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 [<commit>] [--] [<path>...]
11 git diff [<options>] <commit> <commit> [--] [<path>...]
12 git diff [<options>] <blob> <blob>
13 git diff [<options>] --no-index [--] <path> <path>
14
16 Show changes between the working tree and the index or a tree, changes
17 between the index and a tree, changes between two trees, changes
18 between two blob objects, or changes between two files on disk.
19 git diff [<options>] [--] [<path>...]
21 This form is to view the changes you made relative to the index
22 (staging area for the next commit). In other words, the differences
23 are what you could tell Git to further add to the index but you
24 still haven’t. You can stage these changes by using git-add(1).
25 git diff [<options>] --no-index [--] <path> <path>
27 This form is to compare the given two paths on the filesystem. You
28 can omit the --no-index option when running the command in a
29 working tree controlled by Git and at least one of the paths points
30 outside the working tree, or when running the command outside a
31 working tree controlled by Git. This form implies --exit-code.
32 git diff [<options>] --cached [<commit>] [--] [<path>...]
34 This form is to view the changes you staged for the next commit
35 relative to the named <commit>. Typically you would want comparison
36 with the latest commit, so if you do not give <commit>, it defaults
37 to HEAD. If HEAD does not exist (e.g. unborn branches) and <commit>
38 is not given, it shows all staged changes. --staged is a synonym of
39 --cached.
40 git diff [<options>] <commit> [--] [<path>...]
42 This form is to view the changes you have in your working tree
43 relative to the named <commit>. You can use HEAD to compare it with
44 the latest commit, or a branch name to compare with the tip of a
45 different branch.
46 git diff [<options>] <commit> <commit> [--] [<path>...]
48 This is to view the changes between two arbitrary <commit>.
49 git diff [<options>] <commit>..<commit> [--] [<path>...]
51 This is synonymous to the previous form. If <commit> on one side is
52 omitted, it will have the same effect as using HEAD instead.
53 git diff [<options>] <commit>...<commit> [--] [<path>...]
55 This form is to view the changes on the branch containing and up to
56 the second <commit>, starting at a common ancestor of both
57 <commit>. "git diff A...B" is equivalent to "git diff $(git
58 merge-base A B) B". You can omit any one of <commit>, which has the
59 same effect as using HEAD instead.
60 Just in case you are doing something exotic, it should be noted that
62 all of the <commit> in the above description, except in the last two
63 forms that use ".." notations, can be any <tree>.
64 For a more complete list of ways to spell <commit>, see "SPECIFYING
66 REVISIONS" section in gitrevisions(7). However, "diff" is about
67 comparing two endpoints, not ranges, and the range notations
68 ("<commit>..<commit>" and "<commit>...<commit>") do not mean a range as
69 defined in the "SPECIFYING RANGES" section in gitrevisions(7).
70 git diff [<options>] <blob> <blob>
72 This form is to view the differences between the raw contents of
73 two blob objects.
74
76 -p, -u, --patch
77 Generate patch (see section on generating patches). This is the
78 default.
79 -s, --no-patch
81 Suppress diff output. Useful for commands like git show that show
82 the patch by default, or to cancel the effect of --patch.
83 -U<n>, --unified=<n>
85 Generate diffs with <n> lines of context instead of the usual
86 three. Implies --patch. Implies -p.
87 --output=<file>
89 Output to a specific file instead of stdout.
90 --output-indicator-new=<char>, --output-indicator-old=<char>,
92 --output-indicator-context=<char>
93 Specify the character used to indicate new, old or context lines in
94 the generated patch. Normally they are +, - and ' ' respectively.
95 --raw
97 Generate the diff in raw format.
98 --patch-with-raw
100 Synonym for -p --raw.
101 --indent-heuristic
103 Enable the heuristic that shifts diff hunk boundaries to make
104 patches easier to read. This is the default.
105 --no-indent-heuristic
107 Disable the indent heuristic.
108 --minimal
110 Spend extra time to make sure the smallest possible diff is
111 produced.
112 --patience
114 Generate a diff using the "patience diff" algorithm.
115 --histogram
117 Generate a diff using the "histogram diff" algorithm.
118 --anchored=<text>
120 Generate a diff using the "anchored diff" algorithm.
121 This option may be specified more than once.
123 If a line exists in both the source and destination, exists only
125 once, and starts with this text, this algorithm attempts to prevent
126 it from appearing as a deletion or addition in the output. It uses
127 the "patience diff" algorithm internally.
