.2.3.x86_64.rpm
7052 @node General output formatting
7053 @subsection General output formatting
7055 These options affect the appearance of the overall output.
7060 @itemx --format=single-column
7063 @opindex single-column @r{output of files}
7064 List one file per line. This is the default for @command{ls} when standard
7065 output is not a terminal.
7068 @itemx --format=vertical
7071 @opindex vertical @r{sorted files in columns}
7072 List files in columns, sorted vertically. This is the default for
7073 @command{ls} if standard output is a terminal. It is always the default
7074 for the @command{dir} program.
7075 @sc{gnu} @command{ls} uses variable width columns to display as many files as
7076 possible in the fewest lines.
7078 @item --color [=@var{when}]
7080 @cindex color, distinguishing file types with
7081 Specify whether to use color for distinguishing file types. @var{when}
7082 may be omitted, or one of:
7085 @vindex none @r{color option}
7086 - Do not use color at all. This is the default.
7088 @vindex auto @r{color option}
7089 @cindex terminal, using color iff
7090 - Only use color if standard output is a terminal.
7092 @vindex always @r{color option}
7095 Specifying @option{--color} and no @var{when} is equivalent to
7096 @option{--color=always}.
7097 Piping a colorized listing through a pager like @command{more} or
7098 @command{less} usually produces unreadable results. However, using
7099 @code{more -f} does seem to work.
7102 @vindex SHELL @r{environment variable, and color}
7103 Note that using the @option{--color} option may incur a noticeable
7104 performance penalty when run in a directory with very many entries,
7105 because the default settings require that @command{ls} @code{stat} every
7106 single file it lists.
7107 However, if you would like most of the file-type coloring
7108 but can live without the other coloring options (e.g.,
7109 executable, orphan, sticky, other-writable, capability), use
7110 @command{dircolors} to set the @env{LS_COLORS} environment variable like this,
7112 eval $(dircolors -p | perl -pe \
7113 's/^((CAP|S[ET]|O[TR]|M|E)\w+).*/$1 00/' | dircolors -)
7115 and on a @code{dirent.d_type}-capable file system, @command{ls}
7116 will perform only one @code{stat} call per command line argument.
7120 @itemx --indicator-style=classify
7123 @opindex --indicator-style
7124 @cindex file type and executables, marking
7125 @cindex executables and file type, marking
7126 Append a character to each file name indicating the file type. Also,
7127 for regular files that are executable, append @samp{*}. The file type
7128 indicators are @samp{/} for directories, @samp{@@} for symbolic links,
7129 @samp{|} for FIFOs, @samp{=} for sockets, @samp{>} for doors,
7130 and nothing for regular files.
7131 @c The following sentence is the same as the one for -d.
7132 Do not follow symbolic links listed on the
7133 command line unless the @option{--dereference-command-line} (@option{-H}),
7134 @option{--dereference} (@option{-L}), or
7135 @option{--dereference-command-line-symlink-to-dir} options are specified.
7138 @itemx --indicator-style=file-type
7139 @opindex --file-type
7140 @opindex --indicator-style
7141 @cindex file type, marking
7142 Append a character to each file name indicating the file type. This is
7143 like @option{-F}, except that executables are not marked.
7145 @item --indicator-style=@var{word}
7146 @opindex --indicator-style
7147 Append a character indicator with style @var{word} to entry names,
7152 Do not append any character indicator; this is the default.
7154 Append @samp{/} for directories. This is the same as the @option{-p}
7157 Append @samp{/} for directories, @samp{@@} for symbolic links, @samp{|}
7158 for FIFOs, @samp{=} for sockets, and nothing for regular files. This is
7159 the same as the @option{--file-type} option.
7161 Append @samp{*} for executable regular files, otherwise behave as for
7162 @samp{file-type}. This is the same as the @option{-F} or
7163 @option{--classify} option.
7169 @opindex --kibibytes
7170 Set the default block size to its normal value of 1024 bytes,
7171 overriding any contrary specification in environment variables
7172 (@pxref{Block size}). This option is in turn overridden by the
7173 @option{--block-size}, @option{-h} or @option{--human-readable}, and
7174 @option{--si} options.
7176 The @option{-k} or @option{--kibibytes} option affects the
7177 per-directory block count written by the @option{-l} and similar
7178 options, and the size written by the @option{-s} or @option{--size}
7179 option. It does not affect the file size written by @option{-l}.
7182 @itemx --format=commas
7185 @opindex commas@r{, outputting between files}
7186 List files horizontally, with as many as will fit on each line,
7187 separated by @samp{, } (a comma and a space).
7190 @itemx --indicator-style=slash
7192 @opindex --indicator-style
7193 @cindex file type, marking
7194 Append a @samp{/} to directory names.
7197 @itemx --format=across
7198 @itemx --format=horizontal
7201 @opindex across@r{, listing files}
7202 @opindex horizontal@r{, listing files}
7203 List the files in columns, sorted horizontally.
7206 @itemx --tabsize=@var{cols}
7209 Assume that each tab stop is @var{cols} columns wide. The default is 8.
7210 @command{ls} uses tabs where possible in the output, for efficiency. If
7211 @var{cols} is zero, do not use tabs at all.
7213 @c FIXME: remove in 2009, if Apple Terminal has been fixed for long enough.
7214 Some terminal emulators (at least Apple Terminal 1.5 (133) from Mac OS X 10.4.8)
7215 do not properly align columns to the right of a TAB following a
7216 non-ASCII byte. If you use such a terminal emulator, use the
7217 @option{-T0} option or put @code{TABSIZE=0} in your environment to tell
7218 @command{ls} to align using spaces, not tabs.
7221 @itemx --width=@var{cols}
7225 Assume the screen is @var{cols} columns wide. The default is taken
7226 from the terminal settings if possible; otherwise the environment
7227 variable @env{COLUMNS} is used if it is set; otherwise the default
7233 @node Formatting file timestamps
7234 @subsection Formatting file timestamps
7236 By default, file timestamps are listed in abbreviated form, using
7237 a date like @samp{Mar 30@ @ 2002} for non-recent timestamps, and a
7238 date-without-year and time like @samp{Mar 30 23:45} for recent timestamps.
7239 This format can change depending on the current locale as detailed below.
7242 A timestamp is considered to be @dfn{recent} if it is less than six
7243 months old, and is not dated in the future. If a timestamp dated
7244 today is not listed in recent form, the timestamp is in the future,
7245 which means you probably have clock skew problems which may break
7246 programs like @command{make} that rely on file timestamps.
7249 Time stamps are listed according to the time zone rules specified by
7250 the @env{TZ} environment variable, or by the system default rules if
7251 @env{TZ} is not set. @xref{TZ Variable,, Specifying the Time Zone
7252 with @env{TZ}, libc, The GNU C Library Reference Manual}.
7254 The following option changes how file timestamps are printed.
7257 @item --time-style=@var{style}
7258 @opindex --time-style
7260 List timestamps in style @var{style}. The @var{style} should
7261 be one of the following:
7266 List timestamps using @var{format}, where @var{format} is interpreted
7267 like the format argument of @command{date} (@pxref{date invocation}).
7268 For example, @option{--time-style="+%Y-%m-%d %H:%M:%S"} causes
7269 @command{ls} to list timestamps like @samp{2002-03-30 23:45:56}. As
7270 with @command{date}, @var{format}'s interpretation is affected by the
7271 @env{LC_TIME} locale category.
7273 If @var{format} contains two format strings separated by a newline,
7274 the former is used for non-recent files and the latter for recent
7275 files; if you want output columns to line up, you may need to insert
7276 spaces in one of the two formats.
7279 List timestamps in full using ISO 8601 date, time, and time zone
7280 format with nanosecond precision, e.g., @samp{2002-03-30
7281 23:45:56.477817180 -0700}. This style is equivalent to
7282 @samp{+%Y-%m-%d %H:%M:%S.%N %z}.
7284 This is useful because the time output includes all the information that
7285 is available from the operating system. For example, this can help
7286 explain @command{make}'s behavior, since GNU @command{make}
7287 uses the full timestamp to determine whether a file is out of date.
7290 List ISO 8601 date and time in minutes, e.g.,
7291 @samp{2002-03-30 23:45}. These timestamps are shorter than
7292 @samp{full-iso} timestamps, and are usually good enough for everyday
7293 work. This style is equivalent to @samp{+%Y-%m-%d %H:%M}.
7296 List ISO 8601 dates for non-recent timestamps (e.g.,
7297 @samp{2002-03-30@ }), and ISO 8601 month, day, hour, and
7298 minute for recent timestamps (e.g., @samp{03-30 23:45}). These
7299 timestamps are uglier than @samp{long-iso} timestamps, but they carry
7300 nearly the same information in a smaller space and their brevity helps
7301 @command{ls} output fit within traditional 80-column output lines.
7302 The following two @command{ls} invocations are equivalent:
7307 ls -l --time-style="+%Y-%m-%d $newline%m-%d %H:%M"
7308 ls -l --time-style="iso"
7313 List timestamps in a locale-dependent form. For example, a Finnish
7314 locale might list non-recent timestamps like @samp{maalis 30@ @ 2002}
7315 and recent timestamps like @samp{maalis 30 23:45}. Locale-dependent
7316 timestamps typically consume more space than @samp{iso} timestamps and
7317 are harder for programs to parse because locale conventions vary so
7318 widely, but they are easier for many people to read.
7320 The @env{LC_TIME} locale category specifies the timestamp format. The
7321 default POSIX locale uses timestamps like @samp{Mar 30@
7322 @ 2002} and @samp{Mar 30 23:45}; in this locale, the following two
7323 @command{ls} invocations are equivalent:
7328 ls -l --time-style="+%b %e %Y$newline%b %e %H:%M"
7329 ls -l --time-style="locale"
7332 Other locales behave differently. For example, in a German locale,
7333 @option{--time-style="locale"} might be equivalent to
7334 @option{--time-style="+%e. %b %Y $newline%e. %b %H:%M"}
7335 and might generate timestamps like @samp{30. M@"ar 2002@ } and
7336 @samp{30. M@"ar 23:45}.
7338 @item posix-@var{style}
7340 List POSIX-locale timestamps if the @env{LC_TIME} locale
7341 category is POSIX, @var{style} timestamps otherwise. For
7342 example, the @samp{posix-long-iso} style lists
7343 timestamps like @samp{Mar 30@ @ 2002} and @samp{Mar 30 23:45} when in
7344 the POSIX locale, and like @samp{2002-03-30 23:45} otherwise.
7349 You can specify the default value of the @option{--time-style} option
7350 with the environment variable @env{TIME_STYLE}; if @env{TIME_STYLE} is not set
7351 the default style is @samp{locale}. GNU Emacs 21.3 and
7352 later use the @option{--dired} option and therefore can parse any date
7353 format, but if you are using Emacs 21.1 or 21.2 and specify a
7354 non-POSIX locale you may need to set
7355 @samp{TIME_STYLE="posix-long-iso"}.
7357 To avoid certain denial-of-service attacks, timestamps that would be
7358 longer than 1000 bytes may be treated as errors.
7361 @node Formatting the file names
7362 @subsection Formatting the file names
7364 These options change how file names themselves are printed.
7370 @itemx --quoting-style=escape
7373 @opindex --quoting-style
7374 @cindex backslash sequences for file names
7375 Quote nongraphic characters in file names using alphabetic and octal
7376 backslash sequences like those used in C.
7380 @itemx --quoting-style=literal
7383 @opindex --quoting-style
7384 Do not quote file names. However, with @command{ls} nongraphic
7385 characters are still printed as question marks if the output is a
7386 terminal and you do not specify the @option{--show-control-chars}
7390 @itemx --hide-control-chars
7392 @opindex --hide-control-chars
7393 Print question marks instead of nongraphic characters in file names.
7394 This is the default if the output is a terminal and the program is
7399 @itemx --quoting-style=c
7401 @opindex --quote-name
7402 @opindex --quoting-style
7403 Enclose file names in double quotes and quote nongraphic characters as
7406 @item --quoting-style=@var{word}
7407 @opindex --quoting-style
7408 @cindex quoting style
7409 Use style @var{word} to quote file names and other strings that may
7410 contain arbitrary characters. The @var{word} should
7411 be one of the following:
7415 Output strings as-is; this is the same as the @option{-N} or
7416 @option{--literal} option.
7418 Quote strings for the shell if they contain shell metacharacters or would
7419 cause ambiguous output.
7420 The quoting is suitable for POSIX-compatible shells like
7421 @command{bash}, but it does not always work for incompatible shells
7424 Quote strings for the shell, even if they would normally not require quoting.
7426 Quote strings as for C character string literals, including the
7427 surrounding double-quote characters; this is the same as the
7428 @option{-Q} or @option{--quote-name} option.
7430 Quote strings as for C character string literals, except omit the
7431 surrounding double-quote
7432 characters; this is the same as the @option{-b} or @option{--escape} option.
7434 Quote strings as for C character string literals, except use
7435 surrounding quotation marks appropriate for the
7438 @c Use @t instead of @samp to avoid duplicate quoting in some output styles.
7439 Quote strings as for C character string literals, except use
7440 surrounding quotation marks appropriate for the locale, and quote
7441 @t{'like this'} instead of @t{"like
7442 this"} in the default C locale. This looks nicer on many displays.
7445 You can specify the default value of the @option{--quoting-style} option
7446 with the environment variable @env{QUOTING_STYLE}@. If that environment
7447 variable is not set, the default value is @samp{literal}, but this
7448 default may change to @samp{shell} in a future version of this package.
7450 @item --show-control-chars
7451 @opindex --show-control-chars
7452 Print nongraphic characters as-is in file names.
7453 This is the default unless the output is a terminal and the program is
7459 @node dir invocation
7460 @section @command{dir}: Briefly list directory contents
7463 @cindex directory listing, brief
7465 @command{dir} is equivalent to @code{ls -C
7466 -b}; that is, by default files are listed in columns, sorted vertically,
7467 and special characters are represented by backslash escape sequences.
7469 @xref{ls invocation, @command{ls}}.
7472 @node vdir invocation
7473 @section @command{vdir}: Verbosely list directory contents
7476 @cindex directory listing, verbose
7478 @command{vdir} is equivalent to @code{ls -l
7479 -b}; that is, by default files are listed in long format and special
7480 characters are represented by backslash escape sequences.
7482 @node dircolors invocation
7483 @section @command{dircolors}: Color setup for @command{ls}
7487 @cindex setup for color
7489 @command{dircolors} outputs a sequence of shell commands to set up the
7490 terminal for color output from @command{ls} (and @command{dir}, etc.).
7494 eval "$(dircolors [@var{option}]@dots{} [@var{file}])"
7497 If @var{file} is specified, @command{dircolors} reads it to determine which
7498 colors to use for which file types and extensions. Otherwise, a
7499 precompiled database is used. For details on the format of these files,
7500 run @samp{dircolors --print-database}.
7502 To make @command{dircolors} read a @file{~/.dircolors} file if it
7503 exists, you can put the following lines in your @file{~/.bashrc} (or
7504 adapt them to your favorite shell):
7508 test -r $d && eval "$(dircolors $d)"
7512 @vindex SHELL @r{environment variable, and color}
7513 The output is a shell command to set the @env{LS_COLORS} environment
7514 variable. You can specify the shell syntax to use on the command line,
7515 or @command{dircolors} will guess it from the value of the @env{SHELL}
7516 environment variable.
7518 The program accepts the following options. Also see @ref{Common options}.
7523 @itemx --bourne-shell
7526 @opindex --bourne-shell
7527 @cindex Bourne shell syntax for color setup
7528 @cindex @command{sh} syntax for color setup
7529 Output Bourne shell commands. This is the default if the @env{SHELL}
7530 environment variable is set and does not end with @samp{csh} or
7539 @cindex C shell syntax for color setup
7540 @cindex @command{csh} syntax for color setup
7541 Output C shell commands. This is the default if @code{SHELL} ends with
7542 @command{csh} or @command{tcsh}.
7545 @itemx --print-database
7547 @opindex --print-database
7548 @cindex color database, printing
7549 @cindex database for color setup, printing
7550 @cindex printing color database
7551 Print the (compiled-in) default color configuration database. This
7552 output is itself a valid configuration file, and is fairly descriptive
7553 of the possibilities.
7560 @node Basic operations
7561 @chapter Basic operations
7563 @cindex manipulating files
7565 This chapter describes the commands for basic file manipulation:
7566 copying, moving (renaming), and deleting (removing).
7569 * cp invocation:: Copy files.
7570 * dd invocation:: Convert and copy a file.
7571 * install invocation:: Copy files and set attributes.
7572 * mv invocation:: Move (rename) files.
7573 * rm invocation:: Remove files or directories.
7574 * shred invocation:: Remove files more securely.
7579 @section @command{cp}: Copy files and directories
7582 @cindex copying files and directories
7583 @cindex files, copying
7584 @cindex directories, copying
7586 @command{cp} copies files (or, optionally, directories). The copy is
7587 completely independent of the original. You can either copy one file to
7588 another, or copy arbitrarily many files to a destination directory.
7592 cp [@var{option}]@dots{} [-T] @var{source} @var{dest}
7593 cp [@var{option}]@dots{} @var{source}@dots{} @var{directory}
7594 cp [@var{option}]@dots{} -t @var{directory} @var{source}@dots{}
7599 If two file names are given, @command{cp} copies the first file to the
7603 If the @option{--target-directory} (@option{-t}) option is given, or
7604 failing that if the last file is a directory and the
7605 @option{--no-target-directory} (@option{-T}) option is not given,
7606 @command{cp} copies each @var{source} file to the specified directory,
7607 using the @var{source}s' names.
7610 Generally, files are written just as they are read. For exceptions,
7611 see the @option{--sparse} option below.
7613 By default, @command{cp} does not copy directories. However, the
7614 @option{-R}, @option{-a}, and @option{-r} options cause @command{cp} to
7615 copy recursively by descending into source directories and copying files
7616 to corresponding destination directories.
7618 When copying from a symbolic link, @command{cp} normally follows the
7619 link only when not copying
7620 recursively. This default can be overridden with the
7621 @option{--archive} (@option{-a}), @option{-d}, @option{--dereference}
7622 (@option{-L}), @option{--no-dereference} (@option{-P}), and
7623 @option{-H} options. If more than one of these options is specified,
7624 the last one silently overrides the others.
7626 When copying to a symbolic link, @command{cp} follows the
7627 link only when it refers to an existing regular file.
7628 However, when copying to a dangling symbolic link, @command{cp}
7629 refuses by default, and fails with a diagnostic, since the operation
7630 is inherently dangerous. This behavior is contrary to historical
7631 practice and to POSIX@.
7632 Set @env{POSIXLY_CORRECT} to make @command{cp} attempt to create
7633 the target of a dangling destination symlink, in spite of the possible risk.
7634 Also, when an option like
7635 @option{--backup} or @option{--link} acts to rename or remove the
7636 destination before copying, @command{cp} renames or removes the
7637 symbolic link rather than the file it points to.
7639 By default, @command{cp} copies the contents of special files only
7640 when not copying recursively. This default can be overridden with the
7641 @option{--copy-contents} option.
7643 @cindex self-backups
7644 @cindex backups, making only
7645 @command{cp} generally refuses to copy a file onto itself, with the
7646 following exception: if @option{--force --backup} is specified with
7647 @var{source} and @var{dest} identical, and referring to a regular file,
7648 @command{cp} will make a backup file, either regular or numbered, as
7649 specified in the usual ways (@pxref{Backup options}). This is useful when
7650 you simply want to make a backup of an existing file before changing it.
7652 The program accepts the following options. Also see @ref{Common options}.
7659 Preserve as much as possible of the structure and attributes of the
7660 original files in the copy (but do not attempt to preserve internal
7661 directory structure; i.e., @samp{ls -U} may list the entries in a copied
7662 directory in a different order).
7663 Try to preserve SELinux security context and extended attributes (xattr),
7664 but ignore any failure to do that and print no corresponding diagnostic.
7665 Equivalent to @option{-dR --preserve=all} with the reduced diagnostics.
7667 @item --attributes-only
7668 @opindex --attributes-only
7669 Copy only the specified attributes of the source file to the destination.
7670 If the destination already exists, do not alter its contents.
7671 See the @option{--preserve} option for controlling which attributes to copy.
7674 @itemx @w{@kbd{--backup}[=@var{method}]}
7677 @vindex VERSION_CONTROL
7678 @cindex backups, making
7679 @xref{Backup options}.
7680 Make a backup of each file that would otherwise be overwritten or removed.
7681 As a special case, @command{cp} makes a backup of @var{source} when the force
7682 and backup options are given and @var{source} and @var{dest} are the same
7683 name for an existing, regular file. One useful application of this
7684 combination of options is this tiny Bourne shell script:
7688 # Usage: backup FILE...
7689 # Create a @sc{gnu}-style backup of each listed FILE.
7692 cp --backup --force --preserve=all -- "$i" "$i" || fail=1
7697 @item --copy-contents
7698 @cindex directories, copying recursively
7699 @cindex copying directories recursively
7700 @cindex recursively copying directories
7701 @cindex non-directories, copying as special files
7702 If copying recursively, copy the contents of any special files (e.g.,
7703 FIFOs and device files) as if they were regular files. This means
7704 trying to read the data in each source file and writing it to the
7705 destination. It is usually a mistake to use this option, as it
7706 normally has undesirable effects on special files like FIFOs and the
7707 ones typically found in the @file{/dev} directory. In most cases,
7708 @code{cp -R --copy-contents} will hang indefinitely trying to read
7709 from FIFOs and special files like @file{/dev/console}, and it will
7710 fill up your destination disk if you use it to copy @file{/dev/zero}.
7711 This option has no effect unless copying recursively, and it does not
7712 affect the copying of symbolic links.
7716 @cindex symbolic links, copying
7717 @cindex hard links, preserving
7718 Copy symbolic links as symbolic links rather than copying the files that
7719 they point to, and preserve hard links between source files in the copies.
7720 Equivalent to @option{--no-dereference --preserve=links}.
7726 When copying without this option and an existing destination file cannot
7727 be opened for writing, the copy fails. However, with @option{--force},
7728 when a destination file cannot be opened, @command{cp} then removes it and
7729 tries to open it again. Contrast this behavior with that enabled by
7730 @option{--link} and @option{--symbolic-link}, whereby the destination file
7731 is never opened but rather is removed unconditionally. Also see the
7732 description of @option{--remove-destination}.
7734 This option is independent of the @option{--interactive} or
7735 @option{-i} option: neither cancels the effect of the other.
7737 This option is ignored when the @option{--no-clobber} or @option{-n} option
7742 If a command line argument specifies a symbolic link, then copy the
7743 file it points to rather than the symbolic link itself. However,
7744 copy (preserving its nature) any symbolic link that is encountered
7745 via recursive traversal.
7748 @itemx --interactive
7750 @opindex --interactive
7751 When copying a file other than a directory, prompt whether to
7752 overwrite an existing destination file. The @option{-i} option overrides
7753 a previous @option{-n} option.
7759 Make hard links instead of copies of non-directories.
7762 @itemx --dereference
7764 @opindex --dereference
7765 Follow symbolic links when copying from them.
7766 With this option, @command{cp} cannot create a symbolic link.
7767 For example, a symlink (to regular file) in the source tree will be copied to
7768 a regular file in the destination tree.
7773 @opindex --no-clobber
7774 Do not overwrite an existing file. The @option{-n} option overrides a previous
7775 @option{-i} option. This option is mutually exclusive with @option{-b} or
7776 @option{--backup} option.
7779 @itemx --no-dereference
7781 @opindex --no-dereference
7782 @cindex symbolic links, copying
7783 Copy symbolic links as symbolic links rather than copying the files that
7784 they point to. This option affects only symbolic links in the source;
7785 symbolic links in the destination are always followed if possible.
7788 @itemx @w{@kbd{--preserve}[=@var{attribute_list}]}
7791 @cindex file information, preserving, extended attributes, xattr
7792 Preserve the specified attributes of the original files.
7793 If specified, the @var{attribute_list} must be a comma-separated list
7794 of one or more of the following strings:
7798 Preserve the file mode bits and access control lists.
7800 Preserve the owner and group. On most modern systems,
7801 only users with appropriate privileges may change the owner of a file,
7803 may preserve the group ownership of a file only if they happen to be
7804 a member of the desired group.
7806 Preserve the times of last access and last modification, when possible.
7807 On older systems, it is not possible to preserve these attributes
7808 when the affected file is a symbolic link.
7809 However, many systems now provide the @code{utimensat} function,
7810 which makes it possible even for symbolic links.
7812 Preserve in the destination files
7813 any links between corresponding source files.
7814 Note that with @option{-L} or @option{-H}, this option can convert
7815 symbolic links to hard links. For example,
7817 $ mkdir c; : > a; ln -s a b; cp -aH a b c; ls -i1 c
7822 Note the inputs: @file{b} is a symlink to regular file @file{a},
7823 yet the files in destination directory, @file{c/}, are hard-linked.
7824 Since @option{-a} implies @option{--preserve=links}, and since @option{-H}
7825 tells @command{cp} to dereference command line arguments, it sees two files
7826 with the same inode number, and preserves the perceived hard link.
7828 Here is a similar example that exercises @command{cp}'s @option{-L} option:
7830 $ mkdir b c; (cd b; : > a; ln -s a b); cp -aL b c; ls -i1 c/b
7836 Preserve SELinux security context of the file, or fail with full diagnostics.
7838 Preserve extended attributes of the file, or fail with full diagnostics.
7839 If @command{cp} is built without xattr support, ignore this option.
7840 If SELinux context, ACLs or Capabilities are implemented using xattrs,
7841 they are preserved implicitly by this option as well, i.e., even without
7842 specifying @option{--preserve=mode} or @option{--preserve=context}.
7844 Preserve all file attributes.
7845 Equivalent to specifying all of the above, but with the difference
7846 that failure to preserve SELinux security context or extended attributes
7847 does not change @command{cp}'s exit status. In contrast to @option{-a},
7848 all but @samp{Operation not supported} warnings are output.
7851 Using @option{--preserve} with no @var{attribute_list} is equivalent
7852 to @option{--preserve=mode,ownership,timestamps}.
7854 In the absence of this option, each destination file is created with the
7855 mode bits of the corresponding source file, minus the bits set in the
7856 umask and minus the set-user-ID and set-group-ID bits.
7857 @xref{File permissions}.
7859 @item @w{@kbd{--no-preserve}=@var{attribute_list}}
7860 @cindex file information, preserving
7861 Do not preserve the specified attributes. The @var{attribute_list}
7862 has the same form as for @option{--preserve}.
7866 @cindex parent directories and @command{cp}
7867 Form the name of each destination file by appending to the target
7868 directory a slash and the specified name of the source file. The last
7869 argument given to @command{cp} must be the name of an existing directory.
7870 For example, the command:
7873 cp --parents a/b/c existing_dir
7877 copies the file @file{a/b/c} to @file{existing_dir/a/b/c}, creating
7878 any missing intermediate directories.
7885 @opindex --recursive
7886 @cindex directories, copying recursively
7887 @cindex copying directories recursively
7888 @cindex recursively copying directories
7889 @cindex non-directories, copying as special files
7890 Copy directories recursively. By default, do not follow symbolic
7891 links in the source; see the @option{--archive} (@option{-a}), @option{-d},
7892 @option{--dereference} (@option{-L}), @option{--no-dereference}
7893 (@option{-P}), and @option{-H} options. Special files are copied by
7894 creating a destination file of the same type as the source; see the
7895 @option{--copy-contents} option. It is not portable to use
7896 @option{-r} to copy symbolic links or special files. On some
7897 non-@sc{gnu} systems, @option{-r} implies the equivalent of
7898 @option{-L} and @option{--copy-contents} for historical reasons.
7899 Also, it is not portable to use @option{-R} to copy symbolic links
7900 unless you also specify @option{-P}, as POSIX allows
7901 implementations that dereference symbolic links by default.
7903 @item --reflink[=@var{when}]
7904 @opindex --reflink[=@var{when}]
7907 @cindex copy on write
7908 Perform a lightweight, copy-on-write (COW) copy, if supported by the
7909 file system. Once it has succeeded, beware that the source and destination
7910 files share the same disk data blocks as long as they remain unmodified.
7911 Thus, if a disk I/O error affects data blocks of one of the files,
7912 the other suffers the same fate.
7914 The @var{when} value can be one of the following:
7918 The default behavior: if the copy-on-write operation is not supported
7919 then report the failure for each file and exit with a failure status.
7922 If the copy-on-write operation is not supported then fall back
7923 to the standard copy behaviour.
7926 This option is overridden by the @option{--link}, @option{--symbolic-link}
7927 and @option{--attributes-only} options, thus allowing it to be used
7928 to configure the default data copying behavior for @command{cp}.
7929 For example, with the following alias, @command{cp} will use the
7930 minimum amount of space supported by the file system.
7933 alias cp='cp --reflink=auto --sparse=always'
7936 @item --remove-destination
7937 @opindex --remove-destination
7938 Remove each existing destination file before attempting to open it
7939 (contrast with @option{-f} above).
7941 @item --sparse=@var{when}
7942 @opindex --sparse=@var{when}
7943 @cindex sparse files, copying
7944 @cindex holes, copying files with
7945 @findex read @r{system call, and holes}
7946 A @dfn{sparse file} contains @dfn{holes}---a sequence of zero bytes that
7947 does not occupy any physical disk blocks; the @samp{read} system call
7948 reads these as zeros. This can both save considerable disk space and
7949 increase speed, since many binary files contain lots of consecutive zero
7950 bytes. By default, @command{cp} detects holes in input source files via a crude
7951 heuristic and makes the corresponding output file sparse as well.
7952 Only regular files may be sparse.
7954 The @var{when} value can be one of the following:
7958 The default behavior: if the input file is sparse, attempt to make
7959 the output file sparse, too. However, if an output file exists but
7960 refers to a non-regular file, then do not attempt to make it sparse.
7963 For each sufficiently long sequence of zero bytes in the input file,
7964 attempt to create a corresponding hole in the output file, even if the
7965 input file does not appear to be sparse.
7966 This is useful when the input file resides on a file system
7967 that does not support sparse files
7968 (for example, @samp{efs} file systems in SGI IRIX 5.3 and earlier),
7969 but the output file is on a type of file system that does support them.
7970 Holes may be created only in regular files, so if the destination file
7971 is of some other type, @command{cp} does not even try to make it sparse.
7974 Never make the output file sparse.
7975 This is useful in creating a file for use with the @command{mkswap} command,
7976 since such a file must not have any holes.
7979 @optStripTrailingSlashes
7982 @itemx --symbolic-link
7984 @opindex --symbolic-link
7985 @cindex symbolic links, copying with
7986 Make symbolic links instead of copies of non-directories. All source
7987 file names must be absolute (starting with @samp{/}) unless the
7988 destination files are in the current directory. This option merely
7989 results in an error message on systems that do not support symbolic links.
7995 @optNoTargetDirectory
8001 @cindex newer files, copying only
8002 Do not copy a non-directory that has an existing destination with the
8003 same or newer modification time. If time stamps are being preserved,
8004 the comparison is to the source time stamp truncated to the
8005 resolutions of the destination file system and of the system calls
8006 used to update time stamps; this avoids duplicate work if several
8007 @samp{cp -pu} commands are executed with the same source and destination.
8008 If @option{--preserve=links} is also specified (like with @samp{cp -au}
8009 for example), that will take precedence. Consequently, depending on the
8010 order that files are processed from the source, newer files in the destination
8011 may be replaced, to mirror hard links in the source.
8017 Print the name of each file before copying it.
8020 @itemx --one-file-system
8022 @opindex --one-file-system
8023 @cindex file systems, omitting copying to different
8024 Skip subdirectories that are on different file systems from the one that
8025 the copy started on.
8026 However, mount point directories @emph{are} copied.
8034 @section @command{dd}: Convert and copy a file
8037 @cindex converting while copying a file
8039 @command{dd} copies a file (from standard input to standard output, by
8040 default) with a changeable I/O block size, while optionally performing
8041 conversions on it. Synopses:
8044 dd [@var{operand}]@dots{}
8048 The only options are @option{--help} and @option{--version}.
8049 @xref{Common options}. @command{dd} accepts the following operands,
8050 whose syntax was inspired by the DD (data definition) statement of
8057 Read from @var{file} instead of standard input.
8061 Write to @var{file} instead of standard output. Unless
8062 @samp{conv=notrunc} is given, @command{dd} truncates @var{file} to zero
8063 bytes (or the size specified with @samp{seek=}).
8065 @item ibs=@var{bytes}
8067 @cindex block size of input
8068 @cindex input block size
8069 Set the input block size to @var{bytes}.
8070 This makes @command{dd} read @var{bytes} per block.
8071 The default is 512 bytes.
8073 @item obs=@var{bytes}
8075 @cindex block size of output
8076 @cindex output block size
8077 Set the output block size to @var{bytes}.
8078 This makes @command{dd} write @var{bytes} per block.
8079 The default is 512 bytes.
8081 @item bs=@var{bytes}
8084 Set both input and output block sizes to @var{bytes}.
8085 This makes @command{dd} read and write @var{bytes} per block,
8086 overriding any @samp{ibs} and @samp{obs} settings.
8087 In addition, if no data-transforming @option{conv} option is specified,
8088 input is copied to the output as soon as it's read,
8089 even if it is smaller than the block size.
8091 @item cbs=@var{bytes}
8093 @cindex block size of conversion
8094 @cindex conversion block size
8095 @cindex fixed-length records, converting to variable-length
8096 @cindex variable-length records, converting to fixed-length
8097 Set the conversion block size to @var{bytes}.
8098 When converting variable-length records to fixed-length ones
8099 (@option{conv=block}) or the reverse (@option{conv=unblock}),
8100 use @var{bytes} as the fixed record length.
8104 Skip @var{n} @samp{ibs}-byte blocks in the input file before copying.
8105 If @samp{iflag=skip_bytes} is specified, @var{n} is interpreted
8106 as a byte count rather than a block count.
8110 Skip @var{n} @samp{obs}-byte blocks in the output file before copying.
8111 if @samp{oflag=seek_bytes} is specified, @var{n} is interpreted
8112 as a byte count rather than a block count.
8116 Copy @var{n} @samp{ibs}-byte blocks from the input file, instead
8117 of everything until the end of the file.
8118 if @samp{iflag=count_bytes} is specified, @var{n} is interpreted
8119 as a byte count rather than a block count.
8121 @item status=@var{which}
8123 Transfer information is normally output to stderr upon
8124 receipt of the @samp{INFO} signal or when @command{dd} exits.
8125 Specifying @var{which} will identify which information to suppress.
8130 @opindex noxfer @r{dd status=}
8131 Do not print the transfer rate and volume statistics
8132 that normally make up the last status line.
