1GIT-DIFF(1) Git Manual GIT-DIFF(1)
2
6 git-diff - Show changes between commits, commit and working tree, etc
7
9 git diff [<options>] [<commit>] [--] [<path>...]
10 git diff [<options>] --cached [--merge-base] [<commit>] [--] [<path>...]
11 git diff [<options>] [--merge-base] <commit> [<commit>...] <commit> [--] [<path>...]
12 git diff [<options>] <commit>...<commit> [--] [<path>...]
13 git diff [<options>] <blob> <blob>
14 git diff [<options>] --no-index [--] <path> <path>
15
17 Show changes between the working tree and the index or a tree, changes
18 between the index and a tree, changes between two trees, changes
19 resulting from a merge, changes between two blob objects, or changes
20 between two files on disk.
21 git diff [<options>] [--] [<path>...]
23 This form is to view the changes you made relative to the index
24 (staging area for the next commit). In other words, the differences
25 are what you could tell Git to further add to the index but you
26 still haven’t. You can stage these changes by using git-add(1).
27 git diff [<options>] --no-index [--] <path> <path>
29 This form is to compare the given two paths on the filesystem. You
30 can omit the --no-index option when running the command in a
31 working tree controlled by Git and at least one of the paths points
32 outside the working tree, or when running the command outside a
33 working tree controlled by Git. This form implies --exit-code.
34 git diff [<options>] --cached [--merge-base] [<commit>] [--] [<path>...
36 ]
37 This form is to view the changes you staged for the next commit
38 relative to the named <commit>. Typically you would want comparison
39 with the latest commit, so if you do not give <commit>, it defaults
40 to HEAD. If HEAD does not exist (e.g. unborn branches) and <commit>
41 is not given, it shows all staged changes. --staged is a synonym of
42 --cached.
43 If --merge-base is given, instead of using <commit>, use the merge
45 base of <commit> and HEAD. git diff --cached --merge-base A is
46 equivalent to git diff --cached $(git merge-base A HEAD).
47 git diff [<options>] [--merge-base] <commit> [--] [<path>...]
49 This form is to view the changes you have in your working tree
50 relative to the named <commit>. You can use HEAD to compare it with
51 the latest commit, or a branch name to compare with the tip of a
52 different branch.
53 If --merge-base is given, instead of using <commit>, use the merge
55 base of <commit> and HEAD. git diff --merge-base A is equivalent
56 to git diff $(git merge-base A HEAD).
57 git diff [<options>] [--merge-base] <commit> <commit> [--] [<path>...]
59 This is to view the changes between two arbitrary <commit>.
60 If --merge-base is given, use the merge base of the two commits for
62 the "before" side. git diff --merge-base A B is equivalent to git
63 diff $(git merge-base A B) B.
64 git diff [<options>] <commit> <commit>... <commit> [--] [<path>...]
66 This form is to view the results of a merge commit. The first
67 listed <commit> must be the merge itself; the remaining two or more
68 commits should be its parents. A convenient way to produce the
69 desired set of revisions is to use the ^@ suffix. For instance, if
70 master names a merge commit, git diff master master^@ gives the
71 same combined diff as git show master.
72 git diff [<options>] <commit>..<commit> [--] [<path>...]
74 This is synonymous to the earlier form (without the ..) for viewing
75 the changes between two arbitrary <commit>. If <commit> on one side
76 is omitted, it will have the same effect as using HEAD instead.
77 git diff [<options>] <commit>...<commit> [--] [<path>...]
79 This form is to view the changes on the branch containing and up to
80 the second <commit>, starting at a common ancestor of both
81 <commit>. git diff A...B is equivalent to git diff $(git
82 merge-base A B) B. You can omit any one of <commit>, which has the
83 same effect as using HEAD instead.
84 Just in case you are doing something exotic, it should be noted that
86 all of the <commit> in the above description, except in the
87 --merge-base case and in the last two forms that use .. notations, can
88 be any <tree>.
89 For a more complete list of ways to spell <commit>, see "SPECIFYING
91 REVISIONS" section in gitrevisions(7). However, "diff" is about
92 comparing two endpoints, not ranges, and the range notations
93 (<commit>..<commit> and <commit>...<commit>) do not mean a range as
94 defined in the "SPECIFYING RANGES" section in gitrevisions(7).
95 git diff [<options>] <blob> <blob>
97 This form is to view the differences between the raw contents of
98 two blob objects.
99
101 -p, -u, --patch
102 Generate patch (see section on generating patches). This is the
103 default.
104 -s, --no-patch
106 Suppress diff output. Useful for commands like git show that show
107 the patch by default, or to cancel the effect of --patch.
108 -U<n>, --unified=<n>
110 Generate diffs with <n> lines of context instead of the usual
111 three. Implies --patch.
112 --output=<file>
114 Output to a specific file instead of stdout.
115 --output-indicator-new=<char>, --output-indicator-old=<char>,
117 --output-indicator-context=<char>
118 Specify the character used to indicate new, old or context lines in
119 the generated patch. Normally they are +, - and ' ' respectively.
120 --raw
122 Generate the diff in raw format.
123 --patch-with-raw
125 Synonym for -p --raw.
126 --indent-heuristic
128 Enable the heuristic that shifts diff hunk boundaries to make
129 patches easier to read. This is the default.
130 --no-indent-heuristic
132 Disable the indent heuristic.
133 --minimal
135 Spend extra time to make sure the smallest possible diff is
136 produced.