128 --diff-algorithm={patience|minimal|histogram|myers}
130 Choose a diff algorithm. The variants are as follows:
131 default, myers
133 The basic greedy diff algorithm. Currently, this is the
134 default.
135 minimal
137 Spend extra time to make sure the smallest possible diff is
138 produced.
139 patience
141 Use "patience diff" algorithm when generating patches.
142 histogram
144 This algorithm extends the patience algorithm to "support
145 low-occurrence common elements".
146 For instance, if you configured the diff.algorithm variable to a
148 non-default value and want to use the default one, then you have to
149 use --diff-algorithm=default option.
150 --stat[=<width>[,<name-width>[,<count>]]]
152 Generate a diffstat. By default, as much space as necessary will be
153 used for the filename part, and the rest for the graph part.
154 Maximum width defaults to terminal width, or 80 columns if not
155 connected to a terminal, and can be overridden by <width>. The
156 width of the filename part can be limited by giving another width
157 <name-width> after a comma. The width of the graph part can be
158 limited by using --stat-graph-width=<width> (affects all commands
159 generating a stat graph) or by setting diff.statGraphWidth=<width>
160 (does not affect git format-patch). By giving a third parameter
161 <count>, you can limit the output to the first <count> lines,
162 followed by ... if there are more.
163 These parameters can also be set individually with
165 --stat-width=<width>, --stat-name-width=<name-width> and
166 --stat-count=<count>.
167 --compact-summary
169 Output a condensed summary of extended header information such as
170 file creations or deletions ("new" or "gone", optionally "+l" if
171 it’s a symlink) and mode changes ("+x" or "-x" for adding or
172 removing executable bit respectively) in diffstat. The information
173 is put between the filename part and the graph part. Implies
174 --stat.
175 --numstat
177 Similar to --stat, but shows number of added and deleted lines in
178 decimal notation and pathname without abbreviation, to make it more
179 machine friendly. For binary files, outputs two - instead of saying
180 0 0.
181 --shortstat
183 Output only the last line of the --stat format containing total
184 number of modified files, as well as number of added and deleted
185 lines.
186 -X[<param1,param2,...>], --dirstat[=<param1,param2,...>]
188 Output the distribution of relative amount of changes for each
189 sub-directory. The behavior of --dirstat can be customized by
190 passing it a comma separated list of parameters. The defaults are
191 controlled by the diff.dirstat configuration variable (see git-
192 config(1)). The following parameters are available:
193 changes
195 Compute the dirstat numbers by counting the lines that have
196 been removed from the source, or added to the destination. This
197 ignores the amount of pure code movements within a file. In
198 other words, rearranging lines in a file is not counted as much
199 as other changes. This is the default behavior when no
200 parameter is given.
201 lines
203 Compute the dirstat numbers by doing the regular line-based
204 diff analysis, and summing the removed/added line counts. (For
205 binary files, count 64-byte chunks instead, since binary files
206 have no natural concept of lines). This is a more expensive
207 --dirstat behavior than the changes behavior, but it does count
208 rearranged lines within a file as much as other changes. The
209 resulting output is consistent with what you get from the other
210 --*stat options.
211 files
213 Compute the dirstat numbers by counting the number of files
214 changed. Each changed file counts equally in the dirstat
215 analysis. This is the computationally cheapest --dirstat
216 behavior, since it does not have to look at the file contents
217 at all.
218 cumulative
220 Count changes in a child directory for the parent directory as
221 well. Note that when using cumulative, the sum of the
222 percentages reported may exceed 100%. The default
223 (non-cumulative) behavior can be specified with the
224 noncumulative parameter.
225 <limit>
227 An integer parameter specifies a cut-off percent (3% by
228 default). Directories contributing less than this percentage of
229 the changes are not shown in the output.
230 Example: The following will count changed files, while ignoring
232 directories with less than 10% of the total amount of changed
233 files, and accumulating child directory counts in the parent
234 directories: --dirstat=files,10,cumulative.
235 --cumulative
237 Synonym for --dirstat=cumulative
238 --dirstat-by-file[=<param1,param2>...]
240 Synonym for --dirstat=files,param1,param2...
241 --summary
243 Output a condensed summary of extended header information such as
244 creations, renames and mode changes.