8135 @opindex none @r{dd status=}
8136 Do not print any informational messages to stderr.
8137 Error messages are output as normal.
8141 @item conv=@var{conversion}[,@var{conversion}]@dots{}
8143 Convert the file as specified by the @var{conversion} argument(s).
8144 (No spaces around any comma(s).)
8151 @opindex ascii@r{, converting to}
8152 Convert EBCDIC to ASCII,
8153 using the conversion table specified by POSIX@.
8154 This provides a 1:1 translation for all 256 bytes.
8157 @opindex ebcdic@r{, converting to}
8158 Convert ASCII to EBCDIC@.
8159 This is the inverse of the @samp{ascii} conversion.
8162 @opindex alternate ebcdic@r{, converting to}
8163 Convert ASCII to alternate EBCDIC,
8164 using the alternate conversion table specified by POSIX@.
8165 This is not a 1:1 translation, but reflects common historical practice
8166 for @samp{~}, @samp{[}, and @samp{]}.
8168 The @samp{ascii}, @samp{ebcdic}, and @samp{ibm} conversions are
8172 @opindex block @r{(space-padding)}
8173 For each line in the input, output @samp{cbs} bytes, replacing the
8174 input newline with a space and padding with spaces as necessary.
8178 Remove any trailing spaces in each @samp{cbs}-sized input block,
8179 and append a newline.
8181 The @samp{block} and @samp{unblock} conversions are mutually exclusive.
8184 @opindex lcase@r{, converting to}
8185 Change uppercase letters to lowercase.
8188 @opindex ucase@r{, converting to}
8189 Change lowercase letters to uppercase.
8191 The @samp{lcase} and @samp{ucase} conversions are mutually exclusive.
8195 Try to seek rather than write @sc{nul} output blocks.
8196 On a file system that supports sparse files, this will create
8197 sparse output when extending the output file.
8198 Be careful when using this option in conjunction with
8199 @samp{conv=notrunc} or @samp{oflag=append}.
8200 With @samp{conv=notrunc}, existing data in the output file
8201 corresponding to @sc{nul} blocks from the input, will be untouched.
8202 With @samp{oflag=append} the seeks performed will be ineffective.
8203 Similarly, when the output is a device rather than a file,
8204 @sc{nul} input blocks are not copied, and therefore this option
8205 is most useful with virtual or pre zeroed devices.
8208 @opindex swab @r{(byte-swapping)}
8209 @cindex byte-swapping
8210 Swap every pair of input bytes. @sc{gnu} @command{dd}, unlike others, works
8211 when an odd number of bytes are read---the last byte is simply copied
8212 (since there is nothing to swap it with).
8215 @opindex sync @r{(padding with ASCII @sc{nul}s)}
8216 Pad every input block to size of @samp{ibs} with trailing zero bytes.
8217 When used with @samp{block} or @samp{unblock}, pad with spaces instead of
8222 The following ``conversions'' are really file flags
8223 and don't affect internal processing:
8228 @cindex creating output file, requiring
8229 Fail if the output file already exists; @command{dd} must create the
8234 @cindex creating output file, avoiding
8235 Do not create the output file; the output file must already exist.
8237 The @samp{excl} and @samp{nocreat} conversions are mutually exclusive.
8241 @cindex truncating output file, avoiding
8242 Do not truncate the output file.
8246 @cindex read errors, ignoring
8247 Continue after read errors.
8251 @cindex synchronized data writes, before finishing
8252 Synchronize output data just before finishing. This forces a physical
8253 write of output data.
8257 @cindex synchronized data and metadata writes, before finishing
8258 Synchronize output data and metadata just before finishing. This
8259 forces a physical write of output data and metadata.
8263 @item iflag=@var{flag}[,@var{flag}]@dots{}
8265 Access the input file using the flags specified by the @var{flag}
8266 argument(s). (No spaces around any comma(s).)
8268 @item oflag=@var{flag}[,@var{flag}]@dots{}
8270 Access the output file using the flags specified by the @var{flag}
8271 argument(s). (No spaces around any comma(s).)
8273 Here are the flags. Not every flag is supported on every operating
8280 @cindex appending to the output file
8281 Write in append mode, so that even if some other process is writing to
8282 this file, every @command{dd} write will append to the current
8283 contents of the file. This flag makes sense only for output.
8284 If you combine this flag with the @samp{of=@var{file}} operand,
8285 you should also specify @samp{conv=notrunc} unless you want the
8286 output file to be truncated before being appended to.
8290 @cindex concurrent I/O
8291 Use concurrent I/O mode for data. This mode performs direct I/O
8292 and drops the POSIX requirement to serialize all I/O to the same file.
8293 A file cannot be opened in CIO mode and with a standard open at the
8299 Use direct I/O for data, avoiding the buffer cache.
8300 Note that the kernel may impose restrictions on read or write buffer sizes.
8301 For example, with an ext4 destination file system and a linux-based kernel,
8302 using @samp{oflag=direct} will cause writes to fail with @code{EINVAL} if the
8303 output buffer size is not a multiple of 512.
8307 @cindex directory I/O
8309 Fail unless the file is a directory. Most operating systems do not
8310 allow I/O to a directory, so this flag has limited utility.
8314 @cindex synchronized data reads
8315 Use synchronized I/O for data. For the output file, this forces a
8316 physical write of output data on each write. For the input file,
8317 this flag can matter when reading from a remote file that has been
8318 written to synchronously by some other process. Metadata (e.g.,
8319 last-access and last-modified time) is not necessarily synchronized.
8323 @cindex synchronized data and metadata I/O
8324 Use synchronized I/O for both data and metadata.
8328 @cindex discarding file cache
8329 Discard the data cache for a file.
8330 When count=0 all cache is discarded,
8331 otherwise the cache is dropped for the processed
8332 portion of the file. Also when count=0
8333 failure to discard the cache is diagnosed
8334 and reflected in the exit status.
8335 Here as some usage examples:
8338 # Advise to drop cache for whole file
8339 dd if=ifile iflag=nocache count=0
8341 # Ensure drop cache for the whole file
8342 dd of=ofile oflag=nocache conv=notrunc,fdatasync count=0
8344 # Drop cache for part of file
8345 dd if=ifile iflag=nocache skip=10 count=10 of=/dev/null
8347 # Stream data using just the read-ahead cache
8348 dd if=ifile of=ofile iflag=nocache oflag=nocache
8353 @cindex nonblocking I/O
8354 Use non-blocking I/O.
8359 Do not update the file's access time.
8360 Some older file systems silently ignore this flag, so it is a good
8361 idea to test it on your files before relying on it.
8365 @cindex controlling terminal
8366 Do not assign the file to be a controlling terminal for @command{dd}.
8367 This has no effect when the file is not a terminal.
8368 On many hosts (e.g., GNU/Linux hosts), this option has no effect
8373 @cindex symbolic links, following
8374 Do not follow symbolic links.
8379 Fail if the file has multiple hard links.
8384 Use binary I/O@. This option has an effect only on nonstandard
8385 platforms that distinguish binary from text I/O.
8390 Use text I/O@. Like @samp{binary}, this option has no effect on
8395 Accumulate full blocks from input. The @code{read} system call
8396 may return early if a full block is not available.
8397 When that happens, continue calling @code{read} to fill the remainder
8399 This flag can be used only with @code{iflag}.
8402 @opindex count_bytes
8403 Interpret the @samp{count=} operand as a byte count,
8404 rather than a block count, which allows specifying
8405 a length that is not a multiple of the I/O block size.
8406 This flag can be used only with @code{iflag}.
8410 Interpret the @samp{skip=} operand as a byte count,
8411 rather than a block count, which allows specifying
8412 an offset that is not a multiple of the I/O block size.
8413 This flag can be used only with @code{iflag}.
8417 Interpret the @samp{seek=} operand as a byte count,
8418 rather than a block count, which allows specifying
8419 an offset that is not a multiple of the I/O block size.
8420 This flag can be used only with @code{oflag}.
8424 These flags are not supported on all systems, and @samp{dd} rejects
8425 attempts to use them when they are not supported. When reading from
8426 standard input or writing to standard output, the @samp{nofollow} and
8427 @samp{noctty} flags should not be specified, and the other flags
8428 (e.g., @samp{nonblock}) can affect how other processes behave with the
8429 affected file descriptors, even after @command{dd} exits.
8433 @cindex multipliers after numbers
8434 The numeric-valued strings above (@var{n} and @var{bytes})
8435 can be followed by a multiplier: @samp{b}=512, @samp{c}=1,
8436 @samp{w}=2, @samp{x@var{m}}=@var{m}, or any of the
8437 standard block size suffixes like @samp{k}=1024 (@pxref{Block size}).
8439 Any block size you specify via @samp{bs=}, @samp{ibs=}, @samp{obs=}, @samp{cbs=}
8440 should not be too large---values larger than a few megabytes
8441 are generally wasteful or (as in the gigabyte..exabyte case) downright
8442 counterproductive or error-inducing.
8444 To process data that is at an offset or size that is not a
8445 multiple of the I/O@ block size, you can use the @samp{skip_bytes},
8446 @samp{seek_bytes} and @samp{count_bytes} flags. Alternatively
8447 the traditional method of separate @command{dd} invocations can be used.
8448 For example, the following shell commands copy data
8449 in 512 KiB blocks between a disk and a tape, but do not save
8450 or restore a 4 KiB label at the start of the disk:
8453 disk=/dev/rdsk/c0t1d0s2
8456 # Copy all but the label from disk to tape.
8457 (dd bs=4k skip=1 count=0 && dd bs=512k) <$disk >$tape
8459 # Copy from tape back to disk, but leave the disk label alone.
8460 (dd bs=4k seek=1 count=0 && dd bs=512k) <$tape >$disk
8463 Sending an @samp{INFO} signal to a running @command{dd}
8464 process makes it print I/O statistics to standard error
8465 and then resume copying. In the example below,
8466 @command{dd} is run in the background to copy 10 million blocks.
8467 The @command{kill} command makes it output intermediate I/O statistics,
8468 and when @command{dd} completes normally or is killed by the
8469 @code{SIGINT} signal, it outputs the final statistics.
8472 $ dd if=/dev/zero of=/dev/null count=10MB & pid=$!
8473 $ kill -s INFO $pid; wait $pid
8474 3385223+0 records in
8475 3385223+0 records out
8476 1733234176 bytes (1.7 GB) copied, 6.42173 seconds, 270 MB/s
8477 10000000+0 records in
8478 10000000+0 records out
8479 5120000000 bytes (5.1 GB) copied, 18.913 seconds, 271 MB/s
8482 @vindex POSIXLY_CORRECT
8483 On systems lacking the @samp{INFO} signal @command{dd} responds to the
8484 @samp{USR1} signal instead, unless the @env{POSIXLY_CORRECT}
8485 environment variable is set.
8490 @node install invocation
8491 @section @command{install}: Copy files and set attributes
8494 @cindex copying files and setting attributes
8496 @command{install} copies files while setting their file mode bits and, if
8497 possible, their owner and group. Synopses:
8500 install [@var{option}]@dots{} [-T] @var{source} @var{dest}
8501 install [@var{option}]@dots{} @var{source}@dots{} @var{directory}
8502 install [@var{option}]@dots{} -t @var{directory} @var{source}@dots{}
8503 install [@var{option}]@dots{} -d @var{directory}@dots{}
8508 If two file names are given, @command{install} copies the first file to the
8512 If the @option{--target-directory} (@option{-t}) option is given, or
8513 failing that if the last file is a directory and the
8514 @option{--no-target-directory} (@option{-T}) option is not given,
8515 @command{install} copies each @var{source} file to the specified
8516 directory, using the @var{source}s' names.
8519 If the @option{--directory} (@option{-d}) option is given,
8520 @command{install} creates each @var{directory} and any missing parent
8521 directories. Parent directories are created with mode
8522 @samp{u=rwx,go=rx} (755), regardless of the @option{-m} option or the
8523 current umask. @xref{Directory Setuid and Setgid}, for how the
8524 set-user-ID and set-group-ID bits of parent directories are inherited.
8527 @cindex Makefiles, installing programs in
8528 @command{install} is similar to @command{cp}, but allows you to control the
8529 attributes of destination files. It is typically used in Makefiles to
8530 copy programs into their destination directories. It refuses to copy
8531 files onto themselves.
8533 @cindex extended attributes, xattr
8534 @command{install} never preserves extended attributes (xattr).
8536 The program accepts the following options. Also see @ref{Common options}.
8546 Compare each pair of source and destination files, and if the destination has
8547 identical content and any specified owner, group, permissions, and possibly
8548 SELinux context, then do not modify the destination at all.
8552 Ignored; for compatibility with old Unix versions of @command{install}.
8556 Create any missing parent directories of @var{dest},
8557 then copy @var{source} to @var{dest}.
8558 This option is ignored if a destination directory is specified
8559 via @option{--target-directory=DIR}.
8564 @opindex --directory
8565 @cindex directories, creating with given attributes
8566 @cindex parent directories, creating missing
8567 @cindex leading directories, creating missing
8568 Create any missing parent directories, giving them the default
8569 attributes. Then create each given directory, setting their owner,
8570 group and mode as given on the command line or to the defaults.
8572 @item -g @var{group}
8573 @itemx --group=@var{group}
8576 @cindex group ownership of installed files, setting
8577 Set the group ownership of installed files or directories to
8578 @var{group}. The default is the process's current group. @var{group}
8579 may be either a group name or a numeric group ID.
8582 @itemx --mode=@var{mode}
8585 @cindex permissions of installed files, setting
8586 Set the file mode bits for the installed file or directory to @var{mode},
8587 which can be either an octal number, or a symbolic mode as in
8588 @command{chmod}, with @samp{a=} (no access allowed to anyone) as the
8589 point of departure (@pxref{File permissions}).
8590 The default mode is @samp{u=rwx,go=rx,a-s}---read, write, and
8591 execute for the owner, read and execute for group and other, and with
8592 set-user-ID and set-group-ID disabled.
8593 This default is not quite the same as @samp{755}, since it disables
8594 instead of preserving set-user-ID and set-group-ID on directories.
8595 @xref{Directory Setuid and Setgid}.
8597 @item -o @var{owner}
8598 @itemx --owner=@var{owner}
8601 @cindex ownership of installed files, setting
8602 @cindex appropriate privileges
8603 @vindex root @r{as default owner}
8604 If @command{install} has appropriate privileges (is run as root), set the
8605 ownership of installed files or directories to @var{owner}. The default
8606 is @code{root}. @var{owner} may be either a user name or a numeric user
8609 @item --preserve-context
8610 @opindex --preserve-context
8612 @cindex security context
8613 Preserve the SELinux security context of files and directories.
8614 Failure to preserve the context in all of the files or directories
8615 will result in an exit status of 1. If SELinux is disabled then
8616 print a warning and ignore the option.
8619 @itemx --preserve-timestamps
8621 @opindex --preserve-timestamps
8622 @cindex timestamps of installed files, preserving
8623 Set the time of last access and the time of last modification of each
8624 installed file to match those of each corresponding original file.
8625 When a file is installed without this option, its last access and
8626 last modification times are both set to the time of installation.
8627 This option is useful if you want to use the last modification times
8628 of installed files to keep track of when they were last built as opposed
8629 to when they were last installed.
8635 @cindex symbol table information, stripping
8636 @cindex stripping symbol table information
8637 Strip the symbol tables from installed binary executables.
8639 @item --strip-program=@var{program}
8640 @opindex --strip-program
8641 @cindex symbol table information, stripping, program
8642 Program used to strip binaries.
8648 @optNoTargetDirectory
8654 Print the name of each file before copying it.
8656 @item -Z @var{context}
8657 @itemx --context=@var{context}
8661 @cindex security context
8662 Set the default SELinux security context to be used for any
8663 created files and directories. If SELinux is disabled then
8664 print a warning and ignore the option.
8672 @section @command{mv}: Move (rename) files
8676 @command{mv} moves or renames files (or directories). Synopses:
8679 mv [@var{option}]@dots{} [-T] @var{source} @var{dest}
8680 mv [@var{option}]@dots{} @var{source}@dots{} @var{directory}
8681 mv [@var{option}]@dots{} -t @var{directory} @var{source}@dots{}
8686 If two file names are given, @command{mv} moves the first file to the
8690 If the @option{--target-directory} (@option{-t}) option is given, or
8691 failing that if the last file is a directory and the
8692 @option{--no-target-directory} (@option{-T}) option is not given,
8693 @command{mv} moves each @var{source} file to the specified
8694 directory, using the @var{source}s' names.
8697 @command{mv} can move any type of file from one file system to another.
8698 Prior to version @code{4.0} of the fileutils,
8699 @command{mv} could move only regular files between file systems.
8700 For example, now @command{mv} can move an entire directory hierarchy
8701 including special device files from one partition to another. It first
8702 uses some of the same code that's used by @code{cp -a} to copy the
8703 requested directories and files, then (assuming the copy succeeded)
8704 it removes the originals. If the copy fails, then the part that was
8705 copied to the destination partition is removed. If you were to copy
8706 three directories from one partition to another and the copy of the first
8707 directory succeeded, but the second didn't, the first would be left on
8708 the destination partition and the second and third would be left on the
8711 @cindex extended attributes, xattr
8712 @command{mv} always tries to copy extended attributes (xattr), which may
8713 include SELinux context, ACLs or Capabilities.
8714 Upon failure all but @samp{Operation not supported} warnings are output.
8716 @cindex prompting, and @command{mv}
8717 If a destination file exists but is normally unwritable, standard input
8718 is a terminal, and the @option{-f} or @option{--force} option is not given,
8719 @command{mv} prompts the user for whether to replace the file. (You might
8720 own the file, or have write permission on its directory.) If the
8721 response is not affirmative, the file is skipped.
8723 @emph{Warning}: Avoid specifying a source name with a trailing slash,
8724 when it might be a symlink to a directory.
8725 Otherwise, @command{mv} may do something very surprising, since
8726 its behavior depends on the underlying rename system call.
8727 On a system with a modern Linux-based kernel, it fails with
8728 @code{errno=ENOTDIR}@.
8729 However, on other systems (at least FreeBSD 6.1 and Solaris 10) it silently
8730 renames not the symlink but rather the directory referenced by the symlink.
8731 @xref{Trailing slashes}.
8733 The program accepts the following options. Also see @ref{Common options}.
8743 @cindex prompts, omitting
8744 Do not prompt the user before removing a destination file.
8746 If you specify more than one of the @option{-i}, @option{-f}, @option{-n}
8747 options, only the final one takes effect.
8752 @itemx --interactive
8754 @opindex --interactive
8755 @cindex prompts, forcing
8756 Prompt whether to overwrite each existing destination file, regardless
8758 If the response is not affirmative, the file is skipped.
8764 @opindex --no-clobber
8765 @cindex prompts, omitting
8766 Do not overwrite an existing file.
8768 This option is mutually exclusive with @option{-b} or @option{--backup} option.
8774 @cindex newer files, moving only
8775 Do not move a non-directory that has an existing destination with the
8776 same or newer modification time.
8777 If the move is across file system boundaries, the comparison is to the
8778 source time stamp truncated to the resolutions of the destination file
8779 system and of the system calls used to update time stamps; this avoids
8780 duplicate work if several @samp{mv -u} commands are executed with the
8781 same source and destination.
8787 Print the name of each file before moving it.
8789 @optStripTrailingSlashes
8795 @optNoTargetDirectory
8803 @section @command{rm}: Remove files or directories
8806 @cindex removing files or directories
8808 @command{rm} removes each given @var{file}. By default, it does not remove
8809 directories. Synopsis:
8812 rm [@var{option}]@dots{} [@var{file}]@dots{}
8815 @cindex prompting, and @command{rm}
8816 If the @option{-I} or @option{--interactive=once} option is given,
8817 and there are more than three files or the @option{-r}, @option{-R},
8818 or @option{--recursive} are given, then @command{rm} prompts the user
8819 for whether to proceed with the entire operation. If the response is
8820 not affirmative, the entire command is aborted.
8822 Otherwise, if a file is unwritable, standard input is a terminal, and
8823 the @option{-f} or @option{--force} option is not given, or the
8824 @option{-i} or @option{--interactive=always} option @emph{is} given,
8825 @command{rm} prompts the user for whether to remove the file.
8826 If the response is not affirmative, the file is skipped.
8828 Any attempt to remove a file whose last file name component is
8829 @file{.} or @file{..} is rejected without any prompting.
8831 @emph{Warning}: If you use @command{rm} to remove a file, it is usually
8832 possible to recover the contents of that file. If you want more assurance
8833 that the contents are truly unrecoverable, consider using @command{shred}.
8835 The program accepts the following options. Also see @ref{Common options}.
8843 @cindex directories, removing
8844 Remove the listed directories if they are empty.
8850 Ignore nonexistent files and missing operands, and never prompt the user.
8851 Ignore any previous @option{--interactive} (@option{-i}) option.
8855 Prompt whether to remove each file.
8856 If the response is not affirmative, the file is skipped.
8857 Ignore any previous @option{--force} (@option{-f}) option.
8858 Equivalent to @option{--interactive=always}.
8862 Prompt once whether to proceed with the command, if more than three
8863 files are named or if a recursive removal is requested. Ignore any
8864 previous @option{--force} (@option{-f}) option. Equivalent to
8865 @option{--interactive=once}.
8867 @item --interactive [=@var{when}]
8868 @opindex --interactive
8869 Specify when to issue an interactive prompt. @var{when} may be
8873 @vindex never @r{interactive option}
8874 - Do not prompt at all.
8876 @vindex once @r{interactive option}
8877 - Prompt once if more than three files are named or if a recursive
8878 removal is requested. Equivalent to @option{-I}.
8880 @vindex always @r{interactive option}
8881 - Prompt for every file being removed. Equivalent to @option{-i}.
8883 @option{--interactive} with no @var{when} is equivalent to
8884 @option{--interactive=always}.
8886 @item --one-file-system
8887 @opindex --one-file-system
8888 @cindex one file system, restricting @command{rm} to
8889 When removing a hierarchy recursively, skip any directory that is on a
8890 file system different from that of the corresponding command line argument.
8893 This option is useful when removing a build ``chroot'' hierarchy,
8894 which normally contains no valuable data. However, it is not uncommon
8895 to bind-mount @file{/home} into such a hierarchy, to make it easier to
8896 use one's start-up file. The catch is that it's easy to forget to
8897 unmount @file{/home}. Then, when you use @command{rm -rf} to remove
8898 your normally throw-away chroot, that command will remove everything
8899 under @file{/home}, too.
8900 Use the @option{--one-file-system} option, and it will
8901 warn about and skip directories on other file systems.
8902 Of course, this will not save your @file{/home} if it and your
8903 chroot happen to be on the same file system.
8905 @item --preserve-root
8906 @opindex --preserve-root
8907 @cindex root directory, disallow recursive destruction
8908 Fail upon any attempt to remove the root directory, @file{/},
8909 when used with the @option{--recursive} option.
8910 This is the default behavior.
8911 @xref{Treating / specially}.
8913 @item --no-preserve-root
8914 @opindex --no-preserve-root
8915 @cindex root directory, allow recursive destruction
8916 Do not treat @file{/} specially when removing recursively.
8917 This option is not recommended unless you really want to
8918 remove all the files on your computer.
8919 @xref{Treating / specially}.
8926 @opindex --recursive
8927 @cindex directories, removing (recursively)
8928 Remove the listed directories and their contents recursively.
8934 Print the name of each file before removing it.
8938 @cindex files beginning with @samp{-}, removing
8939 @cindex @samp{-}, removing files beginning with
8940 One common question is how to remove files whose names begin with a
8941 @samp{-}. @sc{gnu} @command{rm}, like every program that uses the @code{getopt}
8942 function to parse its arguments, lets you use the @samp{--} option to
8943 indicate that all following arguments are non-options. To remove a file
8944 called @file{-f} in the current directory, you could type either:
8957 @opindex - @r{and Unix @command{rm}}
8958 The Unix @command{rm} program's use of a single @samp{-} for this purpose
8959 predates the development of the getopt standard syntax.
8964 @node shred invocation
8965 @section @command{shred}: Remove files more securely
8968 @cindex data, erasing
8969 @cindex erasing data
8971 @command{shred} overwrites devices or files, to help prevent even
8972 very expensive hardware from recovering the data.
8974 Ordinarily when you remove a file (@pxref{rm invocation}), the data is
8975 not actually destroyed. Only the index listing where the file is
8976 stored is destroyed, and the storage is made available for reuse.
8977 There are undelete utilities that will attempt to reconstruct the index
8978 and can bring the file back if the parts were not reused.
8980 On a busy system with a nearly-full drive, space can get reused in a few
8981 seconds. But there is no way to know for sure. If you have sensitive
8982 data, you may want to be sure that recovery is not possible by actually
8983 overwriting the file with non-sensitive data.
8985 However, even after doing that, it is possible to take the disk back
8986 to a laboratory and use a lot of sensitive (and expensive) equipment
8987 to look for the faint ``echoes'' of the original data underneath the
8988 overwritten data. If the data has only been overwritten once, it's not
8991 The best way to remove something irretrievably is to destroy the media
8992 it's on with acid, melt it down, or the like. For cheap removable media
8993 like floppy disks, this is the preferred method. However, hard drives
8994 are expensive and hard to melt, so the @command{shred} utility tries
8995 to achieve a similar effect non-destructively.
8997 This uses many overwrite passes, with the data patterns chosen to
8998 maximize the damage they do to the old data. While this will work on
8999 floppies, the patterns are designed for best effect on hard drives.
9000 For more details, see the source code and Peter Gutmann's paper
9001 @uref{http://www.cs.auckland.ac.nz/~pgut001/pubs/secure_del.html,
9002 @cite{Secure Deletion of Data from Magnetic and Solid-State Memory}},
9003 from the proceedings of the Sixth USENIX Security Symposium (San Jose,
9004 California, July 22--25, 1996).
9006 @strong{Please note} that @command{shred} relies on a very important assumption:
9007 that the file system overwrites data in place. This is the traditional
9008 way to do things, but many modern file system designs do not satisfy this
9009 assumption. Exceptions include:
9014 Log-structured or journaled file systems, such as those supplied with
9015 AIX and Solaris, and JFS, ReiserFS, XFS, Ext3 (in @code{data=journal} mode),
9016 BFS, NTFS, etc., when they are configured to journal @emph{data}.
9019 File systems that write redundant data and carry on even if some writes
9020 fail, such as RAID-based file systems.
9023 File systems that make snapshots, such as Network Appliance's NFS server.
9026 File systems that cache in temporary locations, such as NFS version 3
9030 Compressed file systems.
9033 In the particular case of ext3 file systems, the above disclaimer applies (and
9034 @command{shred} is thus of limited effectiveness) only in @code{data=journal}
9035 mode, which journals file data in addition to just metadata. In both
9036 the @code{data=ordered} (default) and @code{data=writeback} modes,
9037 @command{shred} works as usual. Ext3 journaling modes can be changed
9038 by adding the @code{data=something} option to the mount options for a
9039 particular file system in the @file{/etc/fstab} file, as documented in
9040 the mount man page (man mount).
9042 If you are not sure how your file system operates, then you should assume
9043 that it does not overwrite data in place, which means that shred cannot
9044 reliably operate on regular files in your file system.
9046 Generally speaking, it is more reliable to shred a device than a file,
9047 since this bypasses the problem of file system design mentioned above.
9048 However, even shredding devices is not always completely reliable. For
9049 example, most disks map out bad sectors invisibly to the application; if
9050 the bad sectors contain sensitive data, @command{shred} won't be able to
9053 @command{shred} makes no attempt to detect or report this problem, just as
9054 it makes no attempt to do anything about backups. However, since it is
9055 more reliable to shred devices than files, @command{shred} by default does
9056 not truncate or remove the output file. This default is more suitable
9057 for devices, which typically cannot be truncated and should not be
9060 Finally, consider the risk of backups and mirrors.
9061 File system backups and remote mirrors may contain copies of the
9062 file that cannot be removed, and that will allow a shredded file
9063 to be recovered later. So if you keep any data you may later want
9064 to destroy using @command{shred}, be sure that it is not backed up or mirrored.
9067 shred [@var{option}]@dots{} @var{file}[@dots{}]
9070 The program accepts the following options. Also see @ref{Common options}.
9078 @cindex force deletion
9079 Override file permissions if necessary to allow overwriting.
9082 @itemx -n @var{number}
9083 @itemx --iterations=@var{number}
9084 @opindex -n @var{number}
9085 @opindex --iterations=@var{number}
9086 @cindex iterations, selecting the number of
9087 By default, @command{shred} uses @value{SHRED_DEFAULT_PASSES} passes of
9088 overwrite. You can reduce this to save time, or increase it if you think it's
9089 appropriate. After 25 passes all of the internal overwrite patterns will have
9090 been used at least once.
9092 @item --random-source=@var{file}
9093 @opindex --random-source
9094 @cindex random source for shredding
9095 Use @var{file} as a source of random data used to overwrite and to
9096 choose pass ordering. @xref{Random sources}.
9098 @item -s @var{bytes}
9099 @itemx --size=@var{bytes}
9100 @opindex -s @var{bytes}
9101 @opindex --size=@var{bytes}
9102 @cindex size of file to shred
9103 Shred the first @var{bytes} bytes of the file. The default is to shred
9104 the whole file. @var{bytes} can be followed by a size specification like
9105 @samp{K}, @samp{M}, or @samp{G} to specify a multiple. @xref{Block size}.
9111 @cindex removing files after shredding
9112 After shredding a file, truncate it (if possible) and then remove it.
9113 If a file has multiple links, only the named links will be removed.
9119 Display to standard error all status updates as sterilization proceeds.
9125 By default, @command{shred} rounds the size of a regular file up to the next
9126 multiple of the file system block size to fully erase the last block
9128 Use @option{--exact} to suppress that behavior.
9129 Thus, by default if you shred a 10-byte regular file on a system with 512-byte
9130 blocks, the resulting file will be 512 bytes long. With this option,
9131 shred does not increase the apparent size of the file.
9137 Normally, the last pass that @command{shred} writes is made up of
9138 random data. If this would be conspicuous on your hard drive (for
9139 example, because it looks like encrypted data), or you just think
9140 it's tidier, the @option{--zero} option adds an additional overwrite pass with
9141 all zero bits. This is in addition to the number of passes specified
9142 by the @option{--iterations} option.
9146 You might use the following command to erase all trace of the
9147 file system you'd created on the floppy disk in your first drive.
9148 That command takes about 20 minutes to erase a ``1.44MB'' (actually
9152 shred --verbose /dev/fd0
9155 Similarly, to erase all data on a selected partition of
9156 your hard disk, you could give a command like this:
9159 shred --verbose /dev/sda5
9162 On modern disks, a single pass should be adequate,
9163 and it will take one third the time of the default three-pass approach.
9166 # 1 pass, write pseudo-random data; 3x faster than the default
9167 shred --verbose -n1 /dev/sda5
9170 To be on the safe side, use at least one pass that overwrites using
9171 pseudo-random data. I.e., don't be tempted to use @samp{-n0 --zero},
9172 in case some disk controller optimizes the process of writing blocks
9173 of all zeros, and thereby does not clear all bytes in a block.
9174 Some SSDs may do just that.
9176 A @var{file} of @samp{-} denotes standard output.
9177 The intended use of this is to shred a removed temporary file.
9184 echo "Hello, world" >&3
9189 However, the command @samp{shred - >file} does not shred the contents
9190 of @var{file}, since the shell truncates @var{file} before invoking
9191 @command{shred}. Use the command @samp{shred file} or (if using a
9192 Bourne-compatible shell) the command @samp{shred - 1<>file} instead.
9197 @node Special file types
9198 @chapter Special file types
9200 @cindex special file types
9201 @cindex file types, special
9203 This chapter describes commands which create special types of files (and
9204 @command{rmdir}, which removes directories, one special file type).
9206 @cindex special file types
9208 Although Unix-like operating systems have markedly fewer special file
9209 types than others, not @emph{everything} can be treated only as the
9210 undifferentiated byte stream of @dfn{normal files}. For example, when a
9211 file is created or removed, the system must record this information,
9212 which it does in a @dfn{directory}---a special type of file. Although
9213 you can read directories as normal files, if you're curious, in order
9214 for the system to do its job it must impose a structure, a certain
9215 order, on the bytes of the file. Thus it is a ``special'' type of file.
9217 Besides directories, other special file types include named pipes
9218 (FIFOs), symbolic links, sockets, and so-called @dfn{special files}.
9221 * link invocation:: Make a hard link via the link syscall
9222 * ln invocation:: Make links between files.
9223 * mkdir invocation:: Make directories.
9224 * mkfifo invocation:: Make FIFOs (named pipes).
9225 * mknod invocation:: Make block or character special files.
9226 * readlink invocation:: Print value of a symlink or canonical file name.
9227 * rmdir invocation:: Remove empty directories.
9228 * unlink invocation:: Remove files via the unlink syscall
9232 @node link invocation
9233 @section @command{link}: Make a hard link via the link syscall
9236 @cindex links, creating
9237 @cindex hard links, creating
9238 @cindex creating links (hard only)
9240 @command{link} creates a single hard link at a time.
9241 It is a minimalist interface to the system-provided
9242 @code{link} function. @xref{Hard Links, , , libc,
9243 The GNU C Library Reference Manual}.
9244 It avoids the bells and whistles of the more commonly-used
9245 @command{ln} command (@pxref{ln invocation}).
9249 link @var{filename} @var{linkname}
9252 @var{filename} must specify an existing file, and @var{linkname}
9253 must specify a nonexistent entry in an existing directory.
9254 @command{link} simply calls @code{link (@var{filename}, @var{linkname})}
9257 On a GNU system, this command acts like @samp{ln --directory
9258 --no-target-directory @var{filename} @var{linkname}}. However, the
9259 @option{--directory} and @option{--no-target-directory} options are
9260 not specified by POSIX, and the @command{link} command is
9261 more portable in practice.
9263 If @var{filename} is a symbolic link, it is unspecified whether
9264 @var{linkname} will be a hard link to the symbolic link or to the
9265 target of the symbolic link. Use @command{ln -P} or @command{ln -L}
9266 to specify which behavior is desired.
9272 @section @command{ln}: Make links between files
9275 @cindex links, creating
9276 @cindex hard links, creating
9277 @cindex symbolic (soft) links, creating
9278 @cindex creating links (hard or soft)
9280 @cindex file systems and hard links
9281 @command{ln} makes links between files. By default, it makes hard links;
9282 with the @option{-s} option, it makes symbolic (or @dfn{soft}) links.
9286 ln [@var{option}]@dots{} [-T] @var{target} @var{linkname}
9287 ln [@var{option}]@dots{} @var{target}
9288 ln [@var{option}]@dots{} @var{target}@dots{} @var{directory}
9289 ln [@var{option}]@dots{} -t @var{directory} @var{target}@dots{}
9295 If two file names are given, @command{ln} creates a link to the first
9296 file from the second.
