1 This is diffutils.info, produced by makeinfo version 6.5 from
4 This manual is for GNU Diffutils (version 3.7, 7 January 2018), and
5 documents the GNU 'diff', 'diff3', 'sdiff', and 'cmp' commands for
6 showing the differences between files and the GNU 'patch' command for
7 using their output to update files.
9 Copyright (C) 1992-1994, 1998, 2001-2002, 2004, 2006, 2009-2018 Free
10 Software Foundation, Inc.
12 Permission is granted to copy, distribute and/or modify this
13 document under the terms of the GNU Free Documentation License,
14 Version 1.3 or any later version published by the Free Software
15 Foundation; with no Invariant Sections, no Front-Cover Texts, and
16 no Back-Cover Texts. A copy of the license is included in the
17 section entitled "GNU Free Documentation License."
18 INFO-DIR-SECTION Individual utilities
20 * cmp: (diffutils)Invoking cmp. Compare 2 files byte by byte.
21 * diff: (diffutils)Invoking diff. Compare 2 files line by line.
22 * diff3: (diffutils)Invoking diff3. Compare 3 files line by line.
23 * patch: (diffutils)Invoking patch. Apply a patch to a file.
24 * sdiff: (diffutils)Invoking sdiff. Merge 2 files side-by-side.
27 INFO-DIR-SECTION Text creation and manipulation
29 * Diffutils: (diffutils). Comparing and merging files.
33 File: diffutils.info, Node: Top, Next: Overview, Up: (dir)
35 Comparing and Merging Files
36 ***************************
38 This manual is for GNU Diffutils (version 3.7, 7 January 2018), and
39 documents the GNU 'diff', 'diff3', 'sdiff', and 'cmp' commands for
40 showing the differences between files and the GNU 'patch' command for
41 using their output to update files.
43 Copyright (C) 1992-1994, 1998, 2001-2002, 2004, 2006, 2009-2018 Free
44 Software Foundation, Inc.
46 Permission is granted to copy, distribute and/or modify this
47 document under the terms of the GNU Free Documentation License,
48 Version 1.3 or any later version published by the Free Software
49 Foundation; with no Invariant Sections, no Front-Cover Texts, and
50 no Back-Cover Texts. A copy of the license is included in the
51 section entitled "GNU Free Documentation License."
55 * Overview:: Preliminary information.
56 * Comparison:: What file comparison means.
58 * Output Formats:: Formats for two-way difference reports.
59 * Incomplete Lines:: Lines that lack trailing newlines.
60 * Comparing Directories:: Comparing files and directories.
61 * Adjusting Output:: Making 'diff' output prettier.
62 * diff Performance:: Making 'diff' smarter or faster.
64 * Comparing Three Files:: Formats for three-way difference reports.
65 * diff3 Merging:: Merging from a common ancestor.
67 * Interactive Merging:: Interactive merging with 'sdiff'.
69 * Merging with patch:: Using 'patch' to change old files into new ones.
70 * Making Patches:: Tips for making and using patch distributions.
72 * Invoking cmp:: Compare two files byte by byte.
73 * Invoking diff:: Compare two files line by line.
74 * Invoking diff3:: Compare three files line by line.
75 * Invoking patch:: Apply a diff file to an original.
76 * Invoking sdiff:: Side-by-side merge of file differences.
78 * Standards conformance:: Conformance to the POSIX standard.
79 * Projects:: If you've found a bug or other shortcoming.
81 * Copying This Manual:: How to make copies of this manual.
82 * Translations:: Available translations of this manual.
86 File: diffutils.info, Node: Overview, Next: Comparison, Prev: Top, Up: Top
91 Computer users often find occasion to ask how two files differ. Perhaps
92 one file is a newer version of the other file. Or maybe the two files
93 started out as identical copies but were changed by different people.
95 You can use the 'diff' command to show differences between two files,
96 or each corresponding file in two directories. 'diff' outputs
97 differences between files line by line in any of several formats,
98 selectable by command line options. This set of differences is often
99 called a "diff" or "patch". For files that are identical, 'diff'
100 normally produces no output; for binary (non-text) files, 'diff'
101 normally reports only that they are different.
103 You can use the 'cmp' command to show the byte and line numbers where
104 two files differ. 'cmp' can also show all the bytes that differ between
105 the two files, side by side. A way to compare two files character by
106 character is the Emacs command 'M-x compare-windows'. *Note Other
107 Window: (emacs)Other Window, for more information on that command.
109 You can use the 'diff3' command to show differences among three
110 files. When two people have made independent changes to a common
111 original, 'diff3' can report the differences between the original and
112 the two changed versions, and can produce a merged file that contains
113 both persons' changes together with warnings about conflicts.
115 You can use the 'sdiff' command to merge two files interactively.
117 You can use the set of differences produced by 'diff' to distribute
118 updates to text files (such as program source code) to other people.
119 This method is especially useful when the differences are small compared
120 to the complete files. Given 'diff' output, you can use the 'patch'
121 program to update, or "patch", a copy of the file. If you think of
122 'diff' as subtracting one file from another to produce their difference,
123 you can think of 'patch' as adding the difference to one file to
126 This manual first concentrates on making diffs, and later shows how
127 to use diffs to update files.
129 GNU 'diff' was written by Paul Eggert, Mike Haertel, David Hayes,
130 Richard Stallman, and Len Tower. Wayne Davison designed and implemented
131 the unified output format. The basic algorithm is described by Eugene
132 W. Myers in "An O(ND) Difference Algorithm and its Variations",
133 'Algorithmica' Vol. 1, 1986, pp. 251-266,
134 <http://dx.doi.org/10.1007/BF01840446>; and in "A File Comparison
135 Program", Webb Miller and Eugene W. Myers, 'Software--Practice and
136 Experience' Vol. 15, 1985, pp. 1025-1040,
137 <http://dx.doi.org/10.1002/spe.4380151102>. The algorithm was
138 independently discovered as described by Esko Ukkonen in "Algorithms for
139 Approximate String Matching", 'Information and Control' Vol. 64, 1985,
140 pp. 100-118, <http://dx.doi.org/10.1016/S0019-9958(85)80046-2>. Unless
141 the '--minimal' option is used, 'diff' uses a heuristic by Paul Eggert
142 that limits the cost to O(N^1.5 log N) at the price of producing
143 suboptimal output for large inputs with many differences. Related
144 algorithms are surveyed by Alfred V. Aho in section 6.3 of "Algorithms
145 for Finding Patterns in Strings", 'Handbook of Theoretical Computer
146 Science' (Jan Van Leeuwen, ed.), Vol. A, 'Algorithms and Complexity',
147 Elsevier/MIT Press, 1990, pp. 255-300.
149 GNU 'diff3' was written by Randy Smith. GNU 'sdiff' was written by
150 Thomas Lord. GNU 'cmp' was written by Torbjo"rn Granlund and David
153 GNU 'patch' was written mainly by Larry Wall and Paul Eggert; several
154 GNU enhancements were contributed by Wayne Davison and David MacKenzie.
155 Parts of this manual are adapted from a manual page written by Larry
156 Wall, with his permission.
159 File: diffutils.info, Node: Comparison, Next: Output Formats, Prev: Overview, Up: Top
161 1 What Comparison Means
162 ***********************
164 There are several ways to think about the differences between two files.
165 One way to think of the differences is as a series of lines that were
166 deleted from, inserted in, or changed in one file to produce the other
167 file. 'diff' compares two files line by line, finds groups of lines
168 that differ, and reports each group of differing lines. It can report
169 the differing lines in several formats, which have different purposes.
171 GNU 'diff' can show whether files are different without detailing the
172 differences. It also provides ways to suppress certain kinds of
173 differences that are not important to you. Most commonly, such
174 differences are changes in the amount of white space between words or
175 lines. 'diff' also provides ways to suppress differences in alphabetic
176 case or in lines that match a regular expression that you provide.
177 These options can accumulate; for example, you can ignore changes in
178 both white space and alphabetic case.
180 Another way to think of the differences between two files is as a
181 sequence of pairs of bytes that can be either identical or different.
182 'cmp' reports the differences between two files byte by byte, instead of
183 line by line. As a result, it is often more useful than 'diff' for
184 comparing binary files. For text files, 'cmp' is useful mainly when you
185 want to know only whether two files are identical, or whether one file
186 is a prefix of the other.
188 To illustrate the effect that considering changes byte by byte can
189 have compared with considering them line by line, think of what happens
190 if a single newline character is added to the beginning of a file. If
191 that file is then compared with an otherwise identical file that lacks
192 the newline at the beginning, 'diff' will report that a blank line has
193 been added to the file, while 'cmp' will report that almost every byte
194 of the two files differs.
196 'diff3' normally compares three input files line by line, finds
197 groups of lines that differ, and reports each group of differing lines.
198 Its output is designed to make it easy to inspect two different sets of
199 changes to the same file.
201 These commands compare input files without necessarily reading them.
202 For example, if 'diff' is asked simply to report whether two files
203 differ, and it discovers that the files have different sizes, it need
204 not read them to do its job.
208 * Hunks:: Groups of differing lines.
209 * White Space:: Suppressing differences in white space.
210 * Blank Lines:: Suppressing differences whose lines are all blank.
211 * Specified Lines:: Suppressing differences whose lines all match a pattern.
212 * Case Folding:: Suppressing differences in alphabetic case.
213 * Brief:: Summarizing which files are different.
214 * Binary:: Comparing binary files or forcing text comparisons.
217 File: diffutils.info, Node: Hunks, Next: White Space, Up: Comparison
222 When comparing two files, 'diff' finds sequences of lines common to both
223 files, interspersed with groups of differing lines called "hunks".
224 Comparing two identical files yields one sequence of common lines and no
225 hunks, because no lines differ. Comparing two entirely different files
226 yields no common lines and one large hunk that contains all lines of
227 both files. In general, there are many ways to match up lines between
228 two given files. 'diff' tries to minimize the total hunk size by
229 finding large sequences of common lines interspersed with small hunks of
232 For example, suppose the file 'F' contains the three lines 'a', 'b',
233 'c', and the file 'G' contains the same three lines in reverse order
234 'c', 'b', 'a'. If 'diff' finds the line 'c' as common, then the command
235 'diff F G' produces this output:
244 But if 'diff' notices the common line 'b' instead, it produces this
256 It is also possible to find 'a' as the common line. 'diff' does not
257 always find an optimal matching between the files; it takes shortcuts to
258 run faster. But its output is usually close to the shortest possible.
259 You can adjust this tradeoff with the '--minimal' ('-d') option (*note
263 File: diffutils.info, Node: White Space, Next: Blank Lines, Prev: Hunks, Up: Comparison
265 1.2 Suppressing Differences in Blank and Tab Spacing
266 ====================================================
268 The '--ignore-tab-expansion' ('-E') option ignores the distinction
269 between tabs and spaces on input. A tab is considered to be equivalent
270 to the number of spaces to the next tab stop (*note Tabs::).
272 The '--ignore-trailing-space' ('-Z') option ignores white space at
275 The '--ignore-space-change' ('-b') option is stronger than '-E' and
276 '-Z' combined. It ignores white space at line end, and considers all
277 other sequences of one or more white space characters within a line to
278 be equivalent. With this option, 'diff' considers the following two
279 lines to be equivalent, where '$' denotes the line end:
281 Here lyeth muche rychnesse in lytell space. -- John Heywood$
282 Here lyeth muche rychnesse in lytell space. -- John Heywood $
284 The '--ignore-all-space' ('-w') option is stronger still. It ignores
285 differences even if one line has white space where the other line has
286 none. "White space" characters include tab, vertical tab, form feed,
287 carriage return, and space; some locales may define additional
288 characters to be white space. With this option, 'diff' considers the
289 following two lines to be equivalent, where '$' denotes the line end and
290 '^M' denotes a carriage return:
292 Here lyeth muche rychnesse in lytell space.-- John Heywood$
293 He relyeth much erychnes seinly tells pace. --John Heywood ^M$
295 For many other programs newline is also a white space character, but
296 'diff' is a line-oriented program and a newline character always ends a
297 line. Hence the '-w' or '--ignore-all-space' option does not ignore
298 newline-related changes; it ignores only other white space changes.
301 File: diffutils.info, Node: Blank Lines, Next: Specified Lines, Prev: White Space, Up: Comparison
303 1.3 Suppressing Differences Whose Lines Are All Blank
304 =====================================================
306 The '--ignore-blank-lines' ('-B') option ignores changes that consist
307 entirely of blank lines. With this option, for example, a file
309 1. A point is that which has no part.
311 2. A line is breadthless length.
312 -- Euclid, The Elements, I
313 is considered identical to a file containing
314 1. A point is that which has no part.
315 2. A line is breadthless length.
318 -- Euclid, The Elements, I
320 Normally this option affects only lines that are completely empty,
321 but if you also specify an option that ignores trailing spaces, lines
322 are also affected if they look empty but contain white space. In other
323 words, '-B' is equivalent to '-I '^$'' by default, but it is equivalent
324 to '-I '^[[:space:]]*$'' if '-b', '-w' or '-Z' is also specified.
327 File: diffutils.info, Node: Specified Lines, Next: Case Folding, Prev: Blank Lines, Up: Comparison
329 1.4 Suppressing Differences Whose Lines All Match a Regular Expression
330 ======================================================================
332 To ignore insertions and deletions of lines that match a 'grep'-style
333 regular expression, use the '--ignore-matching-lines=REGEXP' ('-I
334 REGEXP') option. You should escape regular expressions that contain
335 shell metacharacters to prevent the shell from expanding them. For
336 example, 'diff -I '^[[:digit:]]'' ignores all changes to lines beginning
339 However, '-I' only ignores the insertion or deletion of lines that
340 contain the regular expression if every changed line in the hunk--every
341 insertion and every deletion--matches the regular expression. In other
342 words, for each nonignorable change, 'diff' prints the complete set of
343 changes in its vicinity, including the ignorable ones.
345 You can specify more than one regular expression for lines to ignore
346 by using more than one '-I' option. 'diff' tries to match each line
347 against each regular expression.
350 File: diffutils.info, Node: Case Folding, Next: Brief, Prev: Specified Lines, Up: Comparison
352 1.5 Suppressing Case Differences
353 ================================
355 GNU 'diff' can treat lower case letters as equivalent to their upper
356 case counterparts, so that, for example, it considers 'Funky Stuff',
357 'funky STUFF', and 'fUNKy stuFf' to all be the same. To request this,
358 use the '-i' or '--ignore-case' option.
361 File: diffutils.info, Node: Brief, Next: Binary, Prev: Case Folding, Up: Comparison
363 1.6 Summarizing Which Files Differ
364 ==================================
366 When you only want to find out whether files are different, and you
367 don't care what the differences are, you can use the summary output
368 format. In this format, instead of showing the differences between the
369 files, 'diff' simply reports whether files differ. The '--brief' ('-q')
370 option selects this output format.
372 This format is especially useful when comparing the contents of two
373 directories. It is also much faster than doing the normal line by line
374 comparisons, because 'diff' can stop analyzing the files as soon as it
375 knows that there are any differences.
377 You can also get a brief indication of whether two files differ by
378 using 'cmp'. For files that are identical, 'cmp' produces no output.
379 When the files differ, by default, 'cmp' outputs the byte and line
380 number where the first difference occurs, or reports that one file is a
381 prefix of the other. You can use the '-s', '--quiet', or '--silent'
382 option to suppress that information, so that 'cmp' produces no output
383 and reports whether the files differ using only its exit status (*note
386 Unlike 'diff', 'cmp' cannot compare directories; it can only compare
390 File: diffutils.info, Node: Binary, Prev: Brief, Up: Comparison
392 1.7 Binary Files and Forcing Text Comparisons
393 =============================================
395 If 'diff' thinks that either of the two files it is comparing is binary
396 (a non-text file), it normally treats that pair of files much as if the
397 summary output format had been selected (*note Brief::), and reports
398 only that the binary files are different. This is because line by line
399 comparisons are usually not meaningful for binary files. This does not
400 count as trouble, even though the resulting output does not capture all
403 'diff' determines whether a file is text or binary by checking the
404 first few bytes in the file; the exact number of bytes is system
405 dependent, but it is typically several thousand. If every byte in that
406 part of the file is non-null, 'diff' considers the file to be text;
407 otherwise it considers the file to be binary.
409 Sometimes you might want to force 'diff' to consider files to be
410 text. For example, you might be comparing text files that contain null
411 characters; 'diff' would erroneously decide that those are non-text
412 files. Or you might be comparing documents that are in a format used by
413 a word processing system that uses null characters to indicate special
414 formatting. You can force 'diff' to consider all files to be text
415 files, and compare them line by line, by using the '--text' ('-a')
416 option. If the files you compare using this option do not in fact
417 contain text, they will probably contain few newline characters, and the
418 'diff' output will consist of hunks showing differences between long
419 lines of whatever characters the files contain.
421 You can also force 'diff' to report only whether files differ (but
422 not how). Use the '--brief' ('-q') option for this.
424 In operating systems that distinguish between text and binary files,
425 'diff' normally reads and writes all data as text. Use the '--binary'
426 option to force 'diff' to read and write binary data instead. This
427 option has no effect on a POSIX-compliant system like GNU or traditional
428 Unix. However, many personal computer operating systems represent the
429 end of a line with a carriage return followed by a newline. On such
430 systems, 'diff' normally ignores these carriage returns on input and
431 generates them at the end of each output line, but with the '--binary'
432 option 'diff' treats each carriage return as just another input
433 character, and does not generate a carriage return at the end of each
434 output line. This can be useful when dealing with non-text files that
435 are meant to be interchanged with POSIX-compliant systems.
437 The '--strip-trailing-cr' causes 'diff' to treat input lines that end
438 in carriage return followed by newline as if they end in plain newline.
439 This can be useful when comparing text that is imperfectly imported from
440 many personal computer operating systems. This option affects how lines
441 are read, which in turn affects how they are compared and output.
443 If you want to compare two files byte by byte, you can use the 'cmp'
444 program with the '--verbose' ('-l') option to show the values of each
445 differing byte in the two files. With GNU 'cmp', you can also use the
446 '-b' or '--print-bytes' option to show the ASCII representation of those
447 bytes. *Note Invoking cmp::, for more information.
449 If 'diff3' thinks that any of the files it is comparing is binary (a
450 non-text file), it normally reports an error, because such comparisons
451 are usually not useful. 'diff3' uses the same test as 'diff' to decide
452 whether a file is binary. As with 'diff', if the input files contain a
453 few non-text bytes but otherwise are like text files, you can force
454 'diff3' to consider all files to be text files and compare them line by
455 line by using the '-a' or '--text' option.
458 File: diffutils.info, Node: Output Formats, Next: Incomplete Lines, Prev: Comparison, Up: Top
460 2 'diff' Output Formats
461 ***********************
463 'diff' has several mutually exclusive options for output format. The
464 following sections describe each format, illustrating how 'diff' reports
465 the differences between two sample input files.
469 * Sample diff Input:: Sample 'diff' input files for examples.
470 * Context:: Showing differences with the surrounding text.
471 * Side by Side:: Showing differences in two columns.
472 * Normal:: Showing differences without surrounding text.
473 * Scripts:: Generating scripts for other programs.
474 * If-then-else:: Merging files with if-then-else.
477 File: diffutils.info, Node: Sample diff Input, Next: Context, Up: Output Formats
479 2.1 Two Sample Input Files
480 ==========================
482 Here are two sample files that we will use in numerous examples to
483 illustrate the output of 'diff' and how various options can change it.
485 This is the file 'lao':
487 The Way that can be told of is not the eternal Way;
488 The name that can be named is not the eternal name.
489 The Nameless is the origin of Heaven and Earth;
490 The Named is the mother of all things.
491 Therefore let there always be non-being,
492 so we may see their subtlety,
493 And let there always be being,
494 so we may see their outcome.
495 The two are the same,
496 But after they are produced,
497 they have different names.
499 This is the file 'tzu':
501 The Nameless is the origin of Heaven and Earth;
502 The named is the mother of all things.
504 Therefore let there always be non-being,
505 so we may see their subtlety,
506 And let there always be being,
507 so we may see their outcome.
508 The two are the same,
509 But after they are produced,
510 they have different names.
511 They both may be called deep and profound.
512 Deeper and more profound,
513 The door of all subtleties!
515 In this example, the first hunk contains just the first two lines of
516 'lao', the second hunk contains the fourth line of 'lao' opposing the
517 second and third lines of 'tzu', and the last hunk contains just the
518 last three lines of 'tzu'.
521 File: diffutils.info, Node: Context, Next: Side by Side, Prev: Sample diff Input, Up: Output Formats
523 2.2 Showing Differences in Their Context
524 ========================================
526 Usually, when you are looking at the differences between files, you will
527 also want to see the parts of the files near the lines that differ, to
528 help you understand exactly what has changed. These nearby parts of the
529 files are called the "context".
531 GNU 'diff' provides two output formats that show context around the
532 differing lines: "context format" and "unified format". It can
533 optionally show in which function or section of the file the differing
536 If you are distributing new versions of files to other people in the
537 form of 'diff' output, you should use one of the output formats that
538 show context so that they can apply the diffs even if they have made
539 small changes of their own to the files. 'patch' can apply the diffs in
540 this case by searching in the files for the lines of context around the
541 differing lines; if those lines are actually a few lines away from where
542 the diff says they are, 'patch' can adjust the line numbers accordingly
543 and still apply the diff correctly. *Note Imperfect::, for more
544 information on using 'patch' to apply imperfect diffs.
548 * Context Format:: An output format that shows surrounding lines.
549 * Unified Format:: A more compact output format that shows context.
550 * Sections:: Showing which sections of the files differences are in.
551 * Alternate Names:: Showing alternate file names in context headers.
554 File: diffutils.info, Node: Context Format, Next: Unified Format, Up: Context
559 The context output format shows several lines of context around the
560 lines that differ. It is the standard format for distributing updates
563 To select this output format, use the '--context[=LINES]' ('-C
564 LINES') or '-c' option. The argument LINES that some of these options
565 take is the number of lines of context to show. If you do not specify
566 LINES, it defaults to three. For proper operation, 'patch' typically
567 needs at least two lines of context.
571 * Example Context:: Sample output in context format.
572 * Less Context:: Another sample with less context.
573 * Detailed Context:: A detailed description of the context output format.
576 File: diffutils.info, Node: Example Context, Next: Less Context, Up: Context Format
578 2.2.1.1 An Example of Context Format
579 ....................................
581 Here is the output of 'diff -c lao tzu' (*note Sample diff Input::, for
582 the complete contents of the two files). Notice that up to three lines
583 that are not different are shown around each line that is different;
584 they are the context lines. Also notice that the first two hunks have
585 run together, because their contents overlap.
587 *** lao 2002-02-21 23:30:39.942229878 -0800
588 --- tzu 2002-02-21 23:30:50.442260588 -0800
591 - The Way that can be told of is not the eternal Way;
592 - The name that can be named is not the eternal name.
593 The Nameless is the origin of Heaven and Earth;
594 ! The Named is the mother of all things.
595 Therefore let there always be non-being,
596 so we may see their subtlety,
597 And let there always be being,
599 The Nameless is the origin of Heaven and Earth;
600 ! The named is the mother of all things.
602 Therefore let there always be non-being,
603 so we may see their subtlety,
604 And let there always be being,
608 The two are the same,
609 But after they are produced,
610 they have different names.
611 + They both may be called deep and profound.
612 + Deeper and more profound,
613 + The door of all subtleties!
616 File: diffutils.info, Node: Less Context, Next: Detailed Context, Prev: Example Context, Up: Context Format
618 2.2.1.2 An Example of Context Format with Less Context
619 ......................................................
621 Here is the output of 'diff -C 1 lao tzu' (*note Sample diff Input::,
622 for the complete contents of the two files). Notice that at most one
623 context line is reported here.
625 *** lao 2002-02-21 23:30:39.942229878 -0800
626 --- tzu 2002-02-21 23:30:50.442260588 -0800
629 - The Way that can be told of is not the eternal Way;
630 - The name that can be named is not the eternal name.
