1 \input texinfo @c -*-texinfo-*-
5 @settitle GNU Grep @value{VERSION}
16 @documentencoding UTF-8
19 This manual is for @command{grep}, a pattern matching engine.
21 Copyright @copyright{} 1999-2002, 2005, 2008-2015 Free Software Foundation,
25 Permission is granted to copy, distribute and/or modify this document
26 under the terms of the GNU Free Documentation License, Version 1.3 or
27 any later version published by the Free Software Foundation; with no
28 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
29 Texts. A copy of the license is included in the section entitled
30 ``GNU Free Documentation License''.
34 @dircategory Text creation and manipulation
36 * grep: (grep). Print lines matching a pattern.
40 @title GNU Grep: Print lines matching a pattern
41 @subtitle version @value{VERSION}, @value{UPDATED}
42 @author Alain Magloire et al.
44 @vskip 0pt plus 1filll
55 @command{grep} prints lines that contain a match for a pattern.
57 This manual is for version @value{VERSION} of GNU Grep.
63 * Introduction:: Introduction.
64 * Invoking:: Command-line options, environment, exit status.
65 * Regular Expressions:: Regular Expressions.
67 * Reporting Bugs:: Reporting Bugs.
68 * Copying:: License terms for this manual.
69 * Index:: Combined index.
76 @cindex searching for a pattern
78 @command{grep} searches input files
79 for lines containing a match to a given pattern list.
80 When it finds a match in a line,
81 it copies the line to standard output (by default),
82 or produces whatever other sort of output you have requested with options.
84 Though @command{grep} expects to do the matching on text,
85 it has no limits on input line length other than available memory,
86 and it can match arbitrary characters within a line.
87 If the final byte of an input file is not a newline,
88 @command{grep} silently supplies one.
89 Since newline is also a separator for the list of patterns,
90 there is no way to match newline characters in a text.
94 @chapter Invoking @command{grep}
96 The general synopsis of the @command{grep} command line is
99 grep @var{options} @var{pattern} @var{input_file_names}
103 There can be zero or more @var{options}.
104 @var{pattern} will only be seen as such
105 (and not as an @var{input_file_name})
106 if it wasn't already specified within @var{options}
107 (by using the @samp{-e@ @var{pattern}}
108 or @samp{-f@ @var{file}} options).
109 There can be zero or more @var{input_file_names}.
112 * Command-line Options:: Short and long names, grouped by category.
113 * Environment Variables:: POSIX, GNU generic, and GNU grep specific.
114 * Exit Status:: Exit status returned by @command{grep}.
115 * grep Programs:: @command{grep} programs.
118 @node Command-line Options
119 @section Command-line Options
121 @command{grep} comes with a rich set of options:
122 some from POSIX and some being GNU extensions.
123 Long option names are always a GNU extension,
124 even for options that are from POSIX specifications.
125 Options that are specified by POSIX,
126 under their short names,
127 are explicitly marked as such
128 to facilitate POSIX-portable programming.
129 A few option names are provided
130 for compatibility with older or more exotic implementations.
133 * Generic Program Information::
135 * General Output Control::
136 * Output Line Prefix Control::
137 * Context Line Control::
138 * File and Directory Selection::
142 Several additional options control
143 which variant of the @command{grep} matching engine is used.
144 @xref{grep Programs}.
146 @node Generic Program Information
147 @subsection Generic Program Information
153 @cindex usage summary, printing
154 Print a usage message briefly summarizing the command-line options
155 and the bug-reporting address, then exit.
161 @cindex version, printing
162 Print the version number of @command{grep} to the standard output stream.
163 This version number should be included in all bug reports.
167 @node Matching Control
168 @subsection Matching Control
172 @item -e @var{pattern}
173 @itemx --regexp=@var{pattern}
175 @opindex --regexp=@var{pattern}
177 Use @var{pattern} as the pattern.
178 This can be used to specify multiple search patterns,
179 or to protect a pattern beginning with a @samp{-}.
180 (@option{-e} is specified by POSIX.)
183 @itemx --file=@var{file}
186 @cindex pattern from file
187 Obtain patterns from @var{file}, one per line.
188 The empty file contains zero patterns, and therefore matches nothing.
189 (@option{-f} is specified by POSIX.)
196 @opindex --ignore-case
197 @cindex case insensitive search
198 Ignore case distinctions, so that characters that differ only in case
199 match each other. Although this is straightforward when letters
200 differ in case only via lowercase-uppercase pairs, the behavior is
201 unspecified in other situations. For example, uppercase ``S'' has an
202 unusual lowercase counterpart ``ſ'' (Unicode character U+017F, LATIN
203 SMALL LETTER LONG S) in many locales, and it is unspecified whether
204 this unusual character matches ``S'' or ``s'' even though uppercasing
205 it yields ``S''. Another example: the lowercase German letter ``ß''
206 (U+00DF, LATIN SMALL LETTER SHARP S) is normally capitalized as the
207 two-character string ``SS'' but it does not match ``SS'', and it might
208 not match the uppercase letter ``ẞ'' (U+1E9E, LATIN CAPITAL LETTER
209 SHARP S) even though lowercasing the latter yields the former.
211 @option{-y} is an obsolete synonym that is provided for compatibility.
212 (@option{-i} is specified by POSIX.)
215 @itemx --invert-match
217 @opindex --invert-match
218 @cindex invert matching
219 @cindex print non-matching lines
220 Invert the sense of matching, to select non-matching lines.
221 (@option{-v} is specified by POSIX.)
226 @opindex --word-regexp
227 @cindex matching whole words
228 Select only those lines containing matches that form whole words.
229 The test is that the matching substring must either
230 be at the beginning of the line,
231 or preceded by a non-word constituent character.
233 it must be either at the end of the line
234 or followed by a non-word constituent character.
235 Word-constituent characters are letters, digits, and the underscore.
240 @opindex --line-regexp
241 @cindex match the whole line
242 Select only those matches that exactly match the whole line.
243 (@option{-x} is specified by POSIX.)
247 @node General Output Control
248 @subsection General Output Control
256 @cindex counting lines
257 Suppress normal output;
258 instead print a count of matching lines for each input file.
259 With the @option{-v} (@option{--invert-match}) option,
260 count non-matching lines.
261 (@option{-c} is specified by POSIX.)
263 @item --color[=@var{WHEN}]
264 @itemx --colour[=@var{WHEN}]
267 @cindex highlight, color, colour
268 Surround the matched (non-empty) strings, matching lines, context lines,
269 file names, line numbers, byte offsets, and separators (for fields and
270 groups of context lines) with escape sequences to display them in color
272 The colors are defined by the environment variable @env{GREP_COLORS}
273 and default to @samp{ms=01;31:mc=01;31:sl=:cx=:fn=35:ln=32:bn=32:se=36}
274 for bold red matched text, magenta file names, green line numbers,
275 green byte offsets, cyan separators, and default terminal colors otherwise.