9299 If one @var{target} is given, @command{ln} creates a link to that file
9300 in the current directory.
9303 If the @option{--target-directory} (@option{-t}) option is given, or
9304 failing that if the last file is a directory and the
9305 @option{--no-target-directory} (@option{-T}) option is not given,
9306 @command{ln} creates a link to each @var{target} file in the specified
9307 directory, using the @var{target}s' names.
9311 Normally @command{ln} does not remove existing files. Use the
9312 @option{--force} (@option{-f}) option to remove them unconditionally,
9313 the @option{--interactive} (@option{-i}) option to remove them
9314 conditionally, and the @option{--backup} (@option{-b}) option to
9317 @cindex hard link, defined
9318 @cindex inode, and hard links
9319 A @dfn{hard link} is another name for an existing file; the link and the
9320 original are indistinguishable. Technically speaking, they share the
9321 same inode, and the inode contains all the information about a
9322 file---indeed, it is not incorrect to say that the inode @emph{is} the
9323 file. Most systems prohibit making a hard link to
9324 a directory; on those where it is allowed, only the super-user can do
9325 so (and with caution, since creating a cycle will cause problems to many
9326 other utilities). Hard links cannot cross file system boundaries. (These
9327 restrictions are not mandated by POSIX, however.)
9329 @cindex dereferencing symbolic links
9330 @cindex symbolic link, defined
9331 @dfn{Symbolic links} (@dfn{symlinks} for short), on the other hand, are
9332 a special file type (which not all kernels support: System V release 3
9333 (and older) systems lack symlinks) in which the link file actually
9334 refers to a different file, by name. When most operations (opening,
9335 reading, writing, and so on) are passed the symbolic link file, the
9336 kernel automatically @dfn{dereferences} the link and operates on the
9337 target of the link. But some operations (e.g., removing) work on the
9338 link file itself, rather than on its target. The owner and group of a
9339 symlink are not significant to file access performed through
9340 the link, but do have implications on deleting a symbolic link from a
9341 directory with the restricted deletion bit set. On the GNU system,
9342 the mode of a symlink has no significance and cannot be changed, but
9343 on some BSD systems, the mode can be changed and will affect whether
9344 the symlink will be traversed in file name resolution. @xref{Symbolic Links,,,
9345 libc, The GNU C Library Reference Manual}.
9347 Symbolic links can contain arbitrary strings; a @dfn{dangling symlink}
9348 occurs when the string in the symlink does not resolve to a file.
9349 There are no restrictions against creating dangling symbolic links.
9350 There are trade-offs to using absolute or relative symlinks. An
9351 absolute symlink always points to the same file, even if the directory
9352 containing the link is moved. However, if the symlink is visible from
9353 more than one machine (such as on a networked file system), the file
9354 pointed to might not always be the same. A relative symbolic link is
9355 resolved in relation to the directory that contains the link, and is
9356 often useful in referring to files on the same device without regards
9357 to what name that device is mounted on when accessed via networked
9360 When creating a relative symlink in a different location than the
9361 current directory, the resolution of the symlink will be different
9362 than the resolution of the same string from the current directory.
9363 Therefore, many users prefer to first change directories to the
9364 location where the relative symlink will be created, so that
9365 tab-completion or other file resolution will find the same target as
9366 what will be placed in the symlink.
9368 The program accepts the following options. Also see @ref{Common options}.
9379 @opindex --directory
9380 @cindex hard links to directories
9381 Allow users with appropriate privileges to attempt to make hard links
9383 However, note that this will probably fail due to
9384 system restrictions, even for the super-user.
9390 Remove existing destination files.
9393 @itemx --interactive
9395 @opindex --interactive
9396 @cindex prompting, and @command{ln}
9397 Prompt whether to remove existing destination files.
9403 If @option{-s} is not in effect, and the source file is a symbolic
9404 link, create the hard link to the file referred to by the symbolic
9405 link, rather than the symbolic link itself.
9408 @itemx --no-dereference
9410 @opindex --no-dereference
9411 Do not treat the last operand specially when it is a symbolic link to
9412 a directory. Instead, treat it as if it were a normal file.
9414 When the destination is an actual directory (not a symlink to one),
9415 there is no ambiguity. The link is created in that directory.
9416 But when the specified destination is a symlink to a directory,
9417 there are two ways to treat the user's request. @command{ln} can
9418 treat the destination just as it would a normal directory and create
9419 the link in it. On the other hand, the destination can be viewed as a
9420 non-directory---as the symlink itself. In that case, @command{ln}
9421 must delete or backup that symlink before creating the new link.
9422 The default is to treat a destination that is a symlink to a directory
9423 just like a directory.
9425 This option is weaker than the @option{--no-target-directory}
9426 (@option{-T}) option, so it has no effect if both options are given.
9432 If @option{-s} is not in effect, and the source file is a symbolic
9433 link, create the hard link to the symbolic link itself. On platforms
9434 where this is not supported by the kernel, this option creates a
9435 symbolic link with identical contents; since symbolic link contents
9436 cannot be edited, any file name resolution performed through either
9437 link will be the same as if a hard link had been created.
9443 Make symbolic links relative to the link location.
9448 ln -srv /a/file /tmp
9449 '/tmp/file' -> '../a/file'
9452 @xref{realpath invocation}, which gives greater control
9453 over relative file name generation.
9459 Make symbolic links instead of hard links. This option merely produces
9460 an error message on systems that do not support symbolic links.
9466 @optNoTargetDirectory
9472 Print the name of each file after linking it successfully.
9476 @cindex hard links to symbolic links
9477 @cindex symbolic links and @command{ln}
9478 If @option{-L} and @option{-P} are both given, the last one takes
9479 precedence. If @option{-s} is also given, @option{-L} and @option{-P}
9480 are silently ignored. If neither option is given, then this
9481 implementation defaults to @option{-P} if the system @code{link} supports
9482 hard links to symbolic links (such as the GNU system), and @option{-L}
9483 if @code{link} follows symbolic links (such as on BSD).
9492 # Create link ../a pointing to a in that directory.
9493 # Not really useful because it points to itself.
9498 # Change to the target before creating symlinks to avoid being confused.
9504 # Hard coded file names don't move well.
9505 ln -s $(pwd)/a /some/dir/
9509 # Relative file names survive directory moves and also
9510 # work across networked file systems.
9511 ln -s afile anotherfile
9512 ln -s ../adir/afile yetanotherfile
9516 @node mkdir invocation
9517 @section @command{mkdir}: Make directories
9520 @cindex directories, creating
9521 @cindex creating directories
9523 @command{mkdir} creates directories with the specified names. Synopsis:
9526 mkdir [@var{option}]@dots{} @var{name}@dots{}
9529 @command{mkdir} creates each directory @var{name} in the order given.
9530 It reports an error if @var{name} already exists, unless the
9531 @option{-p} option is given and @var{name} is a directory.
9533 The program accepts the following options. Also see @ref{Common options}.
9538 @itemx --mode=@var{mode}
9541 @cindex modes of created directories, setting
9542 Set the file permission bits of created directories to @var{mode},
9543 which uses the same syntax as
9544 in @command{chmod} and uses @samp{a=rwx} (read, write and execute allowed for
9545 everyone) for the point of the departure. @xref{File permissions}.
9547 Normally the directory has the desired file mode bits at the moment it
9548 is created. As a GNU extension, @var{mode} may also mention
9549 special mode bits, but in this case there may be a temporary window
9550 during which the directory exists but its special mode bits are
9551 incorrect. @xref{Directory Setuid and Setgid}, for how the
9552 set-user-ID and set-group-ID bits of directories are inherited unless
9553 overridden in this way.
9559 @cindex parent directories, creating
9560 Make any missing parent directories for each argument, setting their
9561 file permission bits to the umask modified by @samp{u+wx}. Ignore
9562 existing parent directories, and do not change their file permission
9565 To set the file permission bits of any newly-created parent
9566 directories to a value that includes @samp{u+wx}, you can set the
9567 umask before invoking @command{mkdir}. For example, if the shell
9568 command @samp{(umask u=rwx,go=rx; mkdir -p P/Q)} creates the parent
9569 @file{P} it sets the parent's permission bits to @samp{u=rwx,go=rx}.
9570 To set a parent's special mode bits as well, you can invoke
9571 @command{chmod} after @command{mkdir}. @xref{Directory Setuid and
9572 Setgid}, for how the set-user-ID and set-group-ID bits of
9573 newly-created parent directories are inherited.
9579 Print a message for each created directory. This is most useful with
9582 @item -Z @var{context}
9583 @itemx --context=@var{context}
9587 @cindex security context
9588 Set the default SELinux security context to be used for created directories.
9595 @node mkfifo invocation
9596 @section @command{mkfifo}: Make FIFOs (named pipes)
9599 @cindex FIFOs, creating
9600 @cindex named pipes, creating
9601 @cindex creating FIFOs (named pipes)
9603 @command{mkfifo} creates FIFOs (also called @dfn{named pipes}) with the
9604 specified names. Synopsis:
9607 mkfifo [@var{option}] @var{name}@dots{}
9610 A @dfn{FIFO} is a special file type that permits independent processes
9611 to communicate. One process opens the FIFO file for writing, and
9612 another for reading, after which data can flow as with the usual
9613 anonymous pipe in shells or elsewhere.
9615 The program accepts the following option. Also see @ref{Common options}.
9620 @itemx --mode=@var{mode}
9623 @cindex modes of created FIFOs, setting
9624 Set the mode of created FIFOs to @var{mode}, which is symbolic as in
9625 @command{chmod} and uses @samp{a=rw} (read and write allowed for everyone)
9626 for the point of departure. @var{mode} should specify only file
9627 permission bits. @xref{File permissions}.
9629 @item -Z @var{context}
9630 @itemx --context=@var{context}
9634 @cindex security context
9635 Set the default SELinux security context to be used for created FIFOs.
9642 @node mknod invocation
9643 @section @command{mknod}: Make block or character special files
9646 @cindex block special files, creating
9647 @cindex character special files, creating
9649 @command{mknod} creates a FIFO, character special file, or block special
9650 file with the specified name. Synopsis:
9653 mknod [@var{option}]@dots{} @var{name} @var{type} [@var{major} @var{minor}]
9656 @cindex special files
9657 @cindex block special files
9658 @cindex character special files
9659 Unlike the phrase ``special file type'' above, the term @dfn{special
9660 file} has a technical meaning on Unix: something that can generate or
9661 receive data. Usually this corresponds to a physical piece of hardware,
9662 e.g., a printer or a disk. (These files are typically created at
9663 system-configuration time.) The @command{mknod} command is what creates
9664 files of this type. Such devices can be read either a character at a
9665 time or a ``block'' (many characters) at a time, hence we say there are
9666 @dfn{block special} files and @dfn{character special} files.
9668 @c mknod is a shell built-in at least with OpenBSD's /bin/sh
9669 @mayConflictWithShellBuiltIn{mknod}
9671 The arguments after @var{name} specify the type of file to make:
9676 @opindex p @r{for FIFO file}
9680 @opindex b @r{for block special file}
9681 for a block special file
9684 @c Don't document the 'u' option -- it's just a synonym for 'c'.
9685 @c Do *any* versions of mknod still use it?
9687 @opindex c @r{for character special file}
9688 @c @opindex u @r{for character special file}
9689 for a character special file
9693 When making a block or character special file, the major and minor
9694 device numbers must be given after the file type.
9695 If a major or minor device number begins with @samp{0x} or @samp{0X},
9696 it is interpreted as hexadecimal; otherwise, if it begins with @samp{0},
9697 as octal; otherwise, as decimal.
9699 The program accepts the following option. Also see @ref{Common options}.
9704 @itemx --mode=@var{mode}
9707 Set the mode of created files to @var{mode}, which is symbolic as in
9708 @command{chmod} and uses @samp{a=rw} as the point of departure.
9709 @var{mode} should specify only file permission bits.
9710 @xref{File permissions}.
9712 @item -Z @var{context}
9713 @itemx --context=@var{context}
9717 @cindex security context
9718 Set the default SELinux security context to be used for created files.
9725 @node readlink invocation
9726 @section @command{readlink}: Print value of a symlink or canonical file name
9729 @cindex displaying value of a symbolic link
9730 @cindex canonical file name
9731 @cindex canonicalize a file name
9734 @command{readlink} may work in one of two supported modes:
9740 @command{readlink} outputs the value of the given symbolic link.
9741 If @command{readlink} is invoked with an argument other than the name
9742 of a symbolic link, it produces no output and exits with a nonzero exit code.
9744 @item Canonicalize mode
9746 @command{readlink} outputs the absolute name of the given file which contains
9747 no @file{.}, @file{..} components nor any repeated separators
9748 (@file{/}) or symbolic links.
9753 readlink [@var{option}] @var{file}
9756 By default, @command{readlink} operates in readlink mode.
9758 The program accepts the following options. Also see @ref{Common options}.
9763 @itemx --canonicalize
9765 @opindex --canonicalize
9766 Activate canonicalize mode.
9767 If any component of the file name except the last one is missing or unavailable,
9768 @command{readlink} produces no output and exits with a nonzero exit
9769 code. A trailing slash is ignored.
9772 @itemx --canonicalize-existing
9774 @opindex --canonicalize-existing
9775 Activate canonicalize mode.
9776 If any component is missing or unavailable, @command{readlink} produces
9777 no output and exits with a nonzero exit code. A trailing slash
9778 requires that the name resolve to a directory.
9781 @itemx --canonicalize-missing
9783 @opindex --canonicalize-missing
9784 Activate canonicalize mode.
9785 If any component is missing or unavailable, @command{readlink} treats it
9791 @opindex --no-newline
9792 Do not output the trailing newline.
9802 Suppress most error messages.
9808 Report error messages.
9812 The @command{readlink} utility first appeared in OpenBSD 2.1.
9814 The @command{realpath} command without options, operates like
9815 @command{readlink} in canonicalize mode.
9820 @node rmdir invocation
9821 @section @command{rmdir}: Remove empty directories
9824 @cindex removing empty directories
9825 @cindex directories, removing empty
9827 @command{rmdir} removes empty directories. Synopsis:
9830 rmdir [@var{option}]@dots{} @var{directory}@dots{}
9833 If any @var{directory} argument does not refer to an existing empty
9834 directory, it is an error.
9836 The program accepts the following options. Also see @ref{Common options}.
9840 @item --ignore-fail-on-non-empty
9841 @opindex --ignore-fail-on-non-empty
9842 @cindex directory deletion, ignoring failures
9843 Ignore each failure to remove a directory that is solely because
9844 the directory is non-empty.
9850 @cindex parent directories, removing
9851 Remove @var{directory}, then try to remove each component of @var{directory}.
9852 So, for example, @samp{rmdir -p a/b/c} is similar to @samp{rmdir a/b/c a/b a}.
9853 As such, it fails if any of those directories turns out not to be empty.
9854 Use the @option{--ignore-fail-on-non-empty} option to make it so such
9855 a failure does not evoke a diagnostic and does not cause @command{rmdir} to
9856 exit unsuccessfully.
9862 @cindex directory deletion, reporting
9863 Give a diagnostic for each successful removal.
9864 @var{directory} is removed.
9868 @xref{rm invocation}, for how to remove non-empty directories (recursively).
9873 @node unlink invocation
9874 @section @command{unlink}: Remove files via the unlink syscall
9877 @cindex removing files or directories (via the unlink syscall)
9879 @command{unlink} deletes a single specified file name.
9880 It is a minimalist interface to the system-provided
9881 @code{unlink} function. @xref{Deleting Files, , , libc,
9882 The GNU C Library Reference Manual}. Synopsis:
9883 It avoids the bells and whistles of the more commonly-used
9884 @command{rm} command (@pxref{rm invocation}).
9887 unlink @var{filename}
9890 On some systems @code{unlink} can be used to delete the name of a
9891 directory. On others, it can be used that way only by a privileged user.
9892 In the GNU system @code{unlink} can never delete the name of a directory.
9894 The @command{unlink} command honors the @option{--help} and
9895 @option{--version} options. To remove a file whose name begins with
9896 @samp{-}, prefix the name with @samp{./}, e.g., @samp{unlink ./--help}.
9901 @node Changing file attributes
9902 @chapter Changing file attributes
9904 @cindex changing file attributes
9905 @cindex file attributes, changing
9906 @cindex attributes, file
9908 A file is not merely its contents, a name, and a file type
9909 (@pxref{Special file types}). A file also has an owner (a user ID), a
9910 group (a group ID), permissions (what the owner can do with the file,
9911 what people in the group can do, and what everyone else can do), various
9912 timestamps, and other information. Collectively, we call these a file's
9915 These commands change file attributes.
9918 * chown invocation:: Change file owners and groups.
9919 * chgrp invocation:: Change file groups.
9920 * chmod invocation:: Change access permissions.
9921 * touch invocation:: Change file timestamps.
9925 @node chown invocation
9926 @section @command{chown}: Change file owner and group
9929 @cindex file ownership, changing
9930 @cindex group ownership, changing
9931 @cindex changing file ownership
9932 @cindex changing group ownership
9934 @command{chown} changes the user and/or group ownership of each given @var{file}
9935 to @var{new-owner} or to the user and group of an existing reference file.
9939 chown [@var{option}]@dots{} @{@var{new-owner} | --reference=@var{ref_file}@}@c
9943 If used, @var{new-owner} specifies the new owner and/or group as follows
9944 (with no embedded white space):
9947 [@var{owner}] [ : [@var{group}] ]
9954 If only an @var{owner} (a user name or numeric user ID) is given, that
9955 user is made the owner of each given file, and the files' group is not
9958 @item owner@samp{:}group
9959 If the @var{owner} is followed by a colon and a @var{group} (a
9960 group name or numeric group ID), with no spaces between them, the group
9961 ownership of the files is changed as well (to @var{group}).
9964 If a colon but no group name follows @var{owner}, that user is
9965 made the owner of the files and the group of the files is changed to
9966 @var{owner}'s login group.
9969 If the colon and following @var{group} are given, but the owner
9970 is omitted, only the group of the files is changed; in this case,
9971 @command{chown} performs the same function as @command{chgrp}.
9974 If only a colon is given, or if @var{new-owner} is empty, neither the
9975 owner nor the group is changed.
9979 If @var{owner} or @var{group} is intended to represent a numeric user
9980 or group ID, then you may specify it with a leading @samp{+}.
9981 @xref{Disambiguating names and IDs}.
9983 Some older scripts may still use @samp{.} in place of the @samp{:} separator.
9984 POSIX 1003.1-2001 (@pxref{Standards conformance}) does not
9985 require support for that, but for backward compatibility GNU
9986 @command{chown} supports @samp{.} so long as no ambiguity results.
9987 New scripts should avoid the use of @samp{.} because it is not
9988 portable, and because it has undesirable results if the entire
9989 @var{owner@samp{.}group} happens to identify a user whose name
9992 The @command{chown} command sometimes clears the set-user-ID or
9993 set-group-ID permission bits. This behavior depends on the policy and
9994 functionality of the underlying @code{chown} system call, which may
9995 make system-dependent file mode modifications outside the control of
9996 the @command{chown} command. For example, the @command{chown} command
9997 might not affect those bits when invoked by a user with appropriate
9998 privileges, or when the
9999 bits signify some function other than executable permission (e.g.,
10000 mandatory locking).
10001 When in doubt, check the underlying system behavior.
10003 The program accepts the following options. Also see @ref{Common options}.
10011 @cindex changed owners, verbosely describing
10012 Verbosely describe the action for each @var{file} whose ownership
10021 @cindex error messages, omitting
10022 Do not print error messages about files whose ownership cannot be
10025 @item @w{@kbd{--from}=@var{old-owner}}
10027 @cindex symbolic links, changing owner
10028 Change a @var{file}'s ownership only if it has current attributes specified
10029 by @var{old-owner}. @var{old-owner} has the same form as @var{new-owner}
10031 This option is useful primarily from a security standpoint in that
10032 it narrows considerably the window of potential abuse.
10033 For example, to reflect a user ID numbering change for one user's files
10034 without an option like this, @code{root} might run
10037 find / -owner OLDUSER -print0 | xargs -0 chown -h NEWUSER
10040 But that is dangerous because the interval between when the @command{find}
10041 tests the existing file's owner and when the @command{chown} is actually run
10042 may be quite large.
10043 One way to narrow the gap would be to invoke chown for each file
10047 find / -owner OLDUSER -exec chown -h NEWUSER @{@} \;
10050 But that is very slow if there are many affected files.
10051 With this option, it is safer (the gap is narrower still)
10052 though still not perfect:
10055 chown -h -R --from=OLDUSER NEWUSER /
10058 @item --dereference
10059 @opindex --dereference
10060 @cindex symbolic links, changing owner
10062 Do not act on symbolic links themselves but rather on what they point to.
10063 This is the default.
10066 @itemx --no-dereference
10068 @opindex --no-dereference
10069 @cindex symbolic links, changing owner
10071 Act on symbolic links themselves instead of what they point to.
10072 This mode relies on the @code{lchown} system call.
10073 On systems that do not provide the @code{lchown} system call,
10074 @command{chown} fails when a file specified on the command line
10075 is a symbolic link.
10076 By default, no diagnostic is issued for symbolic links encountered
10077 during a recursive traversal, but see @option{--verbose}.
10079 @item --preserve-root
10080 @opindex --preserve-root
10081 @cindex root directory, disallow recursive modification
10082 Fail upon any attempt to recursively change the root directory, @file{/}.
10083 Without @option{--recursive}, this option has no effect.
10084 @xref{Treating / specially}.
10086 @item --no-preserve-root
10087 @opindex --no-preserve-root
10088 @cindex root directory, allow recursive modification
10089 Cancel the effect of any preceding @option{--preserve-root} option.
10090 @xref{Treating / specially}.
10092 @item --reference=@var{ref_file}
10093 @opindex --reference
10094 Change the user and group of each @var{file} to be the same as those of
10095 @var{ref_file}. If @var{ref_file} is a symbolic link, do not use the
10096 user and group of the symbolic link, but rather those of the file it
10103 Output a diagnostic for every file processed.
10104 If a symbolic link is encountered during a recursive traversal
10105 on a system without the @code{lchown} system call, and @option{--no-dereference}
10106 is in effect, then issue a diagnostic saying neither the symbolic link nor
10107 its referent is being changed.
10112 @opindex --recursive
10113 @cindex recursively changing file ownership
10114 Recursively change ownership of directories and their contents.
10117 @xref{Traversing symlinks}.
10120 @xref{Traversing symlinks}.
10123 @xref{Traversing symlinks}.
10132 # Change the owner of /u to "root".
10135 # Likewise, but also change its group to "staff".
10136 chown root:staff /u
10138 # Change the owner of /u and subfiles to "root".
10143 @node chgrp invocation
10144 @section @command{chgrp}: Change group ownership
10147 @cindex group ownership, changing
10148 @cindex changing group ownership
10150 @command{chgrp} changes the group ownership of each given @var{file}
10151 to @var{group} (which can be either a group name or a numeric group ID)
10152 or to the group of an existing reference file. Synopsis:
10155 chgrp [@var{option}]@dots{} @{@var{group} | --reference=@var{ref_file}@}@c
10159 If @var{group} is intended to represent a
10160 numeric group ID, then you may specify it with a leading @samp{+}.
10161 @xref{Disambiguating names and IDs}.
10163 The program accepts the following options. Also see @ref{Common options}.
10171 @cindex changed files, verbosely describing
10172 Verbosely describe the action for each @var{file} whose group actually
10181 @cindex error messages, omitting
10182 Do not print error messages about files whose group cannot be
10185 @item --dereference
10186 @opindex --dereference
10187 @cindex symbolic links, changing owner
10189 Do not act on symbolic links themselves but rather on what they point to.
10190 This is the default.
10193 @itemx --no-dereference
10195 @opindex --no-dereference
10196 @cindex symbolic links, changing group
10198 Act on symbolic links themselves instead of what they point to.
10199 This mode relies on the @code{lchown} system call.
10200 On systems that do not provide the @code{lchown} system call,
10201 @command{chgrp} fails when a file specified on the command line
10202 is a symbolic link.
10203 By default, no diagnostic is issued for symbolic links encountered
10204 during a recursive traversal, but see @option{--verbose}.
10206 @item --preserve-root
10207 @opindex --preserve-root
10208 @cindex root directory, disallow recursive modification
10209 Fail upon any attempt to recursively change the root directory, @file{/}.
10210 Without @option{--recursive}, this option has no effect.
10211 @xref{Treating / specially}.
10213 @item --no-preserve-root
10214 @opindex --no-preserve-root
10215 @cindex root directory, allow recursive modification
10216 Cancel the effect of any preceding @option{--preserve-root} option.
10217 @xref{Treating / specially}.
10219 @item --reference=@var{ref_file}
10220 @opindex --reference
10221 Change the group of each @var{file} to be the same as that of
10222 @var{ref_file}. If @var{ref_file} is a symbolic link, do not use the
10223 group of the symbolic link, but rather that of the file it refers to.
10229 Output a diagnostic for every file processed.
10230 If a symbolic link is encountered during a recursive traversal
10231 on a system without the @code{lchown} system call, and @option{--no-dereference}
10232 is in effect, then issue a diagnostic saying neither the symbolic link nor
10233 its referent is being changed.
10238 @opindex --recursive
10239 @cindex recursively changing group ownership
10240 Recursively change the group ownership of directories and their contents.
10243 @xref{Traversing symlinks}.
10246 @xref{Traversing symlinks}.
10249 @xref{Traversing symlinks}.
10258 # Change the group of /u to "staff".
10261 # Change the group of /u and subfiles to "staff".
10266 @node chmod invocation
10267 @section @command{chmod}: Change access permissions
10270 @cindex changing access permissions
10271 @cindex access permissions, changing
10272 @cindex permissions, changing access
10274 @command{chmod} changes the access permissions of the named files. Synopsis:
10277 chmod [@var{option}]@dots{} @{@var{mode} | --reference=@var{ref_file}@}@c
10281 @cindex symbolic links, permissions of
10282 @command{chmod} never changes the permissions of symbolic links, since
10283 the @command{chmod} system call cannot change their permissions.
10284 This is not a problem since the permissions of symbolic links are
10285 never used. However, for each symbolic link listed on the command
10286 line, @command{chmod} changes the permissions of the pointed-to file.
10287 In contrast, @command{chmod} ignores symbolic links encountered during
10288 recursive directory traversals.
10290 A successful use of @command{chmod} clears the set-group-ID bit of a
10291 regular file if the file's group ID does not match the user's
10292 effective group ID or one of the user's supplementary group IDs,
10293 unless the user has appropriate privileges. Additional restrictions
10294 may cause the set-user-ID and set-group-ID bits of @var{mode} or
10295 @var{ref_file} to be ignored. This behavior depends on the policy and
10296 functionality of the underlying @code{chmod} system call. When in
10297 doubt, check the underlying system behavior.
10299 If used, @var{mode} specifies the new file mode bits.
10300 For details, see the section on @ref{File permissions}.
10301 If you really want @var{mode} to have a leading @samp{-}, you should
10302 use @option{--} first, e.g., @samp{chmod -- -w file}. Typically,
10303 though, @samp{chmod a-w file} is preferable, and @command{chmod -w
10304 file} (without the @option{--}) complains if it behaves differently
10305 from what @samp{chmod a-w file} would do.
10307 The program accepts the following options. Also see @ref{Common options}.
10315 Verbosely describe the action for each @var{file} whose permissions
10324 @cindex error messages, omitting
10325 Do not print error messages about files whose permissions cannot be
10328 @item --preserve-root
10329 @opindex --preserve-root
10330 @cindex root directory, disallow recursive modification
10331 Fail upon any attempt to recursively change the root directory, @file{/}.
10332 Without @option{--recursive}, this option has no effect.
10333 @xref{Treating / specially}.
10335 @item --no-preserve-root
10336 @opindex --no-preserve-root
10337 @cindex root directory, allow recursive modification
10338 Cancel the effect of any preceding @option{--preserve-root} option.
10339 @xref{Treating / specially}.
10345 Verbosely describe the action or non-action taken for every @var{file}.
10347 @item --reference=@var{ref_file}
10348 @opindex --reference
10349 Change the mode of each @var{file} to be the same as that of @var{ref_file}.
10350 @xref{File permissions}.
10351 If @var{ref_file} is a symbolic link, do not use the mode
10352 of the symbolic link, but rather that of the file it refers to.
10357 @opindex --recursive
10358 @cindex recursively changing access permissions
10359 Recursively change permissions of directories and their contents.
10366 @node touch invocation
10367 @section @command{touch}: Change file timestamps
10370 @cindex changing file timestamps
10371 @cindex file timestamps, changing
10372 @cindex timestamps, changing file
10374 @command{touch} changes the access and/or modification times of the
10375 specified files. Synopsis:
10378 touch [@var{option}]@dots{} @var{file}@dots{}
10381 @cindex empty files, creating
10382 Any @var{file} argument that does not exist is created empty, unless
10383 option @option{--no-create} (@option{-c}) or @option{--no-dereference}
10384 (@option{-h}) was in effect.
10386 A @var{file} argument string of @samp{-} is handled specially and
10387 causes @command{touch} to change the times of the file associated with
10391 By default, @command{touch} sets file timestamps to the current time.
10392 Because @command{touch} acts on its operands left to right, the
10393 resulting timestamps of earlier and later operands may disagree.
10394 Also, the determination of what time is ``current'' depends on the
10395 platform. Platforms with network file systems often use different
10396 clocks for the operating system and for file systems; because
10397 @command{touch} typically uses file systems' clocks by default, clock
10398 skew can cause the resulting file timestamps to appear to be in a
10399 program's ``future'' or ``past''.
10401 @cindex file timestamp resolution
10402 The @command{touch} command sets the file's timestamp to the greatest
10403 representable value that is not greater than the requested time. This
10404 can differ from the requested time for several reasons. First, the
10405 requested time may have a higher resolution than supported. Second, a
10406 file system may use different resolutions for different types of
10407 times. Third, file timestamps may use a different resolution than
10408 operating system timestamps. Fourth, the operating system primitives
10409 used to update timestamps may employ yet a different resolution. For
10410 example, in theory a file system might use 10-microsecond resolution
10411 for access time and 100-nanosecond resolution for modification time,
10412 and the operating system might use nanosecond resolution for the
10413 current time and microsecond resolution for the primitive that
10414 @command{touch} uses to set a file's timestamp to an arbitrary value.
10416 @cindex permissions, for changing file timestamps
10417 When setting file timestamps to the current time, @command{touch} can
10418 change the timestamps for files that the user does not own but has
10419 write permission for. Otherwise, the user must own the files. Some
10420 older systems have a further restriction: the user must own the files
10421 unless both the access and modification times are being set to the
10424 Although @command{touch} provides options for changing two of the times---the
10425 times of last access and modification---of a file, there is actually
10426 a standard third one as well: the inode change time. This is often
10427 referred to as a file's @code{ctime}.
10428 The inode change time represents the time when the file's meta-information
10429 last changed. One common example of this is when the permissions of a
10430 file change. Changing the permissions doesn't access the file, so
10431 the atime doesn't change, nor does it modify the file, so the mtime
10432 doesn't change. Yet, something about the file itself has changed,
10433 and this must be noted somewhere. This is the job of the ctime field.
10434 This is necessary, so that, for example, a backup program can make a
10435 fresh copy of the file, including the new permissions value.
10436 Another operation that modifies a file's ctime without affecting
10437 the others is renaming. In any case, it is not possible, in normal
10438 operations, for a user to change the ctime field to a user-specified value.
10439 Some operating systems and file systems support a fourth time: the
10440 birth time, when the file was first created; by definition, this
10441 timestamp never changes.
10444 Time stamps assume the time zone rules specified by the @env{TZ}
10445 environment variable, or by the system default rules if @env{TZ} is
10446 not set. @xref{TZ Variable,, Specifying the Time Zone with @env{TZ},
10447 libc, The GNU C Library Reference Manual}.
10448 You can avoid ambiguities during
10449 daylight saving transitions by using @sc{utc} time stamps.
10451 The program accepts the following options. Also see @ref{Common options}.
10456 @itemx --time=atime
10457 @itemx --time=access
10461 @opindex atime@r{, changing}
10462 @opindex access @r{time, changing}
10463 @opindex use @r{time, changing}
10464 Change the access time only.
10469 @opindex --no-create
10470 Do not warn about or create files that do not exist.
10473 @itemx --date=@var{time}
10477 Use @var{time} instead of the current time. It can contain month names,
10478 time zones, @samp{am} and @samp{pm}, @samp{yesterday}, etc. For
10479 example, @option{--date="2004-02-27 14:19:13.489392193 +0530"}
10480 specifies the instant of time that is 489,392,193 nanoseconds after
10481 February 27, 2004 at 2:19:13 PM in a time zone that is 5 hours and 30
10482 minutes east of UTC@. @xref{Date input formats}.
10483 File systems that do not support high-resolution time stamps
10484 silently ignore any excess precision here.
10488 @cindex BSD @command{touch} compatibility
10489 Ignored; for compatibility with BSD versions of @command{touch}.
10492 @itemx --no-dereference
10494 @opindex --no-dereference
10495 @cindex symbolic links, changing time
10497 Attempt to change the timestamps of a symbolic link, rather than what
10498 the link refers to. When using this option, empty files are not
10499 created, but option @option{-c} must also be used to avoid warning
10500 about files that do not exist. Not all systems support changing the
10501 timestamps of symlinks, since underlying system support for this
10502 action was not required until POSIX 2008. Also, on some
10503 systems, the mere act of examining a symbolic link changes the access
10504 time, such that only changes to the modification time will persist
10505 long enough to be observable. When coupled with option @option{-r}, a
10506 reference timestamp is taken from a symbolic link rather than the file
10510 @itemx --time=mtime
10511 @itemx --time=modify
10514 @opindex mtime@r{, changing}
10515 @opindex modify @r{time, changing}
10516 Change the modification time only.
10518 @item -r @var{file}
10519 @itemx --reference=@var{file}
10521 @opindex --reference
10522 Use the times of the reference @var{file} instead of the current time.
10523 If this option is combined with the @option{--date=@var{time}}
10524 (@option{-d @var{time}}) option, the reference @var{file}'s time is
10525 the origin for any relative @var{time}s given, but is otherwise ignored.
10526 For example, @samp{-r foo -d '-5 seconds'} specifies a time stamp
10527 equal to five seconds before the corresponding time stamp for @file{foo}.
10528 If @var{file} is a symbolic link, the reference timestamp is taken
10529 from the target of the symlink, unless @option{-h} was also in effect.
10531 @item -t [[@var{cc}]@var{yy}]@var{mmddhhmm}[.@var{ss}]
10532 @cindex leap seconds
10533 Use the argument (optional four-digit or two-digit years, months,
10534 days, hours, minutes, optional seconds) instead of the current time.