631 The Nameless is the origin of Heaven and Earth;
632 ! The Named is the mother of all things.
633 Therefore let there always be non-being,
635 The Nameless is the origin of Heaven and Earth;
636 ! The named is the mother of all things.
638 Therefore let there always be non-being,
642 they have different names.
643 + They both may be called deep and profound.
644 + Deeper and more profound,
645 + The door of all subtleties!
648 File: diffutils.info, Node: Detailed Context, Prev: Less Context, Up: Context Format
650 2.2.1.3 Detailed Description of Context Format
651 ..............................................
653 The context output format starts with a two-line header, which looks
656 *** FROM-FILE FROM-FILE-MODIFICATION-TIME
657 --- TO-FILE TO-FILE-MODIFICATION TIME
659 The timestamp normally looks like '2002-02-21 23:30:39.942229878 -0800'
660 to indicate the date, time with fractional seconds, and time zone in
661 Internet RFC 2822 format (ftp://ftp.isi.edu/in-notes/rfc2822.txt). (The
662 fractional seconds are omitted on hosts that do not support fractional
663 timestamps.) However, a traditional timestamp like 'Thu Feb 21 23:30:39
664 2002' is used if the 'LC_TIME' locale category is either 'C' or 'POSIX'.
666 You can change the header's content with the '--label=LABEL' option;
667 see *note Alternate Names::.
669 Next come one or more hunks of differences; each hunk shows one area
670 where the files differ. Context format hunks look like this:
673 *** FROM-FILE-LINE-NUMBERS ****
676 --- TO-FILE-LINE-NUMBERS ----
680 If a hunk contains two or more lines, its line numbers look like
681 'START,END'. Otherwise only its end line number appears. An empty hunk
682 is considered to end at the line that precedes the hunk.
684 The lines of context around the lines that differ start with two
685 space characters. The lines that differ between the two files start
686 with one of the following indicator characters, followed by a space
690 A line that is part of a group of one or more lines that changed
691 between the two files. There is a corresponding group of lines
692 marked with '!' in the part of this hunk for the other file.
695 An "inserted" line in the second file that corresponds to nothing
699 A "deleted" line in the first file that corresponds to nothing in
702 If all of the changes in a hunk are insertions, the lines of
703 FROM-FILE are omitted. If all of the changes are deletions, the lines
704 of TO-FILE are omitted.
707 File: diffutils.info, Node: Unified Format, Next: Sections, Prev: Context Format, Up: Context
712 The unified output format is a variation on the context format that is
713 more compact because it omits redundant context lines. To select this
714 output format, use the '--unified[=LINES]' ('-U LINES'), or '-u' option.
715 The argument LINES is the number of lines of context to show. When it
716 is not given, it defaults to three.
718 At present, only GNU 'diff' can produce this format and only GNU
719 'patch' can automatically apply diffs in this format. For proper
720 operation, 'patch' typically needs at least three lines of context.
724 * Example Unified:: Sample output in unified format.
725 * Detailed Unified:: A detailed description of unified format.
728 File: diffutils.info, Node: Example Unified, Next: Detailed Unified, Up: Unified Format
730 2.2.2.1 An Example of Unified Format
731 ....................................
733 Here is the output of the command 'diff -u lao tzu' (*note Sample diff
734 Input::, for the complete contents of the two files):
736 --- lao 2002-02-21 23:30:39.942229878 -0800
737 +++ tzu 2002-02-21 23:30:50.442260588 -0800
739 -The Way that can be told of is not the eternal Way;
740 -The name that can be named is not the eternal name.
741 The Nameless is the origin of Heaven and Earth;
742 -The Named is the mother of all things.
743 +The named is the mother of all things.
745 Therefore let there always be non-being,
746 so we may see their subtlety,
747 And let there always be being,
749 The two are the same,
750 But after they are produced,
751 they have different names.
752 +They both may be called deep and profound.
753 +Deeper and more profound,
754 +The door of all subtleties!
757 File: diffutils.info, Node: Detailed Unified, Prev: Example Unified, Up: Unified Format
759 2.2.2.2 Detailed Description of Unified Format
760 ..............................................
762 The unified output format starts with a two-line header, which looks
765 --- FROM-FILE FROM-FILE-MODIFICATION-TIME
766 +++ TO-FILE TO-FILE-MODIFICATION-TIME
768 The timestamp looks like '2002-02-21 23:30:39.942229878 -0800' to
769 indicate the date, time with fractional seconds, and time zone. The
770 fractional seconds are omitted on hosts that do not support fractional
773 You can change the header's content with the '--label=LABEL' option.
774 *Note Alternate Names::.
776 Next come one or more hunks of differences; each hunk shows one area
777 where the files differ. Unified format hunks look like this:
779 @@ FROM-FILE-LINE-NUMBERS TO-FILE-LINE-NUMBERS @@
780 LINE-FROM-EITHER-FILE
781 LINE-FROM-EITHER-FILE...
783 If a hunk contains just one line, only its start line number appears.
784 Otherwise its line numbers look like 'START,COUNT'. An empty hunk is
785 considered to start at the line that follows the hunk.
787 If a hunk and its context contain two or more lines, its line numbers
788 look like 'START,COUNT'. Otherwise only its end line number appears.
789 An empty hunk is considered to end at the line that precedes the hunk.
791 The lines common to both files begin with a space character. The
792 lines that actually differ between the two files have one of the
793 following indicator characters in the left print column:
796 A line was added here to the first file.
799 A line was removed here from the first file.
802 File: diffutils.info, Node: Sections, Next: Alternate Names, Prev: Unified Format, Up: Context
804 2.2.3 Showing Which Sections Differences Are in
805 -----------------------------------------------
807 Sometimes you might want to know which part of the files each change
808 falls in. If the files are source code, this could mean which function
809 was changed. If the files are documents, it could mean which chapter or
810 appendix was changed. GNU 'diff' can show this by displaying the
811 nearest section heading line that precedes the differing lines. Which
812 lines are "section headings" is determined by a regular expression.
816 * Specified Headings:: Showing headings that match regular expressions.
817 * C Function Headings:: Showing headings of C functions.
820 File: diffutils.info, Node: Specified Headings, Next: C Function Headings, Up: Sections
822 2.2.3.1 Showing Lines That Match Regular Expressions
823 ....................................................
825 To show in which sections differences occur for files that are not
826 source code for C or similar languages, use the
827 '--show-function-line=REGEXP' ('-F REGEXP') option. 'diff' considers
828 lines that match the 'grep'-style regular expression REGEXP to be the
829 beginning of a section of the file. Here are suggested regular
830 expressions for some common languages:
839 This option does not automatically select an output format; in order
840 to use it, you must select the context format (*note Context Format::)
841 or unified format (*note Unified Format::). In other output formats it
844 The '--show-function-line' ('-F') option finds the nearest unchanged
845 line that precedes each hunk of differences and matches the given
846 regular expression. Then it adds that line to the end of the line of
847 asterisks in the context format, or to the '@@' line in unified format.
848 If no matching line exists, this option leaves the output for that hunk
849 unchanged. If that line is more than 40 characters long, it outputs
850 only the first 40 characters. You can specify more than one regular
851 expression for such lines; 'diff' tries to match each line against each
852 regular expression, starting with the last one given. This means that
853 you can use '-p' and '-F' together, if you wish.
856 File: diffutils.info, Node: C Function Headings, Prev: Specified Headings, Up: Sections
858 2.2.3.2 Showing C Function Headings
859 ...................................
861 To show in which functions differences occur for C and similar
862 languages, you can use the '--show-c-function' ('-p') option. This
863 option automatically defaults to the context output format (*note
864 Context Format::), with the default number of lines of context. You can
865 override that number with '-C LINES' elsewhere in the command line. You
866 can override both the format and the number with '-U LINES' elsewhere in
869 The '--show-c-function' ('-p') option is equivalent to '-F
870 '^[[:alpha:]$_]'' if the unified format is specified, otherwise '-c -F
871 '^[[:alpha:]$_]'' (*note Specified Headings::). GNU 'diff' provides
872 this option for the sake of convenience.
875 File: diffutils.info, Node: Alternate Names, Prev: Sections, Up: Context
877 2.2.4 Showing Alternate File Names
878 ----------------------------------
880 If you are comparing two files that have meaningless or uninformative
881 names, you might want 'diff' to show alternate names in the header of
882 the context and unified output formats. To do this, use the
883 '--label=LABEL' option. The first time you give this option, its
884 argument replaces the name and date of the first file in the header; the
885 second time, its argument replaces the name and date of the second file.
886 If you give this option more than twice, 'diff' reports an error. The
887 '--label' option does not affect the file names in the 'pr' header when
888 the '-l' or '--paginate' option is used (*note Pagination::).
890 Here are the first two lines of the output from 'diff -C 2
891 --label=original --label=modified lao tzu':
897 File: diffutils.info, Node: Side by Side, Next: Normal, Prev: Context, Up: Output Formats
899 2.3 Showing Differences Side by Side
900 ====================================
902 'diff' can produce a side by side difference listing of two files. The
903 files are listed in two columns with a gutter between them. The gutter
904 contains one of the following markers:
907 The corresponding lines are in common. That is, either the lines
908 are identical, or the difference is ignored because of one of the
909 '--ignore' options (*note White Space::).
912 The corresponding lines differ, and they are either both complete
916 The files differ and only the first file contains the line.
919 The files differ and only the second file contains the line.
922 Only the first file contains the line, but the difference is
926 Only the second file contains the line, but the difference is
930 The corresponding lines differ, and only the first line is
934 The corresponding lines differ, and only the second line is
937 Normally, an output line is incomplete if and only if the lines that
938 it contains are incomplete. *Note Incomplete Lines::. However, when an
939 output line represents two differing lines, one might be incomplete
940 while the other is not. In this case, the output line is complete, but
941 its the gutter is marked '\' if the first line is incomplete, '/' if the
944 Side by side format is sometimes easiest to read, but it has
945 limitations. It generates much wider output than usual, and truncates
946 lines that are too long to fit. Also, it relies on lining up output
947 more heavily than usual, so its output looks particularly bad if you use
948 varying width fonts, nonstandard tab stops, or nonprinting characters.
950 You can use the 'sdiff' command to interactively merge side by side
951 differences. *Note Interactive Merging::, for more information on
956 * Side by Side Format:: Controlling side by side output format.
957 * Example Side by Side:: Sample side by side output.
960 File: diffutils.info, Node: Side by Side Format, Next: Example Side by Side, Up: Side by Side
962 2.3.1 Controlling Side by Side Format
963 -------------------------------------
965 The '--side-by-side' ('-y') option selects side by side format. Because
966 side by side output lines contain two input lines, the output is wider
967 than usual: normally 130 print columns, which can fit onto a traditional
968 printer line. You can set the width of the output with the
969 '--width=COLUMNS' ('-W COLUMNS') option. The output is split into two
970 halves of equal width, separated by a small gutter to mark differences;
971 the right half is aligned to a tab stop so that tabs line up. Input
972 lines that are too long to fit in half of an output line are truncated
975 The '--left-column' option prints only the left column of two common
976 lines. The '--suppress-common-lines' option suppresses common lines
980 File: diffutils.info, Node: Example Side by Side, Prev: Side by Side Format, Up: Side by Side
982 2.3.2 An Example of Side by Side Format
983 ---------------------------------------
985 Here is the output of the command 'diff -y -W 72 lao tzu' (*note Sample
986 diff Input::, for the complete contents of the two files).
988 The Way that can be told of is n <
989 The name that can be named is no <
990 The Nameless is the origin of He The Nameless is the origin of He
991 The Named is the mother of all t | The named is the mother of all t
993 Therefore let there always be no Therefore let there always be no
994 so we may see their subtlety, so we may see their subtlety,
995 And let there always be being, And let there always be being,
996 so we may see their outcome. so we may see their outcome.
997 The two are the same, The two are the same,
998 But after they are produced, But after they are produced,
999 they have different names. they have different names.
1000 > They both may be called deep and
1001 > Deeper and more profound,
1002 > The door of all subtleties!
1005 File: diffutils.info, Node: Normal, Next: Scripts, Prev: Side by Side, Up: Output Formats
1007 2.4 Showing Differences Without Context
1008 =======================================
1010 The "normal" 'diff' output format shows each hunk of differences without
1011 any surrounding context. Sometimes such output is the clearest way to
1012 see how lines have changed, without the clutter of nearby unchanged
1013 lines (although you can get similar results with the context or unified
1014 formats by using 0 lines of context). However, this format is no longer
1015 widely used for sending out patches; for that purpose, the context
1016 format (*note Context Format::) and the unified format (*note Unified
1017 Format::) are superior. Normal format is the default for compatibility
1018 with older versions of 'diff' and the POSIX standard. Use the
1019 '--normal' option to select this output format explicitly.
1023 * Example Normal:: Sample output in the normal format.
1024 * Detailed Normal:: A detailed description of normal output format.
1027 File: diffutils.info, Node: Example Normal, Next: Detailed Normal, Up: Normal
1029 2.4.1 An Example of Normal Format
1030 ---------------------------------
1032 Here is the output of the command 'diff lao tzu' (*note Sample diff
1033 Input::, for the complete contents of the two files). Notice that it
1034 shows only the lines that are different between the two files.
1037 < The Way that can be told of is not the eternal Way;
1038 < The name that can be named is not the eternal name.
1040 < The Named is the mother of all things.
1042 > The named is the mother of all things.
1045 > They both may be called deep and profound.
1046 > Deeper and more profound,
1047 > The door of all subtleties!
1050 File: diffutils.info, Node: Detailed Normal, Prev: Example Normal, Up: Normal
1052 2.4.2 Detailed Description of Normal Format
1053 -------------------------------------------
1055 The normal output format consists of one or more hunks of differences;
1056 each hunk shows one area where the files differ. Normal format hunks
1066 There are three types of change commands. Each consists of a line
1067 number or comma-separated range of lines in the first file, a single
1068 character indicating the kind of change to make, and a line number or
1069 comma-separated range of lines in the second file. All line numbers are
1070 the original line numbers in each file. The types of change commands
1074 Add the lines in range R of the second file after line L of the
1075 first file. For example, '8a12,15' means append lines 12-15 of
1076 file 2 after line 8 of file 1; or, if changing file 2 into file 1,
1077 delete lines 12-15 of file 2.
1080 Replace the lines in range F of the first file with lines in range
1081 T of the second file. This is like a combined add and delete, but
1082 more compact. For example, '5,7c8,10' means change lines 5-7 of
1083 file 1 to read as lines 8-10 of file 2; or, if changing file 2 into
1084 file 1, change lines 8-10 of file 2 to read as lines 5-7 of file 1.
1087 Delete the lines in range R from the first file; line L is where
1088 they would have appeared in the second file had they not been
1089 deleted. For example, '5,7d3' means delete lines 5-7 of file 1;
1090 or, if changing file 2 into file 1, append lines 5-7 of file 1
1091 after line 3 of file 2.
1094 File: diffutils.info, Node: Scripts, Next: If-then-else, Prev: Normal, Up: Output Formats
1096 2.5 Making Edit Scripts
1097 =======================
1099 Several output modes produce command scripts for editing FROM-FILE to
1104 * ed Scripts:: Using 'diff' to produce commands for 'ed'.
1105 * Forward ed:: Making forward 'ed' scripts.
1106 * RCS:: A special 'diff' output format used by RCS.
1109 File: diffutils.info, Node: ed Scripts, Next: Forward ed, Up: Scripts
1114 'diff' can produce commands that direct the 'ed' text editor to change
1115 the first file into the second file. Long ago, this was the only output
1116 mode that was suitable for editing one file into another automatically;
1117 today, with 'patch', it is almost obsolete. Use the '--ed' ('-e')
1118 option to select this output format.
1120 Like the normal format (*note Normal::), this output format does not
1121 show any context; unlike the normal format, it does not include the
1122 information necessary to apply the diff in reverse (to produce the first
1123 file if all you have is the second file and the diff).
1125 If the file 'd' contains the output of 'diff -e old new', then the
1126 command '(cat d && echo w) | ed - old' edits 'old' to make it a copy of
1127 'new'. More generally, if 'd1', 'd2', ..., 'dN' contain the outputs of
1128 'diff -e old new1', 'diff -e new1 new2', ..., 'diff -e newN-1 newN',
1129 respectively, then the command '(cat d1 d2 ... dN && echo w) | ed - old'
1130 edits 'old' to make it a copy of 'newN'.
1134 * Example ed:: A sample 'ed' script.
1135 * Detailed ed:: A detailed description of 'ed' format.
1138 File: diffutils.info, Node: Example ed, Next: Detailed ed, Up: ed Scripts
1140 2.5.1.1 Example 'ed' Script
1141 ...........................
1143 Here is the output of 'diff -e lao tzu' (*note Sample diff Input::, for
1144 the complete contents of the two files):
1147 They both may be called deep and profound.
1148 Deeper and more profound,
1149 The door of all subtleties!
1152 The named is the mother of all things.
1158 File: diffutils.info, Node: Detailed ed, Prev: Example ed, Up: ed Scripts
1160 2.5.1.2 Detailed Description of 'ed' Format
1161 ...........................................
1163 The 'ed' output format consists of one or more hunks of differences.
1164 The changes closest to the ends of the files come first so that commands
1165 that change the number of lines do not affect how 'ed' interprets line
1166 numbers in succeeding commands. 'ed' format hunks look like this:
1173 Because 'ed' uses a single period on a line to indicate the end of
1174 input, GNU 'diff' protects lines of changes that contain a single period
1175 on a line by writing two periods instead, then writing a subsequent 'ed'
1176 command to change the two periods into one. The 'ed' format cannot
1177 represent an incomplete line, so if the second file ends in a changed
1178 incomplete line, 'diff' reports an error and then pretends that a
1179 newline was appended.
1181 There are three types of change commands. Each consists of a line
1182 number or comma-separated range of lines in the first file and a single
1183 character indicating the kind of change to make. All line numbers are
1184 the original line numbers in the file. The types of change commands
1188 Add text from the second file after line L in the first file. For
1189 example, '8a' means to add the following lines after line 8 of file
1193 Replace the lines in range R in the first file with the following
1194 lines. Like a combined add and delete, but more compact. For
1195 example, '5,7c' means change lines 5-7 of file 1 to read as the
1199 Delete the lines in range R from the first file. For example,
1200 '5,7d' means delete lines 5-7 of file 1.
1203 File: diffutils.info, Node: Forward ed, Next: RCS, Prev: ed Scripts, Up: Scripts
1205 2.5.2 Forward 'ed' Scripts
1206 --------------------------
1208 'diff' can produce output that is like an 'ed' script, but with hunks in
1209 forward (front to back) order. The format of the commands is also
1210 changed slightly: command characters precede the lines they modify,
1211 spaces separate line numbers in ranges, and no attempt is made to
1212 disambiguate hunk lines consisting of a single period. Like 'ed'
1213 format, forward 'ed' format cannot represent incomplete lines.
1215 Forward 'ed' format is not very useful, because neither 'ed' nor
1216 'patch' can apply diffs in this format. It exists mainly for
1217 compatibility with older versions of 'diff'. Use the '-f' or
1218 '--forward-ed' option to select it.
1221 File: diffutils.info, Node: RCS, Prev: Forward ed, Up: Scripts
1226 The RCS output format is designed specifically for use by the Revision
1227 Control System, which is a set of free programs used for organizing
1228 different versions and systems of files. Use the '--rcs' ('-n') option
1229 to select this output format. It is like the forward 'ed' format (*note
1230 Forward ed::), but it can represent arbitrary changes to the contents of
1231 a file because it avoids the forward 'ed' format's problems with lines
1232 consisting of a single period and with incomplete lines. Instead of
1233 ending text sections with a line consisting of a single period, each
1234 command specifies the number of lines it affects; a combination of the
1235 'a' and 'd' commands are used instead of 'c'. Also, if the second file
1236 ends in a changed incomplete line, then the output also ends in an
1239 Here is the output of 'diff -n lao tzu' (*note Sample diff Input::,
1240 for the complete contents of the two files):
1245 The named is the mother of all things.
1248 They both may be called deep and profound.
1249 Deeper and more profound,
1250 The door of all subtleties!
1253 File: diffutils.info, Node: If-then-else, Prev: Scripts, Up: Output Formats
1255 2.6 Merging Files with If-then-else
1256 ===================================
1258 You can use 'diff' to merge two files of C source code. The output of
1259 'diff' in this format contains all the lines of both files. Lines
1260 common to both files are output just once; the differing parts are
1261 separated by the C preprocessor directives '#ifdef NAME' or '#ifndef
1262 NAME', '#else', and '#endif'. When compiling the output, you select
1263 which version to use by either defining or leaving undefined the macro
1266 To merge two files, use 'diff' with the '-D NAME' or '--ifdef=NAME'
1267 option. The argument NAME is the C preprocessor identifier to use in
1268 the '#ifdef' and '#ifndef' directives.
1270 For example, if you change an instance of 'wait (&s)' to 'waitpid
1271 (-1, &s, 0)' and then merge the old and new files with the
1272 '--ifdef=HAVE_WAITPID' option, then the affected part of your code might
1276 #ifndef HAVE_WAITPID
1277 if ((w = wait (&s)) < 0 && errno != EINTR)
1278 #else /* HAVE_WAITPID */
1279 if ((w = waitpid (-1, &s, 0)) < 0 && errno != EINTR)
1280 #endif /* HAVE_WAITPID */
1282 } while (w != child);
1284 You can specify formats for languages other than C by using line
1285 group formats and line formats, as described in the next sections.
1289 * Line Group Formats:: Formats for general if-then-else line groups.
1290 * Line Formats:: Formats for each line in a line group.
1291 * Example If-then-else:: Sample if-then-else format output.
1292 * Detailed If-then-else:: A detailed description of if-then-else format.
1295 File: diffutils.info, Node: Line Group Formats, Next: Line Formats, Up: If-then-else
1297 2.6.1 Line Group Formats
1298 ------------------------
1300 Line group formats let you specify formats suitable for many
1301 applications that allow if-then-else input, including programming
1302 languages and text formatting languages. A line group format specifies
1303 the output format for a contiguous group of similar lines.
1305 For example, the following command compares the TeX files 'old' and
1306 'new', and outputs a merged file in which old regions are surrounded by
1307 '\begin{em}'-'\end{em}' lines, and new regions are surrounded by
1308 '\begin{bf}'-'\end{bf}' lines.
1311 --old-group-format='\begin{em}
1314 --new-group-format='\begin{bf}
1319 The following command is equivalent to the above example, but it is a
1320 little more verbose, because it spells out the default line group
1324 --old-group-format='\begin{em}
1327 --new-group-format='\begin{bf}
1330 --unchanged-group-format='%=' \
1331 --changed-group-format='\begin{em}
1338 Here is a more advanced example, which outputs a diff listing with
1339 headers containing line numbers in a "plain English" style.
1342 --unchanged-group-format='' \
1343 --old-group-format='-------- %dn line%(n=1?:s) deleted at %df:
1345 --new-group-format='-------- %dN line%(N=1?:s) added after %de:
1347 --changed-group-format='-------- %dn line%(n=1?:s) changed at %df:
1352 To specify a line group format, use 'diff' with one of the options
1353 listed below. You can specify up to four line group formats, one for
1354 each kind of line group. You should quote FORMAT, because it typically
1355 contains shell metacharacters.
1357 '--old-group-format=FORMAT'
1358 These line groups are hunks containing only lines from the first
1359 file. The default old group format is the same as the changed
1360 group format if it is specified; otherwise it is a format that
1361 outputs the line group as-is.
1363 '--new-group-format=FORMAT'
1364 These line groups are hunks containing only lines from the second
1365 file. The default new group format is same as the changed group
1366 format if it is specified; otherwise it is a format that outputs
1367 the line group as-is.