276 The deprecated environment variable @env{GREP_COLOR} is still supported,
277 but its setting does not have priority;
278 it defaults to @samp{01;31} (bold red)
279 which only covers the color for matched text.
280 @var{WHEN} is @samp{never}, @samp{always}, or @samp{auto}.
283 @itemx --files-without-match
285 @opindex --files-without-match
286 @cindex files which don't match
287 Suppress normal output;
288 instead print the name of each input file from which
289 no output would normally have been printed.
290 The scanning of each file stops on the first match.
293 @itemx --files-with-matches
295 @opindex --files-with-matches
296 @cindex names of matching files
297 Suppress normal output;
298 instead print the name of each input file from which
299 output would normally have been printed.
300 The scanning of each file stops on the first match.
301 (@option{-l} is specified by POSIX.)
304 @itemx --max-count=@var{num}
308 Stop reading a file after @var{num} matching lines.
309 If the input is standard input from a regular file,
310 and @var{num} matching lines are output,
311 @command{grep} ensures that the standard input is positioned
312 just after the last matching line before exiting,
313 regardless of the presence of trailing context lines.
314 This enables a calling process to resume a search.
315 For example, the following shell script makes use of it:
318 while grep -m 1 PATTERN
324 But the following probably will not work because a pipe is not a regular
328 # This probably will not work.
330 while grep -m 1 PATTERN
336 When @command{grep} stops after @var{num} matching lines,
337 it outputs any trailing context lines.
338 Since context does not include matching lines,
339 @command{grep} will stop when it encounters another matching line.
340 When the @option{-c} or @option{--count} option is also used,
341 @command{grep} does not output a count greater than @var{num}.
342 When the @option{-v} or @option{--invert-match} option is also used,
343 @command{grep} stops after outputting @var{num} non-matching lines.
346 @itemx --only-matching
348 @opindex --only-matching
349 @cindex only matching
350 Print only the matched (non-empty) parts of matching lines,
351 with each such part on a separate output line.
359 @cindex quiet, silent
360 Quiet; do not write anything to standard output.
361 Exit immediately with zero status if any match is found,
362 even if an error was detected.
363 Also see the @option{-s} or @option{--no-messages} option.
364 (@option{-q} is specified by POSIX.)
369 @opindex --no-messages
370 @cindex suppress error messages
371 Suppress error messages about nonexistent or unreadable files.
373 unlike GNU @command{grep},
374 7th Edition Unix @command{grep} did not conform to POSIX,
375 because it lacked @option{-q}
376 and its @option{-s} option behaved like
377 GNU @command{grep}'s @option{-q} option.@footnote{Of course, 7th Edition
378 Unix predated POSIX by several years!}
379 USG-style @command{grep} also lacked @option{-q}
380 but its @option{-s} option behaved like GNU @command{grep}'s.
381 Portable shell scripts should avoid both
382 @option{-q} and @option{-s} and should redirect
383 standard and error output to @file{/dev/null} instead.
384 (@option{-s} is specified by POSIX.)
388 @node Output Line Prefix Control
389 @subsection Output Line Prefix Control
391 When several prefix fields are to be output,
392 the order is always file name, line number, and byte offset,
393 regardless of the order in which these options were specified.
400 @opindex --byte-offset
402 Print the 0-based byte offset within the input file
403 before each line of output.
404 If @option{-o} (@option{--only-matching}) is specified,
405 print the offset of the matching part itself.
406 When @command{grep} runs on MS-DOS or MS-Windows,
407 the printed byte offsets depend on whether
408 the @option{-u} (@option{--unix-byte-offsets}) option is used;
412 @itemx --with-filename
414 @opindex --with-filename
415 @cindex with filename prefix
416 Print the file name for each match.
417 This is the default when there is more than one file to search.
422 @opindex --no-filename
423 @cindex no filename prefix
424 Suppress the prefixing of file names on output.
425 This is the default when there is only one file
426 (or only standard input) to search.
428 @item --label=@var{LABEL}
430 @cindex changing name of standard input
431 Display input actually coming from standard input
432 as input coming from file @var{LABEL}. This is
433 especially useful when implementing tools like
434 @command{zgrep}; e.g.:
437 gzip -cd foo.gz | grep --label=foo -H something
443 @opindex --line-number
444 @cindex line numbering
445 Prefix each line of output with the 1-based line number within its input file.
446 (@option{-n} is specified by POSIX.)
451 @opindex --initial-tab
452 @cindex tab-aligned content lines
453 Make sure that the first character of actual line content lies on a tab stop,
454 so that the alignment of tabs looks normal.
455 This is useful with options that prefix their output to the actual content:
456 @option{-H}, @option{-n}, and @option{-b}.
457 In order to improve the probability that lines
458 from a single file will all start at the same column,
459 this also causes the line number and byte offset (if present)
460 to be printed in a minimum-size field width.
463 @itemx --unix-byte-offsets
465 @opindex --unix-byte-offsets
466 @cindex MS-DOS/MS-Windows byte offsets
467 @cindex byte offsets, on MS-DOS/MS-Windows
468 Report Unix-style byte offsets.
469 This option causes @command{grep} to report byte offsets
470 as if the file were a Unix-style text file,
471 i.e., the byte offsets ignore carriage returns that were stripped.
472 This will produce results identical
473 to running @command{grep} on a Unix machine.
474 This option has no effect unless the @option{-b} option is also used;
475 it has no effect on platforms other than MS-DOS and MS-Windows.
481 @cindex zero-terminated file names
482 Output a zero byte (the ASCII NUL character)
483 instead of the character that normally follows a file name.
485 @samp{grep -lZ} outputs a zero byte after each file name
486 instead of the usual newline.
487 This option makes the output unambiguous,
488 even in the presence of file names containing unusual characters like newlines.
489 This option can be used with commands like
490 @samp{find -print0}, @samp{perl -0}, @samp{sort -z}, and @samp{xargs -0}
491 to process arbitrary file names,
492 even those that contain newline characters.
496 @node Context Line Control
497 @subsection Context Line Control
499 Regardless of how these options are set,
500 @command{grep} will never print any given line more than once.
501 If the @option{-o} (@option{--only-matching}) option is specified,
502 these options have no effect and a warning is given upon their use.
507 @itemx --after-context=@var{num}
509 @opindex --after-context
510 @cindex after context
511 @cindex context lines, after match
512 Print @var{num} lines of trailing context after matching lines.
515 @itemx --before-context=@var{num}
517 @opindex --before-context
518 @cindex before context
519 @cindex context lines, before match
520 Print @var{num} lines of leading context before matching lines.
524 @itemx --context=@var{num}
529 Print @var{num} lines of leading and trailing output context.
531 @item --group-separator=@var{string}
532 @opindex --group-separator
533 @cindex group separator
534 When @option{-A}, @option{-B} or @option{-C} are in use,
535 print @var{string} instead of @option{--} between groups of lines.