10535 If the year is specified with only two digits, then @var{cc}
10536 is 20 for years in the range 0 @dots{} 68, and 19 for years in
10537 69 @dots{} 99. If no digits of the year are specified,
10538 the argument is interpreted as a date in the current year.
10539 On the atypical systems that support leap seconds, @var{ss} may be
10544 @vindex _POSIX2_VERSION
10545 On older systems, @command{touch} supports an obsolete syntax, as follows.
10546 If no timestamp is given with any of the @option{-d}, @option{-r}, or
10547 @option{-t} options, and if there are two or more @var{file}s and the
10548 first @var{file} is of the form @samp{@var{mmddhhmm}[@var{yy}]} and this
10549 would be a valid argument to the @option{-t} option (if the @var{yy}, if
10550 any, were moved to the front), and if the represented year
10551 is in the range 1969--1999, that argument is interpreted as the time
10552 for the other files instead of as a file name.
10553 This obsolete behavior can be enabled or disabled with the
10554 @env{_POSIX2_VERSION} environment variable (@pxref{Standards
10555 conformance}), but portable scripts should avoid commands whose
10556 behavior depends on this variable.
10557 For example, use @samp{touch ./12312359 main.c} or @samp{touch -t
10558 12312359 main.c} rather than the ambiguous @samp{touch 12312359 main.c}.
10564 @chapter Disk usage
10568 No disk can hold an infinite amount of data. These commands report
10569 how much disk storage is in use or available, report other file and
10570 file status information, and write buffers to disk.
10573 * df invocation:: Report file system disk space usage.
10574 * du invocation:: Estimate file space usage.
10575 * stat invocation:: Report file or file system status.
10576 * sync invocation:: Synchronize memory and disk.
10577 * truncate invocation:: Shrink or extend the size of a file.
10581 @node df invocation
10582 @section @command{df}: Report file system disk space usage
10585 @cindex file system disk usage
10586 @cindex disk usage by file system
10588 @command{df} reports the amount of disk space used and available on
10589 file systems. Synopsis:
10592 df [@var{option}]@dots{} [@var{file}]@dots{}
10595 With no arguments, @command{df} reports the space used and available on all
10596 currently mounted file systems (of all types). Otherwise, @command{df}
10597 reports on the file system containing each argument @var{file}.
10599 Normally the disk space is printed in units of
10600 1024 bytes, but this can be overridden (@pxref{Block size}).
10601 Non-integer quantities are rounded up to the next higher unit.
10603 For bind mounts and without arguments, @command{df} only outputs the statistics
10604 for the first occurrence of that device in the list of file systems
10605 (@var{mtab}), i.e., it hides duplicate entries, unless the @option{-a} option is
10608 By default, @command{df} omits the early-boot pseudo file system type
10609 @samp{rootfs}, unless the @option{-a} option is specified or that file system
10610 type is explicitly to be included by using the @option{-t} option.
10612 @cindex disk device file
10613 @cindex device file, disk
10614 If an argument @var{file} is a disk device file containing a mounted
10615 file system, @command{df} shows the space available on that file system
10616 rather than on the file system containing the device node (i.e., the root
10617 file system). @sc{gnu} @command{df} does not attempt to determine the
10619 on unmounted file systems, because on most kinds of systems doing so
10620 requires extremely nonportable intimate knowledge of file system
10623 The program accepts the following options. Also see @ref{Common options}.
10631 @cindex automounter file systems
10632 @cindex ignore file systems
10633 Include in the listing dummy file systems, which
10634 are omitted by default. Such file systems are typically special-purpose
10635 pseudo-file-systems, such as automounter entries.
10637 @item -B @var{size}
10638 @itemx --block-size=@var{size}
10640 @opindex --block-size
10641 @cindex file system sizes
10642 Scale sizes by @var{size} before printing them (@pxref{Block size}).
10643 For example, @option{-BG} prints sizes in units of 1,073,741,824 bytes.
10647 @cindex grand total of disk size, usage and available space
10648 Print a grand total of all arguments after all arguments have
10649 been processed. This can be used to find out the total disk size, usage
10650 and available space of all listed devices.
10652 For the grand total line, @command{df} prints @samp{"total"} into the
10653 @var{source} column, and @samp{"-"} into the @var{target} column.
10654 If there is no @var{source} column (see @option{--output}), then
10655 @command{df} prints @samp{"total"} into the @var{target} column,
10662 Equivalent to @option{--si}.
10668 @cindex inode usage
10669 List inode usage information instead of block usage. An inode (short
10670 for index node) contains information about a file such as its owner,
10671 permissions, timestamps, and location on the disk.
10675 @cindex kibibytes for file system sizes
10676 Print sizes in 1024-byte blocks, overriding the default block size
10677 (@pxref{Block size}).
10678 This option is equivalent to @option{--block-size=1K}.
10684 @cindex file system types, limiting output to certain
10685 Limit the listing to local file systems. By default, remote file systems
10690 @cindex file system space, retrieving old data more quickly
10691 Do not invoke the @code{sync} system call before getting any usage data.
10692 This may make @command{df} run significantly faster on systems with many
10693 disks, but on some systems (notably SunOS) the results may be slightly
10694 out of date. This is the default.
10697 @itemx @w{@kbd{--output}[=@var{field_list}]}
10699 Use the output format defined by @var{field_list}, or print all fields if
10700 @var{field_list} is omitted. In the latter case, the order of the columns
10701 conforms to the order of the field descriptions below.
10703 The use of the @option{--output} together with each of the options @option{-i},
10704 @option{-P}, and @option{-T} is mutually exclusive.
10706 FIELD_LIST is a comma-separated list of columns to be included in @command{df}'s
10707 output and therefore effectively controls the order of output columns.
10708 Each field can thus be used at the place of choice, but yet must only be
10711 Valid field names in the @var{field_list} are:
10714 The source of the mount point, usually a device.
10719 Total number of inodes.
10721 Number of used inodes.
10723 Number of available inodes.
10725 Percentage of @var{iused} divided by @var{itotal}.
10728 Total number of blocks.
10730 Number of used blocks.
10732 Number of available blocks.
10734 Percentage of @var{used} divided by @var{size}.
10740 The fields for block and inodes statistics are affected by the scaling
10741 options like @option{-h} as usual.
10743 The definition of the @var{field_list} can even be splitted among several
10744 @option{--output} uses.
10748 # Print the TARGET (i.e., the mount point) along with their percentage
10749 # statistic regarding the blocks and the inodes.
10750 df --out=target --output=pcent,ipcent
10752 # Print all available fields.
10758 @itemx --portability
10760 @opindex --portability
10761 @cindex one-line output format
10762 @cindex POSIX output format
10763 @cindex portable output format
10764 @cindex output format, portable
10765 Use the POSIX output format. This is like the default format except
10770 The information about each file system is always printed on exactly
10771 one line; a mount device is never put on a line by itself. This means
10772 that if the mount device name is more than 20 characters long (e.g., for
10773 some network mounts), the columns are misaligned.
10776 The labels in the header output line are changed to conform to POSIX.
10779 The default block size and output format are unaffected by the
10780 @env{DF_BLOCK_SIZE}, @env{BLOCK_SIZE} and @env{BLOCKSIZE} environment
10781 variables. However, the default block size is still affected by
10782 @env{POSIXLY_CORRECT}: it is 512 if @env{POSIXLY_CORRECT} is set, 1024
10783 otherwise. @xref{Block size}.
10790 @cindex file system space, retrieving current data more slowly
10791 Invoke the @code{sync} system call before getting any usage data. On
10792 some systems (notably SunOS), doing this yields more up to date results,
10793 but in general this option makes @command{df} much slower, especially when
10794 there are many or very busy file systems.
10796 @item -t @var{fstype}
10797 @itemx --type=@var{fstype}
10800 @cindex file system types, limiting output to certain
10801 Limit the listing to file systems of type @var{fstype}. Multiple
10802 file system types can be specified by giving multiple @option{-t} options.
10803 By default, nothing is omitted.
10806 @itemx --print-type
10808 @opindex --print-type
10809 @cindex file system types, printing
10810 Print each file system's type. The types printed here are the same ones
10811 you can include or exclude with @option{-t} and @option{-x}. The particular
10812 types printed are whatever is supported by the system. Here are some of
10813 the common names (this list is certainly not exhaustive):
10818 @cindex NFS file system type
10819 An NFS file system, i.e., one mounted over a network from another
10820 machine. This is the one type name which seems to be used uniformly by
10823 @item 4.2@r{, }ufs@r{, }efs@dots{}
10824 @cindex Linux file system types
10825 @cindex local file system types
10826 @opindex 4.2 @r{file system type}
10827 @opindex ufs @r{file system type}
10828 @opindex efs @r{file system type}
10829 A file system on a locally-mounted hard disk. (The system might even
10830 support more than one type here; Linux does.)
10832 @item hsfs@r{, }cdfs
10833 @cindex CD-ROM file system type
10834 @cindex High Sierra file system
10835 @opindex hsfs @r{file system type}
10836 @opindex cdfs @r{file system type}
10837 A file system on a CD-ROM drive. HP-UX uses @samp{cdfs}, most other
10838 systems use @samp{hsfs} (@samp{hs} for ``High Sierra'').
10841 @cindex PC file system
10842 @cindex DOS file system
10843 @cindex MS-DOS file system
10844 @cindex diskette file system
10846 An MS-DOS file system, usually on a diskette.
10850 @item -x @var{fstype}
10851 @itemx --exclude-type=@var{fstype}
10853 @opindex --exclude-type
10854 Limit the listing to file systems not of type @var{fstype}.
10855 Multiple file system types can be eliminated by giving multiple
10856 @option{-x} options. By default, no file system types are omitted.
10859 Ignored; for compatibility with System V versions of @command{df}.
10864 Failure includes the case where no output is generated, so you can
10865 inspect the exit status of a command like @samp{df -t ext3 -t reiserfs
10866 @var{dir}} to test whether @var{dir} is on a file system of type
10867 @samp{ext3} or @samp{reiserfs}.
10869 Since the list of file systems (@var{mtab}) is needed to determine the
10870 file system type, failure includes the cases when that list cannot
10871 be read and one or more of the options @option{-a}, @option{-l}, @option{-t}
10872 or @option{-x} is used together with a file name argument.
10875 @node du invocation
10876 @section @command{du}: Estimate file space usage
10879 @cindex file space usage
10880 @cindex disk usage for files
10882 @command{du} reports the amount of disk space used by the specified files
10883 and for each subdirectory (of directory arguments). Synopsis:
10886 du [@var{option}]@dots{} [@var{file}]@dots{}
10889 With no arguments, @command{du} reports the disk space for the current
10890 directory. Normally the disk space is printed in units of
10891 1024 bytes, but this can be overridden (@pxref{Block size}).
10892 Non-integer quantities are rounded up to the next higher unit.
10894 If two or more hard links point to the same file, only one of the hard
10895 links is counted. The @var{file} argument order affects which links
10896 are counted, and changing the argument order may change the numbers
10897 that @command{du} outputs.
10899 The program accepts the following options. Also see @ref{Common options}.
10907 Show counts for all files, not just directories.
10909 @item --apparent-size
10910 @opindex --apparent-size
10911 Print apparent sizes, rather than disk usage. The apparent size of a
10912 file is the number of bytes reported by @code{wc -c} on regular files,
10913 or more generally, @code{ls -l --block-size=1} or @code{stat --format=%s}.
10914 For example, a file containing the word @samp{zoo} with no newline would,
10915 of course, have an apparent size of 3. Such a small file may require
10916 anywhere from 0 to 16 KiB or more of disk space, depending on
10917 the type and configuration of the file system on which the file resides.
10918 However, a sparse file created with this command:
10921 dd bs=1 seek=2GiB if=/dev/null of=big
10925 has an apparent size of 2 GiB, yet on most modern
10926 systems, it actually uses almost no disk space.
10932 Equivalent to @code{--apparent-size --block-size=1}.
10934 @item -B @var{size}
10935 @itemx --block-size=@var{size}
10937 @opindex --block-size
10939 Scale sizes by @var{size} before printing them (@pxref{Block size}).
10940 For example, @option{-BG} prints sizes in units of 1,073,741,824 bytes.
10946 @cindex grand total of disk space
10947 Print a grand total of all arguments after all arguments have
10948 been processed. This can be used to find out the total disk usage of
10949 a given set of files or directories.
10952 @itemx --dereference-args
10954 @opindex --dereference-args
10955 Dereference symbolic links that are command line arguments.
10956 Does not affect other symbolic links. This is helpful for finding
10957 out the disk usage of directories, such as @file{/usr/tmp}, which
10958 are often symbolic links.
10960 @c --files0-from=FILE
10961 @filesZeroFromOption{du,, with the @option{--total} (@option{-c}) option}
10967 Equivalent to @option{--dereference-args} (@option{-D}).
10971 @cindex kibibytes for file sizes
10972 Print sizes in 1024-byte blocks, overriding the default block size
10973 (@pxref{Block size}).
10974 This option is equivalent to @option{--block-size=1K}.
10977 @itemx --count-links
10979 @opindex --count-links
10980 @cindex hard links, counting in @command{du}
10981 Count the size of all files, even if they have appeared already (as a
10985 @itemx --dereference
10987 @opindex --dereference
10988 @cindex symbolic links, dereferencing in @command{du}
10989 Dereference symbolic links (show the disk space used by the file
10990 or directory that the link points to instead of the space used by
10995 @cindex mebibytes for file sizes
10996 Print sizes in 1,048,576-byte blocks, overriding the default block size
10997 (@pxref{Block size}).
10998 This option is equivalent to @option{--block-size=1M}.
11001 @itemx --no-dereference
11003 @opindex --no-dereference
11004 @cindex symbolic links, dereferencing in @command{du}
11005 For each symbolic links encountered by @command{du},
11006 consider the disk space used by the symbolic link.
11008 @item -d @var{depth}
11009 @item --max-depth=@var{depth}
11010 @opindex -d @var{depth}
11011 @opindex --max-depth=@var{depth}
11012 @cindex limiting output of @command{du}
11013 Show the total for each directory (and file if --all) that is at
11014 most MAX_DEPTH levels down from the root of the hierarchy. The root
11015 is at level 0, so @code{du --max-depth=0} is equivalent to @code{du -s}.
11024 @opindex --summarize
11025 Display only a total for each argument.
11028 @itemx --separate-dirs
11030 @opindex --separate-dirs
11031 Normally, in the output of @command{du} (when not using @option{--summarize}),
11032 the size listed next to a directory name, @var{d}, represents the sum
11033 of sizes of all entries beneath @var{d} as well as the size of @var{d} itself.
11034 With @option{--separate-dirs}, the size reported for a directory name,
11035 @var{d}, is merely the @code{stat.st_size}-derived size of the directory
11040 @cindex last modified dates, displaying in @command{du}
11041 Show time of the most recent modification of any file in the directory,
11042 or any of its subdirectories.
11045 @itemx --time=status
11048 @opindex ctime@r{, show the most recent}
11049 @opindex status time@r{, show the most recent}
11050 @opindex use time@r{, show the most recent}
11051 Show the most recent status change time (the @samp{ctime} in the inode) of
11052 any file in the directory, instead of the modification time.
11055 @itemx --time=access
11057 @opindex atime@r{, show the most recent}
11058 @opindex access time@r{, show the most recent}
11059 Show the most recent access time (the @samp{atime} in the inode) of
11060 any file in the directory, instead of the modification time.
11062 @item --time-style=@var{style}
11063 @opindex --time-style
11065 List timestamps in style @var{style}. This option has an effect only if
11066 the @option{--time} option is also specified. The @var{style} should
11067 be one of the following:
11070 @item +@var{format}
11072 List timestamps using @var{format}, where @var{format} is interpreted
11073 like the format argument of @command{date} (@pxref{date invocation}).
11074 For example, @option{--time-style="+%Y-%m-%d %H:%M:%S"} causes
11075 @command{du} to list timestamps like @samp{2002-03-30 23:45:56}. As
11076 with @command{date}, @var{format}'s interpretation is affected by the
11077 @env{LC_TIME} locale category.
11080 List timestamps in full using ISO 8601 date, time, and time zone
11081 format with nanosecond precision, e.g., @samp{2002-03-30
11082 23:45:56.477817180 -0700}. This style is equivalent to
11083 @samp{+%Y-%m-%d %H:%M:%S.%N %z}.
11086 List ISO 8601 date and time in minutes, e.g.,
11087 @samp{2002-03-30 23:45}. These timestamps are shorter than
11088 @samp{full-iso} timestamps, and are usually good enough for everyday
11089 work. This style is equivalent to @samp{+%Y-%m-%d %H:%M}.
11092 List ISO 8601 dates for timestamps, e.g., @samp{2002-03-30}.
11093 This style is equivalent to @samp{+%Y-%m-%d}.
11097 You can specify the default value of the @option{--time-style} option
11098 with the environment variable @env{TIME_STYLE}; if @env{TIME_STYLE} is not set
11099 the default style is @samp{long-iso}. For compatibility with @command{ls},
11100 if @env{TIME_STYLE} begins with @samp{+} and contains a newline,
11101 the newline and any later characters are ignored; if @env{TIME_STYLE}
11102 begins with @samp{posix-} the @samp{posix-} is ignored; and if
11103 @env{TIME_STYLE} is @samp{locale} it is ignored.
11106 @itemx --one-file-system
11108 @opindex --one-file-system
11109 @cindex one file system, restricting @command{du} to
11110 Skip directories that are on different file systems from the one that
11111 the argument being processed is on.
11113 @item --exclude=@var{pattern}
11114 @opindex --exclude=@var{pattern}
11115 @cindex excluding files from @command{du}
11116 When recursing, skip subdirectories or files matching @var{pattern}.
11117 For example, @code{du --exclude='*.o'} excludes files whose names
11120 @item -X @var{file}
11121 @itemx --exclude-from=@var{file}
11122 @opindex -X @var{file}
11123 @opindex --exclude-from=@var{file}
11124 @cindex excluding files from @command{du}
11125 Like @option{--exclude}, except take the patterns to exclude from @var{file},
11126 one per line. If @var{file} is @samp{-}, take the patterns from standard
11131 @cindex NFS mounts from BSD to HP-UX
11132 On BSD systems, @command{du} reports sizes that are half the correct
11133 values for files that are NFS-mounted from HP-UX systems. On HP-UX
11134 systems, it reports sizes that are twice the correct values for
11135 files that are NFS-mounted from BSD systems. This is due to a flaw
11136 in HP-UX; it also affects the HP-UX @command{du} program.
11141 @node stat invocation
11142 @section @command{stat}: Report file or file system status
11145 @cindex file status
11146 @cindex file system status
11148 @command{stat} displays information about the specified file(s). Synopsis:
11151 stat [@var{option}]@dots{} [@var{file}]@dots{}
11154 With no option, @command{stat} reports all information about the given files.
11155 But it also can be used to report the information of the file systems the
11156 given files are located on. If the files are links, @command{stat} can
11157 also give information about the files the links point to.
11159 @mayConflictWithShellBuiltIn{stat}
11164 @itemx --dereference
11166 @opindex --dereference
11167 @cindex symbolic links, dereferencing in @command{stat}
11168 Change how @command{stat} treats symbolic links.
11169 With this option, @command{stat} acts on the file referenced
11170 by each symbolic link argument.
11171 Without it, @command{stat} acts on any symbolic link argument directly.
11174 @itemx --file-system
11176 @opindex --file-system
11177 @cindex file systems
11178 Report information about the file systems where the given files are located
11179 instead of information about the files themselves.
11180 This option implies the @option{-L} option.
11183 @itemx --format=@var{format}
11185 @opindex --format=@var{format}
11186 @cindex output format
11187 Use @var{format} rather than the default format.
11188 @var{format} is automatically newline-terminated, so
11189 running a command like the following with two or more @var{file}
11190 operands produces a line of output for each operand:
11192 $ stat --format=%d:%i / /usr
11197 @item --printf=@var{format}
11198 @opindex --printf=@var{format}
11199 @cindex output format
11200 Use @var{format} rather than the default format.
11201 Like @option{--format}, but interpret backslash escapes,
11202 and do not output a mandatory trailing newline.
11203 If you want a newline, include @samp{\n} in the @var{format}.
11204 Here's how you would use @option{--printf} to print the device
11205 and inode numbers of @file{/} and @file{/usr}:
11207 $ stat --printf='%d:%i\n' / /usr
11216 @cindex terse output
11217 Print the information in terse form, suitable for parsing by other programs.
11221 The valid @var{format} directives for files with @option{--format} and
11222 @option{--printf} are:
11225 @item %a - Access rights in octal
11226 @item %A - Access rights in human readable form
11227 @item %b - Number of blocks allocated (see @samp{%B})
11228 @item %B - The size in bytes of each block reported by @samp{%b}
11229 @item %C - The SELinux security context of a file, if available
11230 @item %d - Device number in decimal
11231 @item %D - Device number in hex
11232 @item %f - Raw mode in hex
11233 @item %F - File type
11234 @item %g - Group ID of owner
11235 @item %G - Group name of owner
11236 @item %h - Number of hard links
11237 @item %i - Inode number
11238 @item %m - Mount point (See note below)
11239 @item %n - File name
11240 @item %N - Quoted file name with dereference if symbolic link
11241 @item %o - Optimal I/O transfer size hint
11242 @item %s - Total size, in bytes
11243 @item %t - Major device type in hex
11244 @item %T - Minor device type in hex
11245 @item %u - User ID of owner
11246 @item %U - User name of owner
11247 @item %w - Time of file birth, or @samp{-} if unknown
11248 @item %W - Time of file birth as seconds since Epoch, or @samp{0}
11249 @item %x - Time of last access
11250 @item %X - Time of last access as seconds since Epoch
11251 @item %y - Time of last modification
11252 @item %Y - Time of last modification as seconds since Epoch
11253 @item %z - Time of last change
11254 @item %Z - Time of last change as seconds since Epoch
11257 The @samp{%W}, @samp{%X}, @samp{%Y}, and @samp{%Z} formats accept a
11258 precision preceded by a period to specify the number of digits to
11259 print after the decimal point. For example, @samp{%.3X} outputs the
11260 last access time to millisecond precision. If a period is given but no
11261 precision, @command{stat} uses 9 digits, so @samp{%.X} is equivalent to
11262 @samp{%.9X}@. When discarding excess precision, time stamps are truncated
11263 toward minus infinity.
11267 $ stat -c '[%015Y]' /usr
11270 $ stat -c '[%15Y]' /usr
11272 $ stat -c '[%-15Y]' /usr
11275 $ stat -c '[%.3Y]' /usr
11277 $ stat -c '[%.Y]' /usr
11278 [1288929712.114951834]
11281 The mount point printed by @samp{%m} is similar to that output
11282 by @command{df}, except that:
11285 stat does not dereference symlinks by default
11286 (unless @option{-L} is specified)
11288 stat does not search for specified device nodes in the
11289 file system list, instead operating on them directly
11292 stat outputs the alias for a bind mounted file, rather than
11293 the initial mount point of its backing device.
11294 One can recursively call stat until there is no change in output,
11295 to get the current base mount point
11298 When listing file system information (@option{--file-system} (@option{-f})),
11299 you must use a different set of @var{format} directives:
11302 @item %a - Free blocks available to non-super-user
11303 @item %b - Total data blocks in file system
11304 @item %c - Total file nodes in file system
11305 @item %d - Free file nodes in file system
11306 @item %f - Free blocks in file system
11307 @item %i - File System ID in hex
11308 @item %l - Maximum length of file names
11309 @item %n - File name
11310 @item %s - Block size (for faster transfers)
11311 @item %S - Fundamental block size (for block counts)
11312 @item %t - Type in hex
11313 @item %T - Type in human readable form
11317 Time stamps are listed according to the time zone rules specified by
11318 the @env{TZ} environment variable, or by the system default rules if
11319 @env{TZ} is not set. @xref{TZ Variable,, Specifying the Time Zone
11320 with @env{TZ}, libc, The GNU C Library Reference Manual}.
11325 @node sync invocation
11326 @section @command{sync}: Synchronize data on disk with memory
11329 @cindex synchronize disk and memory
11331 @cindex superblock, writing
11332 @cindex inodes, written buffered
11333 @command{sync} writes any data buffered in memory out to disk. This can
11334 include (but is not limited to) modified superblocks, modified inodes,
11335 and delayed reads and writes. This must be implemented by the kernel;
11336 The @command{sync} program does nothing but exercise the @code{sync} system
11339 @cindex crashes and corruption
11340 The kernel keeps data in memory to avoid doing (relatively slow) disk
11341 reads and writes. This improves performance, but if the computer
11342 crashes, data may be lost or the file system corrupted as a
11343 result. The @command{sync} command ensures everything in memory
11344 is written to disk.
11346 Any arguments are ignored, except for a lone @option{--help} or
11347 @option{--version} (@pxref{Common options}).
11352 @node truncate invocation
11353 @section @command{truncate}: Shrink or extend the size of a file
11356 @cindex truncating, file sizes
11358 @command{truncate} shrinks or extends the size of each @var{file} to the
11359 specified size. Synopsis:
11362 truncate @var{option}@dots{} @var{file}@dots{}
11365 @cindex files, creating
11366 Any @var{file} that does not exist is created.
11368 @cindex sparse files, creating
11369 @cindex holes, creating files with
11370 If a @var{file} is larger than the specified size, the extra data is lost.
11371 If a @var{file} is shorter, it is extended and the extended part (or hole)
11372 reads as zero bytes.
11374 The program accepts the following options. Also see @ref{Common options}.
11381 @opindex --no-create
11382 Do not create files that do not exist.
11387 @opindex --io-blocks
11388 Treat @var{size} as number of I/O blocks of the @var{file} rather than bytes.
11390 @item -r @var{rfile}
11391 @itemx --reference=@var{rfile}
11393 @opindex --reference
11394 Base the size of each @var{file} on the size of @var{rfile}.
11396 @item -s @var{size}
11397 @itemx --size=@var{size}
11400 Set or adjust the size of each @var{file} according to @var{size}.
11401 @multiplierSuffixesNoBlocks{size}
11403 @var{size} may also be prefixed by one of the following to adjust
11404 the size of each @var{file} based on their current size:
11406 @samp{+} => extend by
11407 @samp{-} => reduce by
11408 @samp{<} => at most
11409 @samp{>} => at least
11410 @samp{/} => round down to multiple of
11411 @samp{%} => round up to multiple of
11419 @node Printing text
11420 @chapter Printing text
11422 @cindex printing text, commands for
11423 @cindex commands for printing text
11425 This section describes commands that display text strings.
11428 * echo invocation:: Print a line of text.
11429 * printf invocation:: Format and print data.
11430 * yes invocation:: Print a string until interrupted.
11434 @node echo invocation
11435 @section @command{echo}: Print a line of text
11438 @cindex displaying text
11439 @cindex printing text
11440 @cindex text, displaying
11441 @cindex arbitrary text, displaying
11443 @command{echo} writes each given @var{string} to standard output, with a
11444 space between each and a newline after the last one. Synopsis:
11447 echo [@var{option}]@dots{} [@var{string}]@dots{}
11450 @mayConflictWithShellBuiltIn{echo}
11452 The program accepts the following options. Also see @ref{Common options}.
11453 Options must precede operands, and the normally-special argument
11454 @samp{--} has no special meaning and is treated like any other
11460 Do not output the trailing newline.
11464 @cindex backslash escapes
11465 Enable interpretation of the following backslash-escaped characters in
11474 produce no further output
11490 the eight-bit value that is the octal number @var{nnn}
11491 (zero to three octal digits), if @var{nnn} is
11492 a nine-bit value, the ninth bit is ignored
11494 the eight-bit value that is the octal number @var{nnn}
11495 (one to three octal digits), if @var{nnn} is
11496 a nine-bit value, the ninth bit is ignored
11498 the eight-bit value that is the hexadecimal number @var{hh}
11499 (one or two hexadecimal digits)
11504 @cindex backslash escapes
11505 Disable interpretation of backslash escapes in each @var{string}.
11506 This is the default. If @option{-e} and @option{-E} are both
11507 specified, the last one given takes effect.
11511 @vindex POSIXLY_CORRECT
11512 If the @env{POSIXLY_CORRECT} environment variable is set, then when
11513 @command{echo}'s first argument is not @option{-n} it outputs
11514 option-like arguments instead of treating them as options. For
11515 example, @code{echo -ne hello} outputs @samp{-ne hello} instead of
11516 plain @samp{hello}.
11518 POSIX does not require support for any options, and says
11519 that the behavior of @command{echo} is implementation-defined if any
11520 @var{string} contains a backslash or if the first argument is
11521 @option{-n}. Portable programs can use the @command{printf} command
11522 if they need to omit trailing newlines or output control characters or
11523 backslashes. @xref{printf invocation}.
11528 @node printf invocation
11529 @section @command{printf}: Format and print data
11532 @command{printf} does formatted printing of text. Synopsis:
11535 printf @var{format} [@var{argument}]@dots{}
11538 @command{printf} prints the @var{format} string, interpreting @samp{%}
11539 directives and @samp{\} escapes to format numeric and string arguments
11540 in a way that is mostly similar to the C @samp{printf} function.
11541 @xref{Output Conversion Syntax,, @command{printf} format directives,
11542 libc, The GNU C Library Reference Manual}, for details.
11543 The differences are listed below.
11545 @mayConflictWithShellBuiltIn{printf}
11550 The @var{format} argument is reused as necessary to convert all the
11551 given @var{argument}s. For example, the command @samp{printf %s a b}
11555 Missing @var{argument}s are treated as null strings or as zeros,
11556 depending on whether the context expects a string or a number. For
11557 example, the command @samp{printf %sx%d} prints @samp{x0}.
11561 An additional escape, @samp{\c}, causes @command{printf} to produce no
11562 further output. For example, the command @samp{printf 'A%sC\cD%sF' B
11563 E} prints @samp{ABC}.
11566 The hexadecimal escape sequence @samp{\x@var{hh}} has at most two
11567 digits, as opposed to C where it can have an unlimited number of
11568 digits. For example, the command @samp{printf '\x07e'} prints two
11569 bytes, whereas the C statement @samp{printf ("\x07e")} prints just
11574 @command{printf} has an additional directive, @samp{%b}, which prints its
11575 argument string with @samp{\} escapes interpreted in the same way as in
11576 the @var{format} string, except that octal escapes are of the form
11577 @samp{\0@var{ooo}} where @var{ooo} is 0 to 3 octal digits. If
11578 @samp{\@var{ooo}} is nine-bit value, ignore the ninth bit.
11579 If a precision is also given, it limits the number of bytes printed
11580 from the converted string.
11583 Numeric arguments must be single C constants, possibly with leading
11584 @samp{+} or @samp{-}. For example, @samp{printf %.4d -3} outputs
11588 @vindex POSIXLY_CORRECT
11589 If the leading character of a numeric argument is @samp{"} or @samp{'}
11590 then its value is the numeric value of the immediately following
11591 character. Any remaining characters are silently ignored if the
11592 @env{POSIXLY_CORRECT} environment variable is set; otherwise, a
11593 warning is printed. For example, @samp{printf "%d" "'a"} outputs
11594 @samp{97} on hosts that use the ASCII character set, since
11595 @samp{a} has the numeric value 97 in ASCII.
11600 A floating-point argument must use a period before any fractional
11601 digits, but is printed according to the @env{LC_NUMERIC} category of the
11602 current locale. For example, in a locale whose radix character is a
11603 comma, the command @samp{printf %g 3.14} outputs @samp{3,14} whereas
11604 the command @samp{printf %g 3,14} is an error.
11605 @xref{Floating point}.
11609 @command{printf} interprets @samp{\@var{ooo}} in @var{format} as an octal number
11610 (if @var{ooo} is 1 to 3 octal digits) specifying a byte to print,
11611 and @samp{\x@var{hh}} as a hexadecimal number (if @var{hh} is 1 to 2 hex
11612 digits) specifying a character to print.
11613 Note however that when @samp{\@var{ooo}} specifies a number larger than 255,
11614 @command{printf} ignores the ninth bit.
11615 For example, @samp{printf '\400'} is equivalent to @samp{printf '\0'}.
11620 @cindex ISO/IEC 10646
11622 @command{printf} interprets two character syntaxes introduced in
11624 @samp{\u} for 16-bit Unicode (ISO/IEC 10646)
11625 characters, specified as
11626 four hexadecimal digits @var{hhhh}, and @samp{\U} for 32-bit Unicode
11627 characters, specified as eight hexadecimal digits @var{hhhhhhhh}.
11628 @command{printf} outputs the Unicode characters
11629 according to the @env{LC_CTYPE} locale. Unicode characters in the ranges
11630 U+0000...U+009F, U+D800...U+DFFF cannot be specified by this syntax, except
11631 for U+0024 ($), U+0040 (@@), and U+0060 (@`).
11633 The processing of @samp{\u} and @samp{\U} requires a full-featured
11634 @code{iconv} facility. It is activated on systems with glibc 2.2 (or newer),
11635 or when @code{libiconv} is installed prior to this package. Otherwise
11636 @samp{\u} and @samp{\U} will print as-is.
11638 The only options are a lone @option{--help} or
11639 @option{--version}. @xref{Common options}.
11640 Options must precede operands.
11642 The Unicode character syntaxes are useful for writing strings in a locale
11643 independent way. For example, a string containing the Euro currency symbol
11646 $ env printf '\u20AC 14.95'
11650 will be output correctly in all locales supporting the Euro symbol
11651 (ISO-8859-15, UTF-8, and others). Similarly, a Chinese string
11654 $ env printf '\u4e2d\u6587'
11658 will be output correctly in all Chinese locales (GB2312, BIG5, UTF-8, etc).
11660 Note that in these examples, the @command{printf} command has been
11661 invoked via @command{env} to ensure that we run the program found via
11662 your shell's search path, and not a shell alias or a built-in function.
11664 For larger strings, you don't need to look up the hexadecimal code
11665 values of each character one by one. ASCII characters mixed with \u
11666 escape sequences is also known as the JAVA source file encoding. You can
11667 use GNU recode 3.5c (or newer) to convert strings to this encoding. Here
11668 is how to convert a piece of text into a shell script which will output
11669 this text in a locale-independent way:
11672 $ LC_CTYPE=zh_CN.big5 /usr/local/bin/printf \
11673 '\u4e2d\u6587\n' > sample.txt
11674 $ recode BIG5..JAVA < sample.txt \
11675 | sed -e "s|^|/usr/local/bin/printf '|" -e "s|$|\\\\n'|" \
11682 @node yes invocation
11683 @section @command{yes}: Print a string until interrupted
11686 @cindex repeated output of a string
11688 @command{yes} prints the command line arguments, separated by spaces and
11689 followed by a newline, forever until it is killed. If no arguments are
11690 given, it prints @samp{y} followed by a newline forever until killed.