1369 '--changed-group-format=FORMAT'
1370 These line groups are hunks containing lines from both files. The
1371 default changed group format is the concatenation of the old and
1374 '--unchanged-group-format=FORMAT'
1375 These line groups contain lines common to both files. The default
1376 unchanged group format is a format that outputs the line group
1379 In a line group format, ordinary characters represent themselves;
1380 conversion specifications start with '%' and have one of the following
1384 stands for the lines from the first file, including the trailing
1385 newline. Each line is formatted according to the old line format
1386 (*note Line Formats::).
1389 stands for the lines from the second file, including the trailing
1390 newline. Each line is formatted according to the new line format.
1393 stands for the lines common to both files, including the trailing
1394 newline. Each line is formatted according to the unchanged line
1401 where C is a single character, stands for C. C may not be a
1402 backslash or an apostrophe. For example, '%c':'' stands for a
1403 colon, even inside the then-part of an if-then-else format, which a
1404 colon would normally terminate.
1407 where O is a string of 1, 2, or 3 octal digits, stands for the
1408 character with octal code O. For example, '%c'\0'' stands for a
1412 where F is a 'printf' conversion specification and N is one of the
1413 following letters, stands for N's value formatted with F.
1416 The line number of the line just before the group in the old
1420 The line number of the first line in the group in the old
1424 The line number of the last line in the group in the old file.
1427 The line number of the line just after the group in the old
1431 The number of lines in the group in the old file; equals L - F
1435 Likewise, for lines in the new file.
1437 The 'printf' conversion specification can be '%d', '%o', '%x', or
1438 '%X', specifying decimal, octal, lower case hexadecimal, or upper
1439 case hexadecimal output respectively. After the '%' the following
1440 options can appear in sequence: a series of zero or more flags; an
1441 integer specifying the minimum field width; and a period followed
1442 by an optional integer specifying the minimum number of digits.
1443 The flags are '-' for left-justification, ''' for separating the
1444 digit into groups as specified by the 'LC_NUMERIC' locale category,
1445 and '0' for padding with zeros instead of spaces. For example,
1446 '%5dN' prints the number of new lines in the group in a field of
1447 width 5 characters, using the 'printf' format '"%5d"'.
1450 If A equals B then T else E. A and B are each either a decimal
1451 constant or a single letter interpreted as above. This format spec
1452 is equivalent to T if A's value equals B's; otherwise it is
1455 For example, '%(N=0?no:%dN) line%(N=1?:s)' is equivalent to 'no
1456 lines' if N (the number of lines in the group in the new file) is
1457 0, to '1 line' if N is 1, and to '%dN lines' otherwise.
1460 File: diffutils.info, Node: Line Formats, Next: Example If-then-else, Prev: Line Group Formats, Up: If-then-else
1465 Line formats control how each line taken from an input file is output as
1466 part of a line group in if-then-else format.
1468 For example, the following command outputs text with a one-character
1469 change indicator to the left of the text. The first character of output
1470 is '-' for deleted lines, '|' for added lines, and a space for unchanged
1471 lines. The formats contain newline characters where newlines are
1475 --old-line-format='-%l
1477 --new-line-format='|%l
1479 --unchanged-line-format=' %l
1483 To specify a line format, use one of the following options. You
1484 should quote FORMAT, since it often contains shell metacharacters.
1486 '--old-line-format=FORMAT'
1487 formats lines just from the first file.
1489 '--new-line-format=FORMAT'
1490 formats lines just from the second file.
1492 '--unchanged-line-format=FORMAT'
1493 formats lines common to both files.
1495 '--line-format=FORMAT'
1496 formats all lines; in effect, it sets all three above options
1499 In a line format, ordinary characters represent themselves;
1500 conversion specifications start with '%' and have one of the following
1504 stands for the contents of the line, not counting its trailing
1505 newline (if any). This format ignores whether the line is
1506 incomplete; *Note Incomplete Lines::.
1509 stands for the contents of the line, including its trailing newline
1510 (if any). If a line is incomplete, this format preserves its
1517 where C is a single character, stands for C. C may not be a
1518 backslash or an apostrophe. For example, '%c':'' stands for a
1522 where O is a string of 1, 2, or 3 octal digits, stands for the
1523 character with octal code O. For example, '%c'\0'' stands for a
1527 where F is a 'printf' conversion specification, stands for the line
1528 number formatted with F. For example, '%.5dn' prints the line
1529 number using the 'printf' format '"%.5d"'. *Note Line Group
1530 Formats::, for more about printf conversion specifications.
1532 The default line format is '%l' followed by a newline character.
1534 If the input contains tab characters and it is important that they
1535 line up on output, you should ensure that '%l' or '%L' in a line format
1536 is just after a tab stop (e.g. by preceding '%l' or '%L' with a tab
1537 character), or you should use the '-t' or '--expand-tabs' option.
1539 Taken together, the line and line group formats let you specify many
1540 different formats. For example, the following command uses a format
1541 similar to normal 'diff' format. You can tailor this command to get
1542 fine control over 'diff' output.
1545 --old-line-format='< %l
1547 --new-line-format='> %l
1549 --old-group-format='%df%(f=l?:,%dl)d%dE
1551 --new-group-format='%dea%dF%(F=L?:,%dL)
1553 --changed-group-format='%df%(f=l?:,%dl)c%dF%(F=L?:,%dL)
1556 --unchanged-group-format='' \
1560 File: diffutils.info, Node: Example If-then-else, Next: Detailed If-then-else, Prev: Line Formats, Up: If-then-else
1562 2.6.3 An Example of If-then-else Format
1563 ---------------------------------------
1565 Here is the output of 'diff -DTWO lao tzu' (*note Sample diff Input::,
1566 for the complete contents of the two files):
1569 The Way that can be told of is not the eternal Way;
1570 The name that can be named is not the eternal name.
1572 The Nameless is the origin of Heaven and Earth;
1574 The Named is the mother of all things.
1576 The named is the mother of all things.
1579 Therefore let there always be non-being,
1580 so we may see their subtlety,
1581 And let there always be being,
1582 so we may see their outcome.
1583 The two are the same,
1584 But after they are produced,
1585 they have different names.
1587 They both may be called deep and profound.
1588 Deeper and more profound,
1589 The door of all subtleties!
1593 File: diffutils.info, Node: Detailed If-then-else, Prev: Example If-then-else, Up: If-then-else
1595 2.6.4 Detailed Description of If-then-else Format
1596 -------------------------------------------------
1598 For lines common to both files, 'diff' uses the unchanged line group
1599 format. For each hunk of differences in the merged output format, if
1600 the hunk contains only lines from the first file, 'diff' uses the old
1601 line group format; if the hunk contains only lines from the second file,
1602 'diff' uses the new group format; otherwise, 'diff' uses the changed
1605 The old, new, and unchanged line formats specify the output format of
1606 lines from the first file, lines from the second file, and lines common
1607 to both files, respectively.
1609 The option '--ifdef=NAME' is equivalent to the following sequence of
1610 options using shell syntax:
1612 --old-group-format='#ifndef NAME
1613 %<#endif /* ! NAME */
1615 --new-group-format='#ifdef NAME
1618 --unchanged-group-format='%=' \
1619 --changed-group-format='#ifndef NAME
1624 You should carefully check the 'diff' output for proper nesting. For
1625 example, when using the '-D NAME' or '--ifdef=NAME' option, you should
1626 check that if the differing lines contain any of the C preprocessor
1627 directives '#ifdef', '#ifndef', '#else', '#elif', or '#endif', they are
1628 nested properly and match. If they don't, you must make corrections
1629 manually. It is a good idea to carefully check the resulting code
1630 anyway to make sure that it really does what you want it to; depending
1631 on how the input files were produced, the output might contain duplicate
1632 or otherwise incorrect code.
1634 The 'patch' '-D NAME' option behaves like the 'diff' '-D NAME'
1635 option, except it operates on a file and a diff to produce a merged
1636 file. *Note patch Options::.
1639 File: diffutils.info, Node: Incomplete Lines, Next: Comparing Directories, Prev: Output Formats, Up: Top
1644 When an input file ends in a non-newline character, its last line is
1645 called an "incomplete line" because its last character is not a newline.
1646 All other lines are called "full lines" and end in a newline character.
1647 Incomplete lines do not match full lines unless differences in white
1648 space are ignored (*note White Space::).
1650 An incomplete line is normally distinguished on output from a full
1651 line by a following line that starts with '\'. However, the RCS format
1652 (*note RCS::) outputs the incomplete line as-is, without any trailing
1653 newline or following line. The side by side format normally represents
1654 incomplete lines as-is, but in some cases uses a '\' or '/' gutter
1655 marker. *Note Side by Side::. The if-then-else line format preserves a
1656 line's incompleteness with '%L', and discards the newline with '%l'.
1657 *Note Line Formats::. Finally, with the 'ed' and forward 'ed' output
1658 formats (*note Output Formats::) 'diff' cannot represent an incomplete
1659 line, so it pretends there was a newline and reports an error.
1661 For example, suppose 'F' and 'G' are one-byte files that contain just
1662 'f' and 'g', respectively. Then 'diff F G' outputs
1666 \ No newline at end of file
1669 \ No newline at end of file
1671 (The exact message may differ in non-English locales.) 'diff -n F G'
1672 outputs the following without a trailing newline:
1678 'diff -e F G' reports two errors and outputs the following:
1685 File: diffutils.info, Node: Comparing Directories, Next: Adjusting Output, Prev: Incomplete Lines, Up: Top
1687 4 Comparing Directories
1688 ***********************
1690 You can use 'diff' to compare some or all of the files in two directory
1691 trees. When both file name arguments to 'diff' are directories, it
1692 compares each file that is contained in both directories, examining file
1693 names in alphabetical order as specified by the 'LC_COLLATE' locale
1694 category. Normally 'diff' is silent about pairs of files that contain
1695 no differences, but if you use the '--report-identical-files' ('-s')
1696 option, it reports pairs of identical files. Normally 'diff' reports
1697 subdirectories common to both directories without comparing
1698 subdirectories' files, but if you use the '-r' or '--recursive' option,
1699 it compares every corresponding pair of files in the directory trees, as
1700 many levels deep as they go.
1702 If only one file exists, 'diff' normally does not show its contents;
1703 it merely reports that one file exists but the other does not. You can
1704 make 'diff' act as though the missing file is empty, so that it outputs
1705 the entire contents of the file that actually exists. (It is output as
1706 either an insertion or a deletion, depending on whether the missing file
1707 is in the first or the second position.) To do this, use the
1708 '--new-file' ('-N') option. This option affects command-line arguments
1709 as well as files found via directory traversal; for example, 'diff -N a
1710 b' treats 'a' as empty if 'a' does not exist but 'b' does, and similarly
1711 'diff -N - b' treats standard input as empty if it is closed but 'b'
1714 If the older directory contains large files that are not in the newer
1715 directory, you can make the patch smaller by using the
1716 '--unidirectional-new-file' option instead of '-N'. This option is like
1717 '-N' except that it inserts the contents only of files that appear in
1718 the second directory but not the first (that is, files that were added).
1719 At the top of the patch, write instructions for the user applying the
1720 patch to remove the files that were deleted before applying the patch.
1721 *Note Making Patches::, for more discussion of making patches for
1724 To ignore some files while comparing directories, use the
1725 '--exclude=PATTERN' ('-x PATTERN') option. This option ignores any
1726 files or subdirectories whose base names match the shell pattern
1727 PATTERN. Unlike in the shell, a period at the start of the base of a
1728 file name matches a wildcard at the start of a pattern. You should
1729 enclose PATTERN in quotes so that the shell does not expand it. For
1730 example, the option '-x '*.[ao]'' ignores any file whose name ends with
1733 This option accumulates if you specify it more than once. For
1734 example, using the options '-x 'RCS' -x '*,v'' ignores any file or
1735 subdirectory whose base name is 'RCS' or ends with ',v'.
1737 If you need to give this option many times, you can instead put the
1738 patterns in a file, one pattern per line, and use the
1739 '--exclude-from=FILE' ('-X FILE') option. Trailing white space and
1740 empty lines are ignored in the pattern file.
1742 If you have been comparing two directories and stopped partway
1743 through, later you might want to continue where you left off. You can
1744 do this by using the '--starting-file=FILE' ('-S FILE') option. This
1745 compares only the file FILE and all alphabetically later files in the
1746 topmost directory level.
1748 If two directories differ only in that file names are lower case in
1749 one directory and upper case in the upper, 'diff' normally reports many
1750 differences because it compares file names in a case sensitive way.
1751 With the '--ignore-file-name-case' option, 'diff' ignores case
1752 differences in file names, so that for example the contents of the file
1753 'Tao' in one directory are compared to the contents of the file 'TAO' in
1754 the other. The '--no-ignore-file-name-case' option cancels the effect
1755 of the '--ignore-file-name-case' option, reverting to the default
1758 If an '--exclude=PATTERN' ('-x PATTERN') option, or an
1759 '--exclude-from=FILE' ('-X FILE') option, is specified while the
1760 '--ignore-file-name-case' option is in effect, case is ignored when
1761 excluding file names matching the specified patterns.
1763 To tell 'diff' not to follow a symbolic link, use the
1764 '--no-dereference' option.
1767 File: diffutils.info, Node: Adjusting Output, Next: diff Performance, Prev: Comparing Directories, Up: Top
1769 5 Making 'diff' Output Prettier
1770 *******************************
1772 'diff' provides several ways to adjust the appearance of its output.
1773 These adjustments can be applied to any output format.
1777 * Tabs:: Preserving the alignment of tab stops.
1778 * Trailing Blanks:: Suppressing blanks before empty output lines.
1779 * Pagination:: Page numbering and time-stamping 'diff' output.
1782 File: diffutils.info, Node: Tabs, Next: Trailing Blanks, Up: Adjusting Output
1784 5.1 Preserving Tab Stop Alignment
1785 =================================
1787 The lines of text in some of the 'diff' output formats are preceded by
1788 one or two characters that indicate whether the text is inserted,
1789 deleted, or changed. The addition of those characters can cause tabs to
1790 move to the next tab stop, throwing off the alignment of columns in the
1791 line. GNU 'diff' provides two ways to make tab-aligned columns line up
1794 The first way is to have 'diff' convert all tabs into the correct
1795 number of spaces before outputting them; select this method with the
1796 '--expand-tabs' ('-t') option. To use this form of output with 'patch',
1797 you must give 'patch' the '-l' or '--ignore-white-space' option (*note
1798 Changed White Space::, for more information). 'diff' normally assumes
1799 that tab stops are set every 8 print columns, but this can be altered by
1800 the '--tabsize=COLUMNS' option.
1802 The other method for making tabs line up correctly is to add a tab
1803 character instead of a space after the indicator character at the
1804 beginning of the line. This ensures that all following tab characters
1805 are in the same position relative to tab stops that they were in the
1806 original files, so that the output is aligned correctly. Its
1807 disadvantage is that it can make long lines too long to fit on one line
1808 of the screen or the paper. It also does not work with the unified
1809 output format, which does not have a space character after the change
1810 type indicator character. Select this method with the '-T' or
1811 '--initial-tab' option.
1814 File: diffutils.info, Node: Trailing Blanks, Next: Pagination, Prev: Tabs, Up: Adjusting Output
1816 5.2 Omitting trailing blanks
1817 ============================
1819 When outputting lines in normal or context format, or outputting an
1820 unchanged line in unified format, 'diff' normally outputs a blank just
1821 before each line. If the line is empty, the output of 'diff' therefore
1822 contains trailing blanks even though the input does not contain them.
1823 For example, when outputting an unchanged empty line in context format,
1824 'diff' normally outputs a line with two leading spaces.
1826 Some text editors and email agents routinely delete trailing blanks,
1827 so it can be a problem to deal with diff output files that contain them.
1828 You can avoid this problem with the '--suppress-blank-empty' option. It
1829 causes 'diff' to omit trailing blanks at the end of output lines in
1830 normal, context, and unified format, unless the trailing blanks were
1831 already present in the input. This changes the output format slightly,
1832 so that output lines are guaranteed to never end in a blank unless an
1833 input line ends in a blank. This format is less likely to be munged by
1834 text editors or by transmission via email. It is accepted by GNU
1838 File: diffutils.info, Node: Pagination, Prev: Trailing Blanks, Up: Adjusting Output
1840 5.3 Paginating 'diff' Output
1841 ============================
1843 It can be convenient to have long output page-numbered and time-stamped.
1844 The '--paginate' ('-l') option does this by sending the 'diff' output
1845 through the 'pr' program. Here is what the page header might look like
1846 for 'diff -lc lao tzu':
1848 2002-02-22 14:20 diff -lc lao tzu Page 1
1851 File: diffutils.info, Node: diff Performance, Next: Comparing Three Files, Prev: Adjusting Output, Up: Top
1853 6 'diff' Performance Tradeoffs
1854 ******************************
1856 GNU 'diff' runs quite efficiently; however, in some circumstances you
1857 can cause it to run faster or produce a more compact set of changes.
1859 One way to improve 'diff' performance is to use hard or symbolic
1860 links to files instead of copies. This improves performance because
1861 'diff' normally does not need to read two hard or symbolic links to the
1862 same file, since their contents must be identical. For example, suppose
1863 you copy a large directory hierarchy, make a few changes to the copy,
1864 and then often use 'diff -r' to compare the original to the copy. If
1865 the original files are read-only, you can greatly improve performance by
1866 creating the copy using hard or symbolic links (e.g., with GNU 'cp -lR'
1867 or 'cp -sR'). Before editing a file in the copy for the first time, you
1868 should break the link and replace it with a regular copy.
1870 You can also affect the performance of GNU 'diff' by giving it
1871 options that change the way it compares files. Performance has more
1872 than one dimension. These options improve one aspect of performance at
1873 the cost of another, or they improve performance in some cases while
1874 hurting it in others.
1876 The way that GNU 'diff' determines which lines have changed always
1877 comes up with a near-minimal set of differences. Usually it is good
1878 enough for practical purposes. If the 'diff' output is large, you might
1879 want 'diff' to use a modified algorithm that sometimes produces a
1880 smaller set of differences. The '--minimal' ('-d') option does this;
1881 however, it can also cause 'diff' to run more slowly than usual, so it
1882 is not the default behavior.
1884 When the files you are comparing are large and have small groups of
1885 changes scattered throughout them, you can use the '--speed-large-files'
1886 option to make a different modification to the algorithm that 'diff'
1887 uses. If the input files have a constant small density of changes, this
1888 option speeds up the comparisons without changing the output. If not,
1889 'diff' might produce a larger set of differences; however, the output
1890 will still be correct.
1892 Normally 'diff' discards the prefix and suffix that is common to both
1893 files before it attempts to find a minimal set of differences. This
1894 makes 'diff' run faster, but occasionally it may produce non-minimal
1895 output. The '--horizon-lines=LINES' option prevents 'diff' from
1896 discarding the last LINES lines of the prefix and the first LINES lines
1897 of the suffix. This gives 'diff' further opportunities to find a
1900 Suppose a run of changed lines includes a sequence of lines at one
1901 end and there is an identical sequence of lines just outside the other
1902 end. The 'diff' command is free to choose which identical sequence is
1903 included in the hunk. In this case, 'diff' normally shifts the hunk's
1904 boundaries when this merges adjacent hunks, or shifts a hunk's lines
1905 towards the end of the file. Merging hunks can make the output look
1906 nicer in some cases.
1909 File: diffutils.info, Node: Comparing Three Files, Next: diff3 Merging, Prev: diff Performance, Up: Top
1911 7 Comparing Three Files
1912 ***********************
1914 Use the program 'diff3' to compare three files and show any differences
1915 among them. ('diff3' can also merge files; see *note diff3 Merging::).
1917 The "normal" 'diff3' output format shows each hunk of differences
1918 without surrounding context. Hunks are labeled depending on whether
1919 they are two-way or three-way, and lines are annotated by their location
1922 *Note Invoking diff3::, for more information on how to run 'diff3'.
1926 * Sample diff3 Input:: Sample 'diff3' input for examples.
1927 * Example diff3 Normal:: Sample output in the normal format.
1928 * Detailed diff3 Normal:: A detailed description of normal output format.
1929 * diff3 Hunks:: The format of normal output format.
1932 File: diffutils.info, Node: Sample diff3 Input, Next: Example diff3 Normal, Up: Comparing Three Files
1934 7.1 A Third Sample Input File
1935 =============================
1937 Here is a third sample file that will be used in examples to illustrate
1938 the output of 'diff3' and how various options can change it. The first
1939 two files are the same that we used for 'diff' (*note Sample diff
1940 Input::). This is the third sample file, called 'tao':
1942 The Way that can be told of is not the eternal Way;
1943 The name that can be named is not the eternal name.
1944 The Nameless is the origin of Heaven and Earth;
1945 The named is the mother of all things.
1947 Therefore let there always be non-being,
1948 so we may see their subtlety,
1949 And let there always be being,
1950 so we may see their result.
1951 The two are the same,
1952 But after they are produced,
1953 they have different names.
1955 -- The Way of Lao-Tzu, tr. Wing-tsit Chan
1958 File: diffutils.info, Node: Example diff3 Normal, Next: Detailed diff3 Normal, Prev: Sample diff3 Input, Up: Comparing Three Files
1960 7.2 An Example of 'diff3' Normal Format
1961 =======================================
1963 Here is the output of the command 'diff3 lao tzu tao' (*note Sample
1964 diff3 Input::, for the complete contents of the files). Notice that it
1965 shows only the lines that are different among the three files.
1970 The Way that can be told of is not the eternal Way;
1971 The name that can be named is not the eternal name.
1975 The Named is the mother of all things.
1978 The named is the mother of all things.
1983 so we may see their outcome.
1985 so we may see their result.
1989 They both may be called deep and profound.
1990 Deeper and more profound,
1991 The door of all subtleties!
1994 -- The Way of Lao-Tzu, tr. Wing-tsit Chan
1997 File: diffutils.info, Node: Detailed diff3 Normal, Next: diff3 Hunks, Prev: Example diff3 Normal, Up: Comparing Three Files
1999 7.3 Detailed Description of 'diff3' Normal Format
2000 =================================================
2002 Each hunk begins with a line marked '===='. Three-way hunks have plain
2003 '====' lines, and two-way hunks have '1', '2', or '3' appended to
2004 specify which of the three input files differ in that hunk. The hunks
2005 contain copies of two or three sets of input lines each preceded by one
2006 or two commands identifying where the lines came from.
2008 Normally, two spaces precede each copy of an input line to
2009 distinguish it from the commands. But with the '--initial-tab' ('-T')
2010 option, 'diff3' uses a tab instead of two spaces; this lines up tabs
2011 correctly. *Note Tabs::, for more information.
2013 Commands take the following forms:
2016 This hunk appears after line L of file FILE, and contains no lines
2017 in that file. To edit this file to yield the other files, one must
2018 append hunk lines taken from the other files. For example, '1:11a'
2019 means that the hunk follows line 11 in the first file and contains
2020 no lines from that file.
2023 This hunk contains the lines in the range R of file FILE. The
2024 range R is a comma-separated pair of line numbers, or just one
2025 number if there is only one line. To edit this file to yield the
2026 other files, one must change the specified lines to be the lines
2027 taken from the other files. For example, '2:11,13c' means that the
2028 hunk contains lines 11 through 13 from the second file.
2030 If the last line in a set of input lines is incomplete (*note
2031 Incomplete Lines::), it is distinguished on output from a full line by a
2032 following line that starts with '\'.
2035 File: diffutils.info, Node: diff3 Hunks, Prev: Detailed diff3 Normal, Up: Comparing Three Files
2040 Groups of lines that differ in two or three of the input files are
2041 called "diff3 hunks", by analogy with 'diff' hunks (*note Hunks::). If
2042 all three input files differ in a 'diff3' hunk, the hunk is called a
2043 "three-way hunk"; if just two input files differ, it is a "two-way
2046 As with 'diff', several solutions are possible. When comparing the
2047 files 'A', 'B', and 'C', 'diff3' normally finds 'diff3' hunks by merging
2048 the two-way hunks output by the two commands 'diff A B' and 'diff A C'.