137 --patience
139 Generate a diff using the "patience diff" algorithm.
140 --histogram
142 Generate a diff using the "histogram diff" algorithm.
143 --anchored=<text>
145 Generate a diff using the "anchored diff" algorithm.
146 This option may be specified more than once.
148 If a line exists in both the source and destination, exists only
150 once, and starts with this text, this algorithm attempts to prevent
151 it from appearing as a deletion or addition in the output. It uses
152 the "patience diff" algorithm internally.
153 --diff-algorithm={patience|minimal|histogram|myers}
155 Choose a diff algorithm. The variants are as follows:
156 default, myers
158 The basic greedy diff algorithm. Currently, this is the
159 default.
160 minimal
162 Spend extra time to make sure the smallest possible diff is
163 produced.
164 patience
166 Use "patience diff" algorithm when generating patches.
167 histogram
169 This algorithm extends the patience algorithm to "support
170 low-occurrence common elements".
171 For instance, if you configured the diff.algorithm variable to a
173 non-default value and want to use the default one, then you have to
174 use --diff-algorithm=default option.
175 --stat[=<width>[,<name-width>[,<count>]]]
177 Generate a diffstat. By default, as much space as necessary will be
178 used for the filename part, and the rest for the graph part.
179 Maximum width defaults to terminal width, or 80 columns if not
180 connected to a terminal, and can be overridden by <width>. The
181 width of the filename part can be limited by giving another width
182 <name-width> after a comma. The width of the graph part can be
183 limited by using --stat-graph-width=<width> (affects all commands
184 generating a stat graph) or by setting diff.statGraphWidth=<width>
185 (does not affect git format-patch). By giving a third parameter
186 <count>, you can limit the output to the first <count> lines,
187 followed by ... if there are more.
188 These parameters can also be set individually with
190 --stat-width=<width>, --stat-name-width=<name-width> and
191 --stat-count=<count>.
192 --compact-summary
194 Output a condensed summary of extended header information such as
195 file creations or deletions ("new" or "gone", optionally "+l" if
196 it’s a symlink) and mode changes ("+x" or "-x" for adding or
197 removing executable bit respectively) in diffstat. The information
198 is put between the filename part and the graph part. Implies
199 --stat.
200 --numstat
202 Similar to --stat, but shows number of added and deleted lines in
203 decimal notation and pathname without abbreviation, to make it more
204 machine friendly. For binary files, outputs two - instead of saying
205 0 0.
206 --shortstat
208 Output only the last line of the --stat format containing total
209 number of modified files, as well as number of added and deleted
210 lines.
211 -X[<param1,param2,...>], --dirstat[=<param1,param2,...>]
213 Output the distribution of relative amount of changes for each
214 sub-directory. The behavior of --dirstat can be customized by
215 passing it a comma separated list of parameters. The defaults are
216 controlled by the diff.dirstat configuration variable (see git-
217 config(1)). The following parameters are available:
218 changes
220 Compute the dirstat numbers by counting the lines that have
221 been removed from the source, or added to the destination. This
222 ignores the amount of pure code movements within a file. In
223 other words, rearranging lines in a file is not counted as much
224 as other changes. This is the default behavior when no
225 parameter is given.
226 lines
228 Compute the dirstat numbers by doing the regular line-based
229 diff analysis, and summing the removed/added line counts. (For
230 binary files, count 64-byte chunks instead, since binary files
231 have no natural concept of lines). This is a more expensive
232 --dirstat behavior than the changes behavior, but it does count
233 rearranged lines within a file as much as other changes. The
234 resulting output is consistent with what you get from the other
235 --*stat options.
236 files
238 Compute the dirstat numbers by counting the number of files
239 changed. Each changed file counts equally in the dirstat
240 analysis. This is the computationally cheapest --dirstat
241 behavior, since it does not have to look at the file contents
242 at all.
243 cumulative
245 Count changes in a child directory for the parent directory as
246 well. Note that when using cumulative, the sum of the
247 percentages reported may exceed 100%. The default
248 (non-cumulative) behavior can be specified with the
249 noncumulative parameter.
250 <limit>
252 An integer parameter specifies a cut-off percent (3% by
253 default). Directories contributing less than this percentage of
254 the changes are not shown in the output.
255 Example: The following will count changed files, while ignoring
257 directories with less than 10% of the total amount of changed
258 files, and accumulating child directory counts in the parent
259 directories: --dirstat=files,10,cumulative.
260 --cumulative
262 Synonym for --dirstat=cumulative
263 --dirstat-by-file[=<param1,param2>...]
265 Synonym for --dirstat=files,param1,param2...
266 --summary
268 Output a condensed summary of extended header information such as
269 creations, renames and mode changes.
270 --patch-with-stat
272 Synonym for -p --stat.
273 -z
275 When --raw, --numstat, --name-only or --name-status has been given,
276 do not munge pathnames and use NULs as output field terminators.
277 Without this option, pathnames with "unusual" characters are quoted
279 as explained for the configuration variable core.quotePath (see
280 git-config(1)).
281 --name-only
283 Show only names of changed files. The file names are often encoded
284 in UTF-8. For more information see the discussion about encoding in
285 the git-log(1) manual page.
286 --name-status
288 Show only names and status of changed files. See the description of
289 the --diff-filter option on what the status letters mean. Just like
290 --name-only the file names are often encoded in UTF-8.
291 --submodule[=<format>]
293 Specify how differences in submodules are shown. When specifying
294 --submodule=short the short format is used. This format just shows
295 the names of the commits at the beginning and end of the range.
296 When --submodule or --submodule=log is specified, the log format is
297 used. This format lists the commits in the range like git-
298 submodule(1) summary does. When --submodule=diff is specified, the
299 diff format is used. This format shows an inline diff of the
300 changes in the submodule contents between the commit range.
301 Defaults to diff.submodule or the short format if the config option
302 is unset.