11692 Upon a write error, @command{yes} exits with status @samp{1}.
11694 The only options are a lone @option{--help} or @option{--version}.
11695 To output an argument that begins with
11696 @samp{-}, precede it with @option{--}, e.g., @samp{yes -- --help}.
11697 @xref{Common options}.
11701 @chapter Conditions
11704 @cindex commands for exit status
11705 @cindex exit status commands
11707 This section describes commands that are primarily useful for their exit
11708 status, rather than their output. Thus, they are often used as the
11709 condition of shell @code{if} statements, or as the last command in a
11713 * false invocation:: Do nothing, unsuccessfully.
11714 * true invocation:: Do nothing, successfully.
11715 * test invocation:: Check file types and compare values.
11716 * expr invocation:: Evaluate expressions.
11720 @node false invocation
11721 @section @command{false}: Do nothing, unsuccessfully
11724 @cindex do nothing, unsuccessfully
11725 @cindex failure exit status
11726 @cindex exit status of @command{false}
11728 @command{false} does nothing except return an exit status of 1, meaning
11729 @dfn{failure}. It can be used as a place holder in shell scripts
11730 where an unsuccessful command is needed.
11731 In most modern shells, @command{false} is a built-in command, so when
11732 you use @samp{false} in a script, you're probably using the built-in
11733 command, not the one documented here.
11735 @command{false} honors the @option{--help} and @option{--version} options.
11737 This version of @command{false} is implemented as a C program, and is thus
11738 more secure and faster than a shell script implementation, and may safely
11739 be used as a dummy shell for the purpose of disabling accounts.
11741 Note that @command{false} (unlike all other programs documented herein)
11742 exits unsuccessfully, even when invoked with
11743 @option{--help} or @option{--version}.
11745 Portable programs should not assume that the exit status of
11746 @command{false} is 1, as it is greater than 1 on some
11750 @node true invocation
11751 @section @command{true}: Do nothing, successfully
11754 @cindex do nothing, successfully
11756 @cindex successful exit
11757 @cindex exit status of @command{true}
11759 @command{true} does nothing except return an exit status of 0, meaning
11760 @dfn{success}. It can be used as a place holder in shell scripts
11761 where a successful command is needed, although the shell built-in
11762 command @code{:} (colon) may do the same thing faster.
11763 In most modern shells, @command{true} is a built-in command, so when
11764 you use @samp{true} in a script, you're probably using the built-in
11765 command, not the one documented here.
11767 @command{true} honors the @option{--help} and @option{--version} options.
11769 Note, however, that it is possible to cause @command{true}
11770 to exit with nonzero status: with the @option{--help} or @option{--version}
11771 option, and with standard
11772 output already closed or redirected to a file that evokes an I/O error.
11773 For example, using a Bourne-compatible shell:
11776 $ ./true --version >&-
11777 ./true: write error: Bad file number
11778 $ ./true --version > /dev/full
11779 ./true: write error: No space left on device
11782 This version of @command{true} is implemented as a C program, and is thus
11783 more secure and faster than a shell script implementation, and may safely
11784 be used as a dummy shell for the purpose of disabling accounts.
11786 @node test invocation
11787 @section @command{test}: Check file types and compare values
11790 @cindex check file types
11791 @cindex compare values
11792 @cindex expression evaluation
11794 @command{test} returns a status of 0 (true) or 1 (false) depending on the
11795 evaluation of the conditional expression @var{expr}. Each part of the
11796 expression must be a separate argument.
11798 @command{test} has file status checks, string operators, and numeric
11799 comparison operators.
11801 @command{test} has an alternate form that uses opening and closing
11802 square brackets instead a leading @samp{test}. For example, instead
11803 of @samp{test -d /}, you can write @samp{[ -d / ]}. The square
11804 brackets must be separate arguments; for example, @samp{[-d /]} does
11805 not have the desired effect. Since @samp{test @var{expr}} and @samp{[
11806 @var{expr} ]} have the same meaning, only the former form is discussed
11812 test @var{expression}
11814 [ @var{expression} ]
11819 @mayConflictWithShellBuiltIn{test}
11821 If @var{expression} is omitted, @command{test} returns false.
11822 If @var{expression} is a single argument,
11823 @command{test} returns false if the argument is null and true
11824 otherwise. The argument
11825 can be any string, including strings like @samp{-d}, @samp{-1},
11826 @samp{--}, @samp{--help}, and @samp{--version} that most other
11827 programs would treat as options. To get help and version information,
11828 invoke the commands @samp{[ --help} and @samp{[ --version}, without
11829 the usual closing brackets. @xref{Common options}.
11831 @cindex exit status of @command{test}
11835 0 if the expression is true,
11836 1 if the expression is false,
11837 2 if an error occurred.
11841 * File type tests:: -[bcdfhLpSt]
11842 * Access permission tests:: -[gkruwxOG]
11843 * File characteristic tests:: -e -s -nt -ot -ef
11844 * String tests:: -z -n = == !=
11845 * Numeric tests:: -eq -ne -lt -le -gt -ge
11846 * Connectives for test:: ! -a -o
11850 @node File type tests
11851 @subsection File type tests
11853 @cindex file type tests
11855 These options test for particular types of files. (Everything's a file,
11856 but not all files are the same!)
11860 @item -b @var{file}
11862 @cindex block special check
11863 True if @var{file} exists and is a block special device.
11865 @item -c @var{file}
11867 @cindex character special check
11868 True if @var{file} exists and is a character special device.
11870 @item -d @var{file}
11872 @cindex directory check
11873 True if @var{file} exists and is a directory.
11875 @item -f @var{file}
11877 @cindex regular file check
11878 True if @var{file} exists and is a regular file.
11880 @item -h @var{file}
11881 @itemx -L @var{file}
11884 @cindex symbolic link check
11885 True if @var{file} exists and is a symbolic link.
11886 Unlike all other file-related tests, this test does not dereference
11887 @var{file} if it is a symbolic link.
11889 @item -p @var{file}
11891 @cindex named pipe check
11892 True if @var{file} exists and is a named pipe.
11894 @item -S @var{file}
11896 @cindex socket check
11897 True if @var{file} exists and is a socket.
11901 @cindex terminal check
11902 True if @var{fd} is a file descriptor that is associated with a
11908 @node Access permission tests
11909 @subsection Access permission tests
11911 @cindex access permission tests
11912 @cindex permission tests
11914 These options test for particular access permissions.
11918 @item -g @var{file}
11920 @cindex set-group-ID check
11921 True if @var{file} exists and has its set-group-ID bit set.
11923 @item -k @var{file}
11925 @cindex sticky bit check
11926 True if @var{file} exists and has its @dfn{sticky} bit set.
11928 @item -r @var{file}
11930 @cindex readable file check
11931 True if @var{file} exists and read permission is granted.
11933 @item -u @var{file}
11935 @cindex set-user-ID check
11936 True if @var{file} exists and has its set-user-ID bit set.
11938 @item -w @var{file}
11940 @cindex writable file check
11941 True if @var{file} exists and write permission is granted.
11943 @item -x @var{file}
11945 @cindex executable file check
11946 True if @var{file} exists and execute permission is granted
11947 (or search permission, if it is a directory).
11949 @item -O @var{file}
11951 @cindex owned by effective user ID check
11952 True if @var{file} exists and is owned by the current effective user ID.
11954 @item -G @var{file}
11956 @cindex owned by effective group ID check
11957 True if @var{file} exists and is owned by the current effective group ID.
11961 @node File characteristic tests
11962 @subsection File characteristic tests
11964 @cindex file characteristic tests
11966 These options test other file characteristics.
11970 @item -e @var{file}
11972 @cindex existence-of-file check
11973 True if @var{file} exists.
11975 @item -s @var{file}
11977 @cindex nonempty file check
11978 True if @var{file} exists and has a size greater than zero.
11980 @item @var{file1} -nt @var{file2}
11982 @cindex newer-than file check
11983 True if @var{file1} is newer (according to modification date) than
11984 @var{file2}, or if @var{file1} exists and @var{file2} does not.
11986 @item @var{file1} -ot @var{file2}
11988 @cindex older-than file check
11989 True if @var{file1} is older (according to modification date) than
11990 @var{file2}, or if @var{file2} exists and @var{file1} does not.
11992 @item @var{file1} -ef @var{file2}
11994 @cindex same file check
11995 @cindex hard link check
11996 True if @var{file1} and @var{file2} have the same device and inode
11997 numbers, i.e., if they are hard links to each other.
12003 @subsection String tests
12005 @cindex string tests
12007 These options test string characteristics. You may need to quote
12008 @var{string} arguments for the shell. For example:
12014 The quotes here prevent the wrong arguments from being passed to
12015 @command{test} if @samp{$V} is empty or contains special characters.
12019 @item -z @var{string}
12021 @cindex zero-length string check
12022 True if the length of @var{string} is zero.
12024 @item -n @var{string}
12025 @itemx @var{string}
12027 @cindex nonzero-length string check
12028 True if the length of @var{string} is nonzero.
12030 @item @var{string1} = @var{string2}
12032 @cindex equal string check
12033 True if the strings are equal.
12035 @item @var{string1} == @var{string2}
12037 @cindex equal string check
12038 True if the strings are equal (synonym for =).
12040 @item @var{string1} != @var{string2}
12042 @cindex not-equal string check
12043 True if the strings are not equal.
12048 @node Numeric tests
12049 @subsection Numeric tests
12051 @cindex numeric tests
12052 @cindex arithmetic tests
12054 Numeric relational operators. The arguments must be entirely numeric
12055 (possibly negative), or the special expression @w{@code{-l @var{string}}},
12056 which evaluates to the length of @var{string}.
12060 @item @var{arg1} -eq @var{arg2}
12061 @itemx @var{arg1} -ne @var{arg2}
12062 @itemx @var{arg1} -lt @var{arg2}
12063 @itemx @var{arg1} -le @var{arg2}
12064 @itemx @var{arg1} -gt @var{arg2}
12065 @itemx @var{arg1} -ge @var{arg2}
12072 These arithmetic binary operators return true if @var{arg1} is equal,
12073 not-equal, less-than, less-than-or-equal, greater-than, or
12074 greater-than-or-equal than @var{arg2}, respectively.
12081 test -1 -gt -2 && echo yes
12083 test -l abc -gt 1 && echo yes
12086 @error{} test: integer expression expected before -eq
12090 @node Connectives for test
12091 @subsection Connectives for @command{test}
12093 @cindex logical connectives
12094 @cindex connectives, logical
12096 The usual logical connectives.
12102 True if @var{expr} is false.
12104 @item @var{expr1} -a @var{expr2}
12106 @cindex logical and operator
12107 @cindex and operator
12108 True if both @var{expr1} and @var{expr2} are true.
12110 @item @var{expr1} -o @var{expr2}
12112 @cindex logical or operator
12113 @cindex or operator
12114 True if either @var{expr1} or @var{expr2} is true.
12119 @node expr invocation
12120 @section @command{expr}: Evaluate expressions
12123 @cindex expression evaluation
12124 @cindex evaluation of expressions
12126 @command{expr} evaluates an expression and writes the result on standard
12127 output. Each token of the expression must be a separate argument.
12129 Operands are either integers or strings. Integers consist of one or
12130 more decimal digits, with an optional leading @samp{-}.
12131 @command{expr} converts
12132 anything appearing in an operand position to an integer or a string
12133 depending on the operation being applied to it.
12135 Strings are not quoted for @command{expr} itself, though you may need to
12136 quote them to protect characters with special meaning to the shell,
12137 e.g., spaces. However, regardless of whether it is quoted, a string
12138 operand should not be a parenthesis or any of @command{expr}'s
12139 operators like @code{+}, so you cannot safely pass an arbitrary string
12140 @code{$str} to expr merely by quoting it to the shell. One way to
12141 work around this is to use the @sc{gnu} extension @code{+},
12142 (e.g., @code{+ "$str" = foo}); a more portable way is to use
12143 @code{@w{" $str"}} and to adjust the rest of the expression to take
12144 the leading space into account (e.g., @code{@w{" $str" = " foo"}}).
12146 You should not pass a negative integer or a string with leading
12147 @samp{-} as @command{expr}'s first argument, as it might be
12148 misinterpreted as an option; this can be avoided by parenthesization.
12149 Also, portable scripts should not use a string operand that happens to
12150 take the form of an integer; this can be worked around by inserting
12151 leading spaces as mentioned above.
12153 @cindex parentheses for grouping
12154 Operators may be given as infix symbols or prefix keywords. Parentheses
12155 may be used for grouping in the usual manner. You must quote
12156 parentheses and many operators to avoid the shell evaluating them,
12159 When built with support for the GNU MP library, @command{expr} uses
12160 arbitrary-precision arithmetic; otherwise, it uses native arithmetic
12161 types and may fail due to arithmetic overflow.
12163 The only options are @option{--help} and @option{--version}. @xref{Common
12164 options}. Options must precede operands.
12166 @cindex exit status of @command{expr}
12170 0 if the expression is neither null nor 0,
12171 1 if the expression is null or 0,
12172 2 if the expression is invalid,
12173 3 if an internal error occurred (e.g., arithmetic overflow).
12177 * String expressions:: + : match substr index length
12178 * Numeric expressions:: + - * / %
12179 * Relations for expr:: | & < <= = == != >= >
12180 * Examples of expr:: Examples.
12184 @node String expressions
12185 @subsection String expressions
12187 @cindex string expressions
12188 @cindex expressions, string
12190 @command{expr} supports pattern matching and other string operators. These
12191 have higher precedence than both the numeric and relational operators (in
12192 the next sections).
12196 @item @var{string} : @var{regex}
12197 @cindex pattern matching
12198 @cindex regular expression matching
12199 @cindex matching patterns
12200 Perform pattern matching. The arguments are converted to strings and the
12201 second is considered to be a (basic, a la GNU @code{grep}) regular
12202 expression, with a @code{^} implicitly prepended. The first argument is
12203 then matched against this regular expression.
12205 If the match succeeds and @var{regex} uses @samp{\(} and @samp{\)}, the
12206 @code{:} expression returns the part of @var{string} that matched the
12207 subexpression; otherwise, it returns the number of characters matched.
12209 If the match fails, the @code{:} operator returns the null string if
12210 @samp{\(} and @samp{\)} are used in @var{regex}, otherwise 0.
12212 @kindex \( @r{regexp operator}
12213 Only the first @samp{\( @dots{} \)} pair is relevant to the return
12214 value; additional pairs are meaningful only for grouping the regular
12215 expression operators.
12217 @kindex \+ @r{regexp operator}
12218 @kindex \? @r{regexp operator}
12219 @kindex \| @r{regexp operator}
12220 In the regular expression, @code{\+}, @code{\?}, and @code{\|} are
12221 operators which respectively match one or more, zero or one, or separate
12222 alternatives. SunOS and other @command{expr}'s treat these as regular
12223 characters. (POSIX allows either behavior.)
12224 @xref{Top, , Regular Expression Library, regex, Regex}, for details of
12225 regular expression syntax. Some examples are in @ref{Examples of expr}.
12227 @item match @var{string} @var{regex}
12229 An alternative way to do pattern matching. This is the same as
12230 @w{@samp{@var{string} : @var{regex}}}.
12232 @item substr @var{string} @var{position} @var{length}
12234 Returns the substring of @var{string} beginning at @var{position}
12235 with length at most @var{length}. If either @var{position} or
12236 @var{length} is negative, zero, or non-numeric, returns the null string.
12238 @item index @var{string} @var{charset}
12240 Returns the first position in @var{string} where the first character in
12241 @var{charset} was found. If no character in @var{charset} is found in
12242 @var{string}, return 0.
12244 @item length @var{string}
12246 Returns the length of @var{string}.
12248 @item + @var{token}
12250 Interpret @var{token} as a string, even if it is a keyword like @var{match}
12251 or an operator like @code{/}.
12252 This makes it possible to test @code{expr length + "$x"} or
12253 @code{expr + "$x" : '.*/\(.\)'} and have it do the right thing even if
12254 the value of @var{$x} happens to be (for example) @code{/} or @code{index}.
12255 This operator is a GNU extension. Portable shell scripts should use
12256 @code{@w{" $token"} : @w{' \(.*\)'}} instead of @code{+ "$token"}.
12260 To make @command{expr} interpret keywords as strings, you must use the
12261 @code{quote} operator.
12264 @node Numeric expressions
12265 @subsection Numeric expressions
12267 @cindex numeric expressions
12268 @cindex expressions, numeric
12270 @command{expr} supports the usual numeric operators, in order of increasing
12271 precedence. These numeric operators have lower precedence than the
12272 string operators described in the previous section, and higher precedence
12273 than the connectives (next section).
12281 @cindex subtraction
12282 Addition and subtraction. Both arguments are converted to integers;
12283 an error occurs if this cannot be done.
12289 @cindex multiplication
12292 Multiplication, division, remainder. Both arguments are converted to
12293 integers; an error occurs if this cannot be done.
12298 @node Relations for expr
12299 @subsection Relations for @command{expr}
12301 @cindex connectives, logical
12302 @cindex logical connectives
12303 @cindex relations, numeric or string
12305 @command{expr} supports the usual logical connectives and relations. These
12306 have lower precedence than the string and numeric operators
12307 (previous sections). Here is the list, lowest-precedence operator first.
12313 @cindex logical or operator
12314 @cindex or operator
12315 Returns its first argument if that is neither null nor zero, otherwise
12316 its second argument if it is neither null nor zero, otherwise 0. It
12317 does not evaluate its second argument if its first argument is neither
12322 @cindex logical and operator
12323 @cindex and operator
12324 Return its first argument if neither argument is null or zero, otherwise
12325 0. It does not evaluate its second argument if its first argument is
12328 @item < <= = == != >= >
12335 @cindex comparison operators
12337 Compare the arguments and return 1 if the relation is true, 0 otherwise.
12338 @code{==} is a synonym for @code{=}. @command{expr} first tries to convert
12339 both arguments to integers and do a numeric comparison; if either
12340 conversion fails, it does a lexicographic comparison using the character
12341 collating sequence specified by the @env{LC_COLLATE} locale.
12346 @node Examples of expr
12347 @subsection Examples of using @command{expr}
12349 @cindex examples of @command{expr}
12350 Here are a few examples, including quoting for shell metacharacters.
12352 To add 1 to the shell variable @code{foo}, in Bourne-compatible shells:
12355 foo=$(expr $foo + 1)
12358 To print the non-directory part of the file name stored in
12359 @code{$fname}, which need not contain a @code{/}:
12362 expr $fname : '.*/\(.*\)' '|' $fname
12365 An example showing that @code{\+} is an operator:
12373 expr abc : 'a\(.\)c'
12375 expr index abcdef cz
12378 @error{} expr: syntax error
12379 expr index + index a
12385 @chapter Redirection
12387 @cindex redirection
12388 @cindex commands for redirection
12390 Unix shells commonly provide several forms of @dfn{redirection}---ways
12391 to change the input source or output destination of a command. But one
12392 useful redirection is performed by a separate command, not by the shell;
12393 it's described here.
12396 * tee invocation:: Redirect output to multiple files or processes.
12400 @node tee invocation
12401 @section @command{tee}: Redirect output to multiple files or processes
12404 @cindex pipe fitting
12405 @cindex destinations, multiple output
12406 @cindex read from stdin and write to stdout and files
12408 The @command{tee} command copies standard input to standard output and also
12409 to any files given as arguments. This is useful when you want not only
12410 to send some data down a pipe, but also to save a copy. Synopsis:
12413 tee [@var{option}]@dots{} [@var{file}]@dots{}
12416 If a file being written to does not already exist, it is created. If a
12417 file being written to already exists, the data it previously contained
12418 is overwritten unless the @option{-a} option is used.
12420 A @var{file} of @samp{-} causes @command{tee} to send another copy of
12421 input to standard output, but this is typically not that useful as the
12422 copies are interleaved.
12424 The program accepts the following options. Also see @ref{Common options}.
12431 Append standard input to the given files rather than overwriting
12435 @itemx --ignore-interrupts
12437 @opindex --ignore-interrupts
12438 Ignore interrupt signals.
12442 The @command{tee} command is useful when you happen to be transferring a large
12443 amount of data and also want to summarize that data without reading
12444 it a second time. For example, when you are downloading a DVD image,
12445 you often want to verify its signature or checksum right away.
12446 The inefficient way to do it is simply:
12449 wget http://example.com/some.iso && sha1sum some.iso
12452 One problem with the above is that it makes you wait for the
12453 download to complete before starting the time-consuming SHA1 computation.
12454 Perhaps even more importantly, the above requires reading
12455 the DVD image a second time (the first was from the network).
12457 The efficient way to do it is to interleave the download
12458 and SHA1 computation. Then, you'll get the checksum for
12459 free, because the entire process parallelizes so well:
12462 # slightly contrived, to demonstrate process substitution
12463 wget -O - http://example.com/dvd.iso \
12464 | tee >(sha1sum > dvd.sha1) > dvd.iso
12467 That makes @command{tee} write not just to the expected output file,
12468 but also to a pipe running @command{sha1sum} and saving the final
12469 checksum in a file named @file{dvd.sha1}.
12471 Note, however, that this example relies on a feature of modern shells
12472 called @dfn{process substitution}
12473 (the @samp{>(command)} syntax, above;
12474 @xref{Process Substitution,,Process Substitution, bash,
12475 The Bash Reference Manual}.),
12476 so it works with @command{zsh}, @command{bash}, and @command{ksh},
12477 but not with @command{/bin/sh}. So if you write code like this
12478 in a shell script, be sure to start the script with @samp{#!/bin/bash}.
12480 Since the above example writes to one file and one process,
12481 a more conventional and portable use of @command{tee} is even better:
12484 wget -O - http://example.com/dvd.iso \
12485 | tee dvd.iso | sha1sum > dvd.sha1
12488 You can extend this example to make @command{tee} write to two processes,
12489 computing MD5 and SHA1 checksums in parallel. In this case,
12490 process substitution is required:
12493 wget -O - http://example.com/dvd.iso \
12494 | tee >(sha1sum > dvd.sha1) \
12495 >(md5sum > dvd.md5) \
12499 This technique is also useful when you want to make a @emph{compressed}
12500 copy of the contents of a pipe.
12501 Consider a tool to graphically summarize disk usage data from @samp{du -ak}.
12502 For a large hierarchy, @samp{du -ak} can run for a long time,
12503 and can easily produce terabytes of data, so you won't want to
12504 rerun the command unnecessarily. Nor will you want to save
12505 the uncompressed output.
12507 Doing it the inefficient way, you can't even start the GUI
12508 until after you've compressed all of the @command{du} output:
12511 du -ak | gzip -9 > /tmp/du.gz
12512 gzip -d /tmp/du.gz | xdiskusage -a
12515 With @command{tee} and process substitution, you start the GUI
12516 right away and eliminate the decompression completely:
12519 du -ak | tee >(gzip -9 > /tmp/du.gz) | xdiskusage -a
12522 Finally, if you regularly create more than one type of
12523 compressed tarball at once, for example when @code{make dist} creates
12524 both @command{gzip}-compressed and @command{bzip2}-compressed tarballs,
12525 there may be a better way.
12526 Typical @command{automake}-generated @file{Makefile} rules create
12527 the two compressed tar archives with commands in sequence, like this
12528 (slightly simplified):
12531 tardir=your-pkg-M.N
12532 tar chof - "$tardir" | gzip -9 -c > your-pkg-M.N.tar.gz
12533 tar chof - "$tardir" | bzip2 -9 -c > your-pkg-M.N.tar.bz2
12536 However, if the hierarchy you are archiving and compressing is larger
12537 than a couple megabytes, and especially if you are using a multi-processor
12538 system with plenty of memory, then you can do much better by reading the
12539 directory contents only once and running the compression programs in parallel:
12542 tardir=your-pkg-M.N
12543 tar chof - "$tardir" \
12544 | tee >(gzip -9 -c > your-pkg-M.N.tar.gz) \
12545 | bzip2 -9 -c > your-pkg-M.N.tar.bz2
12551 @node File name manipulation
12552 @chapter File name manipulation
12554 @cindex file name manipulation
12555 @cindex manipulation of file names
12556 @cindex commands for file name manipulation
12558 This section describes commands that manipulate file names.
12561 * basename invocation:: Strip directory and suffix from a file name.
12562 * dirname invocation:: Strip last file name component.
12563 * pathchk invocation:: Check file name validity and portability.
12564 * mktemp invocation:: Create temporary file or directory.
12565 * realpath invocation:: Print resolved file names.
12569 @node basename invocation
12570 @section @command{basename}: Strip directory and suffix from a file name
12573 @cindex strip directory and suffix from file names
12574 @cindex directory, stripping from file names
12575 @cindex suffix, stripping from file names
12576 @cindex file names, stripping directory and suffix
12577 @cindex leading directory components, stripping
12579 @command{basename} removes any leading directory components from
12580 @var{name}. Synopsis:
12583 basename @var{name} [@var{suffix}]
12584 basename @var{option}... @var{name}...
12587 If @var{suffix} is specified and is identical to the end of @var{name},
12588 it is removed from @var{name} as well. Note that since trailing slashes
12589 are removed prior to suffix matching, @var{suffix} will do nothing if it
12590 contains slashes. @command{basename} prints the result on standard
12593 @c This test is used both here and in the section on dirname.
12594 @macro basenameAndDirname
12595 Together, @command{basename} and @command{dirname} are designed such
12596 that if @samp{ls "$name"} succeeds, then the command sequence @samp{cd
12597 "$(dirname "$name")"; ls "$(basename "$name")"} will, too. This works
12598 for everything except file names containing a trailing newline.
12600 @basenameAndDirname
12602 POSIX allows the implementation to define the results if
12603 @var{name} is empty or @samp{//}. In the former case, GNU
12604 @command{basename} returns the empty string. In the latter case, the
12605 result is @samp{//} on platforms where @var{//} is distinct from
12606 @var{/}, and @samp{/} on platforms where there is no difference.
12608 The program accepts the following options. Also see @ref{Common options}.
12609 Options must precede operands.
12616 @opindex --multiple
12617 Support more than one argument. Treat every argument as a @var{name}.
12618 With this, an optional @var{suffix} must be specified using the
12619 @option{-s} option.
12621 @item -s @var{suffix}
12622 @itemx --suffix=@var{suffix}
12625 Remove a trailing @var{suffix}.
12626 This option implies the @option{-a} option.
12632 Separate output items with @sc{nul} characters.
12642 basename /usr/bin/sort
12645 basename include/stdio.h .h
12648 basename -s .h include/stdio.h
12650 # Output "stdio" followed by "stdlib"
12651 basename -a -s .h include/stdio.h include/stdlib.h
12655 @node dirname invocation
12656 @section @command{dirname}: Strip last file name component
12659 @cindex directory components, printing
12660 @cindex stripping non-directory suffix
12661 @cindex non-directory suffix, stripping
12663 @command{dirname} prints all but the final slash-delimited component
12664 of each @var{name}. Slashes on either side of the final component are
12665 also removed. If the string contains no slash, @command{dirname}
12666 prints @samp{.} (meaning the current directory). Synopsis:
12669 dirname [@var{option}] @var{name}...
12672 @var{name} need not be a file name, but if it is, this operation
12673 effectively lists the directory that contains the final component,
12674 including the case when the final component is itself a directory.
12676 @basenameAndDirname
12678 POSIX allows the implementation to define the results if
12679 @var{name} is @samp{//}. With GNU @command{dirname}, the
12680 result is @samp{//} on platforms where @var{//} is distinct from
12681 @var{/}, and @samp{/} on platforms where there is no difference.
12683 The program accepts the following option. Also see @ref{Common options}.
12691 Separate output items with @sc{nul} characters.
12700 # Output "/usr/bin".
12701 dirname /usr/bin/sort
12702 dirname /usr/bin//.//
12704 # Output "dir1" followed by "dir2"
12705 dirname dir1/str dir2/str
12712 @node pathchk invocation
12713 @section @command{pathchk}: Check file name validity and portability
12716 @cindex file names, checking validity and portability
12717 @cindex valid file names, checking for
12718 @cindex portable file names, checking for
12720 @command{pathchk} checks validity and portability of file names. Synopsis:
12723 pathchk [@var{option}]@dots{} @var{name}@dots{}
12726 For each @var{name}, @command{pathchk} prints an error message if any of
12727 these conditions is true:
12731 One of the existing directories in @var{name} does not have search
12732 (execute) permission,
12734 The length of @var{name} is larger than the maximum supported by the
12737 The length of one component of @var{name} is longer than
12738 its file system's maximum.
12741 A nonexistent @var{name} is not an error, so long a file with that
12742 name could be created under the above conditions.
12744 The program accepts the following options. Also see @ref{Common options}.
12745 Options must precede operands.
12751 Instead of performing checks based on the underlying file system,
12752 print an error message if any of these conditions is true:
12756 A file name is empty.
12759 A file name contains a character outside the POSIX portable file
12760 name character set, namely, the ASCII letters and digits, @samp{.},
12761 @samp{_}, @samp{-}, and @samp{/}.
12764 The length of a file name or one of its components exceeds the
12765 POSIX minimum limits for portability.
12770 Print an error message if a file name is empty, or if it contains a component
12771 that begins with @samp{-}.
12773 @item --portability
12774 @opindex --portability
12775 Print an error message if a file name is not portable to all POSIX
12776 hosts. This option is equivalent to @samp{-p -P}.
12780 @cindex exit status of @command{pathchk}
12784 0 if all specified file names passed all checks,
12788 @node mktemp invocation
12789 @section @command{mktemp}: Create temporary file or directory
12792 @cindex file names, creating temporary
12793 @cindex directory, creating temporary
12794 @cindex temporary files and directories
12796 @command{mktemp} manages the creation of temporary files and
12797 directories. Synopsis:
12800 mktemp [@var{option}]@dots{} [@var{template}]
12803 Safely create a temporary file or directory based on @var{template},
12804 and print its name. If given, @var{template} must include at least
12805 three consecutive @samp{X}s in the last component. If omitted, the template
12806 @samp{tmp.XXXXXXXXXX} is used, and option @option{--tmpdir} is
12807 implied. The final run of @samp{X}s in the @var{template} will be replaced
12808 by alpha-numeric characters; thus, on a case-sensitive file system,
12809 and with a @var{template} including a run of @var{n} instances of @samp{X},
12810 there are @samp{62**@var{n}} potential file names.
12812 Older scripts used to create temporary files by simply joining the
12813 name of the program with the process id (@samp{$$}) as a suffix.
12814 However, that naming scheme is easily predictable, and suffers from a
12815 race condition where the attacker can create an appropriately named
12816 symbolic link, such that when the script then opens a handle to what
12817 it thought was an unused file, it is instead modifying an existing
12818 file. Using the same scheme to create a directory is slightly safer,
12819 since the @command{mkdir} will fail if the target already exists, but
12820 it is still inferior because it allows for denial of service attacks.
12821 Therefore, modern scripts should use the @command{mktemp} command to
12822 guarantee that the generated name will be unpredictable, and that
12823 knowledge of the temporary file name implies that the file was created
12824 by the current script and cannot be modified by other users.
12826 When creating a file, the resulting file has read and write
12827 permissions for the current user, but no permissions for the group or
12828 others; these permissions are reduced if the current umask is more
12831 Here are some examples (although note that if you repeat them, you
12832 will most likely get different file names):
12837 Create a temporary file in the current directory.
12844 Create a temporary file with a known suffix.
12846 $ mktemp --suffix=.txt file-XXXX
12848 $ mktemp file-XXXX-XXXX.txt
12853 Create a secure fifo relative to the user's choice of @env{TMPDIR},
12854 but falling back to the current directory rather than @file{/tmp}.
12855 Note that @command{mktemp} does not create fifos, but can create a
12856 secure directory in which the fifo can live. Exit the shell if the
12857 directory or fifo could not be created.
12859 $ dir=$(mktemp -p "$@{TMPDIR:-.@}" -d dir-XXXX) || exit 1
12861 $ mkfifo "$fifo" || @{ rmdir "$dir"; exit 1; @}
12865 Create and use a temporary file if possible, but ignore failure. The
12866 file will reside in the directory named by @env{TMPDIR}, if specified,
12867 or else in @file{/tmp}.
12869 $ file=$(mktemp -q) && @{
12870 > # Safe to use $file only within this block. Use quotes,
12871 > # since $TMPDIR, and thus $file, may contain whitespace.
12872 > echo ... > "$file"
12878 Act as a semi-random character generator (it is not fully random,
12879 since it is impacted by the contents of the current directory). To
12880 avoid security holes, do not use the resulting names to create a file.
12890 The program accepts the following options. Also see @ref{Common options}.
12897 @opindex --directory
12898 Create a directory rather than a file. The directory will have read,
12899 write, and search permissions for the current user, but no permissions
12900 for the group or others; these permissions are reduced if the current
12901 umask is more restrictive.
12907 Suppress diagnostics about failure to create a file or directory. The
12908 exit status will still reflect whether a file was created.
12914 Generate a temporary name that does not name an existing file, without
12915 changing the file system contents. Using the output of this command
12916 to create a new file is inherently unsafe, as there is a window of
12917 time between generating the name and using it where another process
12918 can create an object by the same name.
12921 @itemx --tmpdir[=@var{dir}]
12924 Treat @var{template} relative to the directory @var{dir}. If
12925 @var{dir} is not specified (only possible with the long option
12926 @option{--tmpdir}) or is the empty string, use the value of
12927 @env{TMPDIR} if available, otherwise use @samp{/tmp}. If this is
12928 specified, @var{template} must not be absolute. However,
12929 @var{template} can still contain slashes, although intermediate
12930 directories must already exist.
12932 @item --suffix=@var{suffix}
12934 Append @var{suffix} to the @var{template}. @var{suffix} must not
12935 contain slash. If @option{--suffix} is specified, @var{template} must
12936 end in @samp{X}; if it is not specified, then an appropriate
12937 @option{--suffix} is inferred by finding the last @samp{X} in
12938 @var{template}. This option exists for use with the default
12939 @var{template} and for the creation of a @var{suffix} that starts with
12944 Treat @var{template} as a single file relative to the value of
12945 @env{TMPDIR} if available, or to the directory specified by
12946 @option{-p}, otherwise to @samp{/tmp}. @var{template} must not
12947 contain slashes. This option is deprecated; the use of @option{-p}
12948 without @option{-t} offers better defaults (by favoring the command
12949 line over @env{TMPDIR}) and more flexibility (by allowing intermediate
12954 @cindex exit status of @command{mktemp}
12958 0 if the file was created,
12963 @node realpath invocation
12964 @section @command{realpath}: Print the resolved file name.