2049 This does not necessarily minimize the size of the output, but
2050 exceptions should be rare.
2052 For example, suppose 'F' contains the three lines 'a', 'b', 'f', 'G'
2053 contains the lines 'g', 'b', 'g', and 'H' contains the lines 'a', 'b',
2054 'h'. 'diff3 F G H' might output the following:
2070 because it found a two-way hunk containing 'a' in the first and third
2071 files and 'g' in the second file, then the single line 'b' common to all
2072 three files, then a three-way hunk containing the last line of each
2076 File: diffutils.info, Node: diff3 Merging, Next: Interactive Merging, Prev: Comparing Three Files, Up: Top
2078 8 Merging From a Common Ancestor
2079 ********************************
2081 When two people have made changes to copies of the same file, 'diff3'
2082 can produce a merged output that contains both sets of changes together
2083 with warnings about conflicts.
2085 One might imagine programs with names like 'diff4' and 'diff5' to
2086 compare more than three files simultaneously, but in practice the need
2087 rarely arises. You can use 'diff3' to merge three or more sets of
2088 changes to a file by merging two change sets at a time.
2090 'diff3' can incorporate changes from two modified versions into a
2091 common preceding version. This lets you merge the sets of changes
2092 represented by the two newer files. Specify the common ancestor version
2093 as the second argument and the two newer versions as the first and third
2094 arguments, like this:
2096 diff3 MINE OLDER YOURS
2098 You can remember the order of the arguments by noting that they are in
2101 You can think of this as subtracting OLDER from YOURS and adding the
2102 result to MINE, or as merging into MINE the changes that would turn
2103 OLDER into YOURS. This merging is well-defined as long as MINE and
2104 OLDER match in the neighborhood of each such change. This fails to be
2105 true when all three input files differ or when only OLDER differs; we
2106 call this a "conflict". When all three input files differ, we call the
2107 conflict an "overlap".
2109 'diff3' gives you several ways to handle overlaps and conflicts. You
2110 can omit overlaps or conflicts, or select only overlaps, or mark
2111 conflicts with special '<<<<<<<' and '>>>>>>>' lines.
2113 'diff3' can output the merge results as an 'ed' script that that can
2114 be applied to the first file to yield the merged output. However, it is
2115 usually better to have 'diff3' generate the merged output directly; this
2116 bypasses some problems with 'ed'.
2120 * Which Changes:: Selecting changes to incorporate.
2121 * Marking Conflicts:: Marking conflicts.
2122 * Bypassing ed:: Generating merged output directly.
2123 * Merging Incomplete Lines:: How 'diff3' merges incomplete lines.
2124 * Saving the Changed File:: Emulating System V behavior.
2127 File: diffutils.info, Node: Which Changes, Next: Marking Conflicts, Up: diff3 Merging
2129 8.1 Selecting Which Changes to Incorporate
2130 ==========================================
2132 You can select all unmerged changes from OLDER to YOURS for merging into
2133 MINE with the '--ed' ('-e') option. You can select only the
2134 nonoverlapping unmerged changes with '--easy-only' ('-3'), and you can
2135 select only the overlapping changes with '--overlap-only' ('-x').
2137 The '-e', '-3' and '-x' options select only "unmerged changes", i.e.
2138 changes where MINE and YOURS differ; they ignore changes from OLDER to
2139 YOURS where MINE and YOURS are identical, because they assume that such
2140 changes have already been merged. If this assumption is not a safe one,
2141 you can use the '--show-all' ('-A') option (*note Marking Conflicts::).
2143 Here is the output of the command 'diff3' with each of these three
2144 options (*note Sample diff3 Input::, for the complete contents of the
2145 files). Notice that '-e' outputs the union of the disjoint sets of
2146 changes output by '-3' and '-x'.
2148 Output of 'diff3 -e lao tzu tao':
2151 -- The Way of Lao-Tzu, tr. Wing-tsit Chan
2154 so we may see their result.
2157 Output of 'diff3 -3 lao tzu tao':
2159 so we may see their result.
2162 Output of 'diff3 -x lao tzu tao':
2165 -- The Way of Lao-Tzu, tr. Wing-tsit Chan
2169 File: diffutils.info, Node: Marking Conflicts, Next: Bypassing ed, Prev: Which Changes, Up: diff3 Merging
2171 8.2 Marking Conflicts
2172 =====================
2174 'diff3' can mark conflicts in the merged output by bracketing them with
2175 special marker lines. A conflict that comes from two files A and B is
2184 A conflict that comes from three files A, B and C is marked as
2195 The '--show-all' ('-A') option acts like the '-e' option, except that
2196 it brackets conflicts, and it outputs all changes from OLDER to YOURS,
2197 not just the unmerged changes. Thus, given the sample input files
2198 (*note Sample diff3 Input::), 'diff3 -A lao tzu tao' puts brackets
2199 around the conflict where only 'tzu' differs:
2203 The Way that can be told of is not the eternal Way;
2204 The name that can be named is not the eternal name.
2207 And it outputs the three-way conflict as follows:
2211 They both may be called deep and profound.
2212 Deeper and more profound,
2213 The door of all subtleties!
2216 -- The Way of Lao-Tzu, tr. Wing-tsit Chan
2219 The '--show-overlap' ('-E') option outputs less information than the
2220 '--show-all' ('-A') option, because it outputs only unmerged changes,
2221 and it never outputs the contents of the second file. Thus the '-E'
2222 option acts like the '-e' option, except that it brackets the first and
2223 third files from three-way overlapping changes. Similarly, '-X' acts
2224 like '-x', except it brackets all its (necessarily overlapping) changes.
2225 For example, for the three-way overlapping change above, the '-E' and
2226 '-X' options output the following:
2231 -- The Way of Lao-Tzu, tr. Wing-tsit Chan
2234 If you are comparing files that have meaningless or uninformative
2235 names, you can use the '--label=LABEL' option to show alternate names in
2236 the '<<<<<<<', '|||||||' and '>>>>>>>' brackets. This option can be
2237 given up to three times, once for each input file. Thus 'diff3 -A
2238 --label X --label Y --label Z A B C' acts like 'diff3 -A A B C', except
2239 that the output looks like it came from files named 'X', 'Y' and 'Z'
2240 rather than from files named 'A', 'B' and 'C'.
2243 File: diffutils.info, Node: Bypassing ed, Next: Merging Incomplete Lines, Prev: Marking Conflicts, Up: diff3 Merging
2245 8.3 Generating the Merged Output Directly
2246 =========================================
2248 With the '--merge' ('-m') option, 'diff3' outputs the merged file
2249 directly. This is more efficient than using 'ed' to generate it, and
2250 works even with non-text files that 'ed' would reject. If you specify
2251 '-m' without an 'ed' script option, '-A' is assumed.
2253 For example, the command 'diff3 -m lao tzu tao' (*note Sample diff3
2254 Input:: for a copy of the input files) would output the following:
2258 The Way that can be told of is not the eternal Way;
2259 The name that can be named is not the eternal name.
2261 The Nameless is the origin of Heaven and Earth;
2262 The Named is the mother of all things.
2263 Therefore let there always be non-being,
2264 so we may see their subtlety,
2265 And let there always be being,
2266 so we may see their result.
2267 The two are the same,
2268 But after they are produced,
2269 they have different names.
2272 They both may be called deep and profound.
2273 Deeper and more profound,
2274 The door of all subtleties!
2277 -- The Way of Lao-Tzu, tr. Wing-tsit Chan
2281 File: diffutils.info, Node: Merging Incomplete Lines, Next: Saving the Changed File, Prev: Bypassing ed, Up: diff3 Merging
2283 8.4 How 'diff3' Merges Incomplete Lines
2284 =======================================
2286 With '-m', incomplete lines (*note Incomplete Lines::) are simply copied
2287 to the output as they are found; if the merged output ends in an
2288 conflict and one of the input files ends in an incomplete line,
2289 succeeding '|||||||', '=======' or '>>>>>>>' brackets appear somewhere
2290 other than the start of a line because they are appended to the
2293 Without '-m', if an 'ed' script option is specified and an incomplete
2294 line is found, 'diff3' generates a warning and acts as if a newline had
2298 File: diffutils.info, Node: Saving the Changed File, Prev: Merging Incomplete Lines, Up: diff3 Merging
2300 8.5 Saving the Changed File
2301 ===========================
2303 Traditional Unix 'diff3' generates an 'ed' script without the trailing
2304 'w' and 'q' commands that save the changes. System V 'diff3' generates
2305 these extra commands. GNU 'diff3' normally behaves like traditional
2306 Unix 'diff3', but with the '-i' option it behaves like System V 'diff3'
2307 and appends the 'w' and 'q' commands.
2309 The '-i' option requires one of the 'ed' script options '-AeExX3',
2310 and is incompatible with the merged output option '-m'.
2313 File: diffutils.info, Node: Interactive Merging, Next: Merging with patch, Prev: diff3 Merging, Up: Top
2315 9 Interactive Merging with 'sdiff'
2316 **********************************
2318 With 'sdiff', you can merge two files interactively based on a
2319 side-by-side '-y' format comparison (*note Side by Side::). Use
2320 '--output=FILE' ('-o FILE') to specify where to put the merged text.
2321 *Note Invoking sdiff::, for more details on the options to 'sdiff'.
2323 Another way to merge files interactively is to use the Emacs Lisp
2324 package 'emerge'. *Note Emerge: (emacs)Emerge, for more information.
2328 * sdiff Option Summary:: Summary of 'sdiff' options.
2329 * Merge Commands:: Merging two files interactively.
2332 File: diffutils.info, Node: sdiff Option Summary, Next: Merge Commands, Up: Interactive Merging
2334 9.1 Specifying 'diff' Options to 'sdiff'
2335 ========================================
2337 The following 'sdiff' options have the same meaning as for 'diff'.
2338 *Note diff Options::, for the use of these options.
2344 --ignore-blank-lines --ignore-case
2345 --ignore-matching-lines=REGEXP --ignore-space-change
2346 --ignore-tab-expansion --ignore-trailing-space
2347 --left-column --minimal --speed-large-files
2348 --strip-trailing-cr --suppress-common-lines
2349 --tabsize=COLUMNS --text --version --width=COLUMNS
2351 For historical reasons, 'sdiff' has alternate names for some options.
2352 The '-l' option is equivalent to the '--left-column' option, and
2353 similarly '-s' is equivalent to '--suppress-common-lines'. The meaning
2354 of the 'sdiff' '-w' and '-W' options is interchanged from that of
2355 'diff': with 'sdiff', '-w COLUMNS' is equivalent to '--width=COLUMNS',
2356 and '-W' is equivalent to '--ignore-all-space'. 'sdiff' without the
2357 '-o' option is equivalent to 'diff' with the '--side-by-side' ('-y')
2358 option (*note Side by Side::).
2361 File: diffutils.info, Node: Merge Commands, Prev: sdiff Option Summary, Up: Interactive Merging
2366 Groups of common lines, with a blank gutter, are copied from the first
2367 file to the output. After each group of differing lines, 'sdiff'
2368 prompts with '%' and pauses, waiting for one of the following commands.
2369 Follow each command with <RET>.
2372 Discard both versions. Invoke a text editor on an empty temporary
2373 file, then copy the resulting file to the output.
2376 Concatenate the two versions, edit the result in a temporary file,
2377 then copy the edited result to the output.
2380 Like 'eb', except precede each version with a header that shows
2381 what file and lines the version came from.
2385 Edit a copy of the left version, then copy the result to the
2390 Edit a copy of the right version, then copy the result to the
2395 Copy the left version to the output.
2402 Copy the right version to the output.
2405 Silently copy common lines.
2408 Verbosely copy common lines. This is the default.
2410 The text editor invoked is specified by the 'EDITOR' environment
2411 variable if it is set. The default is system-dependent.
2414 File: diffutils.info, Node: Merging with patch, Next: Making Patches, Prev: Interactive Merging, Up: Top
2416 10 Merging with 'patch'
2417 ***********************
2419 'patch' takes comparison output produced by 'diff' and applies the
2420 differences to a copy of the original file, producing a patched version.
2421 With 'patch', you can distribute just the changes to a set of files
2422 instead of distributing the entire file set; your correspondents can
2423 apply 'patch' to update their copy of the files with your changes.
2424 'patch' automatically determines the diff format, skips any leading or
2425 trailing headers, and uses the headers to determine which file to patch.
2426 This lets your correspondents feed a mail message containing a
2427 difference listing directly to 'patch'.
2429 'patch' detects and warns about common problems like forward patches.
2430 It saves any patches that it could not apply. It can also maintain a
2431 'patchlevel.h' file to ensure that your correspondents apply diffs in
2434 'patch' accepts a series of diffs in its standard input, usually
2435 separated by headers that specify which file to patch. It applies
2436 'diff' hunks (*note Hunks::) one by one. If a hunk does not exactly
2437 match the original file, 'patch' uses heuristics to try to patch the
2438 file as well as it can. If no approximate match can be found, 'patch'
2439 rejects the hunk and skips to the next hunk. 'patch' normally replaces
2440 each file F with its new version, putting reject hunks (if any) into
2443 *Note Invoking patch::, for detailed information on the options to
2448 * patch Input:: Selecting the type of 'patch' input.
2449 * Revision Control:: Getting files from RCS, SCCS, etc.
2450 * Imperfect:: Dealing with imperfect patches.
2451 * Creating and Removing:: Creating and removing files with a patch.
2452 * Patching Timestamps:: Updating timestamps on patched files.
2453 * Multiple Patches:: Handling multiple patches in a file.
2454 * patch Directories:: Changing directory and stripping directories.
2455 * Backups:: Whether backup files are made.
2456 * Backup Names:: Backup file names.
2457 * Reject Names:: Reject file names.
2458 * patch Messages:: Messages and questions 'patch' can produce.
2459 * patch and POSIX:: Conformance to the POSIX standard.
2460 * patch and Tradition:: GNU versus traditional 'patch'.
2463 File: diffutils.info, Node: patch Input, Next: Revision Control, Up: Merging with patch
2465 10.1 Selecting the 'patch' Input Format
2466 =======================================
2468 'patch' normally determines which 'diff' format the patch file uses by
2469 examining its contents. For patch files that contain particularly
2470 confusing leading text, you might need to use one of the following
2471 options to force 'patch' to interpret the patch file as a certain format
2472 of diff. The output formats listed here are the only ones that 'patch'
2492 File: diffutils.info, Node: Revision Control, Next: Imperfect, Prev: patch Input, Up: Merging with patch
2494 10.2 Revision Control
2495 =====================
2497 If a nonexistent input file is under a revision control system supported
2498 by 'patch', 'patch' normally asks the user whether to get (or check out)
2499 the file from the revision control system. Patch currently supports
2500 RCS, ClearCase and SCCS. Under RCS and SCCS, 'patch' also asks when the
2501 input file is read-only and matches the default version in the revision
2504 The '--get=NUM' ('-g NUM') option affects access to files under
2505 supported revision control systems. If NUM is positive, 'patch' gets
2506 the file without asking the user; if zero, 'patch' neither asks the user
2507 nor gets the file; and if negative, 'patch' asks the user before getting
2508 the file. The default value of NUM is given by the value of the
2509 'PATCH_GET' environment variable if it is set; if not, the default value
2510 is zero if 'patch' is conforming to POSIX, negative otherwise. *Note
2513 The choice of revision control system is unaffected by the
2514 'VERSION_CONTROL' environment variable (*note Backup Names::).
2517 File: diffutils.info, Node: Imperfect, Next: Creating and Removing, Prev: Revision Control, Up: Merging with patch
2519 10.3 Applying Imperfect Patches
2520 ===============================
2522 'patch' tries to skip any leading text in the patch file, apply the
2523 diff, and then skip any trailing text. Thus you can feed a mail message
2524 directly to 'patch', and it should work. If the entire diff is indented
2525 by a constant amount of white space, 'patch' automatically ignores the
2526 indentation. If a context diff contains trailing carriage return on
2527 each line, 'patch' automatically ignores the carriage return. If a
2528 context diff has been encapsulated by prepending '- ' to lines beginning
2529 with '-' as per Internet RFC 934
2530 (ftp://ftp.isi.edu/in-notes/rfc934.txt), 'patch' automatically
2531 unencapsulates the input.
2533 However, certain other types of imperfect input require user
2534 intervention or testing.
2538 * Changed White Space:: When tabs and spaces don't match exactly.
2539 * Reversed Patches:: Applying reversed patches correctly.
2540 * Inexact:: Helping 'patch' find close matches.
2541 * Dry Runs:: Predicting what 'patch' will do.
2544 File: diffutils.info, Node: Changed White Space, Next: Reversed Patches, Up: Imperfect
2546 10.3.1 Applying Patches with Changed White Space
2547 ------------------------------------------------
2549 Sometimes mailers, editors, or other programs change spaces into tabs,
2550 or vice versa. If this happens to a patch file or an input file, the
2551 files might look the same, but 'patch' will not be able to match them
2552 properly. If this problem occurs, use the '-l' or
2553 '--ignore-white-space' option, which makes 'patch' compare blank
2554 characters (i.e. spaces and tabs) loosely so that any nonempty sequence
2555 of blanks in the patch file matches any nonempty sequence of blanks in
2556 the input files. Non-blank characters must still match exactly. Each
2557 line of the context must still match a line in the input file.
2560 File: diffutils.info, Node: Reversed Patches, Next: Inexact, Prev: Changed White Space, Up: Imperfect
2562 10.3.2 Applying Reversed Patches
2563 --------------------------------
2565 Sometimes people run 'diff' with the new file first instead of second.
2566 This creates a diff that is "reversed". To apply such patches, give
2567 'patch' the '--reverse' ('-R') option. 'patch' then attempts to swap
2568 each hunk around before applying it. Rejects come out in the swapped
2571 Often 'patch' can guess that the patch is reversed. If the first
2572 hunk of a patch fails, 'patch' reverses the hunk to see if it can apply
2573 it that way. If it can, 'patch' asks you if you want to have the '-R'
2574 option set; if it can't, 'patch' continues to apply the patch normally.
2575 This method cannot detect a reversed patch if it is a normal diff and
2576 the first command is an append (which should have been a delete) since
2577 appends always succeed, because a null context matches anywhere. But
2578 most patches add or change lines rather than delete them, so most
2579 reversed normal diffs begin with a delete, which fails, and 'patch'
2582 If you apply a patch that you have already applied, 'patch' thinks it
2583 is a reversed patch and offers to un-apply the patch. This could be
2584 construed as a feature. If you did this inadvertently and you don't
2585 want to un-apply the patch, just answer 'n' to this offer and to the
2586 subsequent "apply anyway" question--or type 'C-c' to kill the 'patch'
2590 File: diffutils.info, Node: Inexact, Next: Dry Runs, Prev: Reversed Patches, Up: Imperfect
2592 10.3.3 Helping 'patch' Find Inexact Matches
2593 -------------------------------------------
2595 For context diffs, and to a lesser extent normal diffs, 'patch' can
2596 detect when the line numbers mentioned in the patch are incorrect, and
2597 it attempts to find the correct place to apply each hunk of the patch.
2598 As a first guess, it takes the line number mentioned in the hunk, plus
2599 or minus any offset used in applying the previous hunk. If that is not
2600 the correct place, 'patch' scans both forward and backward for a set of
2601 lines matching the context given in the hunk.
2603 First 'patch' looks for a place where all lines of the context match.
2604 If it cannot find such a place, and it is reading a context or unified
2605 diff, and the maximum fuzz factor is set to 1 or more, then 'patch'
2606 makes another scan, ignoring the first and last line of context. If
2607 that fails, and the maximum fuzz factor is set to 2 or more, it makes
2608 another scan, ignoring the first two and last two lines of context are
2609 ignored. It continues similarly if the maximum fuzz factor is larger.
2611 The '--fuzz=LINES' ('-F LINES') option sets the maximum fuzz factor
2612 to LINES. This option only applies to context and unified diffs; it
2613 ignores up to LINES lines while looking for the place to install a hunk.
2614 Note that a larger fuzz factor increases the odds of making a faulty
2615 patch. The default fuzz factor is 2; there is no point to setting it to
2616 more than the number of lines of context in the diff, ordinarily 3.
2618 If 'patch' cannot find a place to install a hunk of the patch, it
2619 writes the hunk out to a reject file (*note Reject Names::, for
2620 information on how reject files are named). It writes out rejected
2621 hunks in context format no matter what form the input patch is in. If
2622 the input is a normal or 'ed' diff, many of the contexts are simply
2623 null. The line numbers on the hunks in the reject file may be different
2624 from those in the patch file: they show the approximate location where
2625 'patch' thinks the failed hunks belong in the new file rather than in
2628 If the '--verbose' option is given, then as it completes each hunk
2629 'patch' tells you whether the hunk succeeded or failed, and if it
2630 failed, on which line (in the new file) 'patch' thinks the hunk should
2631 go. If this is different from the line number specified in the diff, it
2632 tells you the offset. A single large offset _may_ indicate that 'patch'
2633 installed a hunk in the wrong place. 'patch' also tells you if it used
2634 a fuzz factor to make the match, in which case you should also be
2635 slightly suspicious.
2637 'patch' cannot tell if the line numbers are off in an 'ed' script,
2638 and can only detect wrong line numbers in a normal diff when it finds a
2639 change or delete command. It may have the same problem with a context
2640 diff using a fuzz factor equal to or greater than the number of lines of
2641 context shown in the diff (typically 3). In these cases, you should
2642 probably look at a context diff between your original and patched input
2643 files to see if the changes make sense. Compiling without errors is a
2644 pretty good indication that the patch worked, but not a guarantee.
2646 A patch against an empty file applies to a nonexistent file, and vice
2647 versa. *Note Creating and Removing::.
2649 'patch' usually produces the correct results, even when it must make
2650 many guesses. However, the results are guaranteed only when the patch
2651 is applied to an exact copy of the file that the patch was generated
2655 File: diffutils.info, Node: Dry Runs, Prev: Inexact, Up: Imperfect
2657 10.3.4 Predicting what 'patch' will do
2658 --------------------------------------
2660 It may not be obvious in advance what 'patch' will do with a complicated
2661 or poorly formatted patch. If you are concerned that the input might
2662 cause 'patch' to modify the wrong files, you can use the '--dry-run'
2663 option, which causes 'patch' to print the results of applying patches
2664 without actually changing any files. You can then inspect the
2665 diagnostics generated by the dry run to see whether 'patch' will modify
2666 the files that you expect. If the patch does not do what you want, you
2667 can modify the patch (or the other options to 'patch') and try another
2668 dry run. Once you are satisfied with the proposed patch you can apply
2669 it by invoking 'patch' as before, but this time without the '--dry-run'
2673 File: diffutils.info, Node: Creating and Removing, Next: Patching Timestamps, Prev: Imperfect, Up: Merging with patch
2675 10.4 Creating and Removing Files
2676 ================================
2678 Sometimes when comparing two directories, a file may exist in one
2679 directory but not the other. If you give 'diff' the '--new-file' ('-N')
2680 option, or if you supply an old or new file that is named '/dev/null' or
2681 is empty and is dated the Epoch (1970-01-01 00:00:00 UTC), 'diff'
2682 outputs a patch that adds or deletes the contents of this file. When
2683 given such a patch, 'patch' normally creates a new file or removes the
2684 old file. However, when conforming to POSIX (*note patch and POSIX::),
2685 'patch' does not remove the old file, but leaves it empty. The
2686 '--remove-empty-files' ('-E') option causes 'patch' to remove output
2687 files that are empty after applying a patch, even if the patch does not
2688 appear to be one that removed the file.
2690 If the patch appears to create a file that already exists, 'patch'
2691 asks for confirmation before applying the patch.