245 --patch-with-stat
247 Synonym for -p --stat.
248 -z
250 When --raw, --numstat, --name-only or --name-status has been given,
251 do not munge pathnames and use NULs as output field terminators.
252 Without this option, pathnames with "unusual" characters are quoted
254 as explained for the configuration variable core.quotePath (see
255 git-config(1)).
256 --name-only
258 Show only names of changed files.
259 --name-status
261 Show only names and status of changed files. See the description of
262 the --diff-filter option on what the status letters mean.
263 --submodule[=<format>]
265 Specify how differences in submodules are shown. When specifying
266 --submodule=short the short format is used. This format just shows
267 the names of the commits at the beginning and end of the range.
268 When --submodule or --submodule=log is specified, the log format is
269 used. This format lists the commits in the range like git-
270 submodule(1) summary does. When --submodule=diff is specified, the
271 diff format is used. This format shows an inline diff of the
272 changes in the submodule contents between the commit range.
273 Defaults to diff.submodule or the short format if the config option
274 is unset.
275 --color[=<when>]
277 Show colored diff. --color (i.e. without =<when>) is the same as
278 --color=always. <when> can be one of always, never, or auto. It
279 can be changed by the color.ui and color.diff configuration
280 settings.
281 --no-color
283 Turn off colored diff. This can be used to override configuration
284 settings. It is the same as --color=never.
285 --color-moved[=<mode>]
287 Moved lines of code are colored differently. It can be changed by
288 the diff.colorMoved configuration setting. The <mode> defaults to
289 no if the option is not given and to zebra if the option with no
290 mode is given. The mode must be one of:
291 no
293 Moved lines are not highlighted.
294 default
296 Is a synonym for zebra. This may change to a more sensible mode
297 in the future.
298 plain
300 Any line that is added in one location and was removed in
301 another location will be colored with color.diff.newMoved.
302 Similarly color.diff.oldMoved will be used for removed lines
303 that are added somewhere else in the diff. This mode picks up
304 any moved line, but it is not very useful in a review to
305 determine if a block of code was moved without permutation.
306 blocks
308 Blocks of moved text of at least 20 alphanumeric characters are
309 detected greedily. The detected blocks are painted using either
310 the color.diff.{old,new}Moved color. Adjacent blocks cannot be
311 told apart.
312 zebra
314 Blocks of moved text are detected as in blocks mode. The blocks
315 are painted using either the color.diff.{old,new}Moved color or
316 color.diff.{old,new}MovedAlternative. The change between the
317 two colors indicates that a new block was detected.
318 dimmed-zebra
320 Similar to zebra, but additional dimming of uninteresting parts
321 of moved code is performed. The bordering lines of two adjacent
322 blocks are considered interesting, the rest is uninteresting.
323 dimmed_zebra is a deprecated synonym.
324 --no-color-moved
326 Turn off move detection. This can be used to override configuration
327 settings. It is the same as --color-moved=no.
328 --color-moved-ws=<modes>
330 This configures how whitespace is ignored when performing the move
331 detection for --color-moved. It can be set by the diff.colorMovedWS
332 configuration setting. These modes can be given as a comma
333 separated list:
334 no
336 Do not ignore whitespace when performing move detection.
337 ignore-space-at-eol
339 Ignore changes in whitespace at EOL.
340 ignore-space-change
342 Ignore changes in amount of whitespace. This ignores whitespace
343 at line end, and considers all other sequences of one or more
344 whitespace characters to be equivalent.
345 ignore-all-space
347 Ignore whitespace when comparing lines. This ignores
348 differences even if one line has whitespace where the other
349 line has none.
350 allow-indentation-change
352 Initially ignore any whitespace in the move detection, then
353 group the moved code blocks only into a block if the change in
354 whitespace is the same per line. This is incompatible with the
355 other modes.
356 --no-color-moved-ws
358 Do not ignore whitespace when performing move detection. This can
359 be used to override configuration settings. It is the same as
360 --color-moved-ws=no.
361 --word-diff[=<mode>]
363 Show a word diff, using the <mode> to delimit changed words. By
364 default, words are delimited by whitespace; see --word-diff-regex
365 below. The <mode> defaults to plain, and must be one of:
366 color
368 Highlight changed words using only colors. Implies --color.
369 plain
371 Show words as [-removed-] and {+added+}. Makes no attempts to
372 escape the delimiters if they appear in the input, so the
373 output may be ambiguous.