303 --color[=<when>]
305 Show colored diff. --color (i.e. without =<when>) is the same as
306 --color=always. <when> can be one of always, never, or auto. It
307 can be changed by the color.ui and color.diff configuration
308 settings.
309 --no-color
311 Turn off colored diff. This can be used to override configuration
312 settings. It is the same as --color=never.
313 --color-moved[=<mode>]
315 Moved lines of code are colored differently. It can be changed by
316 the diff.colorMoved configuration setting. The <mode> defaults to
317 no if the option is not given and to zebra if the option with no
318 mode is given. The mode must be one of:
319 no
321 Moved lines are not highlighted.
322 default
324 Is a synonym for zebra. This may change to a more sensible mode
325 in the future.
326 plain
328 Any line that is added in one location and was removed in
329 another location will be colored with color.diff.newMoved.
330 Similarly color.diff.oldMoved will be used for removed lines
331 that are added somewhere else in the diff. This mode picks up
332 any moved line, but it is not very useful in a review to
333 determine if a block of code was moved without permutation.
334 blocks
336 Blocks of moved text of at least 20 alphanumeric characters are
337 detected greedily. The detected blocks are painted using either
338 the color.diff.{old,new}Moved color. Adjacent blocks cannot be
339 told apart.
340 zebra
342 Blocks of moved text are detected as in blocks mode. The blocks
343 are painted using either the color.diff.{old,new}Moved color or
344 color.diff.{old,new}MovedAlternative. The change between the
345 two colors indicates that a new block was detected.
346 dimmed-zebra
348 Similar to zebra, but additional dimming of uninteresting parts
349 of moved code is performed. The bordering lines of two adjacent
350 blocks are considered interesting, the rest is uninteresting.
351 dimmed_zebra is a deprecated synonym.
352 --no-color-moved
354 Turn off move detection. This can be used to override configuration
355 settings. It is the same as --color-moved=no.
356 --color-moved-ws=<modes>
358 This configures how whitespace is ignored when performing the move
359 detection for --color-moved. It can be set by the diff.colorMovedWS
360 configuration setting. These modes can be given as a comma
361 separated list:
362 no
364 Do not ignore whitespace when performing move detection.
365 ignore-space-at-eol
367 Ignore changes in whitespace at EOL.
368 ignore-space-change
370 Ignore changes in amount of whitespace. This ignores whitespace
371 at line end, and considers all other sequences of one or more
372 whitespace characters to be equivalent.
373 ignore-all-space
375 Ignore whitespace when comparing lines. This ignores
376 differences even if one line has whitespace where the other
377 line has none.
378 allow-indentation-change
380 Initially ignore any whitespace in the move detection, then
381 group the moved code blocks only into a block if the change in
382 whitespace is the same per line. This is incompatible with the
383 other modes.
384 --no-color-moved-ws
386 Do not ignore whitespace when performing move detection. This can
387 be used to override configuration settings. It is the same as
388 --color-moved-ws=no.
389 --word-diff[=<mode>]
391 Show a word diff, using the <mode> to delimit changed words. By
392 default, words are delimited by whitespace; see --word-diff-regex
393 below. The <mode> defaults to plain, and must be one of:
394 color
396 Highlight changed words using only colors. Implies --color.
397 plain
399 Show words as [-removed-] and {+added+}. Makes no attempts to
400 escape the delimiters if they appear in the input, so the
401 output may be ambiguous.
402 porcelain
404 Use a special line-based format intended for script
405 consumption. Added/removed/unchanged runs are printed in the
406 usual unified diff format, starting with a +/-/` ` character at
407 the beginning of the line and extending to the end of the line.
408 Newlines in the input are represented by a tilde ~ on a line of
409 its own.
410 none
412 Disable word diff again.
413 Note that despite the name of the first mode, color is used to
415 highlight the changed parts in all modes if enabled.
416 --word-diff-regex=<regex>
418 Use <regex> to decide what a word is, instead of considering runs
419 of non-whitespace to be a word. Also implies --word-diff unless it
420 was already enabled.
421 Every non-overlapping match of the <regex> is considered a word.
423 Anything between these matches is considered whitespace and
424 ignored(!) for the purposes of finding differences. You may want to
425 append |[^[:space:]] to your regular expression to make sure that
426 it matches all non-whitespace characters. A match that contains a
427 newline is silently truncated(!) at the newline.
428 For example, --word-diff-regex=. will treat each character as a
430 word and, correspondingly, show differences character by character.
431 The regex can also be set via a diff driver or configuration
433 option, see gitattributes(5) or git-config(1). Giving it explicitly
434 overrides any diff driver or configuration setting. Diff drivers
435 override configuration settings.
436 --color-words[=<regex>]
438 Equivalent to --word-diff=color plus (if a regex was specified)
439 --word-diff-regex=<regex>.
440 --no-renames
442 Turn off rename detection, even when the configuration file gives
443 the default to do so.
444 --[no-]rename-empty
446 Whether to use empty blobs as rename source.
447 --check
449 Warn if changes introduce conflict markers or whitespace errors.
450 What are considered whitespace errors is controlled by
451 core.whitespace configuration. By default, trailing whitespaces
452 (including lines that consist solely of whitespaces) and a space
453 character that is immediately followed by a tab character inside
454 the initial indent of the line are considered whitespace errors.
455 Exits with non-zero status if problems are found. Not compatible
456 with --exit-code.
457 --ws-error-highlight=<kind>
459 Highlight whitespace errors in the context, old or new lines of the
460 diff. Multiple values are separated by comma, none resets previous
461 values, default reset the list to new and all is a shorthand for
462 old,new,context. When this option is not given, and the
463 configuration variable diff.wsErrorHighlight is not set, only
464 whitespace errors in new lines are highlighted. The whitespace
465 errors are colored with color.diff.whitespace.