2694 File: diffutils.info, Node: Patching Timestamps, Next: Multiple Patches, Prev: Creating and Removing, Up: Merging with patch
2696 10.5 Updating Timestamps on Patched Files
2697 =========================================
2699 When 'patch' updates a file, it normally sets the file's last-modified
2700 timestamp to the current time of day. If you are using 'patch' to track
2701 a software distribution, this can cause 'make' to incorrectly conclude
2702 that a patched file is out of date. For example, if 'syntax.c' depends
2703 on 'syntax.y', and 'patch' updates 'syntax.c' and then 'syntax.y', then
2704 'syntax.c' will normally appear to be out of date with respect to
2705 'syntax.y' even though its contents are actually up to date.
2707 The '--set-utc' ('-Z') option causes 'patch' to set a patched file's
2708 modification and access times to the timestamps given in context diff
2709 headers. If the context diff headers do not specify a time zone, they
2710 are assumed to use Coordinated Universal Time (UTC, often known as GMT).
2712 The '--set-time' ('-T') option acts like '-Z' or '--set-utc', except
2713 that it assumes that the context diff headers' timestamps use local time
2714 instead of UTC. This option is not recommended, because patches using
2715 local time cannot easily be used by people in other time zones, and
2716 because local timestamps are ambiguous when local clocks move backwards
2717 during daylight-saving time adjustments. If the context diff headers
2718 specify a time zone, this option is equivalent to '--set-utc' ('-Z').
2720 'patch' normally refrains from setting a file's timestamps if the
2721 file's original last-modified timestamp does not match the time given in
2722 the diff header, of if the file's contents do not exactly match the
2723 patch. However, if the '--force' ('-f') option is given, the file's
2724 timestamps are set regardless.
2726 Due to the limitations of the current 'diff' format, 'patch' cannot
2727 update the times of files whose contents have not changed. Also, if you
2728 set file timestamps to values other than the current time of day, you
2729 should also remove (e.g., with 'make clean') all files that depend on
2730 the patched files, so that later invocations of 'make' do not get
2731 confused by the patched files' times.
2734 File: diffutils.info, Node: Multiple Patches, Next: patch Directories, Prev: Patching Timestamps, Up: Merging with patch
2736 10.6 Multiple Patches in a File
2737 ===============================
2739 If the patch file contains more than one patch, and if you do not
2740 specify an input file on the command line, 'patch' tries to apply each
2741 patch as if they came from separate patch files. This means that it
2742 determines the name of the file to patch for each patch, and that it
2743 examines the leading text before each patch for file names and
2744 prerequisite revision level (*note Making Patches::, for more on that
2747 'patch' uses the following rules to intuit a file name from the
2748 leading text before a patch. First, 'patch' takes an ordered list of
2749 candidate file names as follows:
2751 * If the header is that of a context diff, 'patch' takes the old and
2752 new file names in the header. A name is ignored if it does not
2753 have enough slashes to satisfy the '-pNUM' or '--strip=NUM' option.
2754 The name '/dev/null' is also ignored.
2756 * If there is an 'Index:' line in the leading garbage and if either
2757 the old and new names are both absent or if 'patch' is conforming
2758 to POSIX, 'patch' takes the name in the 'Index:' line.
2760 * For the purpose of the following rules, the candidate file names
2761 are considered to be in the order (old, new, index), regardless of
2762 the order that they appear in the header.
2764 Then 'patch' selects a file name from the candidate list as follows:
2766 * If some of the named files exist, 'patch' selects the first name if
2767 conforming to POSIX, and the best name otherwise.
2769 * If 'patch' is not ignoring RCS, ClearCase, and SCCS (*note Revision
2770 Control::), and no named files exist but an RCS, ClearCase, or SCCS
2771 master is found, 'patch' selects the first named file with an RCS,
2772 ClearCase, or SCCS master.
2774 * If no named files exist, no RCS, ClearCase, or SCCS master was
2775 found, some names are given, 'patch' is not conforming to POSIX,
2776 and the patch appears to create a file, 'patch' selects the best
2777 name requiring the creation of the fewest directories.
2779 * If no file name results from the above heuristics, you are asked
2780 for the name of the file to patch, and 'patch' selects that name.
2782 To determine the "best" of a nonempty list of file names, 'patch'
2783 first takes all the names with the fewest path name components; of
2784 those, it then takes all the names with the shortest basename; of those,
2785 it then takes all the shortest names; finally, it takes the first
2788 *Note patch and POSIX::, to see whether 'patch' is conforming to
2792 File: diffutils.info, Node: patch Directories, Next: Backups, Prev: Multiple Patches, Up: Merging with patch
2794 10.7 Applying Patches in Other Directories
2795 ==========================================
2797 The '--directory=DIRECTORY' ('-d DIRECTORY') option to 'patch' makes
2798 directory DIRECTORY the current directory for interpreting both file
2799 names in the patch file, and file names given as arguments to other
2800 options (such as '-B' and '-o'). For example, while in a mail reading
2801 program, you can patch a file in the '/usr/src/emacs' directory directly
2802 from a message containing the patch like this:
2804 | patch -d /usr/src/emacs
2806 Sometimes the file names given in a patch contain leading
2807 directories, but you keep your files in a directory different from the
2808 one given in the patch. In those cases, you can use the
2809 '--strip=NUMBER' ('-pNUMBER') option to set the file name strip count to
2810 NUMBER. The strip count tells 'patch' how many slashes, along with the
2811 directory names between them, to strip from the front of file names. A
2812 sequence of one or more adjacent slashes is counted as a single slash.
2813 By default, 'patch' strips off all leading directories, leaving just the
2816 For example, suppose the file name in the patch file is
2817 '/gnu/src/emacs/etc/NEWS'. Using '-p0' gives the entire file name
2818 unmodified, '-p1' gives 'gnu/src/emacs/etc/NEWS' (no leading slash),
2819 '-p4' gives 'etc/NEWS', and not specifying '-p' at all gives 'NEWS'.
2821 'patch' looks for each file (after any slashes have been stripped) in
2822 the current directory, or if you used the '-d DIRECTORY' option, in that
2826 File: diffutils.info, Node: Backups, Next: Backup Names, Prev: patch Directories, Up: Merging with patch
2831 Normally, 'patch' creates a backup file if the patch does not exactly
2832 match the original input file, because in that case the original data
2833 might not be recovered if you undo the patch with 'patch -R' (*note
2834 Reversed Patches::). However, when conforming to POSIX, 'patch' does
2835 not create backup files by default. *Note patch and POSIX::.
2837 The '--backup' ('-b') option causes 'patch' to make a backup file
2838 regardless of whether the patch matches the original input. The
2839 '--backup-if-mismatch' option causes 'patch' to create backup files for
2840 mismatches files; this is the default when not conforming to POSIX. The
2841 '--no-backup-if-mismatch' option causes 'patch' to not create backup
2842 files, even for mismatched patches; this is the default when conforming
2845 When backing up a file that does not exist, an empty, unreadable
2846 backup file is created as a placeholder to represent the nonexistent
2850 File: diffutils.info, Node: Backup Names, Next: Reject Names, Prev: Backups, Up: Merging with patch
2852 10.9 Backup File Names
2853 ======================
2855 Normally, 'patch' renames an original input file into a backup file by
2856 appending to its name the extension '.orig', or '~' if using '.orig'
2857 would make the backup file name too long.(1) The '-z BACKUP-SUFFIX' or
2858 '--suffix=BACKUP-SUFFIX' option causes 'patch' to use BACKUP-SUFFIX as
2859 the backup extension instead.
2861 Alternately, you can specify the extension for backup files with the
2862 'SIMPLE_BACKUP_SUFFIX' environment variable, which the options override.
2864 'patch' can also create numbered backup files the way GNU Emacs does.
2865 With this method, instead of having a single backup of each file,
2866 'patch' makes a new backup file name each time it patches a file. For
2867 example, the backups of a file named 'sink' would be called,
2868 successively, 'sink.~1~', 'sink.~2~', 'sink.~3~', etc.
2870 The '-V BACKUP-STYLE' or '--version-control=BACKUP-STYLE' option
2871 takes as an argument a method for creating backup file names. You can
2872 alternately control the type of backups that 'patch' makes with the
2873 'PATCH_VERSION_CONTROL' environment variable, which the '-V' option
2874 overrides. If 'PATCH_VERSION_CONTROL' is not set, the 'VERSION_CONTROL'
2875 environment variable is used instead. Please note that these options
2876 and variables control backup file names; they do not affect the choice
2877 of revision control system (*note Revision Control::).
2879 The values of these environment variables and the argument to the
2880 '-V' option are like the GNU Emacs 'version-control' variable (*note
2881 (emacs)Backup Names::, for more information on backup versions in
2882 Emacs). They also recognize synonyms that are more descriptive. The
2883 valid values are listed below; unique abbreviations are acceptable.
2887 Always make numbered backups.
2891 Make numbered backups of files that already have them, simple
2892 backups of the others. This is the default.
2896 Always make simple backups.
2898 You can also tell 'patch' to prepend a prefix, such as a directory
2899 name, to produce backup file names. The '--prefix=PREFIX' ('-B PREFIX')
2900 option makes backup files by prepending PREFIX to them. The
2901 '--basename-prefix=PREFIX' ('-Y PREFIX') prepends PREFIX to the last
2902 file name component of backup file names instead; for example, '-Y ~'
2903 causes the backup name for 'dir/file.c' to be 'dir/~file.c'. If you use
2904 either of these prefix options, the suffix-based options are ignored.
2906 If you specify the output file with the '-o' option, that file is the
2907 one that is backed up, not the input file.
2909 Options that affect the names of backup files do not affect whether
2910 backups are made. For example, if you specify the
2911 '--no-backup-if-mismatch' option, none of the options described in this
2912 section have any affect, because no backups are made.
2914 ---------- Footnotes ----------
2916 (1) A coding error in GNU 'patch' version 2.5.4 causes it to always
2917 use '~', but this should be fixed in the next release.
2920 File: diffutils.info, Node: Reject Names, Next: patch Messages, Prev: Backup Names, Up: Merging with patch
2922 10.10 Reject File Names
2923 =======================
2925 The names for reject files (files containing patches that 'patch' could
2926 not find a place to apply) are normally the name of the output file with
2927 '.rej' appended (or '#' if using '.rej' would make the backup file name
2930 Alternatively, you can tell 'patch' to place all of the rejected
2931 patches in a single file. The '-r REJECT-FILE' or
2932 '--reject-file=REJECT-FILE' option uses REJECT-FILE as the reject file
2936 File: diffutils.info, Node: patch Messages, Next: patch and POSIX, Prev: Reject Names, Up: Merging with patch
2938 10.11 Messages and Questions from 'patch'
2939 =========================================
2941 'patch' can produce a variety of messages, especially if it has trouble
2942 decoding its input. In a few situations where it's not sure how to
2943 proceed, 'patch' normally prompts you for more information from the
2944 keyboard. There are options to produce more or fewer messages, to have
2945 it not ask for keyboard input, and to affect the way that file names are
2950 * More or Fewer Messages:: Controlling the verbosity of 'patch'.
2951 * patch and Keyboard Input:: Inhibiting keyboard input.
2952 * patch Quoting Style:: Quoting file names in diagnostics.
2954 'patch' exits with status 0 if all hunks are applied successfully, 1
2955 if some hunks cannot be applied, and 2 if there is more serious trouble.
2956 When applying a set of patches in a loop, you should check the exit
2957 status, so you don't apply a later patch to a partially patched file.
2960 File: diffutils.info, Node: More or Fewer Messages, Next: patch and Keyboard Input, Up: patch Messages
2962 10.11.1 Controlling the Verbosity of 'patch'
2963 --------------------------------------------
2965 You can cause 'patch' to produce more messages by using the '--verbose'
2966 option. For example, when you give this option, the message 'Hmm...'
2967 indicates that 'patch' is reading text in the patch file, attempting to
2968 determine whether there is a patch in that text, and if so, what kind of
2971 You can inhibit all terminal output from 'patch', unless an error
2972 occurs, by using the '-s', '--quiet', or '--silent' option.
2975 File: diffutils.info, Node: patch and Keyboard Input, Next: patch Quoting Style, Prev: More or Fewer Messages, Up: patch Messages
2977 10.11.2 Inhibiting Keyboard Input
2978 ---------------------------------
2980 There are two ways you can prevent 'patch' from asking you any
2981 questions. The '--force' ('-f') option assumes that you know what you
2982 are doing. It causes 'patch' to do the following:
2984 * Skip patches that do not contain file names in their headers.
2986 * Patch files even though they have the wrong version for the
2987 'Prereq:' line in the patch;
2989 * Assume that patches are not reversed even if they look like they
2992 The '--batch' ('-t') option is similar to '-f', in that it suppresses
2993 questions, but it makes somewhat different assumptions:
2995 * Skip patches that do not contain file names in their headers (the
2998 * Skip patches for which the file has the wrong version for the
2999 'Prereq:' line in the patch;
3001 * Assume that patches are reversed if they look like they are.
3004 File: diffutils.info, Node: patch Quoting Style, Prev: patch and Keyboard Input, Up: patch Messages
3006 10.11.3 'patch' Quoting Style
3007 -----------------------------
3009 When 'patch' outputs a file name in a diagnostic message, it can format
3010 the name in any of several ways. This can be useful to output file
3011 names unambiguously, even if they contain punctuation or special
3012 characters like newlines. The '--quoting-style=WORD' option controls
3013 how names are output. The WORD should be one of the following:
3018 Quote names for the shell if they contain shell metacharacters or
3019 would cause ambiguous output.
3021 Quote names for the shell, even if they would normally not require
3024 Quote names as for a C language string.
3026 Quote as with 'c' except omit the surrounding double-quote
3029 You can specify the default value of the '--quoting-style' option
3030 with the environment variable 'QUOTING_STYLE'. If that environment
3031 variable is not set, the default value is 'shell', but this default may
3032 change in a future version of 'patch'.
3035 File: diffutils.info, Node: patch and POSIX, Next: patch and Tradition, Prev: patch Messages, Up: Merging with patch
3037 10.12 'patch' and the POSIX Standard
3038 ====================================
3040 If you specify the '--posix' option, or set the 'POSIXLY_CORRECT'
3041 environment variable, 'patch' conforms more strictly to the POSIX
3042 standard, as follows:
3044 * Take the first existing file from the list (old, new, index) when
3045 intuiting file names from diff headers. *Note Multiple Patches::.
3047 * Do not remove files that are removed by a diff. *Note Creating and
3050 * Do not ask whether to get files from RCS, ClearCase, or SCCS. *Note
3053 * Require that all options precede the files in the command line.
3055 * Do not backup files, even when there is a mismatch. *Note
3059 File: diffutils.info, Node: patch and Tradition, Prev: patch and POSIX, Up: Merging with patch
3061 10.13 GNU 'patch' and Traditional 'patch'
3062 =========================================
3064 The current version of GNU 'patch' normally follows the POSIX standard.
3065 *Note patch and POSIX::, for the few exceptions to this general rule.
3067 Unfortunately, POSIX redefined the behavior of 'patch' in several
3068 important ways. You should be aware of the following differences if you
3069 must interoperate with traditional 'patch', or with GNU 'patch' version
3072 * In traditional 'patch', the '-p' option's operand was optional, and
3073 a bare '-p' was equivalent to '-p0'. The '-p' option now requires
3074 an operand, and '-p 0' is now equivalent to '-p0'. For maximum
3075 compatibility, use options like '-p0' and '-p1'.
3077 Also, traditional 'patch' simply counted slashes when stripping
3078 path prefixes; 'patch' now counts pathname components. That is, a
3079 sequence of one or more adjacent slashes now counts as a single
3080 slash. For maximum portability, avoid sending patches containing
3083 * In traditional 'patch', backups were enabled by default. This
3084 behavior is now enabled with the '--backup' ('-b') option.
3086 Conversely, in POSIX 'patch', backups are never made, even when
3087 there is a mismatch. In GNU 'patch', this behavior is enabled with
3088 the '--no-backup-if-mismatch' option, or by conforming to POSIX.
3090 The '-b SUFFIX' option of traditional 'patch' is equivalent to the
3091 '-b -z SUFFIX' options of GNU 'patch'.
3093 * Traditional 'patch' used a complicated (and incompletely
3094 documented) method to intuit the name of the file to be patched
3095 from the patch header. This method did not conform to POSIX, and
3096 had a few gotchas. Now 'patch' uses a different, equally
3097 complicated (but better documented) method that is optionally
3098 POSIX-conforming; we hope it has fewer gotchas. The two methods
3099 are compatible if the file names in the context diff header and the
3100 'Index:' line are all identical after prefix-stripping. Your patch
3101 is normally compatible if each header's file names all contain the
3102 same number of slashes.
3104 * When traditional 'patch' asked the user a question, it sent the
3105 question to standard error and looked for an answer from the first
3106 file in the following list that was a terminal: standard error,
3107 standard output, '/dev/tty', and standard input. Now 'patch' sends
3108 questions to standard output and gets answers from '/dev/tty'.
3109 Defaults for some answers have been changed so that 'patch' never
3110 goes into an infinite loop when using default answers.
3112 * Traditional 'patch' exited with a status value that counted the
3113 number of bad hunks, or with status 1 if there was real trouble.
3114 Now 'patch' exits with status 1 if some hunks failed, or with 2 if
3115 there was real trouble.
3117 * Limit yourself to the following options when sending instructions
3118 meant to be executed by anyone running GNU 'patch', traditional
3119 'patch', or a 'patch' that conforms to POSIX. Spaces are
3120 significant in the following list, and operands are required.
3135 File: diffutils.info, Node: Making Patches, Next: Invoking cmp, Prev: Merging with patch, Up: Top
3137 11 Tips for Making and Using Patches
3138 ************************************
3140 Use some common sense when making and using patches. For example, when
3141 sending bug fixes to a program's maintainer, send several small patches,
3142 one per independent subject, instead of one large, harder-to-digest
3143 patch that covers all the subjects.
3145 Here are some other things you should keep in mind if you are going
3146 to distribute patches for updating a software package.
3150 * Tips for Patch Producers:: Advice for making patches.
3151 * Tips for Patch Consumers:: Advice for using patches.
3152 * Avoiding Common Mistakes:: Avoiding common mistakes when using 'patch'.
3153 * Generating Smaller Patches:: How to generate smaller patches.
3156 File: diffutils.info, Node: Tips for Patch Producers, Next: Tips for Patch Consumers, Up: Making Patches
3158 11.1 Tips for Patch Producers
3159 =============================
3161 To create a patch that changes an older version of a package into a
3162 newer version, first make a copy of the older and newer versions in
3163 adjacent subdirectories. It is common to do that by unpacking 'tar'
3164 archives of the two versions.
3166 To generate the patch, use the command 'diff -Naur OLD NEW' where OLD
3167 and NEW identify the old and new directories. The names OLD and NEW
3168 should not contain any slashes. The '-N' option lets the patch create
3169 and remove files; '-a' lets the patch update non-text files; '-u'
3170 generates useful timestamps and enough context; and '-r' lets the patch
3171 update subdirectories. Here is an example command, using Bourne shell
3174 diff -Naur gcc-3.0.3 gcc-3.0.4
3176 Tell your recipients how to apply the patches. This should include
3177 which working directory to use, and which 'patch' options to use; the
3178 option '-p1' is recommended. Test your procedure by pretending to be a
3179 recipient and applying your patches to a copy of the original files.
3181 *Note Avoiding Common Mistakes::, for how to avoid common mistakes
3182 when generating a patch.
3185 File: diffutils.info, Node: Tips for Patch Consumers, Next: Avoiding Common Mistakes, Prev: Tips for Patch Producers, Up: Making Patches
3187 11.2 Tips for Patch Consumers
3188 =============================
3190 A patch producer should tell recipients how to apply the patches, so the
3191 first rule of thumb for a patch consumer is to follow the instructions
3192 supplied with the patch.
3194 GNU 'diff' can analyze files with arbitrarily long lines and files
3195 that end in incomplete lines. However, older versions of 'patch' cannot
3196 patch such files. If you are having trouble applying such patches, try
3197 upgrading to a recent version of GNU 'patch'.
3200 File: diffutils.info, Node: Avoiding Common Mistakes, Next: Generating Smaller Patches, Prev: Tips for Patch Consumers, Up: Making Patches
3202 11.3 Avoiding Common Mistakes
3203 =============================
3205 When producing a patch for multiple files, apply 'diff' to directories
3206 whose names do not have slashes. This reduces confusion when the patch
3207 consumer specifies the '-pNUMBER' option, since this option can have
3208 surprising results when the old and new file names have different
3209 numbers of slashes. For example, do not send a patch with a header that
3212 diff -Naur v2.0.29/prog/README prog/README
3213 --- v2.0.29/prog/README 2002-03-10 23:30:39.942229878 -0800
3214 +++ prog/README 2002-03-17 20:49:32.442260588 -0800
3216 because the two file names have different numbers of slashes, and
3217 different versions of 'patch' interpret the file names differently. To
3218 avoid confusion, send output that looks like this instead:
3220 diff -Naur v2.0.29/prog/README v2.0.30/prog/README
3221 --- v2.0.29/prog/README 2002-03-10 23:30:39.942229878 -0800
3222 +++ v2.0.30/prog/README 2002-03-17 20:49:32.442260588 -0800
3224 Make sure you have specified the file names correctly, either in a
3225 context diff header or with an 'Index:' line. Take care to not send out
3226 reversed patches, since these make people wonder whether they have
3227 already applied the patch.
3229 Avoid sending patches that compare backup file names like
3230 'README.orig' or 'README~', since this might confuse 'patch' into
3231 patching a backup file instead of the real file. Instead, send patches
3232 that compare the same base file names in different directories, e.g.
3233 'old/README' and 'new/README'.
3235 To save people from partially applying a patch before other patches
3236 that should have gone before it, you can make the first patch in the
3237 patch file update a file with a name like 'patchlevel.h' or 'version.c',
3238 which contains a patch level or version number. If the input file
3239 contains the wrong version number, 'patch' will complain immediately.
3241 An even clearer way to prevent this problem is to put a 'Prereq:'
3242 line before the patch. If the leading text in the patch file contains a
3243 line that starts with 'Prereq:', 'patch' takes the next word from that
3244 line (normally a version number) and checks whether the next input file
3245 contains that word, preceded and followed by either white space or a
3246 newline. If not, 'patch' prompts you for confirmation before
3247 proceeding. This makes it difficult to accidentally apply patches in
3251 File: diffutils.info, Node: Generating Smaller Patches, Prev: Avoiding Common Mistakes, Up: Making Patches
3253 11.4 Generating Smaller Patches
3254 ===============================
3256 The simplest way to generate a patch is to use 'diff -Naur' (*note Tips
3257 for Patch Producers::), but you might be able to reduce the size of the
3258 patch by renaming or removing some files before making the patch. If
3259 the older version of the package contains any files that the newer
3260 version does not, or if any files have been renamed between the two
3261 versions, make a list of 'rm' and 'mv' commands for the user to execute
3262 in the old version directory before applying the patch. Then run those
3263 commands yourself in the scratch directory.
3265 If there are any files that you don't need to include in the patch
3266 because they can easily be rebuilt from other files (for example, 'TAGS'
3267 and output from 'yacc' and 'makeinfo'), exclude them from the patch by
3268 giving 'diff' the '-x PATTERN' option (*note Comparing Directories::).
3269 If you want your patch to modify a derived file because your recipients
3270 lack tools to build it, make sure that the patch for the derived file
3271 follows any patches for files that it depends on, so that the
3272 recipients' timestamps will not confuse 'make'.
3274 Now you can create the patch using 'diff -Naur'. Make sure to
3275 specify the scratch directory first and the newer directory second.
3277 Add to the top of the patch a note telling the user any 'rm' and 'mv'
3278 commands to run before applying the patch. Then you can remove the
3281 You can also shrink the patch size by using fewer lines of context,
3282 but bear in mind that 'patch' typically needs at least two lines for
3283 proper operation when patches do not exactly match the input files.
3286 File: diffutils.info, Node: Invoking cmp, Next: Invoking diff, Prev: Making Patches, Up: Top
3291 The 'cmp' command compares two files, and if they differ, tells the
3292 first byte and line number where they differ or reports that one file is
3293 a prefix of the other. Bytes and lines are numbered starting with 1.