374 porcelain
376 Use a special line-based format intended for script
377 consumption. Added/removed/unchanged runs are printed in the
378 usual unified diff format, starting with a +/-/` ` character at
379 the beginning of the line and extending to the end of the line.
380 Newlines in the input are represented by a tilde ~ on a line of
381 its own.
382 none
384 Disable word diff again.
385 Note that despite the name of the first mode, color is used to
387 highlight the changed parts in all modes if enabled.
388 --word-diff-regex=<regex>
390 Use <regex> to decide what a word is, instead of considering runs
391 of non-whitespace to be a word. Also implies --word-diff unless it
392 was already enabled.
393 Every non-overlapping match of the <regex> is considered a word.
395 Anything between these matches is considered whitespace and
396 ignored(!) for the purposes of finding differences. You may want to
397 append |[^[:space:]] to your regular expression to make sure that
398 it matches all non-whitespace characters. A match that contains a
399 newline is silently truncated(!) at the newline.
400 For example, --word-diff-regex=. will treat each character as a
402 word and, correspondingly, show differences character by character.
403 The regex can also be set via a diff driver or configuration
405 option, see gitattributes(5) or git-config(1). Giving it explicitly
406 overrides any diff driver or configuration setting. Diff drivers
407 override configuration settings.
408 --color-words[=<regex>]
410 Equivalent to --word-diff=color plus (if a regex was specified)
411 --word-diff-regex=<regex>.
412 --no-renames
414 Turn off rename detection, even when the configuration file gives
415 the default to do so.
416 --[no-]rename-empty
418 Whether to use empty blobs as rename source.
419 --check
421 Warn if changes introduce conflict markers or whitespace errors.
422 What are considered whitespace errors is controlled by
423 core.whitespace configuration. By default, trailing whitespaces
424 (including lines that consist solely of whitespaces) and a space
425 character that is immediately followed by a tab character inside
426 the initial indent of the line are considered whitespace errors.
427 Exits with non-zero status if problems are found. Not compatible
428 with --exit-code.
429 --ws-error-highlight=<kind>
431 Highlight whitespace errors in the context, old or new lines of the
432 diff. Multiple values are separated by comma, none resets previous
433 values, default reset the list to new and all is a shorthand for
434 old,new,context. When this option is not given, and the
435 configuration variable diff.wsErrorHighlight is not set, only
436 whitespace errors in new lines are highlighted. The whitespace
437 errors are colored with color.diff.whitespace.
438 --full-index
440 Instead of the first handful of characters, show the full pre- and
441 post-image blob object names on the "index" line when generating
442 patch format output.
443 --binary
445 In addition to --full-index, output a binary diff that can be
446 applied with git-apply. Implies --patch.
447 --abbrev[=<n>]
449 Instead of showing the full 40-byte hexadecimal object name in
450 diff-raw format output and diff-tree header lines, show only a
451 partial prefix. This is independent of the --full-index option
452 above, which controls the diff-patch output format. Non default
453 number of digits can be specified with --abbrev=<n>.
454 -B[<n>][/<m>], --break-rewrites[=[<n>][/<m>]]
456 Break complete rewrite changes into pairs of delete and create.
457 This serves two purposes:
458 It affects the way a change that amounts to a total rewrite of a
460 file not as a series of deletion and insertion mixed together with
461 a very few lines that happen to match textually as the context, but
462 as a single deletion of everything old followed by a single
463 insertion of everything new, and the number m controls this aspect
464 of the -B option (defaults to 60%). -B/70% specifies that less
465 than 30% of the original should remain in the result for Git to
466 consider it a total rewrite (i.e. otherwise the resulting patch
467 will be a series of deletion and insertion mixed together with
468 context lines).
469 When used with -M, a totally-rewritten file is also considered as
471 the source of a rename (usually -M only considers a file that
472 disappeared as the source of a rename), and the number n controls
473 this aspect of the -B option (defaults to 50%). -B20% specifies
474 that a change with addition and deletion compared to 20% or more of
475 the file’s size are eligible for being picked up as a possible
476 source of a rename to another file.