466 --full-index
468 Instead of the first handful of characters, show the full pre- and
469 post-image blob object names on the "index" line when generating
470 patch format output.
471 --binary
473 In addition to --full-index, output a binary diff that can be
474 applied with git-apply. Implies --patch.
475 --abbrev[=<n>]
477 Instead of showing the full 40-byte hexadecimal object name in
478 diff-raw format output and diff-tree header lines, show the
479 shortest prefix that is at least <n> hexdigits long that uniquely
480 refers the object. In diff-patch output format, --full-index takes
481 higher precedence, i.e. if --full-index is specified, full blob
482 names will be shown regardless of --abbrev. Non default number of
483 digits can be specified with --abbrev=<n>.
484 -B[<n>][/<m>], --break-rewrites[=[<n>][/<m>]]
486 Break complete rewrite changes into pairs of delete and create.
487 This serves two purposes:
488 It affects the way a change that amounts to a total rewrite of a
490 file not as a series of deletion and insertion mixed together with
491 a very few lines that happen to match textually as the context, but
492 as a single deletion of everything old followed by a single
493 insertion of everything new, and the number m controls this aspect
494 of the -B option (defaults to 60%). -B/70% specifies that less
495 than 30% of the original should remain in the result for Git to
496 consider it a total rewrite (i.e. otherwise the resulting patch
497 will be a series of deletion and insertion mixed together with
498 context lines).
499 When used with -M, a totally-rewritten file is also considered as
501 the source of a rename (usually -M only considers a file that
502 disappeared as the source of a rename), and the number n controls
503 this aspect of the -B option (defaults to 50%). -B20% specifies
504 that a change with addition and deletion compared to 20% or more of
505 the file’s size are eligible for being picked up as a possible
506 source of a rename to another file.
507 -M[<n>], --find-renames[=<n>]
509 Detect renames. If n is specified, it is a threshold on the
510 similarity index (i.e. amount of addition/deletions compared to the
511 file’s size). For example, -M90% means Git should consider a
512 delete/add pair to be a rename if more than 90% of the file hasn’t
513 changed. Without a % sign, the number is to be read as a fraction,
514 with a decimal point before it. I.e., -M5 becomes 0.5, and is thus
515 the same as -M50%. Similarly, -M05 is the same as -M5%. To limit
516 detection to exact renames, use -M100%. The default similarity
517 index is 50%.
518 -C[<n>], --find-copies[=<n>]
520 Detect copies as well as renames. See also --find-copies-harder. If
521 n is specified, it has the same meaning as for -M<n>.
522 --find-copies-harder
524 For performance reasons, by default, -C option finds copies only if
525 the original file of the copy was modified in the same changeset.
526 This flag makes the command inspect unmodified files as candidates
527 for the source of copy. This is a very expensive operation for
528 large projects, so use it with caution. Giving more than one -C
529 option has the same effect.
530 -D, --irreversible-delete
532 Omit the preimage for deletes, i.e. print only the header but not
533 the diff between the preimage and /dev/null. The resulting patch is
534 not meant to be applied with patch or git apply; this is solely for
535 people who want to just concentrate on reviewing the text after the
536 change. In addition, the output obviously lacks enough information
537 to apply such a patch in reverse, even manually, hence the name of
538 the option.
539 When used together with -B, omit also the preimage in the deletion
541 part of a delete/create pair.
542 -l<num>
544 The -M and -C options involve some preliminary steps that can
545 detect subsets of renames/copies cheaply, followed by an exhaustive
546 fallback portion that compares all remaining unpaired destinations
547 to all relevant sources. (For renames, only remaining unpaired
548 sources are relevant; for copies, all original sources are
549 relevant.) For N sources and destinations, this exhaustive check is
550 O(N^2). This option prevents the exhaustive portion of rename/copy
551 detection from running if the number of source/destination files
552 involved exceeds the specified number. Defaults to
553 diff.renameLimit. Note that a value of 0 is treated as unlimited.
554 --diff-filter=[(A|C|D|M|R|T|U|X|B)...[*]]
556 Select only files that are Added (A), Copied (C), Deleted (D),
557 Modified (M), Renamed (R), have their type (i.e. regular file,
558 symlink, submodule, ...) changed (T), are Unmerged (U), are Unknown
559 (X), or have had their pairing Broken (B). Any combination of the
560 filter characters (including none) can be used. When *
561 (All-or-none) is added to the combination, all paths are selected
562 if there is any file that matches other criteria in the comparison;
563 if there is no file that matches other criteria, nothing is
564 selected.
565 Also, these upper-case letters can be downcased to exclude. E.g.
567 --diff-filter=ad excludes added and deleted paths.
568 Note that not all diffs can feature all types. For instance, diffs
570 from the index to the working tree can never have Added entries
571 (because the set of paths included in the diff is limited by what
572 is in the index). Similarly, copied and renamed entries cannot
573 appear if detection for those types is disabled.
574 -S<string>
576 Look for differences that change the number of occurrences of the
577 specified string (i.e. addition/deletion) in a file. Intended for
578 the scripter’s use.
579 It is useful when you’re looking for an exact block of code (like a
581 struct), and want to know the history of that block since it first
582 came into being: use the feature iteratively to feed the
583 interesting block in the preimage back into -S, and keep going
584 until you get the very first version of the block.
585 Binary files are searched as well.
587 -G<regex>
589 Look for differences whose patch text contains added/removed lines
590 that match <regex>.