12967 @cindex file names, canonicalization
12968 @cindex symlinks, resolution
12969 @cindex canonical file name
12970 @cindex canonicalize a file name
12974 @command{realpath} expands all symbolic links and resolves references to
12975 @samp{/./}, @samp{/../} and extra @samp{/} characters. By default,
12976 all but the last component of the specified files must exist. Synopsis:
12979 realpath [@var{option}]@dots{} @var{file}@dots{}
12982 The program accepts the following options. Also see @ref{Common options}.
12987 @itemx --canonicalize-existing
12989 @opindex --canonicalize-existing
12990 Ensure that all components of the specified file names exist.
12991 If any component is missing or unavailable, @command{realpath} will output
12992 a diagnostic unless the @option{-q} option is specified, and exit with a
12993 nonzero exit code. A trailing slash requires that the name resolve to a
12997 @itemx --canonicalize-missing
12999 @opindex --canonicalize-missing
13000 If any component of a specified file name is missing or unavailable,
13001 treat it as a directory.
13007 Symbolic links are resolved in the specified file names,
13008 but they are resolved after any subsequent @samp{..} components are processed.
13013 @opindex --physical
13014 Symbolic links are resolved in the specified file names,
13015 and they are resolved before any subsequent @samp{..} components are processed.
13016 This is the default mode of operation.
13022 Suppress diagnostic messages for specified file names.
13026 @itemx --no-symlinks
13029 @opindex --no-symlinks
13030 Do not resolve symbolic links. Only resolve references to
13031 @samp{/./}, @samp{/../} and remove extra @samp{/} characters.
13032 When combined with the @option{-m} option, realpath operates
13033 only on the file name, and does not touch any actual file.
13039 Separate output items with @sc{nul} characters.
13041 @item --relative-to=@var{file}
13042 @opindex --relative-to
13044 Print the resolved file names relative to the specified file.
13045 Note this option honors the @option{-m} and @option{-e} options
13046 pertaining to file existence.
13048 @item --relative-base=@var{base}
13049 @opindex --relative-base
13050 This option is valid when used with @option{--relative-to}, and will restrict
13051 the output of @option{--relative-to} so that relative names are output,
13052 only when @var{file}s are descendants of @var{base}. Otherwise output the
13053 absolute file name. If @option{--relative-to} was not specified, then
13054 the descendants of @var{base} are printed relative to @var{base}. If
13055 @option{--relative-to} is specified, then that directory must be a
13056 descendant of @var{base} for this option to have an effect.
13057 Note: this option honors the @option{-m} and @option{-e}
13058 options pertaining to file existence. For example:
13061 realpath --relative-to=/usr /tmp /usr/bin
13064 realpath --relative-base=/usr /tmp /usr/bin
13071 @cindex exit status of @command{realpath}
13075 0 if all file names were printed without issue.
13080 @node Working context
13081 @chapter Working context
13083 @cindex working context
13084 @cindex commands for printing the working context
13086 This section describes commands that display or alter the context in
13087 which you are working: the current directory, the terminal settings, and
13088 so forth. See also the user-related commands in the next section.
13091 * pwd invocation:: Print working directory.
13092 * stty invocation:: Print or change terminal characteristics.
13093 * printenv invocation:: Print environment variables.
13094 * tty invocation:: Print file name of terminal on standard input.
13098 @node pwd invocation
13099 @section @command{pwd}: Print working directory
13102 @cindex print name of current directory
13103 @cindex current working directory, printing
13104 @cindex working directory, printing
13107 @command{pwd} prints the name of the current directory. Synopsis:
13110 pwd [@var{option}]@dots{}
13113 The program accepts the following options. Also see @ref{Common options}.
13120 If the contents of the environment variable @env{PWD} provide an
13121 absolute name of the current directory with no @samp{.} or @samp{..}
13122 components, but possibly with symbolic links, then output those
13123 contents. Otherwise, fall back to default @option{-P} handling.
13128 @opindex --physical
13129 Print a fully resolved name for the current directory. That is, all
13130 components of the printed name will be actual directory names---none
13131 will be symbolic links.
13134 @cindex symbolic links and @command{pwd}
13135 If @option{-L} and @option{-P} are both given, the last one takes
13136 precedence. If neither option is given, then this implementation uses
13137 @option{-P} as the default unless the @env{POSIXLY_CORRECT}
13138 environment variable is set.
13140 @mayConflictWithShellBuiltIn{pwd}
13145 @node stty invocation
13146 @section @command{stty}: Print or change terminal characteristics
13149 @cindex change or print terminal settings
13150 @cindex terminal settings
13151 @cindex line settings of terminal
13153 @command{stty} prints or changes terminal characteristics, such as baud rate.
13157 stty [@var{option}] [@var{setting}]@dots{}
13158 stty [@var{option}]
13161 If given no line settings, @command{stty} prints the baud rate, line
13162 discipline number (on systems that support it), and line settings
13163 that have been changed from the values set by @samp{stty sane}.
13164 By default, mode reading and setting are performed on the tty line
13165 connected to standard input, although this can be modified by the
13166 @option{--file} option.
13168 @command{stty} accepts many non-option arguments that change aspects of
13169 the terminal line operation, as described below.
13171 The program accepts the following options. Also see @ref{Common options}.
13178 Print all current settings in human-readable form. This option may not
13179 be used in combination with any line settings.
13181 @item -F @var{device}
13182 @itemx --file=@var{device}
13185 Set the line opened by the file name specified in @var{device} instead of
13186 the tty line connected to standard input. This option is necessary
13187 because opening a POSIX tty requires use of the
13188 @code{O_NONDELAY} flag to prevent a POSIX tty from blocking
13189 until the carrier detect line is high if
13190 the @code{clocal} flag is not set. Hence, it is not always possible
13191 to allow the shell to open the device in the traditional manner.
13197 @cindex machine-readable @command{stty} output
13198 Print all current settings in a form that can be used as an argument to
13199 another @command{stty} command to restore the current settings. This option
13200 may not be used in combination with any line settings.
13204 Many settings can be turned off by preceding them with a @samp{-}.
13205 Such arguments are marked below with ``May be negated'' in their
13206 description. The descriptions themselves refer to the positive
13207 case, that is, when @emph{not} negated (unless stated otherwise,
13210 Some settings are not available on all POSIX systems, since they use
13211 extensions. Such arguments are marked below with
13212 ``Non-POSIX'' in their description. On non-POSIX
13213 systems, those or other settings also may not
13214 be available, but it's not feasible to document all the variations: just
13220 * Control:: Control settings
13221 * Input:: Input settings
13222 * Output:: Output settings
13223 * Local:: Local settings
13224 * Combination:: Combination settings
13225 * Characters:: Special characters
13226 * Special:: Special settings
13231 @subsection Control settings
13233 @cindex control settings
13239 @cindex two-way parity
13240 Generate parity bit in output and expect parity bit in input.
13246 @cindex even parity
13247 Set odd parity (even if negated). May be negated.
13254 @cindex character size
13255 @cindex eight-bit characters
13256 Set character size to 5, 6, 7, or 8 bits.
13261 Send a hangup signal when the last process closes the tty. May be
13267 Use two stop bits per character (one if negated). May be negated.
13271 Allow input to be received. May be negated.
13275 @cindex modem control
13276 Disable modem control signals. May be negated.
13280 @cindex hardware flow control
13281 @cindex flow control, hardware
13282 @cindex RTS/CTS flow control
13283 Enable RTS/CTS flow control. Non-POSIX@. May be negated.
13288 @subsection Input settings
13290 @cindex input settings
13291 These settings control operations on data received from the terminal.
13296 @cindex breaks, ignoring
13297 Ignore break characters. May be negated.
13301 @cindex breaks, cause interrupts
13302 Make breaks cause an interrupt signal. May be negated.
13306 @cindex parity, ignoring
13307 Ignore characters with parity errors. May be negated.
13311 @cindex parity errors, marking
13312 Mark parity errors (with a 255-0-character sequence). May be negated.
13316 Enable input parity checking. May be negated.
13320 @cindex eight-bit input
13321 Clear high (8th) bit of input characters. May be negated.
13325 @cindex newline, translating to return
13326 Translate newline to carriage return. May be negated.
13330 @cindex return, ignoring
13331 Ignore carriage return. May be negated.
13335 @cindex return, translating to newline
13336 Translate carriage return to newline. May be negated.
13340 @cindex input encoding, UTF-8
13341 Assume input characters are UTF-8 encoded. May be negated.
13345 @kindex C-s/C-q flow control
13346 @cindex XON/XOFF flow control
13347 Enable XON/XOFF flow control (that is, @kbd{CTRL-S}/@kbd{CTRL-Q}). May
13354 @cindex software flow control
13355 @cindex flow control, software
13356 Enable sending of @code{stop} character when the system input buffer
13357 is almost full, and @code{start} character when it becomes almost
13358 empty again. May be negated.
13362 @cindex uppercase, translating to lowercase
13363 Translate uppercase characters to lowercase. Non-POSIX@. May be
13364 negated. Note ilcuc is not implemented, as one would not be able to issue
13365 almost any (lowercase) Unix command, after invoking it.
13369 Allow any character to restart output (only the start character
13370 if negated). Non-POSIX@. May be negated.
13374 @cindex beeping at input buffer full
13375 Enable beeping and not flushing input buffer if a character arrives
13376 when the input buffer is full. Non-POSIX@. May be negated.
13381 @subsection Output settings
13383 @cindex output settings
13384 These settings control operations on data sent to the terminal.
13389 Postprocess output. May be negated.
13393 @cindex lowercase, translating to output
13394 Translate lowercase characters to uppercase. Non-POSIX@. May be
13395 negated. (Note ouclc is not currently implemented.)
13399 @cindex return, translating to newline
13400 Translate carriage return to newline. Non-POSIX@. May be negated.
13404 @cindex newline, translating to crlf
13405 Translate newline to carriage return-newline. Non-POSIX@. May be
13410 Do not print carriage returns in the first column. Non-POSIX@.
13415 Newline performs a carriage return. Non-POSIX@. May be negated.
13419 @cindex pad instead of timing for delaying
13420 Use fill (padding) characters instead of timing for delays.
13426 @cindex pad character
13427 Use ASCII @sc{del} characters for fill instead of
13428 ASCII @sc{nul} characters. Non-POSIX@.
13434 Newline delay style. Non-POSIX.
13441 Carriage return delay style. Non-POSIX.
13447 @opindex tab@var{n}
13448 Horizontal tab delay style. Non-POSIX.
13453 Backspace delay style. Non-POSIX.
13458 Vertical tab delay style. Non-POSIX.
13463 Form feed delay style. Non-POSIX.
13468 @subsection Local settings
13470 @cindex local settings
13475 Enable @code{interrupt}, @code{quit}, and @code{suspend} special
13476 characters. May be negated.
13480 Enable @code{erase}, @code{kill}, @code{werase}, and @code{rprnt}
13481 special characters. May be negated.
13485 Enable non-POSIX special characters. May be negated.
13489 Echo input characters. May be negated.
13495 Echo @code{erase} characters as backspace-space-backspace. May be
13500 @cindex newline echoing after @code{kill}
13501 Echo a newline after a @code{kill} character. May be negated.
13505 @cindex newline, echoing
13506 Echo newline even if not echoing other characters. May be negated.
13510 @cindex flushing, disabling
13511 Disable flushing after @code{interrupt} and @code{quit} special
13512 characters. May be negated.
13516 @cindex case translation
13517 Enable input and output of uppercase characters by preceding their
13518 lowercase equivalents with @samp{\}, when @code{icanon} is set.
13519 Non-POSIX@. May be negated.
13523 @cindex background jobs, stopping at terminal write
13524 Stop background jobs that try to write to the terminal. Non-POSIX@.
13531 Echo erased characters backward, between @samp{\} and @samp{/}.
13532 Non-POSIX@. May be negated.
13538 @cindex control characters, using @samp{^@var{c}}
13539 @cindex hat notation for control characters
13540 Echo control characters in hat notation (@samp{^@var{c}}) instead
13541 of literally. Non-POSIX@. May be negated.
13547 Echo the @code{kill} special character by erasing each character on
13548 the line as indicated by the @code{echoprt} and @code{echoe} settings,
13549 instead of by the @code{echoctl} and @code{echok} settings.
13556 @subsection Combination settings
13558 @cindex combination settings
13559 Combination settings:
13566 Same as @code{parenb -parodd cs7}. May be negated. If negated, same
13567 as @code{-parenb cs8}.
13571 Same as @code{parenb parodd cs7}. May be negated. If negated, same
13572 as @code{-parenb cs8}.
13576 Same as @code{-icrnl -onlcr}. May be negated. If negated, same as
13577 @code{icrnl -inlcr -igncr onlcr -ocrnl -onlret}.
13581 Reset the @code{erase} and @code{kill} special characters to their default
13588 @c This is too long to write inline.
13590 cread -ignbrk brkint -inlcr -igncr icrnl -ixoff
13591 -iuclc -ixany imaxbel opost -olcuc -ocrnl onlcr
13592 -onocr -onlret -ofill -ofdel nl0 cr0 tab0 bs0 vt0
13593 ff0 isig icanon iexten echo echoe echok -echonl
13594 -noflsh -xcase -tostop -echoprt echoctl echoke
13598 and also sets all special characters to their default values.
13602 Same as @code{brkint ignpar istrip icrnl ixon opost isig icanon}, plus
13603 sets the @code{eof} and @code{eol} characters to their default values
13604 if they are the same as the @code{min} and @code{time} characters.
13605 May be negated. If negated, same as @code{raw}.
13612 -ignbrk -brkint -ignpar -parmrk -inpck -istrip
13613 -inlcr -igncr -icrnl -ixon -ixoff -iuclc -ixany
13614 -imaxbel -opost -isig -icanon -xcase min 1 time 0
13618 May be negated. If negated, same as @code{cooked}.
13622 Same as @option{-icanon}. May be negated. If negated, same as
13627 @cindex eight-bit characters
13628 Same as @code{-parenb -istrip cs8}. May be negated. If negated,
13629 same as @code{parenb istrip cs7}.
13633 Same as @option{-parenb -istrip -opost cs8}. May be negated.
13634 If negated, same as @code{parenb istrip opost cs7}.
13638 Same as @option{-ixany}. Non-POSIX@. May be negated.
13642 Same as @code{tab0}. Non-POSIX@. May be negated. If negated, same
13649 Same as @code{xcase iuclc olcuc}. Non-POSIX@. May be negated.
13650 (Used for terminals with uppercase characters only.)
13654 Same as @code{echoe echoctl echoke}.
13658 Same as @code{echoe echoctl echoke -ixany intr ^C erase ^? kill C-u}.
13663 @subsection Special characters
13665 @cindex special characters
13666 @cindex characters, special
13668 The special characters' default values vary from system to system.
13669 They are set with the syntax @samp{name value}, where the names are
13670 listed below and the value can be given either literally, in hat
13671 notation (@samp{^@var{c}}), or as an integer which may start with
13672 @samp{0x} to indicate hexadecimal, @samp{0} to indicate octal, or
13673 any other digit to indicate decimal.
13675 @cindex disabling special characters
13676 @kindex u@r{, and disabling special characters}
13677 For GNU stty, giving a value of @code{^-} or @code{undef} disables that
13678 special character. (This is incompatible with Ultrix @command{stty},
13679 which uses a value of @samp{u} to disable a special character. GNU
13680 @command{stty} treats a value @samp{u} like any other, namely to set that
13681 special character to @key{U}.)
13687 Send an interrupt signal.
13691 Send a quit signal.
13695 Erase the last character typed.
13699 Erase the current line.
13703 Send an end of file (terminate the input).
13711 Alternate character to end the line. Non-POSIX.
13715 Switch to a different shell layer. Non-POSIX.
13719 Restart the output after stopping it.
13727 Send a terminal stop signal.
13731 Send a terminal stop signal after flushing the input. Non-POSIX.
13735 Redraw the current line. Non-POSIX.
13739 Erase the last word typed. Non-POSIX.
13743 Enter the next character typed literally, even if it is a special
13744 character. Non-POSIX.
13749 @subsection Special settings
13751 @cindex special settings
13756 Set the minimum number of characters that will satisfy a read until
13757 the time value has expired, when @option{-icanon} is set.
13761 Set the number of tenths of a second before reads time out if the minimum
13762 number of characters have not been read, when @option{-icanon} is set.
13764 @item ispeed @var{n}
13766 Set the input speed to @var{n}.
13768 @item ospeed @var{n}
13770 Set the output speed to @var{n}.
13774 Tell the tty kernel driver that the terminal has @var{n} rows.
13778 @itemx columns @var{n}
13781 Tell the kernel that the terminal has @var{n} columns. Non-POSIX.
13787 Print the number of rows and columns that the kernel thinks the
13788 terminal has. (Systems that don't support rows and columns in the kernel
13789 typically use the environment variables @env{LINES} and @env{COLUMNS}
13790 instead; however, GNU @command{stty} does not know anything about them.)
13795 Use line discipline @var{n}. Non-POSIX.
13799 Print the terminal speed.
13802 @cindex baud rate, setting
13803 Set the input and output speeds to @var{n}. @var{n} can be one of: 0
13804 50 75 110 134 134.5 150 200 300 600 1200 1800 2400 4800 9600 19200
13805 38400 @code{exta} @code{extb}. @code{exta} is the same as 19200;
13806 @code{extb} is the same as 38400. Many systems, including GNU/Linux,
13807 support higher speeds. The @command{stty} command includes support
13824 4000000 where the system supports these.
13825 0 hangs up the line if @option{-clocal} is set.
13829 @node printenv invocation
13830 @section @command{printenv}: Print all or some environment variables
13833 @cindex printing all or some environment variables
13834 @cindex environment variables, printing
13836 @command{printenv} prints environment variable values. Synopsis:
13839 printenv [@var{option}] [@var{variable}]@dots{}
13842 If no @var{variable}s are specified, @command{printenv} prints the value of
13843 every environment variable. Otherwise, it prints the value of each
13844 @var{variable} that is set, and nothing for those that are not set.
13846 The program accepts the following option. Also see @ref{Common options}.
13854 @cindex exit status of @command{printenv}
13858 0 if all variables specified were found
13859 1 if at least one specified variable was not found
13860 2 if a write error occurred
13864 @node tty invocation
13865 @section @command{tty}: Print file name of terminal on standard input
13868 @cindex print terminal file name
13869 @cindex terminal file name, printing
13871 @command{tty} prints the file name of the terminal connected to its standard
13872 input. It prints @samp{not a tty} if standard input is not a terminal.
13876 tty [@var{option}]@dots{}
13879 The program accepts the following option. Also see @ref{Common options}.
13889 Print nothing; only return an exit status.
13893 @cindex exit status of @command{tty}
13897 0 if standard input is a terminal
13898 1 if standard input is not a terminal
13899 2 if given incorrect arguments
13900 3 if a write error occurs
13904 @node User information
13905 @chapter User information
13907 @cindex user information, commands for
13908 @cindex commands for printing user information
13910 This section describes commands that print user-related information:
13911 logins, groups, and so forth.
13914 * id invocation:: Print user identity.
13915 * logname invocation:: Print current login name.
13916 * whoami invocation:: Print effective user ID.
13917 * groups invocation:: Print group names a user is in.
13918 * users invocation:: Print login names of users currently logged in.
13919 * who invocation:: Print who is currently logged in.
13923 @node id invocation
13924 @section @command{id}: Print user identity
13927 @cindex real user and group IDs, printing
13928 @cindex effective user and group IDs, printing
13929 @cindex printing real and effective user and group IDs
13931 @command{id} prints information about the given user, or the process
13932 running it if no user is specified. Synopsis:
13935 id [@var{option}]@dots{} [@var{username}]
13938 @vindex POSIXLY_CORRECT
13939 By default, it prints the real user ID, real group ID, effective user ID
13940 if different from the real user ID, effective group ID if different from
13941 the real group ID, and supplemental group IDs.
13942 In addition, if SELinux
13943 is enabled and the @env{POSIXLY_CORRECT} environment variable is not set,
13944 then print @samp{context=@var{c}}, where @var{c} is the security context.
13946 Each of these numeric values is preceded by an identifying string and
13947 followed by the corresponding user or group name in parentheses.
13949 The options cause @command{id} to print only part of the above information.
13950 Also see @ref{Common options}.
13957 Print only the group ID.
13963 Print only the group ID and the supplementary groups.
13969 Print the user or group name instead of the ID number. Requires
13970 @option{-u}, @option{-g}, or @option{-G}.
13976 Print the real, instead of effective, user or group ID@. Requires
13977 @option{-u}, @option{-g}, or @option{-G}.
13983 Print only the user ID.
13990 @cindex security context
13991 Print only the security context of the current user.
13992 If SELinux is disabled then print a warning and
13993 set the exit status to 1.
13999 @macro primaryAndSupplementaryGroups{cmd,arg}
14000 Primary and supplementary groups for a process are normally inherited
14001 from its parent and are usually unchanged since login. This means
14002 that if you change the group database after logging in, @command{\cmd\}
14003 will not reflect your changes within your existing login session.
14004 Running @command{\cmd\} with a \arg\ causes the user and group
14005 database to be consulted afresh, and so will give a different result.
14007 @primaryAndSupplementaryGroups{id,user argument}
14009 @node logname invocation
14010 @section @command{logname}: Print current login name
14013 @cindex printing user's login name
14014 @cindex login name, printing
14015 @cindex user name, printing
14018 @command{logname} prints the calling user's name, as found in a
14019 system-maintained file (often @file{/var/run/utmp} or
14020 @file{/etc/utmp}), and exits with a status of 0. If there is no entry
14021 for the calling process, @command{logname} prints
14022 an error message and exits with a status of 1.
14024 The only options are @option{--help} and @option{--version}. @xref{Common
14030 @node whoami invocation
14031 @section @command{whoami}: Print effective user ID
14034 @cindex effective user ID, printing
14035 @cindex printing the effective user ID
14037 @command{whoami} prints the user name associated with the current
14038 effective user ID@. It is equivalent to the command @samp{id -un}.
14040 The only options are @option{--help} and @option{--version}. @xref{Common
14046 @node groups invocation
14047 @section @command{groups}: Print group names a user is in
14050 @cindex printing groups a user is in
14051 @cindex supplementary groups, printing
14053 @command{groups} prints the names of the primary and any supplementary
14054 groups for each given @var{username}, or the current process if no names
14055 are given. If more than one name is given, the name of each user is
14057 the list of that user's groups and the user name is separated from the
14058 group list by a colon. Synopsis:
14061 groups [@var{username}]@dots{}
14064 The group lists are equivalent to the output of the command @samp{id -Gn}.
14066 @primaryAndSupplementaryGroups{groups,list of users}
14068 The only options are @option{--help} and @option{--version}. @xref{Common
14074 @node users invocation
14075 @section @command{users}: Print login names of users currently logged in
14078 @cindex printing current usernames
14079 @cindex usernames, printing current
14081 @cindex login sessions, printing users with
14082 @command{users} prints on a single line a blank-separated list of user
14083 names of users currently logged in to the current host. Each user name
14084 corresponds to a login session, so if a user has more than one login
14085 session, that user's name will appear the same number of times in the
14094 With no @var{file} argument, @command{users} extracts its information from
14095 a system-maintained file (often @file{/var/run/utmp} or
14096 @file{/etc/utmp}). If a file argument is given, @command{users} uses
14097 that file instead. A common choice is @file{/var/log/wtmp}.
14099 The only options are @option{--help} and @option{--version}. @xref{Common
14105 @node who invocation
14106 @section @command{who}: Print who is currently logged in
14109 @cindex printing current user information
14110 @cindex information, about current users
14112 @command{who} prints information about users who are currently logged on.
14116 @command{who} [@var{option}] [@var{file}] [am i]
14119 @cindex terminal lines, currently used
14121 @cindex remote hostname
14122 If given no non-option arguments, @command{who} prints the following
14123 information for each user currently logged on: login name, terminal
14124 line, login time, and remote hostname or X display.
14128 If given one non-option argument, @command{who} uses that instead of
14129 a default system-maintained file (often @file{/var/run/utmp} or
14130 @file{/etc/utmp}) as the name of the file containing the record of
14131 users logged on. @file{/var/log/wtmp} is commonly given as an argument
14132 to @command{who} to look at who has previously logged on.
14136 If given two non-option arguments, @command{who} prints only the entry
14137 for the user running it (determined from its standard input), preceded
14138 by the hostname. Traditionally, the two arguments given are @samp{am
14139 i}, as in @samp{who am i}.
14142 Time stamps are listed according to the time zone rules specified by
14143 the @env{TZ} environment variable, or by the system default rules if
14144 @env{TZ} is not set. @xref{TZ Variable,, Specifying the Time Zone
14145 with @env{TZ}, libc, The GNU C Library Reference Manual}.
14147 The program accepts the following options. Also see @ref{Common options}.
14155 Same as @samp{-b -d --login -p -r -t -T -u}.
14161 Print the date and time of last system boot.
14167 Print information corresponding to dead processes.
14173 Print a line of column headings.
14179 List only the entries that correspond to processes via which the
14180 system is waiting for a user to login. The user name is always @samp{LOGIN}.
14184 Attempt to canonicalize hostnames found in utmp through a DNS lookup. This
14185 is not the default because it can cause significant delays on systems with
14186 automatic dial-up internet access.
14190 Same as @samp{who am i}.
14196 List active processes spawned by init.
14202 Print only the login names and the number of users logged on.
14203 Overrides all other options.
14208 @opindex --runlevel
14209 Print the current (and maybe previous) run-level of the init process.
14213 Ignored; for compatibility with other versions of @command{who}.
14219 Print last system clock change.
14224 After the login time, print the number of hours and minutes that the
14225 user has been idle. @samp{.} means the user was active in the last minute.
14226 @samp{old} means the user has been idle for more than 24 hours.
14237 @opindex --writable
14238 @cindex message status
14239 @pindex write@r{, allowed}
14240 After each login name print a character indicating the user's message status:
14243 @samp{+} allowing @code{write} messages
14244 @samp{-} disallowing @code{write} messages
14245 @samp{?} cannot find terminal device
14253 @node System context
14254 @chapter System context
14256 @cindex system context
14257 @cindex context, system
14258 @cindex commands for system context
14260 This section describes commands that print or change system-wide
14264 * date invocation:: Print or set system date and time.
14265 * arch invocation:: Print machine hardware name.
14266 * nproc invocation:: Print the number of processors.
14267 * uname invocation:: Print system information.
14268 * hostname invocation:: Print or set system name.
14269 * hostid invocation:: Print numeric host identifier.
14270 * uptime invocation:: Print system uptime and load.
14273 @node date invocation
14274 @section @command{date}: Print or set system date and time
14277 @cindex time, printing or setting
14278 @cindex printing the current time
14283 date [@var{option}]@dots{} [+@var{format}]
14284 date [-u|--utc|--universal] @c this avoids a newline in the output
14285 [ MMDDhhmm[[CC]YY][.ss] ]
14289 Invoking @command{date} with no @var{format} argument is equivalent to invoking
14290 it with a default format that depends on the @env{LC_TIME} locale category.
14291 In the default C locale, this format is @samp{'+%a %b %e %H:%M:%S %Z %Y'},
14292 so the output looks like @samp{Thu Mar @ 3 13:47:51 PST 2005}.
14295 Normally, @command{date} uses the time zone rules indicated by the
14296 @env{TZ} environment variable, or the system default rules if @env{TZ}
14297 is not set. @xref{TZ Variable,, Specifying the Time Zone with
14298 @env{TZ}, libc, The GNU C Library Reference Manual}.
14300 @findex strftime @r{and @command{date}}
14301 @cindex time formats
14302 @cindex formatting times
14303 If given an argument that starts with a @samp{+}, @command{date} prints the
14304 current date and time (or the date and time specified by the
14305 @option{--date} option, see below) in the format defined by that argument,
14306 which is similar to that of the @code{strftime} function. Except for
14307 conversion specifiers, which start with @samp{%}, characters in the
14308 format string are printed unchanged. The conversion specifiers are
14314 * Time conversion specifiers:: %[HIklMNpPrRsSTXzZ]
14315 * Date conversion specifiers:: %[aAbBcCdDeFgGhjmuUVwWxyY]
14316 * Literal conversion specifiers:: %[%nt]
14317 * Padding and other flags:: Pad with zeros, spaces, etc.
14318 * Setting the time:: Changing the system clock.
14319 * Options for date:: Instead of the current time.
14321 * Date input formats:: Specifying date strings.
14323 * Examples of date:: Examples.
14326 @node Time conversion specifiers
14327 @subsection Time conversion specifiers
14329 @cindex time conversion specifiers
14330 @cindex conversion specifiers, time
14332 @command{date} conversion specifiers related to times.
14336 hour (@samp{00}@dots{}@samp{23})
14338 hour (@samp{01}@dots{}@samp{12})
14340 hour, space padded (@samp{ 0}@dots{}@samp{23}); equivalent to @samp{%_H}@.
14341 This is a GNU extension.
14343 hour, space padded (@samp{ 1}@dots{}@samp{12}); equivalent to @samp{%_I}@.
14344 This is a GNU extension.
14346 minute (@samp{00}@dots{}@samp{59})
14348 nanoseconds (@samp{000000000}@dots{}@samp{999999999}).
14349 This is a GNU extension.
14351 locale's equivalent of either @samp{AM} or @samp{PM};
14352 blank in many locales.
14353 Noon is treated as @samp{PM} and midnight as @samp{AM}.
14355 like @samp{%p}, except lower case.
14356 This is a GNU extension.
14358 locale's 12-hour clock time (e.g., @samp{11:11:04 PM})
14360 24-hour hour and minute. Same as @samp{%H:%M}.
14362 @cindex epoch, seconds since
14363 @cindex seconds since the epoch
14364 @cindex beginning of time
14365 @cindex leap seconds
14366 seconds since the epoch, i.e., since 1970-01-01 00:00:00 UTC@.
14367 Leap seconds are not counted unless leap second support is available.
14368 @xref{%s-examples}, for examples.
14369 This is a GNU extension.
14371 @cindex leap seconds
14372 second (@samp{00}@dots{}@samp{60}).
14373 This may be @samp{60} if leap seconds are supported.
14375 24-hour hour, minute, and second. Same as @samp{%H:%M:%S}.
14377 locale's time representation (e.g., @samp{23:13:48})
14379 @w{RFC 2822/ISO 8601} style numeric time zone
14380 (e.g., @samp{-0600} or @samp{+0530}), or nothing if no
14381 time zone is determinable. This value reflects the numeric time zone
14382 appropriate for the current time, using the time zone rules specified
14383 by the @env{TZ} environment variable.
14384 The time (and optionally, the time zone rules) can be overridden
14385 by the @option{--date} option.
14387 @w{RFC 3339/ISO 8601} style numeric time zone with
14388 @samp{:} (e.g., @samp{-06:00} or @samp{+05:30}), or nothing if no time
14389 zone is determinable.
14390 This is a GNU extension.
14392 Numeric time zone to the nearest second with @samp{:} (e.g.,
14393 @samp{-06:00:00} or @samp{+05:30:00}), or nothing if no time zone is
14395 This is a GNU extension.
14397 Numeric time zone with @samp{:} using the minimum necessary precision
14398 (e.g., @samp{-06}, @samp{+05:30}, or @samp{-04:56:02}), or nothing if
14399 no time zone is determinable.
14400 This is a GNU extension.
14402 alphabetic time zone abbreviation (e.g., @samp{EDT}), or nothing if no
14403 time zone is determinable. See @samp{%z} for how it is determined.
14407 @node Date conversion specifiers
14408 @subsection Date conversion specifiers
14410 @cindex date conversion specifiers
14411 @cindex conversion specifiers, date
14413 @command{date} conversion specifiers related to dates.
14417 locale's abbreviated weekday name (e.g., @samp{Sun})
14419 locale's full weekday name, variable length (e.g., @samp{Sunday})
14421 locale's abbreviated month name (e.g., @samp{Jan})
14423 locale's full month name, variable length (e.g., @samp{January})
14425 locale's date and time (e.g., @samp{Thu Mar @ 3 23:05:25 2005})
14427 century. This is like @samp{%Y}, except the last two digits are omitted.
14428 For example, it is @samp{20} if @samp{%Y} is @samp{2000},
14429 and is @samp{-0} if @samp{%Y} is @samp{-001}.
14430 It is normally at least two characters, but it may be more.
14432 day of month (e.g., @samp{01})
14434 date; same as @samp{%m/%d/%y}
14436 day of month, space padded; same as @samp{%_d}
14438 full date in ISO 8601 format; same as @samp{%Y-%m-%d}.
14439 This is a good choice for a date format, as it is standard and
14440 is easy to sort in the usual case where years are in the range
14443 year corresponding to the ISO week number, but without the century
14444 (range @samp{00} through @samp{99}). This has the same format and value
14445 as @samp{%y}, except that if the ISO week number (see
14447 to the previous or next year, that year is used instead.
14449 year corresponding to the ISO week number. This has the
14450 same format and value as @samp{%Y}, except that if the ISO
14452 @samp{%V}) belongs to the previous or next year, that year is used
14454 It is normally useful only if @samp{%V} is also used;
14455 for example, the format @samp{%G-%m-%d} is probably a mistake,
14456 since it combines the ISO week number year with the conventional month and day.
14460 day of year (@samp{001}@dots{}@samp{366})
14462 month (@samp{01}@dots{}@samp{12})
14464 day of week (@samp{1}@dots{}@samp{7}) with @samp{1} corresponding to Monday
14466 week number of year, with Sunday as the first day of the week
14467 (@samp{00}@dots{}@samp{53}).
14468 Days in a new year preceding the first Sunday are in week zero.
14470 ISO week number, that is, the
14471 week number of year, with Monday as the first day of the week
14472 (@samp{01}@dots{}@samp{53}).
14473 If the week containing January 1 has four or more days in
14474 the new year, then it is considered week 1; otherwise, it is week 53 of
14475 the previous year, and the next week is week 1. (See the ISO 8601
14478 day of week (@samp{0}@dots{}@samp{6}) with 0 corresponding to Sunday
14480 week number of year, with Monday as first day of week
14481 (@samp{00}@dots{}@samp{53}).
14482 Days in a new year preceding the first Monday are in week zero.
14484 locale's date representation (e.g., @samp{12/31/99})
14486 last two digits of year (@samp{00}@dots{}@samp{99})
14488 year. This is normally at least four characters, but it may be more.
14489 Year @samp{0000} precedes year @samp{0001}, and year @samp{-001}
14490 precedes year @samp{0000}.
14494 @node Literal conversion specifiers
14495 @subsection Literal conversion specifiers
14497 @cindex literal conversion specifiers
14498 @cindex conversion specifiers, literal
14500 @command{date} conversion specifiers that produce literal strings.