477 -M[<n>], --find-renames[=<n>]
479 Detect renames. If n is specified, it is a threshold on the
480 similarity index (i.e. amount of addition/deletions compared to the
481 file’s size). For example, -M90% means Git should consider a
482 delete/add pair to be a rename if more than 90% of the file hasn’t
483 changed. Without a % sign, the number is to be read as a fraction,
484 with a decimal point before it. I.e., -M5 becomes 0.5, and is thus
485 the same as -M50%. Similarly, -M05 is the same as -M5%. To limit
486 detection to exact renames, use -M100%. The default similarity
487 index is 50%.
488 -C[<n>], --find-copies[=<n>]
490 Detect copies as well as renames. See also --find-copies-harder. If
491 n is specified, it has the same meaning as for -M<n>.
492 --find-copies-harder
494 For performance reasons, by default, -C option finds copies only if
495 the original file of the copy was modified in the same changeset.
496 This flag makes the command inspect unmodified files as candidates
497 for the source of copy. This is a very expensive operation for
498 large projects, so use it with caution. Giving more than one -C
499 option has the same effect.
500 -D, --irreversible-delete
502 Omit the preimage for deletes, i.e. print only the header but not
503 the diff between the preimage and /dev/null. The resulting patch is
504 not meant to be applied with patch or git apply; this is solely for
505 people who want to just concentrate on reviewing the text after the
506 change. In addition, the output obviously lacks enough information
507 to apply such a patch in reverse, even manually, hence the name of
508 the option.
509 When used together with -B, omit also the preimage in the deletion
511 part of a delete/create pair.
512 -l<num>
514 The -M and -C options require O(n^2) processing time where n is the
515 number of potential rename/copy targets. This option prevents
516 rename/copy detection from running if the number of rename/copy
517 targets exceeds the specified number.
518 --diff-filter=[(A|C|D|M|R|T|U|X|B)...[*]]
520 Select only files that are Added (A), Copied (C), Deleted (D),
521 Modified (M), Renamed (R), have their type (i.e. regular file,
522 symlink, submodule, ...) changed (T), are Unmerged (U), are Unknown
523 (X), or have had their pairing Broken (B). Any combination of the
524 filter characters (including none) can be used. When *
525 (All-or-none) is added to the combination, all paths are selected
526 if there is any file that matches other criteria in the comparison;
527 if there is no file that matches other criteria, nothing is
528 selected.
529 Also, these upper-case letters can be downcased to exclude. E.g.
531 --diff-filter=ad excludes added and deleted paths.
532 Note that not all diffs can feature all types. For instance, diffs
534 from the index to the working tree can never have Added entries
535 (because the set of paths included in the diff is limited by what
536 is in the index). Similarly, copied and renamed entries cannot
537 appear if detection for those types is disabled.
538 -S<string>
540 Look for differences that change the number of occurrences of the
541 specified string (i.e. addition/deletion) in a file. Intended for
542 the scripter’s use.
543 It is useful when you’re looking for an exact block of code (like a
545 struct), and want to know the history of that block since it first
546 came into being: use the feature iteratively to feed the
547 interesting block in the preimage back into -S, and keep going
548 until you get the very first version of the block.
549 Binary files are searched as well.
551 -G<regex>
553 Look for differences whose patch text contains added/removed lines
554 that match <regex>.
555 To illustrate the difference between -S<regex> --pickaxe-regex and
557 -G<regex>, consider a commit with the following diff in the same
558 file:
559 + return frotz(nitfol, two->ptr, 1, 0);
561 ...
562 - hit = frotz(nitfol, mf2.ptr, 1, 0);
563 While git log -G"frotz\(nitfol" will show this commit, git log
565 -S"frotz\(nitfol" --pickaxe-regex will not (because the number of
566 occurrences of that string did not change).
567 Unless --text is supplied patches of binary files without a
569 textconv filter will be ignored.
570 See the pickaxe entry in gitdiffcore(7) for more information.
572 --find-object=<object-id>
574 Look for differences that change the number of occurrences of the
575 specified object. Similar to -S, just the argument is different in
576 that it doesn’t search for a specific string but for a specific
577 object id.
578 The object can be a blob or a submodule commit. It implies the -t
580 option in git-log to also find trees.
581 --pickaxe-all
583 When -S or -G finds a change, show all the changes in that
584 changeset, not just the files that contain the change in <string>.
585 --pickaxe-regex
587 Treat the <string> given to -S as an extended POSIX regular
588 expression to match.
589 -O<orderfile>
591 Control the order in which files appear in the output. This
592 overrides the diff.orderFile configuration variable (see git-
593 config(1)). To cancel diff.orderFile, use -O/dev/null.