591 To illustrate the difference between -S<regex> --pickaxe-regex and
593 -G<regex>, consider a commit with the following diff in the same
594 file:
595 + return frotz(nitfol, two->ptr, 1, 0);
597 ...
598 - hit = frotz(nitfol, mf2.ptr, 1, 0);
599 While git log -G"frotz\(nitfol" will show this commit, git log
601 -S"frotz\(nitfol" --pickaxe-regex will not (because the number of
602 occurrences of that string did not change).
603 Unless --text is supplied patches of binary files without a
605 textconv filter will be ignored.
606 See the pickaxe entry in gitdiffcore(7) for more information.
608 --find-object=<object-id>
610 Look for differences that change the number of occurrences of the
611 specified object. Similar to -S, just the argument is different in
612 that it doesn’t search for a specific string but for a specific
613 object id.
614 The object can be a blob or a submodule commit. It implies the -t
616 option in git-log to also find trees.
617 --pickaxe-all
619 When -S or -G finds a change, show all the changes in that
620 changeset, not just the files that contain the change in <string>.
621 --pickaxe-regex
623 Treat the <string> given to -S as an extended POSIX regular
624 expression to match.
625 -O<orderfile>
627 Control the order in which files appear in the output. This
628 overrides the diff.orderFile configuration variable (see git-
629 config(1)). To cancel diff.orderFile, use -O/dev/null.
630 The output order is determined by the order of glob patterns in
632 <orderfile>. All files with pathnames that match the first pattern
633 are output first, all files with pathnames that match the second
634 pattern (but not the first) are output next, and so on. All files
635 with pathnames that do not match any pattern are output last, as if
636 there was an implicit match-all pattern at the end of the file. If
637 multiple pathnames have the same rank (they match the same pattern
638 but no earlier patterns), their output order relative to each other
639 is the normal order.
640 <orderfile> is parsed as follows:
642 • Blank lines are ignored, so they can be used as separators for
644 readability.
645 • Lines starting with a hash ("#") are ignored, so they can be
647 used for comments. Add a backslash ("\") to the beginning of
648 the pattern if it starts with a hash.
649 • Each other line contains a single pattern.
651 Patterns have the same syntax and semantics as patterns used for
653 fnmatch(3) without the FNM_PATHNAME flag, except a pathname also
654 matches a pattern if removing any number of the final pathname
655 components matches the pattern. For example, the pattern "foo*bar"
656 matches "fooasdfbar" and "foo/bar/baz/asdf" but not "foobarx".
657 --skip-to=<file>, --rotate-to=<file>
659 Discard the files before the named <file> from the output (i.e.
660 skip to), or move them to the end of the output (i.e. rotate to).
661 These were invented primarily for use of the git difftool command,
662 and may not be very useful otherwise.
663 -R
665 Swap two inputs; that is, show differences from index or on-disk
666 file to tree contents.
667 --relative[=<path>], --no-relative
669 When run from a subdirectory of the project, it can be told to
670 exclude changes outside the directory and show pathnames relative
671 to it with this option. When you are not in a subdirectory (e.g. in
672 a bare repository), you can name which subdirectory to make the
673 output relative to by giving a <path> as an argument.
674 --no-relative can be used to countermand both diff.relative config
675 option and previous --relative.
676 -a, --text
678 Treat all files as text.
679 --ignore-cr-at-eol
681 Ignore carriage-return at the end of line when doing a comparison.
682 --ignore-space-at-eol
684 Ignore changes in whitespace at EOL.
685 -b, --ignore-space-change
687 Ignore changes in amount of whitespace. This ignores whitespace at
688 line end, and considers all other sequences of one or more
689 whitespace characters to be equivalent.
690 -w, --ignore-all-space
692 Ignore whitespace when comparing lines. This ignores differences
693 even if one line has whitespace where the other line has none.
694 --ignore-blank-lines
696 Ignore changes whose lines are all blank.
697 -I<regex>, --ignore-matching-lines=<regex>
699 Ignore changes whose all lines match <regex>. This option may be
700 specified more than once.
701 --inter-hunk-context=<lines>
703 Show the context between diff hunks, up to the specified number of
704 lines, thereby fusing hunks that are close to each other. Defaults
705 to diff.interHunkContext or 0 if the config option is unset.
706 -W, --function-context
708 Show whole function as context lines for each change. The function
709 names are determined in the same way as git diff works out patch
710 hunk headers (see Defining a custom hunk-header in
711 gitattributes(5)).
712 --exit-code
714 Make the program exit with codes similar to diff(1). That is, it
715 exits with 1 if there were differences and 0 means no differences.
716 --quiet
718 Disable all output of the program. Implies --exit-code.
719 --ext-diff
721 Allow an external diff helper to be executed. If you set an
722 external diff driver with gitattributes(5), you need to use this
723 option with git-log(1) and friends.
724 --no-ext-diff
726 Disallow external diff drivers.
727 --textconv, --no-textconv
729 Allow (or disallow) external text conversion filters to be run when
730 comparing binary files. See gitattributes(5) for details. Because
731 textconv filters are typically a one-way conversion, the resulting
732 diff is suitable for human consumption, but cannot be applied. For
733 this reason, textconv filters are enabled by default only for git-
734 diff(1) and git-log(1), but not for git-format-patch(1) or diff
735 plumbing commands.
736 --ignore-submodules[=<when>]
738 Ignore changes to submodules in the diff generation. <when> can be
739 either "none", "untracked", "dirty" or "all", which is the default.
740 Using "none" will consider the submodule modified when it either
741 contains untracked or modified files or its HEAD differs from the
742 commit recorded in the superproject and can be used to override any
743 settings of the ignore option in git-config(1) or gitmodules(5).