3294 The arguments of 'cmp' are as follows:
3296 cmp OPTIONS... FROM-FILE [TO-FILE [FROM-SKIP [TO-SKIP]]]
3298 The file name '-' is always the standard input. 'cmp' also uses the
3299 standard input if one file name is omitted. The FROM-SKIP and TO-SKIP
3300 operands specify how many bytes to ignore at the start of each file;
3301 they are equivalent to the '--ignore-initial=FROM-SKIP:TO-SKIP' option.
3303 By default, 'cmp' outputs nothing if the two files have the same
3304 contents. If the two files have bytes that differ, 'cmp' reports the
3305 location of the first difference to standard output:
3307 FROM-FILE TO-FILE differ: char BYTE-NUMBER, line LINE-NUMBER
3309 If one file is a prefix of the other, 'cmp' reports the shorter file's
3310 name to standard error, followed by a blank and extra information about
3313 cmp: EOF on SHORTER-FILE EXTRA-INFO
3315 The message formats can differ outside the POSIX locale. POSIX
3316 allows but does not require the EOF diagnostic's file name to be
3317 followed by a blank and additional information.
3319 An exit status of 0 means no differences were found, 1 means some
3320 differences were found, and 2 means trouble.
3324 * cmp Options:: Summary of options to 'cmp'.
3327 File: diffutils.info, Node: cmp Options, Up: Invoking cmp
3329 12.1 Options to 'cmp'
3330 =====================
3332 Below is a summary of all of the options that GNU 'cmp' accepts. Most
3333 options have two equivalent names, one of which is a single letter
3334 preceded by '-', and the other of which is a long name preceded by '--'.
3335 Multiple single letter options (unless they take an argument) can be
3336 combined into a single command line word: '-bl' is equivalent to '-b
3341 Print the differing bytes. Display control bytes as a '^' followed
3342 by a letter of the alphabet and precede bytes that have the high
3343 bit set with 'M-' (which stands for "meta").
3346 Output a summary of usage and then exit.
3349 '--ignore-initial=SKIP'
3350 Ignore any differences in the first SKIP bytes of the input files.
3351 Treat files with fewer than SKIP bytes as if they are empty. If
3352 SKIP is of the form 'FROM-SKIP:TO-SKIP', skip the first FROM-SKIP
3353 bytes of the first input file and the first TO-SKIP bytes of the
3358 Output the (decimal) byte numbers and (octal) values of all
3359 differing bytes, instead of the default standard output. Each
3360 output line contains a differing byte's number relative to the
3361 start of the input, followed by the differing byte values. Byte
3362 numbers start at 1. Also, output the EOF message if one file is
3363 shorter than the other.
3367 Compare at most COUNT input bytes.
3372 Do not print anything; only return an exit status indicating
3373 whether the files differ.
3377 Output version information and then exit.
3379 In the above table, operands that are byte counts are normally
3380 decimal, but may be preceded by '0' for octal and '0x' for hexadecimal.
3382 A byte count can be followed by a suffix to specify a multiple of
3383 that count; in this case an omitted integer is understood to be 1. A
3384 bare size letter, or one followed by 'iB', specifies a multiple using
3385 powers of 1024. A size letter followed by 'B' specifies powers of 1000
3386 instead. For example, '-n 4M' and '-n 4MiB' are equivalent to '-n
3387 4194304', whereas '-n 4MB' is equivalent to '-n 4000000'. This notation
3388 is upward compatible with the SI prefixes
3389 (http://www.bipm.fr/enus/3_SI/si-prefixes.html) for decimal multiples
3390 and with the IEC 60027-2 prefixes for binary multiples
3391 (http://physics.nist.gov/cuu/Units/binary.html).
3393 The following suffixes are defined. Large sizes like '1Y' may be
3394 rejected by your computer due to limitations of its arithmetic.
3397 kilobyte: 10^3 = 1000.
3401 kibibyte: 2^10 = 1024. 'K' is special: the SI prefix is 'k' and
3402 the IEC 60027-2 prefix is 'Ki', but tradition and POSIX use 'k' to
3405 megabyte: 10^6 = 1,000,000.
3408 mebibyte: 2^20 = 1,048,576.
3410 gigabyte: 10^9 = 1,000,000,000.
3413 gibibyte: 2^30 = 1,073,741,824.
3415 terabyte: 10^12 = 1,000,000,000,000.
3418 tebibyte: 2^40 = 1,099,511,627,776.
3420 petabyte: 10^15 = 1,000,000,000,000,000.
3423 pebibyte: 2^50 = 1,125,899,906,842,624.
3425 exabyte: 10^18 = 1,000,000,000,000,000,000.
3428 exbibyte: 2^60 = 1,152,921,504,606,846,976.
3430 zettabyte: 10^21 = 1,000,000,000,000,000,000,000
3433 2^70 = 1,180,591,620,717,411,303,424. ('Zi' is a GNU extension to
3436 yottabyte: 10^24 = 1,000,000,000,000,000,000,000,000.
3439 2^80 = 1,208,925,819,614,629,174,706,176. ('Yi' is a GNU extension
3443 File: diffutils.info, Node: Invoking diff, Next: Invoking diff3, Prev: Invoking cmp, Up: Top
3448 The format for running the 'diff' command is:
3450 diff OPTIONS... FILES...
3452 In the simplest case, two file names FROM-FILE and TO-FILE are given,
3453 and 'diff' compares the contents of FROM-FILE and TO-FILE. A file name
3454 of '-' stands for the standard input.
3456 If one file is a directory and the other is not, 'diff' compares the
3457 file in the directory whose name is that of the non-directory. The
3458 non-directory file must not be '-'.
3460 If two file names are given and both are directories, 'diff' compares
3461 corresponding files in both directories, in alphabetical order; this
3462 comparison is not recursive unless the '--recursive' ('-r') option is
3463 given. 'diff' never compares the actual contents of a directory as if
3464 it were a file. The file that is fully specified may not be standard
3465 input, because standard input is nameless and the notion of "file with
3466 the same name" does not apply.
3468 If the '--from-file=FILE' option is given, the number of file names
3469 is arbitrary, and FILE is compared to each named file. Similarly, if
3470 the '--to-file=FILE' option is given, each named file is compared to
3473 'diff' options begin with '-', so normally file names may not begin
3474 with '-'. However, '--' as an argument by itself treats the remaining
3475 arguments as file names even if they begin with '-'.
3477 An exit status of 0 means no differences were found, 1 means some
3478 differences were found, and 2 means trouble.
3482 * diff Options:: Summary of options to 'diff'.
3485 File: diffutils.info, Node: diff Options, Up: Invoking diff
3487 13.1 Options to 'diff'
3488 ======================
3490 Below is a summary of all of the options that GNU 'diff' accepts. Most
3491 options have two equivalent names, one of which is a single letter
3492 preceded by '-', and the other of which is a long name preceded by '--'.
3493 Multiple single letter options (unless they take an argument) can be
3494 combined into a single command line word: '-ac' is equivalent to '-a
3495 -c'. Long named options can be abbreviated to any unique prefix of
3496 their name. Brackets ([ and ]) indicate that an option takes an
3501 Treat all files as text and compare them line-by-line, even if they
3502 do not seem to be text. *Note Binary::.
3505 '--ignore-space-change'
3506 Ignore changes in amount of white space. *Note White Space::.
3509 '--ignore-blank-lines'
3510 Ignore changes that just insert or delete blank lines. *Note Blank
3514 Read and write data in binary mode. *Note Binary::.
3517 Use the context output format, showing three lines of context.
3518 *Note Context Format::.
3521 Specify whether to use color for distinguishing different contexts,
3522 like header, added or removed lines. WHEN may be omitted, or one
3524 * none Do not use color at all. This is the default when no
3525 -color option is specified.
3526 * auto Use color only if standard output is a terminal.
3527 * always Always use color.
3528 Specifying '--color' and no WHEN is equivalent to '--color=auto'.
3532 Use the context output format, showing LINES (an integer) lines of
3533 context, or three if LINES is not given. *Note Context Format::.
3534 For proper operation, 'patch' typically needs at least two lines of
3537 For compatibility 'diff' also supports an obsolete option syntax
3538 '-LINES' that has effect when combined with '-c', '-p', or '-u'.
3539 New scripts should use '-U LINES' ('-C LINES') instead.
3541 '--changed-group-format=FORMAT'
3542 Use FORMAT to output a line group containing differing lines from
3543 both files in if-then-else format. *Note Line Group Formats::.
3547 Change the algorithm perhaps find a smaller set of changes. This
3548 makes 'diff' slower (sometimes much slower). *Note diff
3553 Make merged '#ifdef' format output, conditional on the preprocessor
3554 macro NAME. *Note If-then-else::.
3558 Make output that is a valid 'ed' script. *Note ed Scripts::.
3561 '--ignore-tab-expansion'
3562 Ignore changes due to tab expansion. *Note White Space::.
3566 Make output that looks vaguely like an 'ed' script but has changes
3567 in the order they appear in the file. *Note Forward ed::.
3570 '--show-function-line=REGEXP'
3571 In context and unified format, for each hunk of differences, show
3572 some of the last preceding line that matches REGEXP. *Note
3573 Specified Headings::.
3576 Compare FILE to each operand; FILE may be a directory.
3579 Output a summary of usage and then exit.
3581 '--horizon-lines=LINES'
3582 Do not discard the last LINES lines of the common prefix and the
3583 first LINES lines of the common suffix. *Note diff Performance::.
3587 Ignore changes in case; consider upper- and lower-case letters
3588 equivalent. *Note Case Folding::.
3591 '--ignore-matching-lines=REGEXP'
3592 Ignore changes that just insert or delete lines that match REGEXP.
3593 *Note Specified Lines::.
3595 '--ignore-file-name-case'
3596 Ignore case when comparing file names. For example, recursive
3597 comparison of 'd' to 'e' might compare the contents of 'd/Init' and
3598 'e/inIt'. At the top level, 'diff d inIt' might compare the
3599 contents of 'd/Init' and 'inIt'. *Note Comparing Directories::.
3603 Pass the output through 'pr' to paginate it. *Note Pagination::.
3607 Use LABEL instead of the file name in the context format (*note
3608 Context Format::) and unified format (*note Unified Format::)
3609 headers. *Note RCS::.
3612 Print only the left column of two common lines in side by side
3613 format. *Note Side by Side Format::.
3615 '--line-format=FORMAT'
3616 Use FORMAT to output all input lines in if-then-else format. *Note
3621 Output RCS-format diffs; like '-f' except that each command
3622 specifies the number of lines affected. *Note RCS::.
3626 If one file is missing, treat it as present but empty. *Note
3627 Comparing Directories::.
3629 '--new-group-format=FORMAT'
3630 Use FORMAT to output a group of lines taken from just the second
3631 file in if-then-else format. *Note Line Group Formats::.
3633 '--new-line-format=FORMAT'
3634 Use FORMAT to output a line taken from just the second file in
3635 if-then-else format. *Note Line Formats::.
3638 Act on symbolic links themselves instead of what they point to.
3639 Two symbolic links are deemed equal only when each points to
3640 precisely the same name.
3642 '--old-group-format=FORMAT'
3643 Use FORMAT to output a group of lines taken from just the first
3644 file in if-then-else format. *Note Line Group Formats::.
3646 '--old-line-format=FORMAT'
3647 Use FORMAT to output a line taken from just the first file in
3648 if-then-else format. *Note Line Formats::.
3652 Show which C function each change is in. *Note C Function
3656 Specify what color palette to use when colored output is enabled.
3657 It defaults to 'rs=0:hd=1:ad=32:de=31:ln=36' for red deleted lines,
3658 green added lines, cyan line numbers, bold header.
3660 Supported capabilities are as follows.
3664 SGR substring for added lines. The default is green
3669 SGR substring for deleted lines. The default is red
3674 SGR substring for chunk header. The default is bold
3679 SGR substring for line numbers. The default is cyan
3684 Report only whether the files differ, not the details of the
3685 differences. *Note Brief::.
3689 When comparing directories, recursively compare any subdirectories
3690 found. *Note Comparing Directories::.
3693 '--report-identical-files'
3694 Report when two files are the same. *Note Comparing Directories::.
3697 '--starting-file=FILE'
3698 When comparing directories, start with the file FILE. This is used
3699 for resuming an aborted comparison. *Note Comparing Directories::.
3701 '--speed-large-files'
3702 Use heuristics to speed handling of large files that have numerous
3703 scattered small changes. *Note diff Performance::.
3705 '--strip-trailing-cr'
3706 Strip any trailing carriage return at the end of an input line.
3709 '--suppress-common-lines'
3710 Do not print common lines in side by side format. *Note Side by
3715 Expand tabs to spaces in the output, to preserve the alignment of
3716 tabs in the input files. *Note Tabs::.
3720 Output a tab rather than a space before the text of a line in
3721 normal or context format. This causes the alignment of tabs in the
3722 line to look normal. *Note Tabs::.
3725 Assume that tab stops are set every COLUMNS (default 8) print
3726 columns. *Note Tabs::.
3728 '--suppress-blank-empty'
3729 Suppress any blanks before newlines when printing the
3730 representation of an empty line, when outputting normal, context,
3731 or unified format. *Note Trailing Blanks::.
3734 Compare each operand to FILE; FILE may be a directory.
3737 Use the unified output format, showing three lines of context.
3738 *Note Unified Format::.
3740 '--unchanged-group-format=FORMAT'
3741 Use FORMAT to output a group of common lines taken from both files
3742 in if-then-else format. *Note Line Group Formats::.
3744 '--unchanged-line-format=FORMAT'
3745 Use FORMAT to output a line common to both files in if-then-else
3746 format. *Note Line Formats::.
3748 '--unidirectional-new-file'
3749 If a first file is missing, treat it as present but empty. *Note
3750 Comparing Directories::.
3754 Use the unified output format, showing LINES (an integer) lines of
3755 context, or three if LINES is not given. *Note Unified Format::.
3756 For proper operation, 'patch' typically needs at least two lines of
3759 On older systems, 'diff' supports an obsolete option '-LINES' that
3760 has effect when combined with '-u'. POSIX 1003.1-2001 (*note
3761 Standards conformance::) does not allow this; use '-U LINES'
3766 Output version information and then exit.
3769 '--ignore-all-space'
3770 Ignore white space when comparing lines. *Note White Space::.
3774 Output at most COLUMNS (default 130) print columns per line in side
3775 by side format. *Note Side by Side Format::.
3779 When comparing directories, ignore files and subdirectories whose
3780 basenames match PATTERN. *Note Comparing Directories::.
3783 '--exclude-from=FILE'
3784 When comparing directories, ignore files and subdirectories whose
3785 basenames match any pattern contained in FILE. *Note Comparing
3790 Use the side by side output format. *Note Side by Side Format::.
3793 '--ignore-trailing-space'
3794 Ignore white space at line end. *Note White Space::.
3797 File: diffutils.info, Node: Invoking diff3, Next: Invoking patch, Prev: Invoking diff, Up: Top
3802 The 'diff3' command compares three files and outputs descriptions of
3803 their differences. Its arguments are as follows:
3805 diff3 OPTIONS... MINE OLDER YOURS
3807 The files to compare are MINE, OLDER, and YOURS. At most one of
3808 these three file names may be '-', which tells 'diff3' to read the
3809 standard input for that file.
3811 An exit status of 0 means 'diff3' was successful, 1 means some
3812 conflicts were found, and 2 means trouble.
3816 * diff3 Options:: Summary of options to 'diff3'.
3819 File: diffutils.info, Node: diff3 Options, Up: Invoking diff3
3821 14.1 Options to 'diff3'
3822 =======================
3824 Below is a summary of all of the options that GNU 'diff3' accepts.
3825 Multiple single letter options (unless they take an argument) can be
3826 combined into a single command line argument.
3830 Treat all files as text and compare them line-by-line, even if they
3831 do not appear to be text. *Note Binary::.
3835 Incorporate all unmerged changes from OLDER to YOURS into MINE,
3836 surrounding conflicts with bracket lines. *Note Marking
3839 '--diff-program=PROGRAM'
3840 Use the compatible comparison program PROGRAM to compare files
3845 Generate an 'ed' script that incorporates all the changes from
3846 OLDER to YOURS into MINE. *Note Which Changes::.
3850 Like '-e', except bracket lines from overlapping changes' first and
3851 third files. *Note Marking Conflicts::. With '-E', an overlapping
3852 change looks like this:
3861 Output a summary of usage and then exit.
3864 Generate 'w' and 'q' commands at the end of the 'ed' script for
3865 System V compatibility. This option must be combined with one of
3866 the '-AeExX3' options, and may not be combined with '-m'. *Note
3867 Saving the Changed File::.
3870 Use the label LABEL for the brackets output by the '-A', '-E' and
3871 '-X' options. This option may be given up to three times, one for
3872 each input file. The default labels are the names of the input
3873 files. Thus 'diff3 --label X --label Y --label Z -m A B C' acts
3874 like 'diff3 -m A B C', except that the output looks like it came
3875 from files named 'X', 'Y' and 'Z' rather than from files named 'A',
3876 'B' and 'C'. *Note Marking Conflicts::.
3880 Apply the edit script to the first file and send the result to
3881 standard output. Unlike piping the output from 'diff3' to 'ed',
3882 this works even for binary files and incomplete lines. '-A' is
3883 assumed if no edit script option is specified. *Note Bypassing
3886 '--strip-trailing-cr'
3887 Strip any trailing carriage return at the end of an input line.
3892 Output a tab rather than two spaces before the text of a line in
3893 normal format. This causes the alignment of tabs in the line to
3894 look normal. *Note Tabs::.
3898 Output version information and then exit.
3902 Like '-e', except output only the overlapping changes. *Note Which
3906 Like '-E', except output only the overlapping changes. In other
3907 words, like '-x', except bracket changes as in '-E'. *Note Marking
3912 Like '-e', except output only the nonoverlapping changes. *Note
3916 File: diffutils.info, Node: Invoking patch, Next: Invoking sdiff, Prev: Invoking diff3, Up: Top
3921 Normally 'patch' is invoked like this:
3925 The full format for invoking 'patch' is:
3927 patch OPTIONS... [ORIGFILE [PATCHFILE]]
3929 You can also specify where to read the patch from with the '-i
3930 PATCHFILE' or '--input=PATCHFILE' option. If you do not specify
3931 PATCHFILE, or if PATCHFILE is '-', 'patch' reads the patch (that is, the
3932 'diff' output) from the standard input.
3934 If you do not specify an input file on the command line, 'patch'
3935 tries to intuit from the "leading text" (any text in the patch that
3936 comes before the 'diff' output) which file to edit. *Note Multiple
3939 By default, 'patch' replaces the original input file with the patched
3940 version, possibly after renaming the original file into a backup file
3941 (*note Backup Names::, for a description of how 'patch' names backup
3942 files). You can also specify where to put the output with the '-o FILE'
3943 or '--output=FILE' option; however, do not use this option if FILE is
3944 one of the input files.
3948 * patch Options:: Summary table of options to 'patch'.
3951 File: diffutils.info, Node: patch Options, Up: Invoking patch
3953 15.1 Options to 'patch'
3954 =======================
3956 Here is a summary of all of the options that GNU 'patch' accepts. *Note
3957 patch and Tradition::, for which of these options are safe to use in
3958 older versions of 'patch'.
3960 Multiple single-letter options that do not take an argument can be
3961 combined into a single command line argument with only one dash.
3965 Back up the original contents of each file, even if backups would
3966 normally not be made. *Note Backups::.
3970 Prepend PREFIX to backup file names. *Note Backup Names::.
3972 '--backup-if-mismatch'
3973 Back up the original contents of each file if the patch does not
3974 exactly match the file. This is the default behavior when not
3975 conforming to POSIX. *Note Backups::.
3978 Read and write all files in binary mode, except for standard output
3979 and '/dev/tty'. This option has no effect on POSIX-conforming
3980 systems like GNU/Linux. On systems where this option makes a
3981 difference, the patch should be generated by 'diff -a --binary'.
3986 Interpret the patch file as a context diff. *Note patch Input::.
3989 '--directory=DIRECTORY'
3990 Make directory DIRECTORY the current directory for interpreting
3991 both file names in the patch file, and file names given as
3992 arguments to other options. *Note patch Directories::.
3996 Make merged if-then-else output using NAME. *Note If-then-else::.
3999 Print the results of applying the patches without actually changing
4000 any files. *Note Dry Runs::.
4004 Interpret the patch file as an 'ed' script. *Note patch Input::.
4007 '--remove-empty-files'
4008 Remove output files that are empty after the patches have been
4009 applied. *Note Creating and Removing::.
4013 Assume that the user knows exactly what he or she is doing, and do
4014 not ask any questions. *Note patch Messages::.
4018 Set the maximum fuzz factor to LINES. *Note Inexact::.
4022 If NUM is positive, get input files from a revision control system
4023 as necessary; if zero, do not get the files; if negative, ask the
4024 user whether to get the files. *Note Revision Control::.
4027 Output a summary of usage and then exit.
4031 Read the patch from PATCHFILE rather than from standard input.
4032 *Note patch Options::.
4035 '--ignore-white-space'
4036 Let any sequence of blanks (spaces or tabs) in the patch file match
4037 any sequence of blanks in the input file. *Note Changed White
4042 Interpret the patch file as a normal diff. *Note patch Input::.
4046 Ignore patches that 'patch' thinks are reversed or already applied.
4047 See also '-R'. *Note Reversed Patches::.
4049 '--no-backup-if-mismatch'
4050 Do not back up the original contents of files. This is the default
4051 behavior when conforming to POSIX. *Note Backups::.
4055 Use FILE as the output file name. *Note patch Options::.
4059 Set the file name strip count to NUMBER. *Note patch
4063 Conform to POSIX, as if the 'POSIXLY_CORRECT' environment variable
4064 had been set. *Note patch and POSIX::.
4066 '--quoting-style=WORD'
4067 Use style WORD to quote names in diagnostics, as if the
4068 'QUOTING_STYLE' environment variable had been set to WORD. *Note
4069 patch Quoting Style::.
4072 '--reject-file=REJECT-FILE'
4073 Use REJECT-FILE as the reject file name. *Note Reject Names::.
4077 Assume that this patch was created with the old and new files
4078 swapped. *Note Reversed Patches::.
4083 Work silently unless an error occurs. *Note patch Messages::.
4087 Do not ask any questions. *Note patch Messages::.
4091 Set the modification and access times of patched files from
4092 timestamps given in context diff headers, assuming that the context
4093 diff headers use local time. *Note Patching Timestamps::.
4097 Interpret the patch file as a unified diff. *Note patch Input::.
4101 Output version information and then exit.
4104 '--version=control=BACKUP-STYLE'
4105 Select the naming convention for backup file names. *Note Backup
4109 Print more diagnostics than usual. *Note patch Messages::.
4113 Set internal debugging flags. Of interest only to 'patch'
4117 '--basename-prefix=PREFIX'
4118 Prepend PREFIX to base names of backup files. *Note Backup
4123 Use SUFFIX as the backup extension instead of '.orig' or '~'.
4124 *Note Backup Names::.
4128 Set the modification and access times of patched files from
4129 timestamps given in context diff headers, assuming that the context
4130 diff headers use UTC. *Note Patching Timestamps::.
4133 File: diffutils.info, Node: Invoking sdiff, Next: Standards conformance, Prev: Invoking patch, Up: Top
4138 The 'sdiff' command merges two files and interactively outputs the
4139 results. Its arguments are as follows:
4141 sdiff -o OUTFILE OPTIONS... FROM-FILE TO-FILE
4143 This merges FROM-FILE with TO-FILE, with output to OUTFILE. If
4144 FROM-FILE is a directory and TO-FILE is not, 'sdiff' compares the file
4145 in FROM-FILE whose file name is that of TO-FILE, and vice versa.
4146 FROM-FILE and TO-FILE may not both be directories.
4148 'sdiff' options begin with '-', so normally FROM-FILE and TO-FILE may
4149 not begin with '-'. However, '--' as an argument by itself treats the
4150 remaining arguments as file names even if they begin with '-'. You may
4151 not use '-' as an input file.