594 The output order is determined by the order of glob patterns in
596 <orderfile>. All files with pathnames that match the first pattern
597 are output first, all files with pathnames that match the second
598 pattern (but not the first) are output next, and so on. All files
599 with pathnames that do not match any pattern are output last, as if
600 there was an implicit match-all pattern at the end of the file. If
601 multiple pathnames have the same rank (they match the same pattern
602 but no earlier patterns), their output order relative to each other
603 is the normal order.
604 <orderfile> is parsed as follows:
606 · Blank lines are ignored, so they can be used as separators for
608 readability.
609 · Lines starting with a hash ("#") are ignored, so they can be
611 used for comments. Add a backslash ("\") to the beginning of
612 the pattern if it starts with a hash.
613 · Each other line contains a single pattern.
615 Patterns have the same syntax and semantics as patterns used for
617 fnmatch(3) without the FNM_PATHNAME flag, except a pathname also
618 matches a pattern if removing any number of the final pathname
619 components matches the pattern. For example, the pattern "foo*bar"
620 matches "fooasdfbar" and "foo/bar/baz/asdf" but not "foobarx".
621 -R
623 Swap two inputs; that is, show differences from index or on-disk
624 file to tree contents.
625 --relative[=<path>]
627 When run from a subdirectory of the project, it can be told to
628 exclude changes outside the directory and show pathnames relative
629 to it with this option. When you are not in a subdirectory (e.g. in
630 a bare repository), you can name which subdirectory to make the
631 output relative to by giving a <path> as an argument.
632 -a, --text
634 Treat all files as text.
635 --ignore-cr-at-eol
637 Ignore carriage-return at the end of line when doing a comparison.
638 --ignore-space-at-eol
640 Ignore changes in whitespace at EOL.
641 -b, --ignore-space-change
643 Ignore changes in amount of whitespace. This ignores whitespace at
644 line end, and considers all other sequences of one or more
645 whitespace characters to be equivalent.
646 -w, --ignore-all-space
648 Ignore whitespace when comparing lines. This ignores differences
649 even if one line has whitespace where the other line has none.
650 --ignore-blank-lines
652 Ignore changes whose lines are all blank.
653 --inter-hunk-context=<lines>
655 Show the context between diff hunks, up to the specified number of
656 lines, thereby fusing hunks that are close to each other. Defaults
657 to diff.interHunkContext or 0 if the config option is unset.
658 -W, --function-context
660 Show whole surrounding functions of changes.
661 --exit-code
663 Make the program exit with codes similar to diff(1). That is, it
664 exits with 1 if there were differences and 0 means no differences.
665 --quiet
667 Disable all output of the program. Implies --exit-code.
668 --ext-diff
670 Allow an external diff helper to be executed. If you set an
671 external diff driver with gitattributes(5), you need to use this
672 option with git-log(1) and friends.
673 --no-ext-diff
675 Disallow external diff drivers.
676 --textconv, --no-textconv
678 Allow (or disallow) external text conversion filters to be run when
679 comparing binary files. See gitattributes(5) for details. Because
680 textconv filters are typically a one-way conversion, the resulting
681 diff is suitable for human consumption, but cannot be applied. For
682 this reason, textconv filters are enabled by default only for git-
683 diff(1) and git-log(1), but not for git-format-patch(1) or diff
684 plumbing commands.
685 --ignore-submodules[=<when>]
687 Ignore changes to submodules in the diff generation. <when> can be
688 either "none", "untracked", "dirty" or "all", which is the default.
689 Using "none" will consider the submodule modified when it either
690 contains untracked or modified files or its HEAD differs from the
691 commit recorded in the superproject and can be used to override any
692 settings of the ignore option in git-config(1) or gitmodules(5).
693 When "untracked" is used submodules are not considered dirty when
694 they only contain untracked content (but they are still scanned for
695 modified content). Using "dirty" ignores all changes to the work
696 tree of submodules, only changes to the commits stored in the
697 superproject are shown (this was the behavior until 1.7.0). Using
698 "all" hides all changes to submodules.