537 @item --no-group-separator
538 @opindex --group-separator
539 @cindex group separator
540 When @option{-A}, @option{-B} or @option{-C} are in use,
541 do not print a separator between groups of lines.
545 Here are some points about how @command{grep} chooses
546 the separator to print between prefix fields and line content:
550 Matching lines normally use @samp{:} as a separator
551 between prefix fields and actual line content.
554 Context (i.e., non-matching) lines use @samp{-} instead.
557 When context is not specified,
558 matching lines are simply output one right after another.
561 When context is specified,
562 lines that are adjacent in the input form a group
563 and are output one right after another, while
564 by default a separator appears between non-adjacent groups.
567 The default separator
568 is a @samp{--} line; its presence and appearance
569 can be changed with the options above.
572 Each group may contain
573 several matching lines when they are close enough to each other
574 that two adjacent groups connect and can merge into a single
578 @node File and Directory Selection
579 @subsection File and Directory Selection
587 @cindex suppress binary data
589 Process a binary file as if it were text;
590 this is equivalent to the @samp{--binary-files=text} option.
592 @item --binary-files=@var{type}
593 @opindex --binary-files
595 If a file's allocation metadata,
596 or if its data read before a line is selected for output,
597 indicate that the file contains binary data,
598 assume that the file is of type @var{type}.
599 Non-text bytes indicate binary data; these are either data bytes
600 improperly encoded for the current locale, or null bytes when the
601 @option{-z} (@option{--null-data}) option is not given (@pxref{Other
604 By default, @var{type} is @samp{binary},
605 and @command{grep} normally outputs either
606 a one-line message saying that a binary file matches,
607 or no message if there is no match.
608 When processing binary data, @command{grep} may treat non-text bytes
609 as line terminators; for example, the pattern @samp{.} (period) might
610 not match a null byte, as the null byte might be treated as a line
611 terminator even without the @option{-z} (@option{--null-data}) option.
613 If @var{type} is @samp{without-match},
614 @command{grep} assumes that a binary file does not match;
615 this is equivalent to the @option{-I} option.
617 If @var{type} is @samp{text},
618 @command{grep} processes a binary file as if it were text;
619 this is equivalent to the @option{-a} option.
621 @emph{Warning:} @samp{--binary-files=text} might output binary garbage,
622 which can have nasty side effects
623 if the output is a terminal and
624 if the terminal driver interprets some of it as commands.
626 @item -D @var{action}
627 @itemx --devices=@var{action}
630 @cindex device search
631 If an input file is a device, FIFO, or socket, use @var{action} to process it.
632 If @var{action} is @samp{read},
633 all devices are read just as if they were ordinary files.
634 If @var{action} is @samp{skip},
635 devices, FIFOs, and sockets are silently skipped.
636 By default, devices are read if they are on the command line or if the
637 @option{-R} (@option{--dereference-recursive}) option is used, and are
638 skipped if they are encountered recursively and the @option{-r}
639 (@option{--recursive}) option is used.
640 This option has no effect on a file that is read via standard input.
642 @item -d @var{action}
643 @itemx --directories=@var{action}
645 @opindex --directories
646 @cindex directory search
647 @cindex symbolic links
648 If an input file is a directory, use @var{action} to process it.
649 By default, @var{action} is @samp{read},
650 which means that directories are read just as if they were ordinary files
651 (some operating systems and file systems disallow this,
652 and will cause @command{grep}
653 to print error messages for every directory or silently skip them).
654 If @var{action} is @samp{skip}, directories are silently skipped.
655 If @var{action} is @samp{recurse},
656 @command{grep} reads all files under each directory, recursively,
657 following command-line symbolic links and skipping other symlinks;
658 this is equivalent to the @option{-r} option.
660 @item --exclude=@var{glob}
662 @cindex exclude files
663 @cindex searching directory trees
664 Skip files whose name matches the pattern @var{glob}, using wildcard
665 matching. When searching recursively, skip any subfile whose base
666 name matches @var{glob}; the base name is the part after the last
667 @samp{/}. A pattern can use
668 @samp{*}, @samp{?}, and @samp{[}...@samp{]} as wildcards,
669 and @code{\} to quote a wildcard or backslash character literally.
671 @item --exclude-from=@var{file}
672 @opindex --exclude-from
673 @cindex exclude files
674 @cindex searching directory trees
675 Skip files whose name matches any of the patterns
676 read from @var{file} (using wildcard matching as described
677 under @option{--exclude}).
679 @item --exclude-dir=@var{glob}
680 @opindex --exclude-dir
681 @cindex exclude directories
682 Skip any directory whose name matches the pattern @var{glob}. When
683 searching recursively, skip any subdirectory whose base name matches
684 @var{glob}. Ignore any redundant trailing slashes in @var{glob}.
687 Process a binary file as if it did not contain matching data;
688 this is equivalent to the @samp{--binary-files=without-match} option.
690 @item --include=@var{glob}
692 @cindex include files
693 @cindex searching directory trees
694 Search only files whose name matches @var{glob},
695 using wildcard matching as described under @option{--exclude}.
701 @cindex recursive search
702 @cindex searching directory trees
703 @cindex symbolic links
704 For each directory operand,
705 read and process all files in that directory, recursively.
706 Follow symbolic links on the command line, but skip symlinks
707 that are encountered recursively.
708 Note that if no file operand is given, grep searches the working directory.
709 This is the same as the @samp{--directories=recurse} option.
712 @itemx --dereference-recursive
714 @opindex --dereference-recursive
715 @cindex recursive search
716 @cindex searching directory trees
717 @cindex symbolic links
718 For each directory operand, read and process all files in that
719 directory, recursively, following all symbolic links.
724 @subsection Other Options
728 @item --line-buffered
729 @opindex --line-buffered
730 @cindex line buffering
731 Use line buffering on output.
732 This can cause a performance penalty.
738 @cindex MS-DOS/MS-Windows binary files
739 @cindex binary files, MS-DOS/MS-Windows
740 Treat the file(s) as binary.
741 By default, under MS-DOS and MS-Windows,
742 @command{grep} guesses whether a file is text or binary
743 as described for the @option{--binary-files} option.
744 If @command{grep} decides the file is a text file,
745 it strips carriage returns from the original file contents
746 (to make regular expressions with @code{^} and @code{$} work correctly).
747 Specifying @option{-U} overrules this guesswork,
748 causing all files to be read and passed to the matching mechanism verbatim;
749 if the file is a text file with @code{CR/LF} pairs at the end of each line,
750 this will cause some regular expressions to fail.