14512 @node Padding and other flags
14513 @subsection Padding and other flags
14515 @cindex numeric field padding
14516 @cindex padding of numeric fields
14517 @cindex fields, padding numeric
14519 Unless otherwise specified, @command{date} normally pads numeric fields
14520 with zeros, so that, for
14521 example, numeric months are always output as two digits.
14522 Seconds since the epoch are not padded, though,
14523 since there is no natural width for them.
14525 As a GNU extension, @command{date} recognizes any of the
14526 following optional flags after the @samp{%}:
14530 (hyphen) Do not pad the field; useful if the output is intended for
14533 (underscore) Pad with spaces; useful if you need a fixed
14534 number of characters in the output, but zeros are too distracting.
14536 (zero) Pad with zeros even if the conversion specifier
14537 would normally pad with spaces.
14539 Use upper case characters if possible.
14541 Use opposite case characters if possible.
14542 A field that is normally upper case becomes lower case, and vice versa.
14546 Here are some examples of padding:
14549 date +%d/%m -d "Feb 1"
14551 date +%-d/%-m -d "Feb 1"
14553 date +%_d/%_m -d "Feb 1"
14557 As a GNU extension, you can specify the field width
14558 (after any flag, if present) as a decimal number. If the natural size of the
14559 output of the field has less than the specified number of characters,
14560 the result is written right adjusted and padded to the given
14561 size. For example, @samp{%9B} prints the right adjusted month name in
14562 a field of width 9.
14564 An optional modifier can follow the optional flag and width
14565 specification. The modifiers are:
14569 Use the locale's alternate representation for date and time. This
14570 modifier applies to the @samp{%c}, @samp{%C}, @samp{%x}, @samp{%X},
14571 @samp{%y} and @samp{%Y} conversion specifiers. In a Japanese locale, for
14572 example, @samp{%Ex} might yield a date format based on the Japanese
14576 Use the locale's alternate numeric symbols for numbers. This modifier
14577 applies only to numeric conversion specifiers.
14580 If the format supports the modifier but no alternate representation
14581 is available, it is ignored.
14584 @node Setting the time
14585 @subsection Setting the time
14587 @cindex setting the time
14588 @cindex time setting
14589 @cindex appropriate privileges
14591 If given an argument that does not start with @samp{+}, @command{date} sets
14592 the system clock to the date and time specified by that argument (as
14593 described below). You must have appropriate privileges to set the
14594 system clock. Note for changes to persist across a reboot, the
14595 hardware clock may need to be updated from the system clock, which
14596 might not happen automatically on your system.
14598 The argument must consist entirely of digits, which have the following
14611 first two digits of year (optional)
14613 last two digits of year (optional)
14618 Note, the @option{--date} and @option{--set} options may not be used with an
14619 argument in the above format. The @option{--universal} option may be used
14620 with such an argument to indicate that the specified date and time are
14621 relative to Coordinated Universal Time rather than to the local time zone.
14624 @node Options for date
14625 @subsection Options for @command{date}
14627 @cindex @command{date} options
14628 @cindex options for @command{date}
14630 The program accepts the following options. Also see @ref{Common options}.
14634 @item -d @var{datestr}
14635 @itemx --date=@var{datestr}
14638 @cindex parsing date strings
14639 @cindex date strings, parsing
14640 @cindex arbitrary date strings, parsing
14643 @opindex next @var{day}
14644 @opindex last @var{day}
14645 Display the date and time specified in @var{datestr} instead of the
14646 current date and time. @var{datestr} can be in almost any common
14647 format. It can contain month names, time zones, @samp{am} and @samp{pm},
14648 @samp{yesterday}, etc. For example, @option{--date="2004-02-27
14649 14:19:13.489392193 +0530"} specifies the instant of time that is
14650 489,392,193 nanoseconds after February 27, 2004 at 2:19:13 PM in a
14651 time zone that is 5 hours and 30 minutes east of UTC.@*
14652 Note: input currently must be in locale independent format. E.g., the
14653 LC_TIME=C below is needed to print back the correct date in many locales:
14655 date -d "$(LC_TIME=C date)"
14657 @xref{Date input formats}.
14659 @item -f @var{datefile}
14660 @itemx --file=@var{datefile}
14663 Parse each line in @var{datefile} as with @option{-d} and display the
14664 resulting date and time. If @var{datefile} is @samp{-}, use standard
14665 input. This is useful when you have many dates to process, because the
14666 system overhead of starting up the @command{date} executable many times can
14669 @item -I[@var{timespec}]
14670 @itemx --iso-8601[=@var{timespec}]
14671 @opindex -I[@var{timespec}]
14672 @opindex --iso-8601[=@var{timespec}]
14673 Display the date using the ISO 8601 format, @samp{%Y-%m-%d}.
14675 The argument @var{timespec} specifies the number of additional
14676 terms of the time to include. It can be one of the following:
14679 Print just the date. This is the default if @var{timespec} is omitted.
14682 Append the hour of the day to the date.
14685 Append the hours and minutes.
14688 Append the hours, minutes and seconds.
14691 Append the hours, minutes, seconds and nanoseconds.
14694 If showing any time terms, then include the time zone using the format
14697 @item -r @var{file}
14698 @itemx --reference=@var{file}
14700 @opindex --reference
14701 Display the date and time of the last modification of @var{file},
14702 instead of the current date and time.
14709 @opindex --rfc-2822
14710 Display the date and time using the format @samp{%a, %d %b %Y %H:%M:%S
14711 %z}, evaluated in the C locale so abbreviations are always in English.
14715 Fri, 09 Sep 2005 13:51:39 -0700
14718 This format conforms to
14719 @uref{ftp://ftp.rfc-editor.org/in-notes/rfc2822.txt, Internet
14721 @uref{ftp://ftp.rfc-editor.org/in-notes/rfc822.txt, 822}, the
14722 current and previous standards for Internet email.
14724 @item --rfc-3339=@var{timespec}
14725 @opindex --rfc-3339=@var{timespec}
14726 Display the date using a format specified by
14727 @uref{ftp://ftp.rfc-editor.org/in-notes/rfc3339.txt, Internet
14728 RFC 3339}. This is a subset of the ISO 8601
14729 format, except that it also permits applications to use a space rather
14730 than a @samp{T} to separate dates from times. Unlike the other
14731 standard formats, RFC 3339 format is always suitable as
14732 input for the @option{--date} (@option{-d}) and @option{--file}
14733 (@option{-f}) options, regardless of the current locale.
14735 The argument @var{timespec} specifies how much of the time to include.
14736 It can be one of the following:
14740 Print just the full-date, e.g., @samp{2005-09-14}.
14741 This is equivalent to the format @samp{%Y-%m-%d}.
14744 Print the full-date and full-time separated by a space, e.g.,
14745 @samp{2005-09-14 00:56:06+05:30}. The output ends with a numeric
14746 time-offset; here the @samp{+05:30} means that local time is five
14747 hours and thirty minutes east of UTC@. This is equivalent to
14748 the format @samp{%Y-%m-%d %H:%M:%S%:z}.
14751 Like @samp{seconds}, but also print nanoseconds, e.g.,
14752 @samp{2005-09-14 00:56:06.998458565+05:30}.
14753 This is equivalent to the format @samp{%Y-%m-%d %H:%M:%S.%N%:z}.
14757 @item -s @var{datestr}
14758 @itemx --set=@var{datestr}
14761 Set the date and time to @var{datestr}. See @option{-d} above.
14762 See also @ref{Setting the time}.
14769 @opindex --universal
14770 @cindex Coordinated Universal Time
14772 @cindex Greenwich Mean Time
14774 @cindex leap seconds
14776 Use Coordinated Universal Time (UTC) by operating as if the
14777 @env{TZ} environment variable were set to the string @samp{UTC0}.
14779 Universal Time is often called ``Greenwich Mean Time'' (@sc{gmt}) for
14780 historical reasons.
14781 Typically, systems ignore leap seconds and thus implement an
14782 approximation to UTC rather than true UTC.
14786 @node Examples of date
14787 @subsection Examples of @command{date}
14789 @cindex examples of @command{date}
14791 Here are a few examples. Also see the documentation for the @option{-d}
14792 option in the previous section.
14797 To print the date of the day before yesterday:
14800 date --date='2 days ago'
14804 To print the date of the day three months and one day hence:
14807 date --date='3 months 1 day'
14811 To print the day of year of Christmas in the current year:
14814 date --date='25 Dec' +%j
14818 To print the current full month name and the day of the month:
14824 But this may not be what you want because for the first nine days of
14825 the month, the @samp{%d} expands to a zero-padded two-digit field,
14826 for example @samp{date -d 1may '+%B %d'} will print @samp{May 01}.
14829 To print a date without the leading zero for one-digit days
14830 of the month, you can use the (GNU extension)
14831 @samp{-} flag to suppress
14832 the padding altogether:
14835 date -d 1may '+%B %-d
14839 To print the current date and time in the format required by many
14840 non-GNU versions of @command{date} when setting the system clock:
14843 date +%m%d%H%M%Y.%S
14847 To set the system clock forward by two minutes:
14850 date --set='+2 minutes'
14854 To print the date in RFC 2822 format,
14855 use @samp{date --rfc-2822}. Here is some example output:
14858 Fri, 09 Sep 2005 13:51:39 -0700
14861 @anchor{%s-examples}
14863 To convert a date string to the number of seconds since the epoch
14864 (which is 1970-01-01 00:00:00 UTC), use the @option{--date} option with
14865 the @samp{%s} format. That can be useful in sorting and/or graphing
14866 and/or comparing data by date. The following command outputs the
14867 number of the seconds since the epoch for the time two minutes after the
14871 date --date='1970-01-01 00:02:00 +0000' +%s
14875 If you do not specify time zone information in the date string,
14876 @command{date} uses your computer's idea of the time zone when
14877 interpreting the string. For example, if your computer's time zone is
14878 that of Cambridge, Massachusetts, which was then 5 hours (i.e., 18,000
14879 seconds) behind UTC:
14882 # local time zone used
14883 date --date='1970-01-01 00:02:00' +%s
14888 If you're sorting or graphing dated data, your raw date values may be
14889 represented as seconds since the epoch. But few people can look at
14890 the date @samp{946684800} and casually note ``Oh, that's the first second
14891 of the year 2000 in Greenwich, England.''
14894 date --date='2000-01-01 UTC' +%s
14898 An alternative is to use the @option{--utc} (@option{-u}) option.
14899 Then you may omit @samp{UTC} from the date string. Although this
14900 produces the same result for @samp{%s} and many other format sequences,
14901 with a time zone offset different from zero, it would give a different
14902 result for zone-dependent formats like @samp{%z}.
14905 date -u --date=2000-01-01 +%s
14909 To convert such an unwieldy number of seconds back to
14910 a more readable form, use a command like this:
14913 # local time zone used
14914 date -d '1970-01-01 UTC 946684800 seconds' +"%Y-%m-%d %T %z"
14915 1999-12-31 19:00:00 -0500
14918 Or if you do not mind depending on the @samp{@@} feature present since
14919 coreutils 5.3.0, you could shorten this to:
14922 date -d @@946684800 +"%F %T %z"
14923 1999-12-31 19:00:00 -0500
14926 Often it is better to output UTC-relative date and time:
14929 date -u -d '1970-01-01 946684800 seconds' +"%Y-%m-%d %T %z"
14930 2000-01-01 00:00:00 +0000
14934 @cindex leap seconds
14935 Typically the seconds count omits leap seconds, but some systems are
14936 exceptions. Because leap seconds are not predictable, the mapping
14937 between the seconds count and a future timestamp is not reliable on
14938 the atypical systems that include leap seconds in their counts.
14940 Here is how the two kinds of systems handle the leap second at
14941 2012-06-30 23:59:60 UTC:
14944 # Typical systems ignore leap seconds:
14945 date --date='2012-06-30 23:59:59 +0000' +%s
14947 date --date='2012-06-30 23:59:60 +0000' +%s
14948 date: invalid date '2012-06-30 23:59:60 +0000'
14949 date --date='2012-07-01 00:00:00 +0000' +%s
14954 # Atypical systems count leap seconds:
14955 date --date='2012-06-30 23:59:59 +0000' +%s
14957 date --date='2012-06-30 23:59:60 +0000' +%s
14959 date --date='2012-07-01 00:00:00 +0000' +%s
14966 @node arch invocation
14967 @section @command{arch}: Print machine hardware name
14970 @cindex print machine hardware name
14971 @cindex system information, printing
14973 @command{arch} prints the machine hardware name,
14974 and is equivalent to @samp{uname -m}.
14978 arch [@var{option}]
14981 The program accepts the @ref{Common options} only.
14986 @node nproc invocation
14987 @section @command{nproc}: Print the number of available processors
14990 @cindex Print the number of processors
14991 @cindex system information, printing
14993 Print the number of processing units available to the current process,
14994 which may be less than the number of online processors.
14995 If this information is not accessible, then print the number of
14996 processors installed. If the @env{OMP_NUM_THREADS} environment variable is
14997 set, then it will determine the returned value. The result is guaranteed to be
14998 greater than zero. Synopsis:
15001 nproc [@var{option}]
15004 The program accepts the following options. Also see @ref{Common options}.
15010 Print the number of installed processors on the system, which may
15011 be greater than the number online or available to the current process.
15012 The @env{OMP_NUM_THREADS} environment variable is not honored in this case.
15014 @item --ignore=@var{number}
15016 If possible, exclude this @var{number} of processing units.
15023 @node uname invocation
15024 @section @command{uname}: Print system information
15027 @cindex print system information
15028 @cindex system information, printing
15030 @command{uname} prints information about the machine and operating system
15031 it is run on. If no options are given, @command{uname} acts as if the
15032 @option{-s} option were given. Synopsis:
15035 uname [@var{option}]@dots{}
15038 If multiple options or @option{-a} are given, the selected information is
15039 printed in this order:
15042 @var{kernel-name} @var{nodename} @var{kernel-release} @var{kernel-version}
15043 @var{machine} @var{processor} @var{hardware-platform} @var{operating-system}
15046 The information may contain internal spaces, so such output cannot be
15047 parsed reliably. In the following example, @var{release} is
15048 @samp{2.2.18ss.e820-bda652a #4 SMP Tue Jun 5 11:24:08 PDT 2001}:
15052 @result{} Linux dumdum 2.2.18 #4 SMP Tue Jun 5 11:24:08 PDT 2001 i686@c
15053 unknown unknown GNU/Linux
15057 The program accepts the following options. Also see @ref{Common options}.
15065 Print all of the below information, except omit the processor type
15066 and the hardware platform name if they are unknown.
15069 @itemx --hardware-platform
15071 @opindex --hardware-platform
15072 @cindex implementation, hardware
15073 @cindex hardware platform
15074 @cindex platform, hardware
15075 Print the hardware platform name
15076 (sometimes called the hardware implementation).
15077 Print @samp{unknown} if the kernel does not make this information
15078 easily available, as is the case with Linux kernels.
15084 @cindex machine type
15085 @cindex hardware class
15086 @cindex hardware type
15087 Print the machine hardware name (sometimes called the hardware class
15093 @opindex --nodename
15096 @cindex network node name
15097 Print the network node hostname.
15102 @opindex --processor
15103 @cindex host processor type
15104 Print the processor type (sometimes called the instruction set
15105 architecture or ISA).
15106 Print @samp{unknown} if the kernel does not make this information
15107 easily available, as is the case with Linux kernels.
15110 @itemx --operating-system
15112 @opindex --operating-system
15113 @cindex operating system name
15114 Print the name of the operating system.
15117 @itemx --kernel-release
15119 @opindex --kernel-release
15120 @cindex kernel release
15121 @cindex release of kernel
15122 Print the kernel release.
15125 @itemx --kernel-name
15127 @opindex --kernel-name
15128 @cindex kernel name
15129 @cindex name of kernel
15130 Print the kernel name.
15131 POSIX 1003.1-2001 (@pxref{Standards conformance}) calls this
15132 ``the implementation of the operating system'', because the
15133 POSIX specification itself has no notion of ``kernel''.
15134 The kernel name might be the same as the operating system name printed
15135 by the @option{-o} or @option{--operating-system} option, but it might
15136 differ. Some operating systems (e.g., FreeBSD, HP-UX) have the same
15137 name as their underlying kernels; others (e.g., GNU/Linux, Solaris)
15141 @itemx --kernel-version
15143 @opindex --kernel-version
15144 @cindex kernel version
15145 @cindex version of kernel
15146 Print the kernel version.
15153 @node hostname invocation
15154 @section @command{hostname}: Print or set system name
15157 @cindex setting the hostname
15158 @cindex printing the hostname
15159 @cindex system name, printing
15160 @cindex appropriate privileges
15162 With no arguments, @command{hostname} prints the name of the current host
15163 system. With one argument, it sets the current host name to the
15164 specified string. You must have appropriate privileges to set the host
15168 hostname [@var{name}]
15171 The only options are @option{--help} and @option{--version}. @xref{Common
15177 @node hostid invocation
15178 @section @command{hostid}: Print numeric host identifier
15181 @cindex printing the host identifier
15183 @command{hostid} prints the numeric identifier of the current host
15184 in hexadecimal. This command accepts no arguments.
15185 The only options are @option{--help} and @option{--version}.
15186 @xref{Common options}.
15188 For example, here's what it prints on one system I use:
15195 On that system, the 32-bit quantity happens to be closely
15196 related to the system's Internet address, but that isn't always
15201 @node uptime invocation
15202 @section @command{uptime}: Print system uptime and load
15205 @cindex printing the system uptime and load
15207 @command{uptime} prints the current time, the system's uptime, the
15208 number of logged-in users and the current load average.
15210 If an argument is specified, it is used as the file to be read
15211 to discover how many users are logged in. If no argument is
15212 specified, a system default is used (@command{uptime --help} indicates
15213 the default setting).
15215 The only options are @option{--help} and @option{--version}.
15216 @xref{Common options}.
15218 For example, here's what it prints right now on one system I use:
15222 14:07 up 3:35, 3 users, load average: 1.39, 1.15, 1.04
15225 The precise method of calculation of load average varies somewhat
15226 between systems. Some systems calculate it as the average number of
15227 runnable processes over the last 1, 5 and 15 minutes, but some systems
15228 also include processes in the uninterruptible sleep state (that is,
15229 those processes which are waiting for disk I/O). The Linux kernel
15230 includes uninterruptible processes.
15232 @node SELinux context
15233 @chapter SELinux context
15235 @cindex SELinux context
15236 @cindex SELinux, context
15237 @cindex commands for SELinux context
15239 This section describes commands for operations with SELinux
15243 * chcon invocation:: Change SELinux context of file
15244 * runcon invocation:: Run a command in specified SELinux context
15247 @node chcon invocation
15248 @section @command{chcon}: Change SELinux context of file
15251 @cindex changing security context
15252 @cindex change SELinux context
15254 @command{chcon} changes the SELinux security context of the selected files.
15258 chcon [@var{option}]@dots{} @var{context} @var{file}@dots{}
15259 chcon [@var{option}]@dots{} [-u @var{user}] [-r @var{role}] [-l @var{range}]@c
15260 [-t @var{type}] @var{file}@dots{}
15261 chcon [@var{option}]@dots{} --reference=@var{rfile} @var{file}@dots{}
15264 Change the SELinux security context of each @var{file} to @var{context}.
15265 With @option{--reference}, change the security context of each @var{file}
15266 to that of @var{rfile}.
15268 The program accepts the following options. Also see @ref{Common options}.
15272 @item --dereference
15273 @opindex --dereference
15274 Do not affect symbolic links but what they refer to; this is the default.
15277 @itemx --no-dereference
15279 @opindex --no-dereference
15280 @cindex no dereference
15281 Affect the symbolic links themselves instead of any referenced file.
15283 @item --reference=@var{rfile}
15284 @opindex --reference
15285 @cindex reference file
15286 Use @var{rfile}'s security context rather than specifying a @var{context} value.
15291 @opindex --recursive
15292 Operate on files and directories recursively.
15294 @item --preserve-root
15295 @opindex --preserve-root
15296 Refuse to operate recursively on the root directory, @file{/},
15297 when used together with the @option{--recursive} option.
15298 @xref{Treating / specially}.
15300 @item --no-preserve-root
15301 @opindex --no-preserve-root
15302 Do not treat the root directory, @file{/}, specially when operating
15303 recursively; this is the default.
15304 @xref{Treating / specially}.
15307 @xref{Traversing symlinks}.
15310 @xref{Traversing symlinks}.
15313 @xref{Traversing symlinks}.
15320 Output a diagnostic for every file processed.
15322 @item -u @var{user}
15323 @itemx --user=@var{user}
15326 Set user @var{user} in the target security context.
15328 @item -r @var{role}
15329 @itemx --role=@var{role}
15332 Set role @var{role} in the target security context.
15334 @item -t @var{type}
15335 @itemx --type=@var{type}
15338 Set type @var{type} in the target security context.
15340 @item -l @var{range}
15341 @itemx --range=@var{range}
15344 Set range @var{range} in the target security context.
15350 @node runcon invocation
15351 @section @command{runcon}: Run a command in specified SELinux context
15354 @cindex run with security context
15357 @command{runcon} runs file in specified SELinux security context.
15361 runcon @var{context} @var{command} [@var{args}]
15362 runcon [ -c ] [-u @var{user}] [-r @var{role}] [-t @var{type}]@c
15363 [-l @var{range}] @var{command} [@var{args}]
15366 Run @var{command} with completely-specified @var{context}, or with
15367 current or transitioned security context modified by one or more of @var{level},
15368 @var{role}, @var{type} and @var{user}.
15370 If none of @option{-c}, @option{-t}, @option{-u}, @option{-r}, or @option{-l}
15371 is specified, the first argument is used as the complete context.
15372 Any additional arguments after @var{command}
15373 are interpreted as arguments to the command.
15375 With neither @var{context} nor @var{command}, print the current
15378 The program accepts the following options. Also see @ref{Common options}.
15386 Compute process transition context before modifying.
15388 @item -u @var{user}
15389 @itemx --user=@var{user}
15392 Set user @var{user} in the target security context.
15394 @item -r @var{role}
15395 @itemx --role=@var{role}
15398 Set role @var{role} in the target security context.
15400 @item -t @var{type}
15401 @itemx --type=@var{type}
15404 Set type @var{type} in the target security context.
15406 @item -l @var{range}
15407 @itemx --range=@var{range}
15410 Set range @var{range} in the target security context.
15414 @cindex exit status of @command{runcon}
15418 126 if @var{command} is found but cannot be invoked
15419 127 if @command{runcon} itself fails or if @var{command} cannot be found
15420 the exit status of @var{command} otherwise
15423 @node Modified command invocation
15424 @chapter Modified command invocation
15426 @cindex modified command invocation
15427 @cindex invocation of commands, modified
15428 @cindex commands for invoking other commands
15430 This section describes commands that run other commands in some context
15431 different than the current one: a modified environment, as a different
15435 * chroot invocation:: Modify the root directory.
15436 * env invocation:: Modify environment variables.
15437 * nice invocation:: Modify niceness.
15438 * nohup invocation:: Immunize to hangups.
15439 * stdbuf invocation:: Modify buffering of standard streams.
15440 * timeout invocation:: Run with time limit.
15444 @node chroot invocation
15445 @section @command{chroot}: Run a command with a different root directory
15448 @cindex running a program in a specified root directory
15449 @cindex root directory, running a program in a specified
15451 @command{chroot} runs a command with a specified root directory.
15452 On many systems, only the super-user can do this.@footnote{However,
15453 some systems (e.g., FreeBSD) can be configured to allow certain regular
15454 users to use the @code{chroot} system call, and hence to run this program.
15455 Also, on Cygwin, anyone can run the @command{chroot} command, because the
15456 underlying function is non-privileged due to lack of support in MS-Windows.}
15460 chroot @var{option} @var{newroot} [@var{command} [@var{args}]@dots{}]
15461 chroot @var{option}
15464 Ordinarily, file names are looked up starting at the root of the
15465 directory structure, i.e., @file{/}. @command{chroot} changes the root to
15466 the directory @var{newroot} (which must exist) and then runs
15467 @var{command} with optional @var{args}. If @var{command} is not
15468 specified, the default is the value of the @env{SHELL} environment
15469 variable or @command{/bin/sh} if not set, invoked with the @option{-i} option.
15470 @var{command} must not be a special built-in utility
15471 (@pxref{Special built-in utilities}).
15473 The program accepts the following options. Also see @ref{Common options}.
15474 Options must precede operands.
15478 @item --userspec=@var{user}[:@var{group}]
15479 @opindex --userspec
15480 By default, @var{command} is run with the same credentials
15481 as the invoking process.
15482 Use this option to run it as a different @var{user} and/or with a
15483 different primary @var{group}.
15485 @item --groups=@var{groups}
15487 Use this option to specify the supplementary @var{groups} to be
15488 used by the new process.
15489 The items in the list (names or numeric IDs) must be separated by commas.
15493 Here are a few tips to help avoid common problems in using chroot.
15494 To start with a simple example, make @var{command} refer to a statically
15495 linked binary. If you were to use a dynamically linked executable, then
15496 you'd have to arrange to have the shared libraries in the right place under
15497 your new root directory.
15499 For example, if you create a statically linked @command{ls} executable,
15500 and put it in @file{/tmp/empty}, you can run this command as root:
15503 $ chroot /tmp/empty /ls -Rl /
15506 Then you'll see output like this:
15511 -rwxr-xr-x 1 0 0 1041745 Aug 16 11:17 ls
15514 If you want to use a dynamically linked executable, say @command{bash},
15515 then first run @samp{ldd bash} to see what shared objects it needs.
15516 Then, in addition to copying the actual binary, also copy the listed
15517 files to the required positions under your intended new root directory.
15518 Finally, if the executable requires any other files (e.g., data, state,
15519 device files), copy them into place, too.
15521 @cindex exit status of @command{chroot}
15525 125 if @command{chroot} itself fails
15526 126 if @var{command} is found but cannot be invoked
15527 127 if @var{command} cannot be found
15528 the exit status of @var{command} otherwise
15532 @node env invocation
15533 @section @command{env}: Run a command in a modified environment
15536 @cindex environment, running a program in a modified
15537 @cindex modified environment, running a program in a
15538 @cindex running a program in a modified environment
15540 @command{env} runs a command with a modified environment. Synopses:
15543 env [@var{option}]@dots{} [@var{name}=@var{value}]@dots{} @c
15544 [@var{command} [@var{args}]@dots{}]
15548 Operands of the form @samp{@var{variable}=@var{value}} set
15549 the environment variable @var{variable} to value @var{value}.
15550 @var{value} may be empty (@samp{@var{variable}=}). Setting a variable
15551 to an empty value is different from unsetting it.
15552 These operands are evaluated left-to-right, so if two operands
15553 mention the same variable the earlier is ignored.
15555 Environment variable names can be empty, and can contain any
15556 characters other than @samp{=} and ASCII @sc{nul}.
15557 However, it is wise to limit yourself to names that
15558 consist solely of underscores, digits, and ASCII letters,
15559 and that begin with a non-digit, as applications like the shell do not
15560 work well with other names.
15563 The first operand that does not contain the character @samp{=}
15564 specifies the program to invoke; it is
15565 searched for according to the @env{PATH} environment variable. Any
15566 remaining arguments are passed as arguments to that program.
15567 The program should not be a special built-in utility
15568 (@pxref{Special built-in utilities}).
15570 Modifications to @env{PATH} take effect prior to searching for
15571 @var{command}. Use caution when reducing @env{PATH}; behavior is
15572 not portable when @env{PATH} is undefined or omits key directories
15573 such as @file{/bin}.
15575 In the rare case that a utility contains a @samp{=} in the name, the
15576 only way to disambiguate it from a variable assignment is to use an
15577 intermediate command for @var{command}, and pass the problematic
15578 program name via @var{args}. For example, if @file{./prog=} is an
15579 executable in the current @env{PATH}:
15582 env prog= true # runs 'true', with prog= in environment
15583 env ./prog= true # runs 'true', with ./prog= in environment
15584 env -- prog= true # runs 'true', with prog= in environment
15585 env sh -c '\prog= true' # runs 'prog=' with argument 'true'
15586 env sh -c 'exec "$@@"' sh prog= true # also runs 'prog='
15589 @cindex environment, printing
15591 If no command name is specified following the environment
15592 specifications, the resulting environment is printed. This is like
15593 specifying the @command{printenv} program.
15595 For some examples, suppose the environment passed to @command{env}
15596 contains @samp{LOGNAME=rms}, @samp{EDITOR=emacs}, and
15597 @samp{PATH=.:/gnubin:/hacks}:
15602 Output the current environment.
15604 $ env | LC_ALL=C sort
15607 PATH=.:/gnubin:/hacks
15611 Run @command{foo} with a reduced environment, preserving only the
15612 original @env{PATH} to avoid problems in locating @command{foo}.
15614 env - PATH="$PATH" foo
15618 Run @command{foo} with the environment containing @samp{LOGNAME=rms},
15619 @samp{EDITOR=emacs}, and @samp{PATH=.:/gnubin:/hacks}, and guarantees
15620 that @command{foo} was found in the file system rather than as a shell
15627 Run @command{nemacs} with the environment containing @samp{LOGNAME=foo},
15628 @samp{EDITOR=emacs}, @samp{PATH=.:/gnubin:/hacks}, and
15629 @samp{DISPLAY=gnu:0}.
15631 env DISPLAY=gnu:0 LOGNAME=foo nemacs
15635 Attempt to run the program @command{/energy/--} (as that is the only
15636 possible path search result); if the command exists, the environment
15637 will contain @samp{LOGNAME=rms} and @samp{PATH=/energy}, and the
15638 arguments will be @samp{e=mc2}, @samp{bar}, and @samp{baz}.
15640 env -u EDITOR PATH=/energy -- e=mc2 bar baz
15646 The program accepts the following options. Also see @ref{Common options}.
15647 Options must precede operands.
15653 @item -u @var{name}
15654 @itemx --unset=@var{name}
15657 Remove variable @var{name} from the environment, if it was in the
15662 @itemx --ignore-environment
15665 @opindex --ignore-environment
15666 Start with an empty environment, ignoring the inherited environment.
15670 @cindex exit status of @command{env}
15674 0 if no @var{command} is specified and the environment is output
15675 125 if @command{env} itself fails
15676 126 if @var{command} is found but cannot be invoked
15677 127 if @var{command} cannot be found
15678 the exit status of @var{command} otherwise
15682 @node nice invocation
15683 @section @command{nice}: Run a command with modified niceness
15687 @cindex scheduling, affecting
15688 @cindex appropriate privileges
15690 @command{nice} prints a process's @dfn{niceness}, or runs
15691 a command with modified niceness. @dfn{niceness} affects how
15692 favorably the process is scheduled in the system.
15696 nice [@var{option}]@dots{} [@var{command} [@var{arg}]@dots{}]
15699 If no arguments are given, @command{nice} prints the current niceness.
15700 Otherwise, @command{nice} runs the given @var{command} with its
15701 niceness adjusted. By default, its niceness is incremented by 10.
15703 Niceness values range at least from @minus{}20 (process has high priority
15704 and gets more resources, thus slowing down other processes) through 19
15705 (process has lower priority and runs slowly itself, but has less impact
15706 on the speed of other running processes). Some systems
15707 may have a wider range of niceness values; conversely, other systems may
15708 enforce more restrictive limits. An attempt to set the niceness
15709 outside the supported range is treated as an attempt to use the
15710 minimum or maximum supported value.
15712 A niceness should not be confused with a scheduling priority, which
15713 lets applications determine the order in which threads are scheduled
15714 to run. Unlike a priority, a niceness is merely advice to the
15715 scheduler, which the scheduler is free to ignore. Also, as a point of
15716 terminology, POSIX defines the behavior of @command{nice} in
15717 terms of a @dfn{nice value}, which is the nonnegative difference
15718 between a niceness and the minimum niceness. Though @command{nice}
15719 conforms to POSIX, its documentation and diagnostics use the
15720 term ``niceness'' for compatibility with historical practice.
15722 @var{command} must not be a special built-in utility (@pxref{Special
15723 built-in utilities}).
15725 @mayConflictWithShellBuiltIn{nice}
15727 Note to change the @dfn{niceness} of an existing process,
15728 one needs to use the @command{renice} command.
15730 The program accepts the following option. Also see @ref{Common options}.
15731 Options must precede operands.
15734 @item -n @var{adjustment}
15735 @itemx --adjustment=@var{adjustment}
15737 @opindex --adjustment
15738 Add @var{adjustment} instead of 10 to the command's niceness. If
15739 @var{adjustment} is negative and you lack appropriate privileges,
15740 @command{nice} issues a warning but otherwise acts as if you specified
15743 For compatibility @command{nice} also supports an obsolete
15744 option syntax @option{-@var{adjustment}}. New scripts should use
15745 @option{-n @var{adjustment}} instead.
15749 @cindex exit status of @command{nice}
15753 0 if no @var{command} is specified and the niceness is output
15754 125 if @command{nice} itself fails
15755 126 if @var{command} is found but cannot be invoked
15756 127 if @var{command} cannot be found
15757 the exit status of @var{command} otherwise
15760 It is sometimes useful to run a non-interactive program with reduced niceness.
15763 $ nice factor 4611686018427387903
15766 Since @command{nice} prints the current niceness,
15767 you can invoke it through itself to demonstrate how it works.
15769 The default behavior is to increase the niceness by @samp{10}:
15780 The @var{adjustment} is relative to the current niceness. In the
15781 next example, the first @command{nice} invocation runs the second one
15782 with niceness 10, and it in turn runs the final one with a niceness
15786 $ nice nice -n 3 nice
15790 Specifying a niceness larger than the supported range
15791 is the same as specifying the maximum supported value:
15794 $ nice -n 10000000000 nice
15798 Only a privileged user may run a process with lower niceness:
15802 nice: cannot set niceness: Permission denied
15804 $ sudo nice -n -1 nice
15809 @node nohup invocation
15810 @section @command{nohup}: Run a command immune to hangups
15813 @cindex hangups, immunity to
15814 @cindex immunity to hangups
15815 @cindex logging out and continuing to run
15818 @command{nohup} runs the given @var{command} with hangup signals ignored,
15819 so that the command can continue running in the background after you log
15823 nohup @var{command} [@var{arg}]@dots{}
15826 If standard input is a terminal, it is redirected from
15827 @file{/dev/null} so that terminal sessions do not mistakenly consider
15828 the terminal to be used by the command. This is a GNU
15829 extension; programs intended to be portable to non-GNU hosts
15830 should use @samp{nohup @var{command} [@var{arg}]@dots{} </dev/null}
15834 If standard output is a terminal, the command's standard output is appended
15835 to the file @file{nohup.out}; if that cannot be written to, it is appended
15836 to the file @file{$HOME/nohup.out}; and if that cannot be written to, the
15837 command is not run.