744 When "untracked" is used submodules are not considered dirty when
745 they only contain untracked content (but they are still scanned for
746 modified content). Using "dirty" ignores all changes to the work
747 tree of submodules, only changes to the commits stored in the
748 superproject are shown (this was the behavior until 1.7.0). Using
749 "all" hides all changes to submodules.
750 --src-prefix=<prefix>
752 Show the given source prefix instead of "a/".
753 --dst-prefix=<prefix>
755 Show the given destination prefix instead of "b/".
756 --no-prefix
758 Do not show any source or destination prefix.
759 --line-prefix=<prefix>
761 Prepend an additional prefix to every line of output.
762 --ita-invisible-in-index
764 By default entries added by "git add -N" appear as an existing
765 empty file in "git diff" and a new file in "git diff --cached".
766 This option makes the entry appear as a new file in "git diff" and
767 non-existent in "git diff --cached". This option could be reverted
768 with --ita-visible-in-index. Both options are experimental and
769 could be removed in future.
770 For more detailed explanation on these common options, see also
772 gitdiffcore(7).
773 -1 --base, -2 --ours, -3 --theirs
775 Compare the working tree with the "base" version (stage #1), "our
776 branch" (stage #2) or "their branch" (stage #3). The index contains
777 these stages only for unmerged entries i.e. while resolving
778 conflicts. See git-read-tree(1) section "3-Way Merge" for detailed
779 information.
780 -0
782 Omit diff output for unmerged entries and just show "Unmerged". Can
783 be used only when comparing the working tree with the index.
784 <path>...
786 The <paths> parameters, when given, are used to limit the diff to
787 the named paths (you can give directory names and get diff for all
788 files under them).
789
791 The raw output format from "git-diff-index", "git-diff-tree",
792 "git-diff-files" and "git diff --raw" are very similar.
793 These commands all compare two sets of things; what is compared
795 differs:
796 git-diff-index <tree-ish>
798 compares the <tree-ish> and the files on the filesystem.
799 git-diff-index --cached <tree-ish>
801 compares the <tree-ish> and the index.
802 git-diff-tree [-r] <tree-ish-1> <tree-ish-2> [<pattern>...]
804 compares the trees named by the two arguments.
805 git-diff-files [<pattern>...]
807 compares the index and the files on the filesystem.
808 The "git-diff-tree" command begins its output by printing the hash of
810 what is being compared. After that, all the commands print one output
811 line per changed file.
812 An output line is formatted this way:
814 in-place edit :100644 100644 bcd1234 0123456 M file0
816 copy-edit :100644 100644 abcd123 1234567 C68 file1 file2
817 rename-edit :100644 100644 abcd123 1234567 R86 file1 file3
818 create :000000 100644 0000000 1234567 A file4
819 delete :100644 000000 1234567 0000000 D file5
820 unmerged :000000 000000 0000000 0000000 U file6
821 That is, from the left to the right:
823 1. a colon.
825 2. mode for "src"; 000000 if creation or unmerged.
827 3. a space.
829 4. mode for "dst"; 000000 if deletion or unmerged.
831 5. a space.
833 6. sha1 for "src"; 0{40} if creation or unmerged.
835 7. a space.
837 8. sha1 for "dst"; 0{40} if creation, unmerged or "look at work tree".
839 9. a space.
841 10. status, followed by optional "score" number.
843 11. a tab or a NUL when -z option is used.
845 12. path for "src"
847 13. a tab or a NUL when -z option is used; only exists for C or R.
849 14. path for "dst"; only exists for C or R.
851 15. an LF or a NUL when -z option is used, to terminate the record.
853 Possible status letters are:
855 • A: addition of a file
857 • C: copy of a file into a new one
859 • D: deletion of a file
861 • M: modification of the contents or mode of a file
863 • R: renaming of a file
865 • T: change in the type of the file
867 • U: file is unmerged (you must complete the merge before it can be
869 committed)
870 • X: "unknown" change type (most probably a bug, please report it)
872 Status letters C and R are always followed by a score (denoting the
874 percentage of similarity between the source and target of the move or
875 copy). Status letter M may be followed by a score (denoting the
876 percentage of dissimilarity) for file rewrites.
877 <sha1> is shown as all 0’s if a file is new on the filesystem and it is
879 out of sync with the index.
880 Example:
882 :100644 100644 5be4a4a 0000000 M file.c
884 Without the -z option, pathnames with "unusual" characters are quoted
886 as explained for the configuration variable core.quotePath (see git-
887 config(1)). Using -z the filename is output verbatim and the line is
888 terminated by a NUL byte.
889
891 "git-diff-tree", "git-diff-files" and "git-diff --raw" can take -c or
892 --cc option to generate diff output also for merge commits. The output
893 differs from the format described above in the following way:
894 1. there is a colon for each parent
896 2. there are more "src" modes and "src" sha1
898 3. status is concatenated status characters for each parent
900 4. no optional "score" number
902 5. tab-separated pathname(s) of the file
904 For -c and --cc, only the destination or final path is shown even if
906 the file was renamed on any side of history. With --combined-all-paths,
907 the name of the path in each parent is shown followed by the name of
908 the path in the merge commit.
909 Examples for -c and --cc without --combined-all-paths:
911 ::100644 100644 100644 fabadb8 cc95eb0 4866510 MM desc.c
913 ::100755 100755 100755 52b7a2d 6d1ac04 d2ac7d7 RM bar.sh
914 ::100644 100644 100644 e07d6c5 9042e82 ee91881 RR phooey.c
915 Examples when --combined-all-paths added to either -c or --cc:
917 ::100644 100644 100644 fabadb8 cc95eb0 4866510 MM desc.c desc.c desc.c
919 ::100755 100755 100755 52b7a2d 6d1ac04 d2ac7d7 RM foo.sh bar.sh bar.sh
920 ::100644 100644 100644 e07d6c5 9042e82 ee91881 RR fooey.c fuey.c phooey.c
921 Note that combined diff lists only files which were modified from all
923 parents.