751 This option has no effect
752 on platforms other than MS-DOS and MS-Windows.
758 @cindex zero-terminated lines
759 Treat the input as a set of lines, each terminated by a zero byte (the
760 ASCII NUL character) instead of a newline.
761 Like the @option{-Z} or @option{--null} option,
762 this option can be used with commands like
763 @samp{sort -z} to process arbitrary file names.
767 @node Environment Variables
768 @section Environment Variables
770 The behavior of @command{grep} is affected
771 by the following environment variables.
773 @vindex LANGUAGE @r{environment variable}
774 @vindex LC_ALL @r{environment variable}
775 @vindex LC_MESSAGES @r{environment variable}
776 @vindex LANG @r{environment variable}
777 The locale for category @w{@code{LC_@var{foo}}}
778 is specified by examining the three environment variables
779 @env{LC_ALL}, @w{@env{LC_@var{foo}}}, and @env{LANG},
781 The first of these variables that is set specifies the locale.
782 For example, if @env{LC_ALL} is not set,
783 but @env{LC_COLLATE} is set to @samp{pt_BR},
784 then the Brazilian Portuguese locale is used
785 for the @env{LC_COLLATE} category.
786 As a special case for @env{LC_MESSAGES} only, the environment variable
787 @env{LANGUAGE} can contain a colon-separated list of languages that
788 overrides the three environment variables that ordinarily specify
789 the @env{LC_MESSAGES} category.
790 The @samp{C} locale is used if none of these environment variables are set,
791 if the locale catalog is not installed,
792 or if @command{grep} was not compiled
793 with national language support (NLS).
795 Many of the environment variables in the following list let you
796 control highlighting using
797 Select Graphic Rendition (SGR)
798 commands interpreted by the terminal or terminal emulator.
801 in the documentation of your text terminal
802 for permitted values and their meanings as character attributes.)
803 These substring values are integers in decimal representation
804 and can be concatenated with semicolons.
805 @command{grep} takes care of assembling the result
806 into a complete SGR sequence (@samp{\33[}...@samp{m}).
807 Common values to concatenate include
809 @samp{4} for underline,
811 @samp{7} for inverse,
812 @samp{39} for default foreground color,
813 @samp{30} to @samp{37} for foreground colors,
814 @samp{90} to @samp{97} for 16-color mode foreground colors,
815 @samp{38;5;0} to @samp{38;5;255}
816 for 88-color and 256-color modes foreground colors,
817 @samp{49} for default background color,
818 @samp{40} to @samp{47} for background colors,
819 @samp{100} to @samp{107} for 16-color mode background colors,
820 and @samp{48;5;0} to @samp{48;5;255}
821 for 88-color and 256-color modes background colors.
823 The two-letter names used in the @env{GREP_COLORS} environment variable
824 (and some of the others) refer to terminal ``capabilities,'' the ability
825 of a terminal to highlight text, or change its color, and so on.
826 These capabilities are stored in an online database and accessed by
827 the @code{terminfo} library.
829 @cindex environment variables
834 @vindex GREP_OPTIONS @r{environment variable}
835 @cindex default options environment variable
836 This variable specifies default options to be placed in front of any
838 As this causes problems when writing portable scripts, this feature
839 will be removed in a future release of @command{grep}, and @command{grep}
840 warns if it is used. Please use an alias or script instead.
841 For example, if @command{grep} is in the directory @samp{/usr/bin} you
842 can prepend @file{$HOME/bin} to your @env{PATH} and create an
843 executable script @file{$HOME/bin/grep} containing the following:
848 exec grep --color=auto --devices=skip "$@@"
852 @vindex GREP_COLOR @r{environment variable}
853 @cindex highlight markers
854 This variable specifies the color used to highlight matched (non-empty) text.
855 It is deprecated in favor of @env{GREP_COLORS}, but still supported.
856 The @samp{mt}, @samp{ms}, and @samp{mc} capabilities of @env{GREP_COLORS}
857 have priority over it.
858 It can only specify the color used to highlight
859 the matching non-empty text in any matching line
860 (a selected line when the @option{-v} command-line option is omitted,
861 or a context line when @option{-v} is specified).
862 The default is @samp{01;31},
863 which means a bold red foreground text on the terminal's default background.
866 @vindex GREP_COLORS @r{environment variable}
867 @cindex highlight markers
868 This variable specifies the colors and other attributes
869 used to highlight various parts of the output.
870 Its value is a colon-separated list of @code{terminfo} capabilities
871 that defaults to @samp{ms=01;31:mc=01;31:sl=:cx=:fn=35:ln=32:bn=32:se=36}
872 with the @samp{rv} and @samp{ne} boolean capabilities omitted (i.e., false).
873 Supported capabilities are as follows.
877 @vindex sl GREP_COLORS @r{capability}
878 SGR substring for whole selected lines
880 matching lines when the @option{-v} command-line option is omitted,
881 or non-matching lines when @option{-v} is specified).
882 If however the boolean @samp{rv} capability
883 and the @option{-v} command-line option are both specified,
884 it applies to context matching lines instead.
885 The default is empty (i.e., the terminal's default color pair).
888 @vindex cx GREP_COLORS @r{capability}
889 SGR substring for whole context lines
891 non-matching lines when the @option{-v} command-line option is omitted,
892 or matching lines when @option{-v} is specified).
893 If however the boolean @samp{rv} capability
894 and the @option{-v} command-line option are both specified,
895 it applies to selected non-matching lines instead.
896 The default is empty (i.e., the terminal's default color pair).
899 @vindex rv GREP_COLORS @r{capability}
900 Boolean value that reverses (swaps) the meanings of
901 the @samp{sl=} and @samp{cx=} capabilities
902 when the @option{-v} command-line option is specified.
903 The default is false (i.e., the capability is omitted).
906 @vindex mt GREP_COLORS @r{capability}
907 SGR substring for matching non-empty text in any matching line
909 a selected line when the @option{-v} command-line option is omitted,
910 or a context line when @option{-v} is specified).
911 Setting this is equivalent to setting both @samp{ms=} and @samp{mc=}
912 at once to the same value.
913 The default is a bold red text foreground over the current line background.
916 @vindex ms GREP_COLORS @r{capability}
917 SGR substring for matching non-empty text in a selected line.
918 (This is used only when the @option{-v} command-line option is omitted.)
919 The effect of the @samp{sl=} (or @samp{cx=} if @samp{rv}) capability
920 remains active when this takes effect.
921 The default is a bold red text foreground over the current line background.
924 @vindex mc GREP_COLORS @r{capability}
925 SGR substring for matching non-empty text in a context line.
926 (This is used only when the @option{-v} command-line option is specified.)