15838 Any @file{nohup.out} or @file{$HOME/nohup.out} file created by
15839 @command{nohup} is made readable and writable only to the user,
15840 regardless of the current umask settings.
15842 If standard error is a terminal, it is normally redirected to the same file
15843 descriptor as the (possibly-redirected) standard output.
15844 However, if standard output is closed, standard error terminal output
15845 is instead appended to the file @file{nohup.out} or
15846 @file{$HOME/nohup.out} as above.
15848 To capture the command's output to a file other than @file{nohup.out}
15849 you can redirect it. For example, to capture the output of
15853 nohup make > make.log
15856 @command{nohup} does not automatically put the command it runs in the
15857 background; you must do that explicitly, by ending the command line
15858 with an @samp{&}. Also, @command{nohup} does not alter the
15859 niceness of @var{command}; use @command{nice} for that,
15860 e.g., @samp{nohup nice @var{command}}.
15862 @var{command} must not be a special built-in utility (@pxref{Special
15863 built-in utilities}).
15865 The only options are @option{--help} and @option{--version}. @xref{Common
15866 options}. Options must precede operands.
15868 @cindex exit status of @command{nohup}
15872 125 if @command{nohup} itself fails, and @env{POSIXLY_CORRECT} is not set
15873 126 if @var{command} is found but cannot be invoked
15874 127 if @var{command} cannot be found
15875 the exit status of @var{command} otherwise
15878 If @env{POSIXLY_CORRECT} is set, internal failures give status 127
15882 @node stdbuf invocation
15883 @section @command{stdbuf}: Run a command with modified I/O stream buffering
15886 @cindex standard streams, buffering
15887 @cindex line buffered
15889 @command{stdbuf} allows one to modify the buffering operations of the
15890 three standard I/O streams associated with a program. Synopsis:
15893 stdbuf @var{option}@dots{} @var{command}
15896 @var{command} must start with the name of a program that
15899 uses the ISO C @code{FILE} streams for input/output (note the
15900 programs @command{dd} and @command{cat} don't do that),
15903 does not adjust the buffering of its standard streams (note the
15904 program @command{tee} is not in this category).
15907 Any additional @var{arg}s are passed as additional arguments to the
15910 The program accepts the following options. Also see @ref{Common options}.
15914 @item -i @var{mode}
15915 @itemx --input=@var{mode}
15918 Adjust the standard input stream buffering.
15920 @item -o @var{mode}
15921 @itemx --output=@var{mode}
15924 Adjust the standard output stream buffering.
15926 @item -e @var{mode}
15927 @itemx --error=@var{mode}
15930 Adjust the standard error stream buffering.
15934 The @var{mode} can be specified as follows:
15939 Set the stream to line buffered mode.
15940 In this mode data is coalesced until a newline is output or
15941 input is read from any stream attached to a terminal device.
15942 This option is invalid with standard input.
15945 Disable buffering of the selected stream.
15946 In this mode, data is output immediately and only the
15947 amount of data requested is read from input.
15948 Note the difference in function for input and output.
15949 Disabling buffering for input will not influence the responsiveness
15950 or blocking behavior of the stream input functions.
15951 For example @code{fread} will still block until @code{EOF} or error,
15952 even if the underlying @code{read} returns less data than requested.
15955 Specify the size of the buffer to use in fully buffered mode.
15956 @multiplierSuffixesNoBlocks{size}
15960 @cindex exit status of @command{stdbuf}
15964 125 if @command{stdbuf} itself fails
15965 126 if @var{command} is found but cannot be invoked
15966 127 if @var{command} cannot be found
15967 the exit status of @var{command} otherwise
15971 @node timeout invocation
15972 @section @command{timeout}: Run a command with a time limit
15976 @cindex run commands with bounded time
15978 @command{timeout} runs the given @var{command} and kills it if it is
15979 still running after the specified time interval. Synopsis:
15982 timeout [@var{option}] @var{duration} @var{command} [@var{arg}]@dots{}
15985 @var{command} must not be a special built-in utility (@pxref{Special
15986 built-in utilities}).
15988 The program accepts the following options. Also see @ref{Common options}.
15989 Options must precede operands.
15992 @item --preserve-status
15993 @opindex --preserve-status
15994 Return the exit status of the managed @var{command} on timeout, rather than
15995 a specific exit status indicating a timeout. This is useful if the
15996 managed @var{command} supports running for an indeterminite amount of time.
15999 @opindex --foreground
16000 Don't create a separate background program group, so that
16001 the managed @var{command} can use the foreground TTY normally.
16002 This is needed to support timing out commands not started
16003 directly from an interactive shell, in two situations.
16006 @var{command} is interactive and needs to read from the terminal for example
16008 the user wants to support sending signals directly to @var{command}
16009 from the terminal (like Ctrl-C for example)
16012 Note in this mode of operation, any children of @var{command}
16013 will not be timed out.
16015 @item -k @var{duration}
16016 @itemx --kill-after=@var{duration}
16018 @opindex --kill-after
16019 Ensure the monitored @var{command} is killed by also sending a @samp{KILL}
16020 signal, after the specified @var{duration}. Without this option, if the
16021 selected signal proves not to be fatal, @command{timeout} does not kill
16024 @item -s @var{signal}
16025 @itemx --signal=@var{signal}
16028 Send this @var{signal} to @var{command} on timeout, rather than the
16029 default @samp{TERM} signal. @var{signal} may be a name like @samp{HUP}
16030 or a number. @xref{Signal specifications}.
16034 @var{duration} is a floating point number followed by an optional unit:
16036 @samp{s} for seconds (the default)
16037 @samp{m} for minutes
16041 A duration of 0 disables the associated timeout.
16042 Note that the actual timeout duration is dependent on system conditions,
16043 which should be especially considered when specifying sub-second timeouts.
16045 @cindex exit status of @command{timeout}
16049 124 if @var{command} times out
16050 125 if @command{timeout} itself fails
16051 126 if @var{command} is found but cannot be invoked
16052 127 if @var{command} cannot be found
16053 137 if @var{command} is sent the KILL(9) signal (128+9)
16054 the exit status of @var{command} otherwise
16058 @node Process control
16059 @chapter Process control
16061 @cindex processes, commands for controlling
16062 @cindex commands for controlling processes
16065 * kill invocation:: Sending a signal to processes.
16069 @node kill invocation
16070 @section @command{kill}: Send a signal to processes
16073 @cindex send a signal to processes
16075 The @command{kill} command sends a signal to processes, causing them
16076 to terminate or otherwise act upon receiving the signal in some way.
16077 Alternatively, it lists information about signals. Synopses:
16080 kill [-s @var{signal} | --signal @var{signal} | -@var{signal}] @var{pid}@dots{}
16081 kill [-l | --list | -t | --table] [@var{signal}]@dots{}
16084 @mayConflictWithShellBuiltIn{kill}
16086 The first form of the @command{kill} command sends a signal to all
16087 @var{pid} arguments. The default signal to send if none is specified
16088 is @samp{TERM}@. The special signal number @samp{0} does not denote a
16089 valid signal, but can be used to test whether the @var{pid} arguments
16090 specify processes to which a signal could be sent.
16092 If @var{pid} is positive, the signal is sent to the process with the
16093 process ID @var{pid}. If @var{pid} is zero, the signal is sent to all
16094 processes in the process group of the current process. If @var{pid}
16095 is @minus{}1, the signal is sent to all processes for which the user has
16096 permission to send a signal. If @var{pid} is less than @minus{}1, the signal
16097 is sent to all processes in the process group that equals the absolute
16098 value of @var{pid}.
16100 If @var{pid} is not positive, a system-dependent set of system
16101 processes is excluded from the list of processes to which the signal
16104 If a negative @var{pid} argument is desired as the first one, it
16105 should be preceded by @option{--}. However, as a common extension to
16106 POSIX, @option{--} is not required with @samp{kill
16107 -@var{signal} -@var{pid}}. The following commands are equivalent:
16116 The first form of the @command{kill} command succeeds if every @var{pid}
16117 argument specifies at least one process that the signal was sent to.
16119 The second form of the @command{kill} command lists signal information.
16120 Either the @option{-l} or @option{--list} option, or the @option{-t}
16121 or @option{--table} option must be specified. Without any
16122 @var{signal} argument, all supported signals are listed. The output
16123 of @option{-l} or @option{--list} is a list of the signal names, one
16124 per line; if @var{signal} is already a name, the signal number is
16125 printed instead. The output of @option{-t} or @option{--table} is a
16126 table of signal numbers, names, and descriptions. This form of the
16127 @command{kill} command succeeds if all @var{signal} arguments are valid
16128 and if there is no output error.
16130 The @command{kill} command also supports the @option{--help} and
16131 @option{--version} options. @xref{Common options}.
16133 A @var{signal} may be a signal name like @samp{HUP}, or a signal
16134 number like @samp{1}, or an exit status of a process terminated by the
16135 signal. A signal name can be given in canonical form or prefixed by
16136 @samp{SIG}@. The case of the letters is ignored, except for the
16137 @option{-@var{signal}} option which must use upper case to avoid
16138 ambiguity with lower case option letters.
16139 @xref{Signal specifications}, for a list of supported
16140 signal names and numbers.
16145 @cindex delaying commands
16146 @cindex commands for delaying
16148 @c Perhaps @command{wait} or other commands should be described here also?
16151 * sleep invocation:: Delay for a specified time.
16155 @node sleep invocation
16156 @section @command{sleep}: Delay for a specified time
16159 @cindex delay for a specified time
16161 @command{sleep} pauses for an amount of time specified by the sum of
16162 the values of the command line arguments.
16166 sleep @var{number}[smhd]@dots{}
16170 Each argument is a number followed by an optional unit; the default
16171 is seconds. The units are:
16184 Historical implementations of @command{sleep} have required that
16185 @var{number} be an integer, and only accepted a single argument
16186 without a suffix. However, GNU @command{sleep} accepts
16187 arbitrary floating point numbers. @xref{Floating point}.
16189 The only options are @option{--help} and @option{--version}. @xref{Common
16192 @c sleep is a shell built-in at least with Solaris 11's /bin/sh
16193 @mayConflictWithShellBuiltIn{sleep}
16198 @node Numeric operations
16199 @chapter Numeric operations
16201 @cindex numeric operations
16202 These programs do numerically-related operations.
16205 * factor invocation:: Show factors of numbers.
16206 * seq invocation:: Print sequences of numbers.
16210 @node factor invocation
16211 @section @command{factor}: Print prime factors
16214 @cindex prime factors
16216 @command{factor} prints prime factors. Synopses:
16219 factor [@var{number}]@dots{}
16220 factor @var{option}
16223 If no @var{number} is specified on the command line, @command{factor} reads
16224 numbers from standard input, delimited by newlines, tabs, or spaces.
16226 The @command{factor} command supports only a small number of options:
16230 Print a short help on standard output, then exit without further
16234 Print the program version on standard output, then exit without further
16238 Factoring the product of the eighth and ninth Mersenne primes
16239 takes about 30 milliseconds of CPU time on a 2.2 GHz Athlon.
16242 M8=$(echo 2^31-1|bc)
16243 M9=$(echo 2^61-1|bc)
16244 n=$(echo "$M8 * $M9" | bc)
16245 /usr/bin/time -f %U factor $n
16246 4951760154835678088235319297: 2147483647 2305843009213693951
16250 Similarly, factoring the eighth Fermat number @math{2^{256}+1} takes
16251 about 20 seconds on the same machine.
16253 Factoring large numbers is, in general, hard. The Pollard Rho
16254 algorithm used by @command{factor} is particularly effective for
16255 numbers with relatively small factors. If you wish to factor large
16256 numbers which do not have small factors (for example, numbers which
16257 are the product of two large primes), other methods are far better.
16259 If @command{factor} is built without using GNU MP, only
16260 single-precision arithmetic is available, and so large numbers
16261 (typically @math{2^{64}} and above) will not be supported. The single-precision
16262 code uses an algorithm which is designed for factoring smaller
16268 @node seq invocation
16269 @section @command{seq}: Print numeric sequences
16272 @cindex numeric sequences
16273 @cindex sequence of numbers
16275 @command{seq} prints a sequence of numbers to standard output. Synopses:
16278 seq [@var{option}]@dots{} @var{last}
16279 seq [@var{option}]@dots{} @var{first} @var{last}
16280 seq [@var{option}]@dots{} @var{first} @var{increment} @var{last}
16283 @command{seq} prints the numbers from @var{first} to @var{last} by
16284 @var{increment}. By default, each number is printed on a separate line.
16285 When @var{increment} is not specified, it defaults to @samp{1},
16286 even when @var{first} is larger than @var{last}.
16287 @var{first} also defaults to @samp{1}. So @code{seq 1} prints
16288 @samp{1}, but @code{seq 0} and @code{seq 10 5} produce no output.
16289 Floating-point numbers may be specified. @xref{Floating point}.
16291 The program accepts the following options. Also see @ref{Common options}.
16292 Options must precede operands.
16295 @item -f @var{format}
16296 @itemx --format=@var{format}
16297 @opindex -f @var{format}
16298 @opindex --format=@var{format}
16299 @cindex formatting of numbers in @command{seq}
16300 Print all numbers using @var{format}.
16301 @var{format} must contain exactly one of the @samp{printf}-style
16302 floating point conversion specifications @samp{%a}, @samp{%e},
16303 @samp{%f}, @samp{%g}, @samp{%A}, @samp{%E}, @samp{%F}, @samp{%G}@.
16304 The @samp{%} may be followed by zero or more flags taken from the set
16305 @samp{-+#0 '}, then an optional width containing one or more digits,
16306 then an optional precision consisting of a @samp{.} followed by zero
16307 or more digits. @var{format} may also contain any number of @samp{%%}
16308 conversion specifications. All conversion specifications have the
16309 same meaning as with @samp{printf}.
16311 The default format is derived from @var{first}, @var{step}, and
16312 @var{last}. If these all use a fixed point decimal representation,
16313 the default format is @samp{%.@var{p}f}, where @var{p} is the minimum
16314 precision that can represent the output numbers exactly. Otherwise,
16315 the default format is @samp{%g}.
16317 @item -s @var{string}
16318 @itemx --separator=@var{string}
16319 @cindex separator for numbers in @command{seq}
16320 Separate numbers with @var{string}; default is a newline.
16321 The output always terminates with a newline.
16324 @itemx --equal-width
16325 Print all numbers with the same width, by padding with leading zeros.
16326 @var{first}, @var{step}, and @var{last} should all use a fixed point
16327 decimal representation.
16328 (To have other kinds of padding, use @option{--format}).
16332 You can get finer-grained control over output with @option{-f}:
16335 $ seq -f '(%9.2E)' -9e5 1.1e6 1.3e6
16341 If you want hexadecimal integer output, you can use @command{printf}
16342 to perform the conversion:
16345 $ printf '%x\n' $(seq 1048575 1024 1050623)
16351 For very long lists of numbers, use xargs to avoid
16352 system limitations on the length of an argument list:
16355 $ seq 1000000 | xargs printf '%x\n' | tail -n 3
16361 To generate octal output, use the printf @code{%o} format instead
16364 On most systems, seq can produce whole-number output for values up to
16365 at least @math{2^{53}}. Larger integers are approximated. The details
16366 differ depending on your floating-point implementation.
16367 @xref{Floating point}. A common
16368 case is that @command{seq} works with integers through @math{2^{64}},
16369 and larger integers may not be numerically correct:
16372 $ seq 50000000000000000000 2 50000000000000000004
16373 50000000000000000000
16374 50000000000000000000
16375 50000000000000000004
16378 However, note that when limited to non-negative whole numbers,
16379 an increment of 1 and no format-specifying option, seq can print
16380 arbitrarily large numbers.
16382 Be careful when using @command{seq} with outlandish values: otherwise
16383 you may see surprising results, as @command{seq} uses floating point
16384 internally. For example, on the x86 platform, where the internal
16385 representation uses a 64-bit fraction, the command:
16388 seq 1 0.0000000000000000001 1.0000000000000000009
16391 outputs 1.0000000000000000007 twice and skips 1.0000000000000000008.
16396 @node File permissions
16397 @chapter File permissions
16400 @include parse-datetime.texi
16404 @node Opening the software toolbox
16405 @chapter Opening the Software Toolbox
16407 An earlier version of this chapter appeared in
16408 @uref{http://www.linuxjournal.com/article.php?sid=2762, the
16409 @cite{What's GNU@?} column of the June 1994 @cite{Linux Journal}}.
16410 It was written by Arnold Robbins.
16413 * Toolbox introduction:: Toolbox introduction
16414 * I/O redirection:: I/O redirection
16415 * The who command:: The @command{who} command
16416 * The cut command:: The @command{cut} command
16417 * The sort command:: The @command{sort} command
16418 * The uniq command:: The @command{uniq} command
16419 * Putting the tools together:: Putting the tools together
16423 @node Toolbox introduction
16424 @unnumberedsec Toolbox Introduction
16426 This month's column is only peripherally related to the GNU Project, in
16427 that it describes a number of the GNU tools on your GNU/Linux system
16429 might be used. What it's really about is the ``Software Tools'' philosophy
16430 of program development and usage.
16432 The software tools philosophy was an important and integral concept
16433 in the initial design and development of Unix (of which Linux and GNU are
16434 essentially clones). Unfortunately, in the modern day press of
16435 Internetworking and flashy GUIs, it seems to have fallen by the
16436 wayside. This is a shame, since it provides a powerful mental model
16437 for solving many kinds of problems.
16439 Many people carry a Swiss Army knife around in their pants pockets (or
16440 purse). A Swiss Army knife is a handy tool to have: it has several knife
16441 blades, a screwdriver, tweezers, toothpick, nail file, corkscrew, and perhaps
16442 a number of other things on it. For the everyday, small miscellaneous jobs
16443 where you need a simple, general purpose tool, it's just the thing.
16445 On the other hand, an experienced carpenter doesn't build a house using
16446 a Swiss Army knife. Instead, he has a toolbox chock full of specialized
16447 tools---a saw, a hammer, a screwdriver, a plane, and so on. And he knows
16448 exactly when and where to use each tool; you won't catch him hammering nails
16449 with the handle of his screwdriver.
16451 The Unix developers at Bell Labs were all professional programmers and trained
16452 computer scientists. They had found that while a one-size-fits-all program
16453 might appeal to a user because there's only one program to use, in practice
16458 difficult to write,
16461 difficult to maintain and
16465 difficult to extend to meet new situations.
16468 Instead, they felt that programs should be specialized tools. In short, each
16469 program ``should do one thing well.'' No more and no less. Such programs are
16470 simpler to design, write, and get right---they only do one thing.
16472 Furthermore, they found that with the right machinery for hooking programs
16473 together, that the whole was greater than the sum of the parts. By combining
16474 several special purpose programs, you could accomplish a specific task
16475 that none of the programs was designed for, and accomplish it much more
16476 quickly and easily than if you had to write a special purpose program.
16477 We will see some (classic) examples of this further on in the column.
16478 (An important additional point was that, if necessary, take a detour
16479 and build any software tools you may need first, if you don't already
16480 have something appropriate in the toolbox.)
16482 @node I/O redirection
16483 @unnumberedsec I/O Redirection
16485 Hopefully, you are familiar with the basics of I/O redirection in the
16486 shell, in particular the concepts of ``standard input,'' ``standard output,''
16487 and ``standard error''. Briefly, ``standard input'' is a data source, where
16488 data comes from. A program should not need to either know or care if the
16489 data source is a disk file, a keyboard, a magnetic tape, or even a punched
16490 card reader. Similarly, ``standard output'' is a data sink, where data goes
16491 to. The program should neither know nor care where this might be.
16492 Programs that only read their standard input, do something to the data,
16493 and then send it on, are called @dfn{filters}, by analogy to filters in a
16496 With the Unix shell, it's very easy to set up data pipelines:
16499 program_to_create_data | filter1 | ... | filterN > final.pretty.data
16502 We start out by creating the raw data; each filter applies some successive
16503 transformation to the data, until by the time it comes out of the pipeline,
16504 it is in the desired form.
16506 This is fine and good for standard input and standard output. Where does the
16507 standard error come in to play? Well, think about @command{filter1} in
16508 the pipeline above. What happens if it encounters an error in the data it
16509 sees? If it writes an error message to standard output, it will just
16510 disappear down the pipeline into @command{filter2}'s input, and the
16511 user will probably never see it. So programs need a place where they can send
16512 error messages so that the user will notice them. This is standard error,
16513 and it is usually connected to your console or window, even if you have
16514 redirected standard output of your program away from your screen.
16516 For filter programs to work together, the format of the data has to be
16517 agreed upon. The most straightforward and easiest format to use is simply
16518 lines of text. Unix data files are generally just streams of bytes, with
16519 lines delimited by the ASCII @sc{lf} (Line Feed) character,
16520 conventionally called a ``newline'' in the Unix literature. (This is
16521 @code{'\n'} if you're a C programmer.) This is the format used by all
16522 the traditional filtering programs. (Many earlier operating systems
16523 had elaborate facilities and special purpose programs for managing
16524 binary data. Unix has always shied away from such things, under the
16525 philosophy that it's easiest to simply be able to view and edit your
16526 data with a text editor.)
16528 OK, enough introduction. Let's take a look at some of the tools, and then
16529 we'll see how to hook them together in interesting ways. In the following
16530 discussion, we will only present those command line options that interest
16531 us. As you should always do, double check your system documentation
16532 for the full story.
16534 @node The who command
16535 @unnumberedsec The @command{who} Command
16537 The first program is the @command{who} command. By itself, it generates a
16538 list of the users who are currently logged in. Although I'm writing
16539 this on a single-user system, we'll pretend that several people are
16544 @print{} arnold console Jan 22 19:57
16545 @print{} miriam ttyp0 Jan 23 14:19(:0.0)
16546 @print{} bill ttyp1 Jan 21 09:32(:0.0)
16547 @print{} arnold ttyp2 Jan 23 20:48(:0.0)
16550 Here, the @samp{$} is the usual shell prompt, at which I typed @samp{who}.
16551 There are three people logged in, and I am logged in twice. On traditional
16552 Unix systems, user names are never more than eight characters long. This
16553 little bit of trivia will be useful later. The output of @command{who} is nice,
16554 but the data is not all that exciting.
16556 @node The cut command
16557 @unnumberedsec The @command{cut} Command
16559 The next program we'll look at is the @command{cut} command. This program
16560 cuts out columns or fields of input data. For example, we can tell it
16561 to print just the login name and full name from the @file{/etc/passwd}
16562 file. The @file{/etc/passwd} file has seven fields, separated by
16566 arnold:xyzzy:2076:10:Arnold D. Robbins:/home/arnold:/bin/bash
16569 To get the first and fifth fields, we would use @command{cut} like this:
16572 $ cut -d: -f1,5 /etc/passwd
16573 @print{} root:Operator
16575 @print{} arnold:Arnold D. Robbins
16576 @print{} miriam:Miriam A. Robbins
16580 With the @option{-c} option, @command{cut} will cut out specific characters
16581 (i.e., columns) in the input lines. This is useful for input data
16582 that has fixed width fields, and does not have a field separator. For
16583 example, list the Monday dates for the current month:
16585 @c Is using cal ok? Looked at gcal, but I don't like it.
16596 @node The sort command
16597 @unnumberedsec The @command{sort} Command
16599 Next we'll look at the @command{sort} command. This is one of the most
16600 powerful commands on a Unix-style system; one that you will often find
16601 yourself using when setting up fancy data plumbing.
16604 command reads and sorts each file named on the command line. It then
16605 merges the sorted data and writes it to standard output. It will read
16606 standard input if no files are given on the command line (thus
16607 making it into a filter). The sort is based on the character collating
16608 sequence or based on user-supplied ordering criteria.
16611 @node The uniq command
16612 @unnumberedsec The @command{uniq} Command
16614 Finally (at least for now), we'll look at the @command{uniq} program. When
16615 sorting data, you will often end up with duplicate lines, lines that
16616 are identical. Usually, all you need is one instance of each line.
16617 This is where @command{uniq} comes in. The @command{uniq} program reads its
16618 standard input. It prints only one
16619 copy of each repeated line. It does have several options. Later on,
16620 we'll use the @option{-c} option, which prints each unique line, preceded
16621 by a count of the number of times that line occurred in the input.
16624 @node Putting the tools together
16625 @unnumberedsec Putting the Tools Together
16627 Now, let's suppose this is a large ISP server system with dozens of users
16628 logged in. The management wants the system administrator to write a
16630 generate a sorted list of logged in users. Furthermore, even if a user
16631 is logged in multiple times, his or her name should only show up in the
16634 The administrator could sit down with the system documentation and write a C
16635 program that did this. It would take perhaps a couple of hundred lines
16636 of code and about two hours to write it, test it, and debug it.
16637 However, knowing the software toolbox, the administrator can instead start out
16638 by generating just a list of logged on users:
16648 Next, sort the list:
16651 $ who | cut -c1-8 | sort
16658 Finally, run the sorted list through @command{uniq}, to weed out duplicates:
16661 $ who | cut -c1-8 | sort | uniq
16667 The @command{sort} command actually has a @option{-u} option that does what
16668 @command{uniq} does. However, @command{uniq} has other uses for which one
16669 cannot substitute @samp{sort -u}.
16671 The administrator puts this pipeline into a shell script, and makes it
16673 all the users on the system (@samp{#} is the system administrator,
16674 or @code{root}, prompt):
16677 # cat > /usr/local/bin/listusers
16678 who | cut -c1-8 | sort | uniq
16680 # chmod +x /usr/local/bin/listusers
16683 There are four major points to note here. First, with just four
16684 programs, on one command line, the administrator was able to save about two
16685 hours worth of work. Furthermore, the shell pipeline is just about as
16686 efficient as the C program would be, and it is much more efficient in
16687 terms of programmer time. People time is much more expensive than
16688 computer time, and in our modern ``there's never enough time to do
16689 everything'' society, saving two hours of programmer time is no mean
16692 Second, it is also important to emphasize that with the
16693 @emph{combination} of the tools, it is possible to do a special
16694 purpose job never imagined by the authors of the individual programs.
16696 Third, it is also valuable to build up your pipeline in stages, as we did here.
16697 This allows you to view the data at each stage in the pipeline, which helps
16698 you acquire the confidence that you are indeed using these tools correctly.
16700 Finally, by bundling the pipeline in a shell script, other users can use
16701 your command, without having to remember the fancy plumbing you set up for
16702 them. In terms of how you run them, shell scripts and compiled programs are
16705 After the previous warm-up exercise, we'll look at two additional, more
16706 complicated pipelines. For them, we need to introduce two more tools.
16708 The first is the @command{tr} command, which stands for ``transliterate.''
16709 The @command{tr} command works on a character-by-character basis, changing
16710 characters. Normally it is used for things like mapping upper case to
16714 $ echo ThIs ExAmPlE HaS MIXED case! | tr '[:upper:]' '[:lower:]'
16715 @print{} this example has mixed case!
16718 There are several options of interest:
16722 work on the complement of the listed characters, i.e.,
16723 operations apply to characters not in the given set
16726 delete characters in the first set from the output
16729 squeeze repeated characters in the output into just one character.
16732 We will be using all three options in a moment.
16734 The other command we'll look at is @command{comm}. The @command{comm}
16735 command takes two sorted input files as input data, and prints out the
16736 files' lines in three columns. The output columns are the data lines
16737 unique to the first file, the data lines unique to the second file, and
16738 the data lines that are common to both. The @option{-1}, @option{-2}, and
16739 @option{-3} command line options @emph{omit} the respective columns. (This is
16740 non-intuitive and takes a little getting used to.) For example:
16762 The file name @file{-} tells @command{comm} to read standard input
16763 instead of a regular file.
16765 Now we're ready to build a fancy pipeline. The first application is a word
16766 frequency counter. This helps an author determine if he or she is over-using
16769 The first step is to change the case of all the letters in our input file
16770 to one case. ``The'' and ``the'' are the same word when doing counting.
16773 $ tr '[:upper:]' '[:lower:]' < whats.gnu | ...
16776 The next step is to get rid of punctuation. Quoted words and unquoted words
16777 should be treated identically; it's easiest to just get the punctuation out of
16781 $ tr '[:upper:]' '[:lower:]' < whats.gnu | tr -cd '[:alnum:]_ \n' | ...
16784 The second @command{tr} command operates on the complement of the listed
16785 characters, which are all the letters, the digits, the underscore, and
16786 the blank. The @samp{\n} represents the newline character; it has to
16787 be left alone. (The ASCII tab character should also be included for
16788 good measure in a production script.)
16790 At this point, we have data consisting of words separated by blank space.
16791 The words only contain alphanumeric characters (and the underscore). The
16792 next step is break the data apart so that we have one word per line. This
16793 makes the counting operation much easier, as we will see shortly.
16796 $ tr '[:upper:]' '[:lower:]' < whats.gnu | tr -cd '[:alnum:]_ \n' |
16797 > tr -s ' ' '\n' | ...
16800 This command turns blanks into newlines. The @option{-s} option squeezes
16801 multiple newline characters in the output into just one. This helps us
16802 avoid blank lines. (The @samp{>} is the shell's ``secondary prompt.''
16803 This is what the shell prints when it notices you haven't finished
16804 typing in all of a command.)
16806 We now have data consisting of one word per line, no punctuation, all one
16807 case. We're ready to count each word:
16810 $ tr '[:upper:]' '[:lower:]' < whats.gnu | tr -cd '[:alnum:]_ \n' |
16811 > tr -s ' ' '\n' | sort | uniq -c | ...
16814 At this point, the data might look something like this:
16827 The output is sorted by word, not by count! What we want is the most
16828 frequently used words first. Fortunately, this is easy to accomplish,
16829 with the help of two more @command{sort} options:
16833 do a numeric sort, not a textual one
16836 reverse the order of the sort
16839 The final pipeline looks like this:
16842 $ tr '[:upper:]' '[:lower:]' < whats.gnu | tr -cd '[:alnum:]_ \n' |
16843 > tr -s ' ' '\n' | sort | uniq -c | sort -n -r
16852 Whew! That's a lot to digest. Yet, the same principles apply. With six
16853 commands, on two lines (really one long one split for convenience), we've
16854 created a program that does something interesting and useful, in much
16855 less time than we could have written a C program to do the same thing.
16857 A minor modification to the above pipeline can give us a simple spelling
16858 checker! To determine if you've spelled a word correctly, all you have to
16859 do is look it up in a dictionary. If it is not there, then chances are
16860 that your spelling is incorrect. So, we need a dictionary.
16861 The conventional location for a dictionary is @file{/usr/dict/words}.
16862 On my GNU/Linux system,@footnote{Redhat Linux 6.1, for the November 2000
16863 revision of this article.}
16864 this is a sorted, 45,402 word dictionary.
16866 Now, how to compare our file with the dictionary? As before, we generate
16867 a sorted list of words, one per line:
16870 $ tr '[:upper:]' '[:lower:]' < whats.gnu | tr -cd '[:alnum:]_ \n' |
16871 > tr -s ' ' '\n' | sort -u | ...
16874 Now, all we need is a list of words that are @emph{not} in the
16875 dictionary. Here is where the @command{comm} command comes in.
16878 $ tr '[:upper:]' '[:lower:]' < whats.gnu | tr -cd '[:alnum:]_ \n' |
16879 > tr -s ' ' '\n' | sort -u |
16880 > comm -23 - /usr/dict/words
16883 The @option{-2} and @option{-3} options eliminate lines that are only in the
16884 dictionary (the second file), and lines that are in both files. Lines
16885 only in the first file (standard input, our stream of words), are
16886 words that are not in the dictionary. These are likely candidates for
16887 spelling errors. This pipeline was the first cut at a production
16888 spelling checker on Unix.
16890 There are some other tools that deserve brief mention.
16894 search files for text that matches a regular expression
16897 count lines, words, characters
16900 a T-fitting for data pipes, copies data to files and to standard output
16903 the stream editor, an advanced tool
16906 a data manipulation language, another advanced tool
16909 The software tools philosophy also espoused the following bit of
16910 advice: ``Let someone else do the hard part.'' This means, take
16911 something that gives you most of what you need, and then massage it the
16912 rest of the way until it's in the form that you want.
16918 Each program should do one thing well. No more, no less.
16921 Combining programs with appropriate plumbing leads to results where
16922 the whole is greater than the sum of the parts. It also leads to novel
16923 uses of programs that the authors might never have imagined.
16926 Programs should never print extraneous header or trailer data, since these
16927 could get sent on down a pipeline. (A point we didn't mention earlier.)
16930 Let someone else do the hard part.
16933 Know your toolbox! Use each program appropriately. If you don't have an
16934 appropriate tool, build one.
16937 As of this writing, all the programs we've discussed are available via
16938 anonymous @command{ftp} from: @*
16939 @uref{ftp://gnudist.gnu.org/textutils/textutils-1.22.tar.gz}. (There may
16940 be more recent versions available now.)
16942 None of what I have presented in this column is new. The Software Tools
16943 philosophy was first introduced in the book @cite{Software Tools}, by
16944 Brian Kernighan and P.J. Plauger (Addison-Wesley, ISBN 0-201-03669-X).
16945 This book showed how to write and use software tools. It was written in
16946 1976, using a preprocessor for FORTRAN named @command{ratfor} (RATional
16947 FORtran). At the time, C was not as ubiquitous as it is now; FORTRAN
16948 was. The last chapter presented a @command{ratfor} to FORTRAN
16949 processor, written in @command{ratfor}. @command{ratfor} looks an awful
16950 lot like C; if you know C, you won't have any problem following the
16953 In 1981, the book was updated and made available as @cite{Software Tools
16954 in Pascal} (Addison-Wesley, ISBN 0-201-10342-7). Both books are
16955 still in print and are well worth
16956 reading if you're a programmer. They certainly made a major change in
16957 how I view programming.
16959 The programs in both books are available from
16960 @uref{http://cm.bell-labs.com/who/bwk, Brian Kernighan's home page}.
16961 For a number of years, there was an active
16962 Software Tools Users Group, whose members had ported the original
16963 @command{ratfor} programs to essentially every computer system with a
16964 FORTRAN compiler. The popularity of the group waned in the middle 1980s
16965 as Unix began to spread beyond universities.
16967 With the current proliferation of GNU code and other clones of Unix programs,
16968 these programs now receive little attention; modern C versions are
16969 much more efficient and do more than these programs do. Nevertheless, as
16970 exposition of good programming style, and evangelism for a still-valuable
16971 philosophy, these books are unparalleled, and I recommend them highly.
16973 Acknowledgment: I would like to express my gratitude to Brian Kernighan
16974 of Bell Labs, the original Software Toolsmith, for reviewing this column.
16976 @node GNU Free Documentation License
16977 @appendix GNU Free Documentation License
16981 @node Concept index
16988 @c Local variables:
16989 @c texinfo-column-for-description: 32