4153 'sdiff' without '--output' ('-o') produces a side-by-side difference.
4154 This usage is obsolete; use the '--side-by-side' ('-y') option of 'diff'
4157 An exit status of 0 means no differences were found, 1 means some
4158 differences were found, and 2 means trouble.
4162 * sdiff Options:: Summary of options to 'diff'.
4165 File: diffutils.info, Node: sdiff Options, Up: Invoking sdiff
4167 16.1 Options to 'sdiff'
4168 =======================
4170 Below is a summary of all of the options that GNU 'sdiff' accepts. Each
4171 option has two equivalent names, one of which is a single letter
4172 preceded by '-', and the other of which is a long name preceded by '--'.
4173 Multiple single letter options (unless they take an argument) can be
4174 combined into a single command line argument. Long named options can be
4175 abbreviated to any unique prefix of their name.
4179 Treat all files as text and compare them line-by-line, even if they
4180 do not appear to be text. *Note Binary::.
4183 '--ignore-space-change'
4184 Ignore changes in amount of white space. *Note White Space::.
4187 '--ignore-blank-lines'
4188 Ignore changes that just insert or delete blank lines. *Note Blank
4193 Change the algorithm to perhaps find a smaller set of changes.
4194 This makes 'sdiff' slower (sometimes much slower). *Note diff
4197 '--diff-program=PROGRAM'
4198 Use the compatible comparison program PROGRAM to compare files
4202 '--ignore-tab-expansion'
4203 Ignore changes due to tab expansion. *Note White Space::.
4206 Output a summary of usage and then exit.
4210 Ignore changes in case; consider upper- and lower-case to be the
4211 same. *Note Case Folding::.
4214 '--ignore-matching-lines=REGEXP'
4215 Ignore changes that just insert or delete lines that match REGEXP.
4216 *Note Specified Lines::.
4220 Print only the left column of two common lines. *Note Side by Side
4225 Put merged output into FILE. This option is required for merging.
4228 '--suppress-common-lines'
4229 Do not print common lines. *Note Side by Side Format::.
4231 '--speed-large-files'
4232 Use heuristics to speed handling of large files that have numerous
4233 scattered small changes. *Note diff Performance::.
4235 '--strip-trailing-cr'
4236 Strip any trailing carriage return at the end of an input line.
4241 Expand tabs to spaces in the output, to preserve the alignment of
4242 tabs in the input files. *Note Tabs::.
4245 Assume that tab stops are set every COLUMNS (default 8) print
4246 columns. *Note Tabs::.
4250 Output version information and then exit.
4254 Output at most COLUMNS (default 130) print columns per line. *Note
4255 Side by Side Format::. Note that for historical reasons, this
4256 option is '-W' in 'diff', '-w' in 'sdiff'.
4259 '--ignore-all-space'
4260 Ignore white space when comparing lines. *Note White Space::.
4261 Note that for historical reasons, this option is '-w' in 'diff',
4265 '--ignore-trailing-space'
4266 Ignore white space at line end. *Note White Space::.
4269 File: diffutils.info, Node: Standards conformance, Next: Projects, Prev: Invoking sdiff, Up: Top
4271 17 Standards conformance
4272 ************************
4274 In a few cases, the GNU utilities' default behavior is incompatible with
4275 the POSIX standard. To suppress these incompatibilities, define the
4276 'POSIXLY_CORRECT' environment variable. Unless you are checking for
4277 POSIX conformance, you probably do not need to define 'POSIXLY_CORRECT'.
4279 Normally options and operands can appear in any order, and programs
4280 act as if all the options appear before any operands. For example,
4281 'diff lao tzu -C 2' acts like 'diff -C 2 lao tzu', since '2' is an
4282 option-argument of '-C'. However, if the 'POSIXLY_CORRECT' environment
4283 variable is set, options must appear before operands, unless otherwise
4284 specified for a particular command.
4286 Newer versions of POSIX are occasionally incompatible with older
4287 versions. For example, older versions of POSIX allowed the command
4288 'diff -c -10' to have the same meaning as 'diff -C 10', but POSIX
4289 1003.1-2001 'diff' no longer allows digit-string options like '-10'.
4291 The GNU utilities normally conform to the version of POSIX that is
4292 standard for your system. To cause them to conform to a different
4293 version of POSIX, define the '_POSIX2_VERSION' environment variable to a
4294 value of the form YYYYMM specifying the year and month the standard was
4295 adopted. Two values are currently supported for '_POSIX2_VERSION':
4296 '199209' stands for POSIX 1003.2-1992, and '200112' stands for POSIX
4297 1003.1-2001. For example, if you are running older software that
4298 assumes an older version of POSIX and uses 'diff -c -10', you can work
4299 around the compatibility problems by setting '_POSIX2_VERSION=199209' in
4303 File: diffutils.info, Node: Projects, Next: Copying This Manual, Prev: Standards conformance, Up: Top
4308 Here are some ideas for improving GNU 'diff' and 'patch'. The GNU
4309 project has identified some improvements as potential programming
4310 projects for volunteers. You can also help by reporting any bugs that
4313 If you are a programmer and would like to contribute something to the
4314 GNU project, please consider volunteering for one of these projects. If
4315 you are seriously contemplating work, please write to <gvc@gnu.org> to
4316 coordinate with other volunteers.
4320 * Shortcomings:: Suggested projects for improvements.
4321 * Bugs:: Reporting bugs.
4324 File: diffutils.info, Node: Shortcomings, Next: Bugs, Up: Projects
4326 18.1 Suggested Projects for Improving GNU 'diff' and 'patch'
4327 ============================================================
4329 One should be able to use GNU 'diff' to generate a patch from any pair
4330 of directory trees, and given the patch and a copy of one such tree, use
4331 'patch' to generate a faithful copy of the other. Unfortunately, some
4332 changes to directory trees cannot be expressed using current patch
4333 formats; also, 'patch' does not handle some of the existing formats.
4334 These shortcomings motivate the following suggested projects.
4338 * Internationalization:: Handling multibyte and varying-width characters.
4339 * Changing Structure:: Handling changes to the directory structure.
4340 * Special Files:: Handling symbolic links, device special files, etc.
4341 * Unusual File Names:: Handling file names that contain unusual characters.
4342 * Timestamp Order:: Outputting diffs in timestamp order.
4343 * Ignoring Changes:: Ignoring certain changes while showing others.
4344 * Speedups:: Improving performance.
4347 File: diffutils.info, Node: Internationalization, Next: Changing Structure, Up: Shortcomings
4349 18.1.1 Handling Multibyte and Varying-Width Characters
4350 ------------------------------------------------------
4352 'diff', 'diff3' and 'sdiff' treat each line of input as a string of
4353 unibyte characters. This can mishandle multibyte characters in some
4354 cases. For example, when asked to ignore spaces, 'diff' does not
4355 properly ignore a multibyte space character.
4357 Also, 'diff' currently assumes that each byte is one column wide, and
4358 this assumption is incorrect in some locales, e.g., locales that use
4359 UTF-8 encoding. This causes problems with the '-y' or '--side-by-side'
4362 These problems need to be fixed without unduly affecting the
4363 performance of the utilities in unibyte environments.
4365 The IBM GNU/Linux Technology Center Internationalization Team has
4366 proposed patches to support internationalized 'diff'
4367 (http://oss.software.ibm.com/developer/opensource/linux/patches/i18n/diffutils-2.7.2-i18n-0.1.patch.gz).
4368 Unfortunately, these patches are incomplete and are to an older version
4369 of 'diff', so more work needs to be done in this area.
4372 File: diffutils.info, Node: Changing Structure, Next: Special Files, Prev: Internationalization, Up: Shortcomings
4374 18.1.2 Handling Changes to the Directory Structure
4375 --------------------------------------------------
4377 'diff' and 'patch' do not handle some changes to directory structure.
4378 For example, suppose one directory tree contains a directory named 'D'
4379 with some subsidiary files, and another contains a file with the same
4380 name 'D'. 'diff -r' does not output enough information for 'patch' to
4381 transform the directory subtree into the file.
4383 There should be a way to specify that a file has been removed without
4384 having to include its entire contents in the patch file. There should
4385 also be a way to tell 'patch' that a file was renamed, even if there is
4386 no way for 'diff' to generate such information. There should be a way
4387 to tell 'patch' that a file's timestamp has changed, even if its
4388 contents have not changed.
4390 These problems can be fixed by extending the 'diff' output format to
4391 represent changes in directory structure, and extending 'patch' to
4392 understand these extensions.
4395 File: diffutils.info, Node: Special Files, Next: Unusual File Names, Prev: Changing Structure, Up: Shortcomings
4397 18.1.3 Files that are Neither Directories Nor Regular Files
4398 -----------------------------------------------------------
4400 Some files are neither directories nor regular files: they are unusual
4401 files like symbolic links, device special files, named pipes, and
4402 sockets. Currently, 'diff' treats symbolic links as if they were the
4403 pointed-to files, except that a recursive 'diff' reports an error if it
4404 detects infinite loops of symbolic links (e.g., symbolic links to '..').
4405 'diff' treats other special files like regular files if they are
4406 specified at the top level, but simply reports their presence when
4407 comparing directories. This means that 'patch' cannot represent changes
4408 to such files. For example, if you change which file a symbolic link
4409 points to, 'diff' outputs the difference between the two files, instead
4410 of the change to the symbolic link.
4412 'diff' should optionally report changes to special files specially,
4413 and 'patch' should be extended to understand these extensions.
4416 File: diffutils.info, Node: Unusual File Names, Next: Timestamp Order, Prev: Special Files, Up: Shortcomings
4418 18.1.4 File Names that Contain Unusual Characters
4419 -------------------------------------------------
4421 When a file name contains an unusual character like a newline or white
4422 space, 'diff -r' generates a patch that 'patch' cannot parse. The
4423 problem is with format of 'diff' output, not just with 'patch', because
4424 with odd enough file names one can cause 'diff' to generate a patch that
4425 is syntactically correct but patches the wrong files. The format of
4426 'diff' output should be extended to handle all possible file names.
4429 File: diffutils.info, Node: Timestamp Order, Next: Ignoring Changes, Prev: Unusual File Names, Up: Shortcomings
4431 18.1.5 Outputting Diffs in Timestamp Order
4432 ------------------------------------------
4434 Applying 'patch' to a multiple-file diff can result in files whose
4435 timestamps are out of order. GNU 'patch' has options to restore the
4436 timestamps of the updated files (*note Patching Timestamps::), but
4437 sometimes it is useful to generate a patch that works even if the
4438 recipient does not have GNU patch, or does not use these options. One
4439 way to do this would be to implement a 'diff' option to output diffs in
4443 File: diffutils.info, Node: Ignoring Changes, Next: Speedups, Prev: Timestamp Order, Up: Shortcomings
4445 18.1.6 Ignoring Certain Changes
4446 -------------------------------
4448 It would be nice to have a feature for specifying two strings, one in
4449 FROM-FILE and one in TO-FILE, which should be considered to match.
4450 Thus, if the two strings are 'foo' and 'bar', then if two lines differ
4451 only in that 'foo' in file 1 corresponds to 'bar' in file 2, the lines
4452 are treated as identical.
4454 It is not clear how general this feature can or should be, or what
4455 syntax should be used for it.
4457 A partial substitute is to filter one or both files before comparing,
4460 sed 's/foo/bar/g' file1 | diff - file2
4462 However, this outputs the filtered text, not the original.
4465 File: diffutils.info, Node: Speedups, Prev: Ignoring Changes, Up: Shortcomings
4467 18.1.7 Improving Performance
4468 ----------------------------
4470 When comparing two large directory structures, one of which was
4471 originally copied from the other with timestamps preserved (e.g., with
4472 'cp -pR'), it would greatly improve performance if an option told 'diff'
4473 to assume that two files with the same size and timestamps have the same
4474 content. *Note diff Performance::.
4477 File: diffutils.info, Node: Bugs, Prev: Shortcomings, Up: Projects
4482 If you think you have found a bug in GNU 'cmp', 'diff', 'diff3', or
4483 'sdiff', please report it by electronic mail to the GNU utilities bug
4484 report mailing list (http://mail.gnu.org/mailman/listinfo/bug-diffutils)
4485 <bug-diffutils@gnu.org>. Please send bug reports for GNU 'patch' to
4486 <bug-patch@gnu.org>. Send as precise a description of the problem as
4487 you can, including the output of the '--version' option and sample input
4488 files that produce the bug, if applicable. If you have a nontrivial fix
4489 for the bug, please send it as well. If you have a patch, please send
4490 it too. It may simplify the maintainer's job if the patch is relative
4491 to a recent test release, which you can find in the directory
4492 <ftp://alpha.gnu.org/gnu/diffutils/>.
4495 File: diffutils.info, Node: Copying This Manual, Next: Translations, Prev: Projects, Up: Top
4497 Appendix A Copying This Manual
4498 ******************************
4500 Version 1.3, 3 November 2008
4502 Copyright (C) 2000, 2001, 2002, 2007, 2008 Free Software Foundation, Inc.
4505 Everyone is permitted to copy and distribute verbatim copies
4506 of this license document, but changing it is not allowed.
4510 The purpose of this License is to make a manual, textbook, or other
4511 functional and useful document "free" in the sense of freedom: to
4512 assure everyone the effective freedom to copy and redistribute it,
4513 with or without modifying it, either commercially or
4514 noncommercially. Secondarily, this License preserves for the
4515 author and publisher a way to get credit for their work, while not
4516 being considered responsible for modifications made by others.
4518 This License is a kind of "copyleft", which means that derivative
4519 works of the document must themselves be free in the same sense.
4520 It complements the GNU General Public License, which is a copyleft
4521 license designed for free software.
4523 We have designed this License in order to use it for manuals for
4524 free software, because free software needs free documentation: a
4525 free program should come with manuals providing the same freedoms
4526 that the software does. But this License is not limited to
4527 software manuals; it can be used for any textual work, regardless
4528 of subject matter or whether it is published as a printed book. We
4529 recommend this License principally for works whose purpose is
4530 instruction or reference.
4532 1. APPLICABILITY AND DEFINITIONS
4534 This License applies to any manual or other work, in any medium,
4535 that contains a notice placed by the copyright holder saying it can
4536 be distributed under the terms of this License. Such a notice
4537 grants a world-wide, royalty-free license, unlimited in duration,
4538 to use that work under the conditions stated herein. The
4539 "Document", below, refers to any such manual or work. Any member
4540 of the public is a licensee, and is addressed as "you". You accept
4541 the license if you copy, modify or distribute the work in a way
4542 requiring permission under copyright law.
4544 A "Modified Version" of the Document means any work containing the
4545 Document or a portion of it, either copied verbatim, or with
4546 modifications and/or translated into another language.
4548 A "Secondary Section" is a named appendix or a front-matter section
4549 of the Document that deals exclusively with the relationship of the
4550 publishers or authors of the Document to the Document's overall
4551 subject (or to related matters) and contains nothing that could
4552 fall directly within that overall subject. (Thus, if the Document
4553 is in part a textbook of mathematics, a Secondary Section may not
4554 explain any mathematics.) The relationship could be a matter of
4555 historical connection with the subject or with related matters, or
4556 of legal, commercial, philosophical, ethical or political position
4559 The "Invariant Sections" are certain Secondary Sections whose
4560 titles are designated, as being those of Invariant Sections, in the
4561 notice that says that the Document is released under this License.
4562 If a section does not fit the above definition of Secondary then it
4563 is not allowed to be designated as Invariant. The Document may
4564 contain zero Invariant Sections. If the Document does not identify
4565 any Invariant Sections then there are none.
4567 The "Cover Texts" are certain short passages of text that are
4568 listed, as Front-Cover Texts or Back-Cover Texts, in the notice
4569 that says that the Document is released under this License. A
4570 Front-Cover Text may be at most 5 words, and a Back-Cover Text may
4571 be at most 25 words.
4573 A "Transparent" copy of the Document means a machine-readable copy,
4574 represented in a format whose specification is available to the
4575 general public, that is suitable for revising the document
4576 straightforwardly with generic text editors or (for images composed
4577 of pixels) generic paint programs or (for drawings) some widely
4578 available drawing editor, and that is suitable for input to text
4579 formatters or for automatic translation to a variety of formats
4580 suitable for input to text formatters. A copy made in an otherwise
4581 Transparent file format whose markup, or absence of markup, has
4582 been arranged to thwart or discourage subsequent modification by
4583 readers is not Transparent. An image format is not Transparent if
4584 used for any substantial amount of text. A copy that is not
4585 "Transparent" is called "Opaque".
4587 Examples of suitable formats for Transparent copies include plain
4588 ASCII without markup, Texinfo input format, LaTeX input format,
4589 SGML or XML using a publicly available DTD, and standard-conforming
4590 simple HTML, PostScript or PDF designed for human modification.
4591 Examples of transparent image formats include PNG, XCF and JPG.
4592 Opaque formats include proprietary formats that can be read and
4593 edited only by proprietary word processors, SGML or XML for which
4594 the DTD and/or processing tools are not generally available, and
4595 the machine-generated HTML, PostScript or PDF produced by some word
4596 processors for output purposes only.
4598 The "Title Page" means, for a printed book, the title page itself,
4599 plus such following pages as are needed to hold, legibly, the
4600 material this License requires to appear in the title page. For
4601 works in formats which do not have any title page as such, "Title
4602 Page" means the text near the most prominent appearance of the
4603 work's title, preceding the beginning of the body of the text.
4605 The "publisher" means any person or entity that distributes copies
4606 of the Document to the public.
4608 A section "Entitled XYZ" means a named subunit of the Document
4609 whose title either is precisely XYZ or contains XYZ in parentheses
4610 following text that translates XYZ in another language. (Here XYZ
4611 stands for a specific section name mentioned below, such as
4612 "Acknowledgements", "Dedications", "Endorsements", or "History".)
4613 To "Preserve the Title" of such a section when you modify the
4614 Document means that it remains a section "Entitled XYZ" according
4617 The Document may include Warranty Disclaimers next to the notice
4618 which states that this License applies to the Document. These
4619 Warranty Disclaimers are considered to be included by reference in
4620 this License, but only as regards disclaiming warranties: any other
4621 implication that these Warranty Disclaimers may have is void and
4622 has no effect on the meaning of this License.
4626 You may copy and distribute the Document in any medium, either
4627 commercially or noncommercially, provided that this License, the
4628 copyright notices, and the license notice saying this License
4629 applies to the Document are reproduced in all copies, and that you
4630 add no other conditions whatsoever to those of this License. You
4631 may not use technical measures to obstruct or control the reading
4632 or further copying of the copies you make or distribute. However,
4633 you may accept compensation in exchange for copies. If you
4634 distribute a large enough number of copies you must also follow the
4635 conditions in section 3.
4637 You may also lend copies, under the same conditions stated above,
4638 and you may publicly display copies.
4640 3. COPYING IN QUANTITY
4642 If you publish printed copies (or copies in media that commonly
4643 have printed covers) of the Document, numbering more than 100, and
4644 the Document's license notice requires Cover Texts, you must
4645 enclose the copies in covers that carry, clearly and legibly, all
4646 these Cover Texts: Front-Cover Texts on the front cover, and
4647 Back-Cover Texts on the back cover. Both covers must also clearly
4648 and legibly identify you as the publisher of these copies. The
4649 front cover must present the full title with all words of the title
4650 equally prominent and visible. You may add other material on the
4651 covers in addition. Copying with changes limited to the covers, as
4652 long as they preserve the title of the Document and satisfy these
4653 conditions, can be treated as verbatim copying in other respects.
4655 If the required texts for either cover are too voluminous to fit
4656 legibly, you should put the first ones listed (as many as fit
4657 reasonably) on the actual cover, and continue the rest onto
4660 If you publish or distribute Opaque copies of the Document
4661 numbering more than 100, you must either include a machine-readable
4662 Transparent copy along with each Opaque copy, or state in or with
4663 each Opaque copy a computer-network location from which the general
4664 network-using public has access to download using public-standard
4665 network protocols a complete Transparent copy of the Document, free
4666 of added material. If you use the latter option, you must take
4667 reasonably prudent steps, when you begin distribution of Opaque
4668 copies in quantity, to ensure that this Transparent copy will
4669 remain thus accessible at the stated location until at least one
4670 year after the last time you distribute an Opaque copy (directly or
4671 through your agents or retailers) of that edition to the public.
4673 It is requested, but not required, that you contact the authors of
4674 the Document well before redistributing any large number of copies,
4675 to give them a chance to provide you with an updated version of the
4680 You may copy and distribute a Modified Version of the Document
4681 under the conditions of sections 2 and 3 above, provided that you
4682 release the Modified Version under precisely this License, with the
4683 Modified Version filling the role of the Document, thus licensing
4684 distribution and modification of the Modified Version to whoever
4685 possesses a copy of it. In addition, you must do these things in
4686 the Modified Version:
4688 A. Use in the Title Page (and on the covers, if any) a title
4689 distinct from that of the Document, and from those of previous
4690 versions (which should, if there were any, be listed in the
4691 History section of the Document). You may use the same title
4692 as a previous version if the original publisher of that
4693 version gives permission.
4695 B. List on the Title Page, as authors, one or more persons or
4696 entities responsible for authorship of the modifications in
4697 the Modified Version, together with at least five of the
4698 principal authors of the Document (all of its principal
4699 authors, if it has fewer than five), unless they release you
4700 from this requirement.
4702 C. State on the Title page the name of the publisher of the
4703 Modified Version, as the publisher.
4705 D. Preserve all the copyright notices of the Document.
4707 E. Add an appropriate copyright notice for your modifications
4708 adjacent to the other copyright notices.
4710 F. Include, immediately after the copyright notices, a license
4711 notice giving the public permission to use the Modified
4712 Version under the terms of this License, in the form shown in
4715 G. Preserve in that license notice the full lists of Invariant
4716 Sections and required Cover Texts given in the Document's
4719 H. Include an unaltered copy of this License.
4721 I. Preserve the section Entitled "History", Preserve its Title,
4722 and add to it an item stating at least the title, year, new
4723 authors, and publisher of the Modified Version as given on the
4724 Title Page. If there is no section Entitled "History" in the
4725 Document, create one stating the title, year, authors, and
4726 publisher of the Document as given on its Title Page, then add
4727 an item describing the Modified Version as stated in the
4730 J. Preserve the network location, if any, given in the Document
4731 for public access to a Transparent copy of the Document, and
4732 likewise the network locations given in the Document for
4733 previous versions it was based on. These may be placed in the
4734 "History" section. You may omit a network location for a work
4735 that was published at least four years before the Document
4736 itself, or if the original publisher of the version it refers
4737 to gives permission.
4739 K. For any section Entitled "Acknowledgements" or "Dedications",
4740 Preserve the Title of the section, and preserve in the section
4741 all the substance and tone of each of the contributor
4742 acknowledgements and/or dedications given therein.
4744 L. Preserve all the Invariant Sections of the Document, unaltered
4745 in their text and in their titles. Section numbers or the
4746 equivalent are not considered part of the section titles.
4748 M. Delete any section Entitled "Endorsements". Such a section
4749 may not be included in the Modified Version.
4751 N. Do not retitle any existing section to be Entitled
4752 "Endorsements" or to conflict in title with any Invariant
4755 O. Preserve any Warranty Disclaimers.
4757 If the Modified Version includes new front-matter sections or
4758 appendices that qualify as Secondary Sections and contain no
4759 material copied from the Document, you may at your option designate
4760 some or all of these sections as invariant. To do this, add their
4761 titles to the list of Invariant Sections in the Modified Version's
4762 license notice. These titles must be distinct from any other
4765 You may add a section Entitled "Endorsements", provided it contains
4766 nothing but endorsements of your Modified Version by various
4767 parties--for example, statements of peer review or that the text
4768 has been approved by an organization as the authoritative
4769 definition of a standard.