927 The effect of the @samp{cx=} (or @samp{sl=} if @samp{rv}) capability
928 remains active when this takes effect.
929 The default is a bold red text foreground over the current line background.
932 @vindex fn GREP_COLORS @r{capability}
933 SGR substring for file names prefixing any content line.
934 The default is a magenta text foreground over the terminal's default background.
937 @vindex ln GREP_COLORS @r{capability}
938 SGR substring for line numbers prefixing any content line.
939 The default is a green text foreground over the terminal's default background.
942 @vindex bn GREP_COLORS @r{capability}
943 SGR substring for byte offsets prefixing any content line.
944 The default is a green text foreground over the terminal's default background.
947 @vindex fn GREP_COLORS @r{capability}
948 SGR substring for separators that are inserted
949 between selected line fields (@samp{:}),
950 between context line fields (@samp{-}),
951 and between groups of adjacent lines
952 when nonzero context is specified (@samp{--}).
953 The default is a cyan text foreground over the terminal's default background.
956 @vindex ne GREP_COLORS @r{capability}
957 Boolean value that prevents clearing to the end of line
958 using Erase in Line (EL) to Right (@samp{\33[K})
959 each time a colorized item ends.
960 This is needed on terminals on which EL is not supported.
961 It is otherwise useful on terminals
962 for which the @code{back_color_erase}
963 (@code{bce}) boolean @code{terminfo} capability does not apply,
964 when the chosen highlight colors do not affect the background,
965 or when EL is too slow or causes too much flicker.
966 The default is false (i.e., the capability is omitted).
969 Note that boolean capabilities have no @samp{=}... part.
970 They are omitted (i.e., false) by default and become true when specified.
976 @vindex LC_ALL @r{environment variable}
977 @vindex LC_COLLATE @r{environment variable}
978 @vindex LANG @r{environment variable}
979 @cindex character type
980 @cindex national language support
982 These variables specify the locale for the @env{LC_COLLATE} category,
983 which might affect how range expressions like @samp{[a-z]} are
989 @vindex LC_ALL @r{environment variable}
990 @vindex LC_CTYPE @r{environment variable}
991 @vindex LANG @r{environment variable}
992 These variables specify the locale for the @env{LC_CTYPE} category,
993 which determines the type of characters,
994 e.g., which characters are whitespace.
1000 @vindex LANGUAGE @r{environment variable}
1001 @vindex LC_ALL @r{environment variable}
1002 @vindex LC_MESSAGES @r{environment variable}
1003 @vindex LANG @r{environment variable}
1004 @cindex language of messages
1005 @cindex message language
1006 @cindex national language support
1007 @cindex translation of message language
1008 These variables specify the locale for the @env{LC_MESSAGES} category,
1009 which determines the language that @command{grep} uses for messages.
1010 The default @samp{C} locale uses American English messages.
1012 @item POSIXLY_CORRECT
1013 @vindex POSIXLY_CORRECT @r{environment variable}
1014 If set, @command{grep} behaves as POSIX requires; otherwise,
1015 @command{grep} behaves more like other GNU programs.
1017 requires that options that
1018 follow file names must be treated as file names;
1020 such options are permuted to the front of the operand list
1021 and are treated as options.
1022 Also, @env{POSIXLY_CORRECT} disables special handling of an
1023 invalid bracket expression. @xref{invalid-bracket-expr}.
1025 @item _@var{N}_GNU_nonoption_argv_flags_
1026 @vindex _@var{N}_GNU_nonoption_argv_flags_ @r{environment variable}
1027 (Here @code{@var{N}} is @command{grep}'s numeric process ID.)
1028 If the @var{i}th character of this environment variable's value is @samp{1},
1029 do not consider the @var{i}th operand of @command{grep} to be an option,
1030 even if it appears to be one.
1031 A shell can put this variable in the environment for each command it runs,
1032 specifying which operands are the results of file name wildcard expansion
1033 and therefore should not be treated as options.
1034 This behavior is available only with the GNU C library,
1035 and only when @env{POSIXLY_CORRECT} is not set.
1041 @section Exit Status
1043 @cindex return status
1045 Normally the exit status is 0 if a line is selected, 1 if no lines
1046 were selected, and 2 if an error occurred. However, if the
1047 @option{-q} or @option{--quiet} or @option{--silent} option is used
1048 and a line is selected, the exit status is 0 even if an error
1049 occurred. Other @command{grep} implementations may exit with status
1050 greater than 2 on error.
1053 @section @command{grep} Programs
1054 @cindex @command{grep} programs
1055 @cindex variants of @command{grep}
1057 @command{grep} searches the named input files
1058 for lines containing a match to the given pattern.
1059 By default, @command{grep} prints the matching lines.
1060 A file named @file{-} stands for standard input.
1061 If no input is specified, @command{grep} searches the working
1062 directory @file{.} if given a command-line option specifying
1063 recursion; otherwise, @command{grep} searches standard input.
1064 There are four major variants of @command{grep},
1065 controlled by the following options.
1070 @itemx --basic-regexp
1072 @opindex --basic-regexp
1073 @cindex matching basic regular expressions
1074 Interpret the pattern as a basic regular expression (BRE).
1075 This is the default.
1078 @itemx --extended-regexp
1080 @opindex --extended-regexp
1081 @cindex matching extended regular expressions
1082 Interpret the pattern as an extended regular expression (ERE).
1083 (@option{-E} is specified by POSIX.)
1086 @itemx --fixed-strings
1088 @opindex --fixed-strings
1089 @cindex matching fixed strings
1090 Interpret the pattern as a list of fixed strings, separated
1091 by newlines, any of which is to be matched.
1092 (@option{-F} is specified by POSIX.)
1095 @itemx --perl-regexp
1097 @opindex --perl-regexp
1098 @cindex matching Perl regular expressions
1099 Interpret the pattern as a Perl regular expression.
1100 This is highly experimental and
1101 @samp{grep@ -P} may warn of unimplemented features.
1106 two variant programs @command{egrep} and @command{fgrep} are available.
1107 @command{egrep} is the same as @samp{grep@ -E}.
1108 @command{fgrep} is the same as @samp{grep@ -F}.
1109 Direct invocation as either
1110 @command{egrep} or @command{fgrep} is deprecated,
1111 but is provided to allow historical applications
1112 that rely on them to run unmodified.
1115 @node Regular Expressions
1116 @chapter Regular Expressions
1117 @cindex regular expressions
1119 A @dfn{regular expression} is a pattern that describes a set of strings.
1120 Regular expressions are constructed analogously to arithmetic expressions,
1121 by using various operators to combine smaller expressions.
1122 @command{grep} understands
1123 three different versions of regular expression syntax:
1124 ``basic,'' (BRE) ``extended'' (ERE) and ``perl''.