924
926 Running git-diff(1), git-log(1), git-show(1), git-diff-index(1), git-
927 diff-tree(1), or git-diff-files(1) with the -p option produces patch
928 text. You can customize the creation of patch text via the
929 GIT_EXTERNAL_DIFF and the GIT_DIFF_OPTS environment variables (see
930 git(1)), and the diff attribute (see gitattributes(5)).
931 What the -p option produces is slightly different from the traditional
933 diff format:
934 1. It is preceded with a "git diff" header that looks like this:
936 diff --git a/file1 b/file2
938 The a/ and b/ filenames are the same unless rename/copy is
940 involved. Especially, even for a creation or a deletion, /dev/null
941 is not used in place of the a/ or b/ filenames.
942 When rename/copy is involved, file1 and file2 show the name of the
944 source file of the rename/copy and the name of the file that
945 rename/copy produces, respectively.
946 2. It is followed by one or more extended header lines:
948 old mode <mode>
950 new mode <mode>
951 deleted file mode <mode>
952 new file mode <mode>
953 copy from <path>
954 copy to <path>
955 rename from <path>
956 rename to <path>
957 similarity index <number>
958 dissimilarity index <number>
959 index <hash>..<hash> <mode>
960 File modes are printed as 6-digit octal numbers including the file
962 type and file permission bits.
963 Path names in extended headers do not include the a/ and b/
965 prefixes.
966 The similarity index is the percentage of unchanged lines, and the
968 dissimilarity index is the percentage of changed lines. It is a
969 rounded down integer, followed by a percent sign. The similarity
970 index value of 100% is thus reserved for two equal files, while
971 100% dissimilarity means that no line from the old file made it
972 into the new one.
973 The index line includes the blob object names before and after the
975 change. The <mode> is included if the file mode does not change;
976 otherwise, separate lines indicate the old and the new mode.
977 3. Pathnames with "unusual" characters are quoted as explained for the
979 configuration variable core.quotePath (see git-config(1)).
980 4. All the file1 files in the output refer to files before the commit,
982 and all the file2 files refer to files after the commit. It is
983 incorrect to apply each change to each file sequentially. For
984 example, this patch will swap a and b:
985 diff --git a/a b/b
987 rename from a
988 rename to b
989 diff --git a/b b/a
990 rename from b
991 rename to a
992 5. Hunk headers mention the name of the function to which the hunk
994 applies. See "Defining a custom hunk-header" in gitattributes(5)
995 for details of how to tailor to this to specific languages.
996
998 Any diff-generating command can take the -c or --cc option to produce a
999 combined diff when showing a merge. This is the default format when
1000 showing merges with git-diff(1) or git-show(1). Note also that you can
1001 give suitable --diff-merges option to any of these commands to force
1002 generation of diffs in specific format.
1003 A "combined diff" format looks like this:
1005 diff --combined describe.c
1007 index fabadb8,cc95eb0..4866510
1008 --- a/describe.c
1009 +++ b/describe.c
1010 @@@ -98,20 -98,12 +98,20 @@@
1011 return (a_date > b_date) ? -1 : (a_date == b_date) ? 0 : 1;
1012 }
1013 - static void describe(char *arg)
1015 -static void describe(struct commit *cmit, int last_one)
1016 ++static void describe(char *arg, int last_one)
1017 {
1018 + unsigned char sha1[20];
1019 + struct commit *cmit;
1020 struct commit_list *list;
1021 static int initialized = 0;
1022 struct commit_name *n;
1023 + if (get_sha1(arg, sha1) < 0)
1025 + usage(describe_usage);
1026 + cmit = lookup_commit_reference(sha1);
1027 + if (!cmit)
1028 + usage(describe_usage);
1029 +
1030 if (!initialized) {
1031 initialized = 1;
1032 for_each_ref(get_name);
1033 1. It is preceded with a "git diff" header, that looks like this (when
1035 the -c option is used):
1036 diff --combined file
1038 or like this (when the --cc option is used):
1040 diff --cc file
1042 2. It is followed by one or more extended header lines (this example
1044 shows a merge with two parents):
1045 index <hash>,<hash>..<hash>
1047 mode <mode>,<mode>..<mode>
1048 new file mode <mode>
1049 deleted file mode <mode>,<mode>
1050 The mode <mode>,<mode>..<mode> line appears only if at least one of
1052 the <mode> is different from the rest. Extended headers with
1053 information about detected contents movement (renames and copying
1054 detection) are designed to work with diff of two <tree-ish> and are
1055 not used by combined diff format.
1056 3. It is followed by two-line from-file/to-file header
1058 --- a/file
1060 +++ b/file
1061 Similar to two-line header for traditional unified diff format,
1063 /dev/null is used to signal created or deleted files.
1064 However, if the --combined-all-paths option is provided, instead of
1066 a two-line from-file/to-file you get a N+1 line from-file/to-file
1067 header, where N is the number of parents in the merge commit
1068 --- a/file
1070 --- a/file
1071 --- a/file
1072 +++ b/file
1073 This extended format can be useful if rename or copy detection is
1075 active, to allow you to see the original name of the file in
1076 different parents.
1077 4. Chunk header format is modified to prevent people from accidentally
1079 feeding it to patch -p1. Combined diff format was created for
1080 review of merge commit changes, and was not meant to be applied.