4771 You may add a passage of up to five words as a Front-Cover Text,
4772 and a passage of up to 25 words as a Back-Cover Text, to the end of
4773 the list of Cover Texts in the Modified Version. Only one passage
4774 of Front-Cover Text and one of Back-Cover Text may be added by (or
4775 through arrangements made by) any one entity. If the Document
4776 already includes a cover text for the same cover, previously added
4777 by you or by arrangement made by the same entity you are acting on
4778 behalf of, you may not add another; but you may replace the old
4779 one, on explicit permission from the previous publisher that added
4782 The author(s) and publisher(s) of the Document do not by this
4783 License give permission to use their names for publicity for or to
4784 assert or imply endorsement of any Modified Version.
4786 5. COMBINING DOCUMENTS
4788 You may combine the Document with other documents released under
4789 this License, under the terms defined in section 4 above for
4790 modified versions, provided that you include in the combination all
4791 of the Invariant Sections of all of the original documents,
4792 unmodified, and list them all as Invariant Sections of your
4793 combined work in its license notice, and that you preserve all
4794 their Warranty Disclaimers.
4796 The combined work need only contain one copy of this License, and
4797 multiple identical Invariant Sections may be replaced with a single
4798 copy. If there are multiple Invariant Sections with the same name
4799 but different contents, make the title of each such section unique
4800 by adding at the end of it, in parentheses, the name of the
4801 original author or publisher of that section if known, or else a
4802 unique number. Make the same adjustment to the section titles in
4803 the list of Invariant Sections in the license notice of the
4806 In the combination, you must combine any sections Entitled
4807 "History" in the various original documents, forming one section
4808 Entitled "History"; likewise combine any sections Entitled
4809 "Acknowledgements", and any sections Entitled "Dedications". You
4810 must delete all sections Entitled "Endorsements."
4812 6. COLLECTIONS OF DOCUMENTS
4814 You may make a collection consisting of the Document and other
4815 documents released under this License, and replace the individual
4816 copies of this License in the various documents with a single copy
4817 that is included in the collection, provided that you follow the
4818 rules of this License for verbatim copying of each of the documents
4819 in all other respects.
4821 You may extract a single document from such a collection, and
4822 distribute it individually under this License, provided you insert
4823 a copy of this License into the extracted document, and follow this
4824 License in all other respects regarding verbatim copying of that
4827 7. AGGREGATION WITH INDEPENDENT WORKS
4829 A compilation of the Document or its derivatives with other
4830 separate and independent documents or works, in or on a volume of a
4831 storage or distribution medium, is called an "aggregate" if the
4832 copyright resulting from the compilation is not used to limit the
4833 legal rights of the compilation's users beyond what the individual
4834 works permit. When the Document is included in an aggregate, this
4835 License does not apply to the other works in the aggregate which
4836 are not themselves derivative works of the Document.
4838 If the Cover Text requirement of section 3 is applicable to these
4839 copies of the Document, then if the Document is less than one half
4840 of the entire aggregate, the Document's Cover Texts may be placed
4841 on covers that bracket the Document within the aggregate, or the
4842 electronic equivalent of covers if the Document is in electronic
4843 form. Otherwise they must appear on printed covers that bracket
4844 the whole aggregate.
4848 Translation is considered a kind of modification, so you may
4849 distribute translations of the Document under the terms of section
4850 4. Replacing Invariant Sections with translations requires special
4851 permission from their copyright holders, but you may include
4852 translations of some or all Invariant Sections in addition to the
4853 original versions of these Invariant Sections. You may include a
4854 translation of this License, and all the license notices in the
4855 Document, and any Warranty Disclaimers, provided that you also
4856 include the original English version of this License and the
4857 original versions of those notices and disclaimers. In case of a
4858 disagreement between the translation and the original version of
4859 this License or a notice or disclaimer, the original version will
4862 If a section in the Document is Entitled "Acknowledgements",
4863 "Dedications", or "History", the requirement (section 4) to
4864 Preserve its Title (section 1) will typically require changing the
4869 You may not copy, modify, sublicense, or distribute the Document
4870 except as expressly provided under this License. Any attempt
4871 otherwise to copy, modify, sublicense, or distribute it is void,
4872 and will automatically terminate your rights under this License.
4874 However, if you cease all violation of this License, then your
4875 license from a particular copyright holder is reinstated (a)
4876 provisionally, unless and until the copyright holder explicitly and
4877 finally terminates your license, and (b) permanently, if the
4878 copyright holder fails to notify you of the violation by some
4879 reasonable means prior to 60 days after the cessation.
4881 Moreover, your license from a particular copyright holder is
4882 reinstated permanently if the copyright holder notifies you of the
4883 violation by some reasonable means, this is the first time you have
4884 received notice of violation of this License (for any work) from
4885 that copyright holder, and you cure the violation prior to 30 days
4886 after your receipt of the notice.
4888 Termination of your rights under this section does not terminate
4889 the licenses of parties who have received copies or rights from you
4890 under this License. If your rights have been terminated and not
4891 permanently reinstated, receipt of a copy of some or all of the
4892 same material does not give you any rights to use it.
4894 10. FUTURE REVISIONS OF THIS LICENSE
4896 The Free Software Foundation may publish new, revised versions of
4897 the GNU Free Documentation License from time to time. Such new
4898 versions will be similar in spirit to the present version, but may
4899 differ in detail to address new problems or concerns. See
4900 <https://www.gnu.org/copyleft/>.
4902 Each version of the License is given a distinguishing version
4903 number. If the Document specifies that a particular numbered
4904 version of this License "or any later version" applies to it, you
4905 have the option of following the terms and conditions either of
4906 that specified version or of any later version that has been
4907 published (not as a draft) by the Free Software Foundation. If the
4908 Document does not specify a version number of this License, you may
4909 choose any version ever published (not as a draft) by the Free
4910 Software Foundation. If the Document specifies that a proxy can
4911 decide which future versions of this License can be used, that
4912 proxy's public statement of acceptance of a version permanently
4913 authorizes you to choose that version for the Document.
4917 "Massive Multiauthor Collaboration Site" (or "MMC Site") means any
4918 World Wide Web server that publishes copyrightable works and also
4919 provides prominent facilities for anybody to edit those works. A
4920 public wiki that anybody can edit is an example of such a server.
4921 A "Massive Multiauthor Collaboration" (or "MMC") contained in the
4922 site means any set of copyrightable works thus published on the MMC
4925 "CC-BY-SA" means the Creative Commons Attribution-Share Alike 3.0
4926 license published by Creative Commons Corporation, a not-for-profit
4927 corporation with a principal place of business in San Francisco,
4928 California, as well as future copyleft versions of that license
4929 published by that same organization.
4931 "Incorporate" means to publish or republish a Document, in whole or
4932 in part, as part of another Document.
4934 An MMC is "eligible for relicensing" if it is licensed under this
4935 License, and if all works that were first published under this
4936 License somewhere other than this MMC, and subsequently
4937 incorporated in whole or in part into the MMC, (1) had no cover
4938 texts or invariant sections, and (2) were thus incorporated prior
4939 to November 1, 2008.
4941 The operator of an MMC Site may republish an MMC contained in the
4942 site under CC-BY-SA on the same site at any time before August 1,
4943 2009, provided the MMC is eligible for relicensing.
4945 ADDENDUM: How to use this License for your documents
4946 ====================================================
4948 To use this License in a document you have written, include a copy of
4949 the License in the document and put the following copyright and license
4950 notices just after the title page:
4952 Copyright (C) YEAR YOUR NAME.
4953 Permission is granted to copy, distribute and/or modify this document
4954 under the terms of the GNU Free Documentation License, Version 1.3
4955 or any later version published by the Free Software Foundation;
4956 with no Invariant Sections, no Front-Cover Texts, and no Back-Cover
4957 Texts. A copy of the license is included in the section entitled ``GNU
4958 Free Documentation License''.
4960 If you have Invariant Sections, Front-Cover Texts and Back-Cover
4961 Texts, replace the "with...Texts." line with this:
4963 with the Invariant Sections being LIST THEIR TITLES, with
4964 the Front-Cover Texts being LIST, and with the Back-Cover Texts
4967 If you have Invariant Sections without Cover Texts, or some other
4968 combination of the three, merge those two alternatives to suit the
4971 If your document contains nontrivial examples of program code, we
4972 recommend releasing these examples in parallel under your choice of free
4973 software license, such as the GNU General Public License, to permit
4974 their use in free software.
4977 File: diffutils.info, Node: Translations, Next: Index, Prev: Copying This Manual, Up: Top
4979 Appendix B Translations of This Manual
4980 **************************************
4982 Nishio Futoshi of the GNUjdoc project has prepared a Japanese
4983 translation of this manual. Its most recent version can be found at
4984 <http://openlab.ring.gr.jp/gnujdoc/cvsweb/cvsweb.cgi/gnujdoc/>.
4987 File: diffutils.info, Node: Index, Prev: Translations, Up: Top
4995 * ! output format: Context. (line 6)
4996 * +- output format: Unified Format. (line 6)
4997 * < output format: Normal. (line 6)
4998 * <<<<<<< for marking conflicts: Marking Conflicts. (line 6)
4999 * _POSIX2_VERSION: Standards conformance.
5001 * ad capability: diff Options. (line 179)
5002 * aligning tab stops: Tabs. (line 6)
5003 * alternate file names: Alternate Names. (line 6)
5004 * always color option: diff Options. (line 43)
5005 * auto color option: diff Options. (line 42)
5006 * backup file names: Backup Names. (line 6)
5007 * backup file strategy: Backups. (line 6)
5008 * binary file diff: Binary. (line 6)
5009 * blank and tab difference suppression: White Space. (line 6)
5010 * blank line difference suppression: Blank Lines. (line 6)
5011 * brief difference reports: Brief. (line 6)
5012 * bug reports: Bugs. (line 6)
5013 * C function headings: C Function Headings. (line 6)
5014 * C if-then-else output format: If-then-else. (line 6)
5015 * case difference suppression: Case Folding. (line 6)
5016 * ClearCase: Revision Control. (line 6)
5017 * cmp invocation: Invoking cmp. (line 6)
5018 * cmp options: cmp Options. (line 6)
5019 * color, distinguishing different context: diff Options. (line 37)
5020 * columnar output: Side by Side. (line 6)
5021 * common mistakes with patches: Avoiding Common Mistakes.
5023 * comparing three files: Comparing Three Files.
5025 * conflict: diff3 Merging. (line 26)
5026 * conflict marking: Marking Conflicts. (line 6)
5027 * context output format: Context. (line 6)
5028 * creating files: Creating and Removing.
5030 * de capability: diff Options. (line 184)
5031 * diagnostics from patch: patch Messages. (line 6)
5032 * diff invocation: Invoking diff. (line 6)
5033 * diff merging: Interactive Merging. (line 6)
5034 * diff options: diff Options. (line 6)
5035 * diff sample input: Sample diff Input. (line 6)
5036 * diff3 hunks: diff3 Hunks. (line 6)
5037 * diff3 invocation: Invoking diff3. (line 6)
5038 * diff3 options: diff3 Options. (line 6)
5039 * diff3 sample input: Sample diff3 Input. (line 6)
5040 * directories and patch: patch Directories. (line 6)
5041 * directory structure changes: Changing Structure. (line 6)
5042 * dry runs for patch: Dry Runs. (line 6)
5043 * ed script output format: ed Scripts. (line 6)
5044 * EDITOR: Merge Commands. (line 50)
5045 * empty files, removing: Creating and Removing.
5047 * exabyte, definition of: cmp Options. (line 99)
5048 * exbibyte, definition of: cmp Options. (line 102)
5049 * file name alternates: Alternate Names. (line 6)
5050 * file names with unusual characters: Unusual File Names. (line 6)
5051 * format of diff output: Output Formats. (line 6)
5052 * format of diff3 output: Comparing Three Files.
5054 * formats for if-then-else line groups: Line Group Formats. (line 6)
5055 * forward ed script output format: Forward ed. (line 6)
5056 * full lines: Incomplete Lines. (line 6)
5057 * function headings, C: C Function Headings. (line 6)
5058 * fuzz factor when patching: Inexact. (line 6)
5059 * gibibyte, definition of: cmp Options. (line 87)
5060 * gigabyte, definition of: cmp Options. (line 84)
5061 * hd capability: diff Options. (line 189)
5062 * headings: Sections. (line 6)
5063 * hunks: Hunks. (line 6)
5064 * hunks for diff3: diff3 Hunks. (line 6)
5065 * if-then-else output format: If-then-else. (line 6)
5066 * ifdef output format: If-then-else. (line 6)
5067 * imperfect patch application: Imperfect. (line 6)
5068 * incomplete line merging: Merging Incomplete Lines.
5070 * incomplete lines: Incomplete Lines. (line 6)
5071 * inexact patches: Inexact. (line 6)
5072 * inhibit messages from patch: More or Fewer Messages.
5074 * interactive merging: Interactive Merging. (line 6)
5075 * introduction: Comparison. (line 6)
5076 * intuiting file names from patches: Multiple Patches. (line 6)
5077 * invoking cmp: Invoking cmp. (line 6)
5078 * invoking diff: Invoking diff. (line 6)
5079 * invoking diff3: Invoking diff3. (line 6)
5080 * invoking patch: Invoking patch. (line 6)
5081 * invoking sdiff: Invoking sdiff. (line 6)
5082 * keyboard input to patch: patch and Keyboard Input.
5084 * kibibyte, definition of: cmp Options. (line 75)
5085 * kilobyte, definition of: cmp Options. (line 71)
5086 * LC_COLLATE: Comparing Directories.
5088 * LC_NUMERIC: Line Group Formats. (line 143)
5089 * LC_TIME: Detailed Context. (line 12)
5090 * line formats: Line Formats. (line 6)
5091 * line group formats: Line Group Formats. (line 6)
5092 * ln capability: diff Options. (line 194)
5093 * mebibyte, definition of: cmp Options. (line 82)
5094 * megabyte, definition of: cmp Options. (line 79)
5095 * merge commands: Merge Commands. (line 6)
5096 * merged diff3 format: Bypassing ed. (line 6)
5097 * merged output format: If-then-else. (line 6)
5098 * merging from a common ancestor: diff3 Merging. (line 6)
5099 * merging interactively: Merge Commands. (line 6)
5100 * messages from patch: patch Messages. (line 6)
5101 * multibyte characters: Internationalization.
5103 * multiple patches: Multiple Patches. (line 6)
5104 * newline treatment by diff: Incomplete Lines. (line 6)
5105 * none color option: diff Options. (line 40)
5106 * normal output format: Normal. (line 6)
5107 * options for cmp: cmp Options. (line 6)
5108 * options for diff: diff Options. (line 6)
5109 * options for diff3: diff3 Options. (line 6)
5110 * options for patch: patch Options. (line 6)
5111 * options for sdiff: sdiff Options. (line 6)
5112 * output formats: Output Formats. (line 6)
5113 * overlap: diff3 Merging. (line 26)
5114 * overlapping change, selection of: Which Changes. (line 6)
5115 * overview of diff and patch: Overview. (line 6)
5116 * paginating diff output: Pagination. (line 6)
5117 * patch consumer tips: Tips for Patch Consumers.
5119 * patch input format: patch Input. (line 6)
5120 * patch invocation: Invoking patch. (line 6)
5121 * patch messages and questions: patch Messages. (line 6)
5122 * patch options: patch Options. (line 6)
5123 * patch producer tips: Tips for Patch Producers.
5125 * patch, common mistakes: Avoiding Common Mistakes.
5127 * patches, shrinking: Generating Smaller Patches.
5129 * patching directories: patch Directories. (line 6)
5130 * PATCH_GET: Revision Control. (line 13)
5131 * PATCH_VERSION_CONTROL: Backup Names. (line 21)
5132 * pebibyte, definition of: cmp Options. (line 97)
5133 * performance of diff: diff Performance. (line 6)
5134 * petabyte, definition of: cmp Options. (line 94)
5135 * POSIX: patch and POSIX. (line 6)
5136 * POSIX <1>: Standards conformance.
5138 * POSIXLY_CORRECT: patch and POSIX. (line 6)
5139 * POSIXLY_CORRECT <1>: Standards conformance.
5141 * projects for directories: Shortcomings. (line 6)
5142 * quoting style: patch Quoting Style. (line 6)
5143 * QUOTING_STYLE: patch Quoting Style. (line 26)
5144 * RCS: Revision Control. (line 6)
5145 * RCS script output format: RCS. (line 6)
5146 * regular expression matching headings: Specified Headings. (line 6)
5147 * regular expression suppression: Specified Lines. (line 6)
5148 * reject file names: Reject Names. (line 6)
5149 * removing empty files: Creating and Removing.
5151 * reporting bugs: Bugs. (line 6)
5152 * reversed patches: Reversed Patches. (line 6)
5153 * revision control: Revision Control. (line 6)
5154 * sample input for diff: Sample diff Input. (line 6)
5155 * sample input for diff3: Sample diff3 Input. (line 6)
5156 * SCCS: Revision Control. (line 6)
5157 * script output formats: Scripts. (line 6)
5158 * sdiff invocation: Invoking sdiff. (line 6)
5159 * sdiff options: sdiff Options. (line 6)
5160 * sdiff output format: sdiff Option Summary.
5162 * section headings: Sections. (line 6)
5163 * side by side: Side by Side. (line 6)
5164 * side by side format: Side by Side Format. (line 6)
5165 * SIMPLE_BACKUP_SUFFIX: Backup Names. (line 12)
5166 * special files: Special Files. (line 6)
5167 * specified headings: Specified Headings. (line 6)
5168 * summarizing which files differ: Brief. (line 6)
5169 * System V diff3 compatibility: Saving the Changed File.
5171 * tab and blank difference suppression: White Space. (line 6)
5172 * tab stop alignment: Tabs. (line 6)
5173 * tebibyte, definition of: cmp Options. (line 92)
5174 * terabyte, definition of: cmp Options. (line 89)
5175 * terminal, using color iff: diff Options. (line 42)
5176 * testing patch: Dry Runs. (line 6)
5177 * text versus binary diff: Binary. (line 6)
5178 * timestamp format, context diffs: Detailed Context. (line 12)
5179 * timestamp format, unified diffs: Detailed Unified. (line 12)
5180 * timestamps on patched files: Patching Timestamps. (line 6)
5181 * traditional patch: patch and Tradition. (line 6)
5182 * trailing blanks: Trailing Blanks. (line 6)
5183 * two-column output: Side by Side. (line 6)
5184 * unified output format: Unified Format. (line 6)
5185 * unmerged change: Which Changes. (line 6)
5186 * varying-width characters: Internationalization.
5188 * verbose messages from patch: More or Fewer Messages.
5190 * version control: Revision Control. (line 6)
5191 * VERSION_CONTROL: Revision Control. (line 22)
5192 * VERSION_CONTROL <1>: Backup Names. (line 21)
5193 * white space in patches: Changed White Space. (line 6)
5194 * yottabyte, definition of: cmp Options. (line 110)
5195 * zettabyte, definition of: cmp Options. (line 104)
5201 Node: Overview
\7f3644
5202 Node: Comparison
\7f7398
5204 Node: White Space
\7f11781
5205 Node: Blank Lines
\7f13620
5206 Node: Specified Lines
\7f14609
5207 Node: Case Folding
\7f15735
5209 Node: Binary
\7f17475
5210 Node: Output Formats
\7f21272
5211 Node: Sample diff Input
\7f21997
5212 Node: Context
\7f23496
5213 Node: Context Format
\7f25073
5214 Node: Example Context
\7f25865
5215 Node: Less Context
\7f27372
5216 Node: Detailed Context
\7f28561
5217 Node: Unified Format
\7f30756
5218 Node: Example Unified
\7f31553
5219 Node: Detailed Unified
\7f32591
5220 Node: Sections
\7f34231
5221 Node: Specified Headings
\7f34990
5222 Node: C Function Headings
\7f36539
5223 Node: Alternate Names
\7f37385
5224 Node: Side by Side
\7f38298
5225 Node: Side by Side Format
\7f40448
5226 Node: Example Side by Side
\7f41350
5227 Node: Normal
\7f42690
5228 Node: Example Normal
\7f43691
5229 Node: Detailed Normal
\7f44427
5230 Node: Scripts
\7f46166
5231 Node: ed Scripts
\7f46571
5232 Node: Example ed
\7f47777
5233 Node: Detailed ed
\7f48227
5234 Node: Forward ed
\7f49986
5236 Node: If-then-else
\7f51978
5237 Node: Line Group Formats
\7f53656
5238 Node: Line Formats
\7f59537
5239 Node: Example If-then-else
\7f62806
5240 Node: Detailed If-then-else
\7f63885
5241 Node: Incomplete Lines
\7f65770
5242 Node: Comparing Directories
\7f67406
5243 Node: Adjusting Output
\7f71698
5245 Node: Trailing Blanks
\7f73819
5246 Node: Pagination
\7f75044
5247 Node: diff Performance
\7f75512
5248 Node: Comparing Three Files
\7f78601
5249 Node: Sample diff3 Input
\7f79479
5250 Node: Example diff3 Normal
\7f80427
5251 Node: Detailed diff3 Normal
\7f81487
5252 Node: diff3 Hunks
\7f83275
5253 Node: diff3 Merging
\7f84541
5254 Node: Which Changes
\7f86786
5255 Node: Marking Conflicts
\7f88186
5256 Node: Bypassing ed
\7f90641
5257 Node: Merging Incomplete Lines
\7f91984
5258 Node: Saving the Changed File
\7f92710
5259 Node: Interactive Merging
\7f93326
5260 Node: sdiff Option Summary
\7f94035
5261 Node: Merge Commands
\7f95238
5262 Node: Merging with patch
\7f96527
5263 Node: patch Input
\7f98894
5264 Node: Revision Control
\7f99576
5265 Node: Imperfect
\7f100746
5266 Node: Changed White Space
\7f101895
5267 Node: Reversed Patches
\7f102692
5268 Node: Inexact
\7f104156
5269 Node: Dry Runs
\7f107712
5270 Node: Creating and Removing
\7f108576
5271 Node: Patching Timestamps
\7f109625
5272 Node: Multiple Patches
\7f111812
5273 Node: patch Directories
\7f114473
5274 Node: Backups
\7f116097
5275 Node: Backup Names
\7f117161
5276 Ref: Backup Names-Footnote-1
\7f120116
5277 Node: Reject Names
\7f120243
5278 Node: patch Messages
\7f120832
5279 Node: More or Fewer Messages
\7f121892
5280 Node: patch and Keyboard Input
\7f122523
5281 Node: patch Quoting Style
\7f123552
5282 Node: patch and POSIX
\7f124698
5283 Node: patch and Tradition
\7f125537
5284 Node: Making Patches
\7f128958
5285 Node: Tips for Patch Producers
\7f129784
5286 Node: Tips for Patch Consumers
\7f131039
5287 Node: Avoiding Common Mistakes
\7f131676
5288 Node: Generating Smaller Patches
\7f134201
5289 Node: Invoking cmp
\7f135960
5290 Node: cmp Options
\7f137510
5291 Node: Invoking diff
\7f141154
5292 Node: diff Options
\7f142773
5293 Node: Invoking diff3
\7f152622
5294 Node: diff3 Options
\7f153260
5295 Node: Invoking patch
\7f156293
5296 Node: patch Options
\7f157501
5297 Node: Invoking sdiff
\7f162721
5298 Node: sdiff Options
\7f163863
5299 Node: Standards conformance
\7f166835
5300 Node: Projects
\7f168581
5301 Node: Shortcomings
\7f169293
5302 Node: Internationalization
\7f170393
5303 Node: Changing Structure
\7f171559
5304 Node: Special Files
\7f172662
5305 Node: Unusual File Names
\7f173773
5306 Node: Timestamp Order
\7f174411
5307 Node: Ignoring Changes
\7f175047
5308 Node: Speedups
\7f175816
5310 Node: Copying This Manual
\7f177131
5311 Node: Translations
\7f202252
5312 Node: Index
\7f202624