1125 In GNU @command{grep},
1126 there is no difference in available functionality between the basic and
1128 In other implementations, basic regular expressions are less powerful.
1129 The following description applies to extended regular expressions;
1130 differences for basic regular expressions are summarized afterwards.
1131 Perl regular expressions give additional functionality, and are
1132 documented in the @i{pcresyntax}(3) and @i{pcrepattern}(3) manual pages,
1133 but may not be available on every system.
1136 * Fundamental Structure::
1137 * Character Classes and Bracket Expressions::
1138 * The Backslash Character and Special Expressions::
1140 * Back-references and Subexpressions::
1141 * Basic vs Extended::
1144 @node Fundamental Structure
1145 @section Fundamental Structure
1147 The fundamental building blocks are the regular expressions that match
1149 Most characters, including all letters and digits,
1150 are regular expressions that match themselves.
1152 with special meaning may be quoted by preceding it with a backslash.
1154 A regular expression may be followed by one of several
1155 repetition operators:
1163 The period @samp{.} matches any single character.
1167 @cindex question mark
1168 @cindex match expression at most once
1169 The preceding item is optional and will be matched at most once.
1174 @cindex match expression zero or more times
1175 The preceding item will be matched zero or more times.
1180 @cindex match expression one or more times
1181 The preceding item will be matched one or more times.
1184 @opindex @{@var{n}@}
1185 @cindex braces, one argument
1186 @cindex match expression @var{n} times
1187 The preceding item is matched exactly @var{n} times.
1190 @opindex @{@var{n},@}
1191 @cindex braces, second argument omitted
1192 @cindex match expression @var{n} or more times
1193 The preceding item is matched @var{n} or more times.
1196 @opindex @{,@var{m}@}
1197 @cindex braces, first argument omitted
1198 @cindex match expression at most @var{m} times
1199 The preceding item is matched at most @var{m} times.
1200 This is a GNU extension.
1202 @item @{@var{n},@var{m}@}
1203 @opindex @{@var{n},@var{m}@}
1204 @cindex braces, two arguments
1205 @cindex match expression from @var{n} to @var{m} times
1206 The preceding item is matched at least @var{n} times, but not more than
1211 The empty regular expression matches the empty string.
1212 Two regular expressions may be concatenated;
1213 the resulting regular expression
1214 matches any string formed by concatenating two substrings
1215 that respectively match the concatenated expressions.
1217 Two regular expressions may be joined by the infix operator @samp{|};
1218 the resulting regular expression
1219 matches any string matching either alternate expression.
1221 Repetition takes precedence over concatenation,
1222 which in turn takes precedence over alternation.
1223 A whole expression may be enclosed in parentheses
1224 to override these precedence rules and form a subexpression.
1225 An unmatched @samp{)} matches just itself.
1227 @node Character Classes and Bracket Expressions
1228 @section Character Classes and Bracket Expressions
1230 @cindex bracket expression
1231 @cindex character class
1232 A @dfn{bracket expression} is a list of characters enclosed by @samp{[} and
1234 It matches any single character in that list;
1235 if the first character of the list is the caret @samp{^},
1236 then it matches any character @strong{not} in the list.
1237 For example, the regular expression
1238 @samp{[0123456789]} matches any single digit.
1240 @cindex range expression
1241 Within a bracket expression, a @dfn{range expression} consists of two
1242 characters separated by a hyphen.
1243 It matches any single character that
1244 sorts between the two characters, inclusive.
1245 In the default C locale, the sorting sequence is the native character
1246 order; for example, @samp{[a-d]} is equivalent to @samp{[abcd]}.
1247 In other locales, the sorting sequence is not specified, and
1248 @samp{[a-d]} might be equivalent to @samp{[abcd]} or to
1249 @samp{[aBbCcDd]}, or it might fail to match any character, or the set of
1250 characters that it matches might even be erratic.
1251 To obtain the traditional interpretation
1252 of bracket expressions, you can use the @samp{C} locale by setting the
1253 @env{LC_ALL} environment variable to the value @samp{C}.
1255 Finally, certain named classes of characters are predefined within
1256 bracket expressions, as follows.
1257 Their interpretation depends on the @env{LC_CTYPE} locale;
1258 for example, @samp{[[:alnum:]]} means the character class of numbers and letters
1259 in the current locale.
1261 @cindex classes of characters
1262 @cindex character classes
1266 @opindex alnum @r{character class}
1267 @cindex alphanumeric characters
1268 Alphanumeric characters:
1269 @samp{[:alpha:]} and @samp{[:digit:]}; in the @samp{C} locale and ASCII
1270 character encoding, this is the same as @samp{[0-9A-Za-z]}.
1273 @opindex alpha @r{character class}
1274 @cindex alphabetic characters
1275 Alphabetic characters:
1276 @samp{[:lower:]} and @samp{[:upper:]}; in the @samp{C} locale and ASCII
1277 character encoding, this is the same as @samp{[A-Za-z]}.
1280 @opindex blank @r{character class}
1281 @cindex blank characters
1286 @opindex cntrl @r{character class}
1287 @cindex control characters
1289 In ASCII, these characters have octal codes 000
1290 through 037, and 177 (DEL).
1291 In other character sets, these are
1292 the equivalent characters, if any.
1295 @opindex digit @r{character class}
1296 @cindex digit characters
1297 @cindex numeric characters
1298 Digits: @code{0 1 2 3 4 5 6 7 8 9}.
1301 @opindex graph @r{character class}
1302 @cindex graphic characters
1303 Graphical characters:
1304 @samp{[:alnum:]} and @samp{[:punct:]}.
1307 @opindex lower @r{character class}
1308 @cindex lower-case letters
1309 Lower-case letters; in the @samp{C} locale and ASCII character
1311 @code{a b c d e f g h i j k l m n o p q r s t u v w x y z}.
1314 @opindex print @r{character class}
1315 @cindex printable characters
1316 Printable characters:
1317 @samp{[:alnum:]}, @samp{[:punct:]}, and space.
1320 @opindex punct @r{character class}
1321 @cindex punctuation characters
1322 Punctuation characters; in the @samp{C} locale and ASCII character
1324 @code{!@: " # $ % & ' ( ) * + , - .@: / : ; < = > ?@: @@ [ \ ] ^ _ ` @{ | @} ~}.
1327 @opindex space @r{character class}
1328 @cindex space characters
1329 @cindex whitespace characters
1330 Space characters: in the @samp{C} locale, this is
1331 tab, newline, vertical tab, form feed, carriage return, and space.
1332 @xref{Usage}, for more discussion of matching newlines.
1335 @opindex upper @r{character class}
1336 @cindex upper-case letters
1337 Upper-case letters: in the @samp{C} locale and ASCII character
1339 @code{A B C D E F G H I J K L M N O P Q R S T U V W X Y Z}.