1081 The change is similar to the change in the extended index header:
1082 @@@ <from-file-range> <from-file-range> <to-file-range> @@@
1084 There are (number of parents + 1) @ characters in the chunk header
1086 for combined diff format.
1087 Unlike the traditional unified diff format, which shows two files A and
1089 B with a single column that has - (minus — appears in A but removed in
1090 B), + (plus — missing in A but added to B), or " " (space — unchanged)
1091 prefix, this format compares two or more files file1, file2,... with
1092 one file X, and shows how X differs from each of fileN. One column for
1093 each of fileN is prepended to the output line to note how X’s line is
1094 different from it.
1095 A - character in the column N means that the line appears in fileN but
1097 it does not appear in the result. A + character in the column N means
1098 that the line appears in the result, and fileN does not have that line
1099 (in other words, the line was added, from the point of view of that
1100 parent).
1101 In the above example output, the function signature was changed from
1103 both files (hence two - removals from both file1 and file2, plus ++ to
1104 mean one line that was added does not appear in either file1 or file2).
1105 Also eight other lines are the same from file1 but do not appear in
1106 file2 (hence prefixed with +).
1107 When shown by git diff-tree -c, it compares the parents of a merge
1109 commit with the merge result (i.e. file1..fileN are the parents). When
1110 shown by git diff-files -c, it compares the two unresolved merge
1111 parents with the working tree file (i.e. file1 is stage 2 aka "our
1112 version", file2 is stage 3 aka "their version").
1113
1115 The --summary option describes newly added, deleted, renamed and copied
1116 files. The --stat option adds diffstat(1) graph to the output. These
1117 options can be combined with other options, such as -p, and are meant
1118 for human consumption.
1119 When showing a change that involves a rename or a copy, --stat output
1121 formats the pathnames compactly by combining common prefix and suffix
1122 of the pathnames. For example, a change that moves arch/i386/Makefile
1123 to arch/x86/Makefile while modifying 4 lines will be shown like this:
1124 arch/{i386 => x86}/Makefile | 4 +--
1126 The --numstat option gives the diffstat(1) information but is designed
1128 for easier machine consumption. An entry in --numstat output looks like
1129 this:
1130 1 2 README
1132 3 1 arch/{i386 => x86}/Makefile
1133 That is, from left to right:
1135 1. the number of added lines;
1137 2. a tab;
1139 3. the number of deleted lines;
1141 4. a tab;
1143 5. pathname (possibly with rename/copy information);
1145 6. a newline.
1147 When -z output option is in effect, the output is formatted this way:
1149 1 2 README NUL
1151 3 1 NUL arch/i386/Makefile NUL arch/x86/Makefile NUL
1152 That is:
1154 1. the number of added lines;
1156 2. a tab;
1158 3. the number of deleted lines;
1160 4. a tab;
1162 5. a NUL (only exists if renamed/copied);
1164 6. pathname in preimage;
1166 7. a NUL (only exists if renamed/copied);
1168 8. pathname in postimage (only exists if renamed/copied);
1170 9. a NUL.
1172 The extra NUL before the preimage path in renamed case is to allow
1174 scripts that read the output to tell if the current record being read
1175 is a single-path record or a rename/copy record without reading ahead.
1176 After reading added and deleted lines, reading up to NUL would yield
1177 the pathname, but if that is NUL, the record will show two paths.
1178
1180 Various ways to check your working tree
1181 $ git diff [1m(1)
1183 $ git diff --cached [1m(2)
1184 $ git diff HEAD [1m(3)
1185 1. Changes in the working tree not yet staged for the next
1187 commit.
1188 2. Changes between the index and your last commit; what you
1190 would be committing if you run git commit without -a
1191 option.
1192 3. Changes in the working tree since your last commit; what
1193 you would be committing if you run git commit -a
1194 Comparing with arbitrary commits
1196 $ git diff test [1m(1)
1198 $ git diff HEAD -- ./test [1m(2)
1199 $ git diff HEAD^ HEAD [1m(3)
1200 1. Instead of using the tip of the current branch, compare
1202 with the tip of "test" branch.
1203 2. Instead of comparing with the tip of "test" branch,
1204 compare with the tip of the current branch, but limit the
1205 comparison to the file "test".
1206 3. Compare the version before the last commit and the last
1207 commit.
1208 Comparing branches
1210 $ git diff topic master [1m(1)
1212 $ git diff topic..master [1m(2)
1213 $ git diff topic...master [1m(3)
1214 1. Changes between the tips of the topic and the master
1216 branches.
1217 2. Same as above.
1218 3. Changes that occurred on the master branch since when the
1219 topic branch was started off it.
1220 Limiting the diff output
1222 $ git diff --diff-filter=MRC [1m(1)
1224 $ git diff --name-status [1m(2)
1225 $ git diff arch/i386 include/asm-i386 [1m(3)
1226 1. Show only modification, rename, and copy, but not addition
1228 or deletion.
1229 2. Show only names and the nature of change, but not actual
1230 diff output.
1231 3. Limit diff output to named subtrees.
1232 Munging the diff output
1234 $ git diff --find-copies-harder -B -C [1m(1)
1236 $ git diff -R [1m(2)
1237 1. Spend extra cycles to find renames, copies and complete
1239 rewrites (very expensive).
1240 2. Output diff in reverse.
1241
1243 diff(1), git-difftool(1), git-log(1), gitdiffcore(7), git-format-
1244 patch(1), git-apply(1), git-show(1)
1245
1247 Part of the git(1) suite
1248Git 2.33.1 2021-10-12 GIT-DIFF(1)