1342 @opindex xdigit @r{character class}
1343 @cindex xdigit class
1344 @cindex hexadecimal digits
1346 @code{0 1 2 3 4 5 6 7 8 9 A B C D E F a b c d e f}.
1349 Note that the brackets in these class names are
1350 part of the symbolic names, and must be included in addition to
1351 the brackets delimiting the bracket expression.
1353 @anchor{invalid-bracket-expr}
1354 If you mistakenly omit the outer brackets, and search for say, @samp{[:upper:]},
1355 GNU @command{grep} prints a diagnostic and exits with status 2, on
1356 the assumption that you did not intend to search for the nominally
1357 equivalent regular expression: @samp{[:epru]}.
1358 Set the @env{POSIXLY_CORRECT} environment variable to disable this feature.
1360 Most meta-characters lose their special meaning inside bracket expressions.
1364 ends the bracket expression if it's not the first list item.
1365 So, if you want to make the @samp{]} character a list item,
1366 you must put it first.
1369 represents the open collating symbol.
1372 represents the close collating symbol.
1375 represents the open equivalence class.
1378 represents the close equivalence class.
1381 represents the open character class symbol, and should be followed by a
1382 valid character class name.
1385 represents the close character class symbol.
1388 represents the range if it's not first or last in a list or the ending point
1392 represents the characters not in the list.
1393 If you want to make the @samp{^}
1394 character a list item, place it anywhere but first.
1398 @node The Backslash Character and Special Expressions
1399 @section The Backslash Character and Special Expressions
1402 The @samp{\} character,
1403 when followed by certain ordinary characters,
1404 takes a special meaning:
1409 Match the empty string at the edge of a word.
1412 Match the empty string provided it's not at the edge of a word.
1415 Match the empty string at the beginning of word.
1418 Match the empty string at the end of word.
1421 Match word constituent, it is a synonym for @samp{[_[:alnum:]]}.
1424 Match non-word constituent, it is a synonym for @samp{[^_[:alnum:]]}.
1427 Match whitespace, it is a synonym for @samp{[[:space:]]}.
1430 Match non-whitespace, it is a synonym for @samp{[^[:space:]]}.
1434 For example, @samp{\brat\b} matches the separate word @samp{rat},
1435 @samp{\Brat\B} matches @samp{crate} but not @samp{furry rat}.
1441 The caret @samp{^} and the dollar sign @samp{$} are meta-characters that
1442 respectively match the empty string at the beginning and end of a line.
1443 They are termed @dfn{anchors}, since they force the match to be ``anchored''
1444 to beginning or end of a line, respectively.
1446 @node Back-references and Subexpressions
1447 @section Back-references and Subexpressions
1448 @cindex subexpression
1449 @cindex back-reference
1451 The back-reference @samp{\@var{n}}, where @var{n} is a single digit, matches
1452 the substring previously matched by the @var{n}th parenthesized subexpression
1453 of the regular expression.
1454 For example, @samp{(a)\1} matches @samp{aa}.
1455 When used with alternation, if the group does not participate in the match then
1456 the back-reference makes the whole match fail.
1457 For example, @samp{a(.)|b\1}
1458 will not match @samp{ba}.
1459 When multiple regular expressions are given with
1460 @option{-e} or from a file (@samp{-f @var{file}}),
1461 back-references are local to each expression.
1463 @node Basic vs Extended
1464 @section Basic vs Extended Regular Expressions
1465 @cindex basic regular expressions
1467 In basic regular expressions the meta-characters @samp{?}, @samp{+},
1468 @samp{@{}, @samp{|}, @samp{(}, and @samp{)} lose their special meaning;
1469 instead use the backslashed versions @samp{\?}, @samp{\+}, @samp{\@{},
1470 @samp{\|}, @samp{\(}, and @samp{\)}.
1472 @cindex interval specifications
1473 Traditional @command{egrep} did not support the @samp{@{} meta-character,
1474 and some @command{egrep} implementations support @samp{\@{} instead, so
1475 portable scripts should avoid @samp{@{} in @samp{grep@ -E} patterns and
1476 should use @samp{[@{]} to match a literal @samp{@{}.
1478 GNU @command{grep@ -E} attempts to support traditional usage by
1479 assuming that @samp{@{} is not special if it would be the start of an
1480 invalid interval specification.
1481 For example, the command
1482 @samp{grep@ -E@ '@{1'} searches for the two-character string @samp{@{1}
1483 instead of reporting a syntax error in the regular expression.
1484 POSIX allows this behavior as an extension, but portable scripts
1491 @cindex usage, examples
1492 Here is an example command that invokes GNU @command{grep}:
1495 grep -i 'hello.*world' menu.h main.c
1499 This lists all lines in the files @file{menu.h} and @file{main.c} that
1500 contain the string @samp{hello} followed by the string @samp{world};
1501 this is because @samp{.*} matches zero or more characters within a line.
1502 @xref{Regular Expressions}.
1503 The @option{-i} option causes @command{grep}
1504 to ignore case, causing it to match the line @samp{Hello, world!}, which
1505 it would not otherwise match.
1506 @xref{Invoking}, for more details about
1507 how to invoke @command{grep}.
1509 @cindex using @command{grep}, Q&A
1510 @cindex FAQ about @command{grep} usage
1511 Here are some common questions and answers about @command{grep} usage.
1516 How can I list just the names of matching files?
1523 lists the names of all C files in the current directory whose contents
1524 mention @samp{main}.
1527 How do I search directories recursively?
1530 grep -r 'hello' /home/gigi
1534 searches for @samp{hello} in all files
1535 under the @file{/home/gigi} directory.
1536 For more control over which files are searched,
1537 use @command{find}, @command{grep}, and @command{xargs}.
1538 For example, the following command searches only C files:
1541 find /home/gigi -name '*.c' -print0 | xargs -0r grep -H 'hello'
1544 This differs from the command:
1550 which merely looks for @samp{hello} in all files in the current
1551 directory whose names end in @samp{.c}.
1552 The @samp{find ...} command line above is more similar to the command:
1555 grep -rH --include='*.c' 'hello' /home/gigi
1559 What if a pattern has a leading @samp{-}?
1562 grep -e '--cut here--' *
1566 searches for all lines matching @samp{--cut here--}.
1567 Without @option{-e},
1568 @command{grep} would attempt to parse @samp{--cut here--} as a list of
1572 Suppose I want to search for a whole word, not a part of a word?
1579 searches only for instances of @samp{hello} that are entire words;
1580 it does not match @samp{Othello}.
1581 For more control, use @samp{\<} and
1582 @samp{\>} to match the start and end of words.
1590 searches only for words ending in @samp{hello}, so it matches the word
1594 How do I output context around the matching lines?
1601 prints two lines of context around each matching line.
1604 How do I force @command{grep} to print the name of the file?
1606 Append @file{/dev/null}:
1609 grep 'eli' /etc/passwd /dev/null
1615 /etc/passwd:eli:x:2098:1000:Eli Smith:/home/eli:/bin/bash
1618 Alternatively, use @option{-H}, which is a GNU extension:
1621 grep -H 'eli' /etc/passwd
1625 Why do people use strange regular expressions on @command{ps} output?
1628 ps -ef | grep '[c]ron'
1631 If the pattern had been written without the square brackets, it would
1632 have matched not only the @command{ps} output line for @command{cron},
1633 but also the @command{ps} output line for @command{grep}.
1634 Note that on some platforms,
1635 @command{ps} limits the output to the width of the screen;
1636 @command{grep} does not have any limit on the length of a line
1637 except the available memory.
1640 Why does @command{grep} report ``Binary file matches''?
1642 If @command{grep} listed all matching ``lines'' from a binary file, it
1643 would probably generate output that is not useful, and it might even
1644 muck up your display.
1645 So GNU @command{grep} suppresses output from
1646 files that appear to be binary files.
1647 To force GNU @command{grep}
1648 to output lines even from files that appear to be binary, use the
1649 @option{-a} or @samp{--binary-files=text} option.
1651 ``Binary file matches'' messages, use the @option{-I} or
1652 @samp{--binary-files=without-match} option.
1655 Why doesn't @samp{grep -lv} print non-matching file names?
1657 @samp{grep -lv} lists the names of all files containing one or more
1658 lines that do not match.
1659 To list the names of all files that contain no
1660 matching lines, use the @option{-L} or @option{--files-without-match}
1664 I can do ``OR'' with @samp{|}, but what about ``AND''?
1667 grep 'paul' /etc/motd | grep 'franc,ois'
1671 finds all lines that contain both @samp{paul} and @samp{franc,ois}.
1674 Why does the empty pattern match every input line?
1676 The @command{grep} command searches for lines that contain strings
1677 that match a pattern. Every line contains the empty string, so an
1678 empty pattern causes @command{grep} to find a match on each line. It
1679 is not the only such pattern: @samp{^}, @samp{$}, @samp{.*}, and many
1680 other patterns cause @command{grep} to match every line.
1682 To match empty lines, use the pattern @samp{^$}. To match blank
1683 lines, use the pattern @samp{^[[:blank:]]*$}. To match no lines at
1684 all, use the command @samp{grep -f /dev/null}.
1687 How can I search in both standard input and in files?
1689 Use the special file name @samp{-}:
1692 cat /etc/passwd | grep 'alain' - /etc/motd
1697 How to express palindromes in a regular expression?
1699 It can be done by using back-references;
1701 a palindrome of 4 characters can be written with a BRE:
1704 grep -w -e '\(.\)\(.\).\2\1' file
1707 It matches the word ``radar'' or ``civic.''
1709 Guglielmo Bondioni proposed a single RE
1710 that finds all palindromes up to 19 characters long
1711 using @w{9 subexpressions} and @w{9 back-references}:
1714 grep -E -e '^(.?)(.?)(.?)(.?)(.?)(.?)(.?)(.?)(.?).?\9\8\7\6\5\4\3\2\1$' file
1717 Note this is done by using GNU ERE extensions;
1718 it might not be portable to other implementations of @command{grep}.
1721 Why is this back-reference failing?
1724 echo 'ba' | grep -E '(a)\1|b\1'
1727 This gives no output, because the first alternate @samp{(a)\1} does not match,
1728 as there is no @samp{aa} in the input, so the @samp{\1} in the second alternate
1729 has nothing to refer back to, meaning it will never match anything.
1730 (The second alternate in this example can only match
1731 if the first alternate has matched---making the second one superfluous.)
1734 How can I match across lines?
1736 Standard grep cannot do this, as it is fundamentally line-based.
1737 Therefore, merely using the @code{[:space:]} character class does not
1738 match newlines in the way you might expect.
1740 With the GNU @command{grep} option @option{-z} (@option{--null-data}), each
1741 input ``line'' is terminated by a null byte; @pxref{Other Options}. Thus,
1742 you can match newlines in the input, but typically if there is a match
1743 the entire input is output, so this usage is often combined with
1744 output-suppressing options like @option{-q}, e.g.:
1747 printf 'foo\nbar\n' | grep -z -q 'foo[[:space:]]\+bar'
1750 If this does not suffice, you can transform the input
1751 before giving it to @command{grep}, or turn to @command{awk},
1752 @command{sed}, @command{perl}, or many other utilities that are
1753 designed to operate across lines.
1756 What do @command{grep}, @command{fgrep}, and @command{egrep} stand for?
1758 The name @command{grep} comes from the way line editing was done on Unix.
1760 @command{ed} uses the following syntax
1761 to print a list of matching lines on the screen:
1764 global/regular expression/print
1768 @command{fgrep} stands for Fixed @command{grep};
1769 @command{egrep} stands for Extended @command{grep}.
1774 @node Reporting Bugs
1775 @chapter Reporting bugs
1777 @cindex bugs, reporting
1778 Bug reports can be found at the
1779 @url{http://debbugs.gnu.org/cgi/pkgreport.cgi?package=grep,
1780 GNU bug report logs for @command{grep}}.
1781 If you find a bug not listed there, please email it to
1782 @email{bug-grep@@gnu.org} to create a new bug report.
1787 Large repetition counts in the @samp{@{n,m@}} construct may cause
1788 @command{grep} to use lots of memory.
1789 In addition, certain other
1790 obscure regular expressions require exponential time and
1791 space, and may cause @command{grep} to run out of memory.
1793 Back-references are very slow, and may require exponential time.
1800 GNU @command{grep} is licensed under the GNU GPL, which makes it @dfn{free
1803 The ``free'' in ``free software'' refers to liberty, not price. As
1804 some GNU project advocates like to point out, think of ``free speech''
1805 rather than ``free beer''. In short, you have the right (freedom) to
1806 run and change @command{grep} and distribute it to other people, and---if you
1807 want---charge money for doing either. The important restriction is
1808 that you have to grant your recipients the same rights and impose the
1811 This general method of licensing software is sometimes called
1812 @dfn{open source}. The GNU project prefers the term ``free software''
1813 for reasons outlined at
1814 @url{http://www.gnu.org/philosophy/open-source-misses-the-point.html}.
1816 This manual is free documentation in the same sense. The
1817 documentation license is included below. The license for the program
1818 is available with the source code, or at
1819 @url{http://www.gnu.org/licenses/gpl.html}.
1822 * GNU Free Documentation License::
1825 @node GNU Free Documentation License
1826 @section GNU Free Documentation License