699 --src-prefix=<prefix>
701 Show the given source prefix instead of "a/".
702 --dst-prefix=<prefix>
704 Show the given destination prefix instead of "b/".
705 --no-prefix
707 Do not show any source or destination prefix.
708 --line-prefix=<prefix>
710 Prepend an additional prefix to every line of output.
711 --ita-invisible-in-index
713 By default entries added by "git add -N" appear as an existing
714 empty file in "git diff" and a new file in "git diff --cached".
715 This option makes the entry appear as a new file in "git diff" and
716 non-existent in "git diff --cached". This option could be reverted
717 with --ita-visible-in-index. Both options are experimental and
718 could be removed in future.
719 For more detailed explanation on these common options, see also
721 gitdiffcore(7).
722 -1 --base, -2 --ours, -3 --theirs
724 Compare the working tree with the "base" version (stage #1), "our
725 branch" (stage #2) or "their branch" (stage #3). The index contains
726 these stages only for unmerged entries i.e. while resolving
727 conflicts. See git-read-tree(1) section "3-Way Merge" for detailed
728 information.
729 -0
731 Omit diff output for unmerged entries and just show "Unmerged". Can
732 be used only when comparing the working tree with the index.
733 <path>...
735 The <paths> parameters, when given, are used to limit the diff to
736 the named paths (you can give directory names and get diff for all
737 files under them).
738
740 The raw output format from "git-diff-index", "git-diff-tree",
741 "git-diff-files" and "git diff --raw" are very similar.
742 These commands all compare two sets of things; what is compared
744 differs:
745 git-diff-index <tree-ish>
747 compares the <tree-ish> and the files on the filesystem.
748 git-diff-index --cached <tree-ish>
750 compares the <tree-ish> and the index.
751 git-diff-tree [-r] <tree-ish-1> <tree-ish-2> [<pattern>...]
753 compares the trees named by the two arguments.
754 git-diff-files [<pattern>...]
756 compares the index and the files on the filesystem.
757 The "git-diff-tree" command begins its output by printing the hash of
759 what is being compared. After that, all the commands print one output
760 line per changed file.
761 An output line is formatted this way:
763 in-place edit :100644 100644 bcd1234 0123456 M file0
765 copy-edit :100644 100644 abcd123 1234567 C68 file1 file2
766 rename-edit :100644 100644 abcd123 1234567 R86 file1 file3
767 create :000000 100644 0000000 1234567 A file4
768 delete :100644 000000 1234567 0000000 D file5
769 unmerged :000000 000000 0000000 0000000 U file6
770 That is, from the left to the right:
772 1. a colon.
774 2. mode for "src"; 000000 if creation or unmerged.
776 3. a space.
778 4. mode for "dst"; 000000 if deletion or unmerged.
780 5. a space.
782 6. sha1 for "src"; 0{40} if creation or unmerged.
784 7. a space.
786 8. sha1 for "dst"; 0{40} if creation, unmerged or "look at work tree".
788 9. a space.
790 10. status, followed by optional "score" number.
792 11. a tab or a NUL when -z option is used.
794 12. path for "src"
796 13. a tab or a NUL when -z option is used; only exists for C or R.
798 14. path for "dst"; only exists for C or R.
800 15. an LF or a NUL when -z option is used, to terminate the record.
802 Possible status letters are:
804 · A: addition of a file
806 · C: copy of a file into a new one
808 · D: deletion of a file
810 · M: modification of the contents or mode of a file
812 · R: renaming of a file
814 · T: change in the type of the file
816 · U: file is unmerged (you must complete the merge before it can be
818 committed)
819 · X: "unknown" change type (most probably a bug, please report it)
821 Status letters C and R are always followed by a score (denoting the
823 percentage of similarity between the source and target of the move or
824 copy). Status letter M may be followed by a score (denoting the
825 percentage of dissimilarity) for file rewrites.
826 <sha1> is shown as all 0’s if a file is new on the filesystem and it is
828 out of sync with the index.
829 Example:
831 :100644 100644 5be4a4a 0000000 M file.c
833 Without the -z option, pathnames with "unusual" characters are quoted
835 as explained for the configuration variable core.quotePath (see git-
836 config(1)). Using -z the filename is output verbatim and the line is
837 terminated by a NUL byte.
838
840 "git-diff-tree", "git-diff-files" and "git-diff --raw" can take -c or
841 --cc option to generate diff output also for merge commits. The output
842 differs from the format described above in the following way:
843 1. there is a colon for each parent
845 2. there are more "src" modes and "src" sha1
847 3. status is concatenated status characters for each parent
849 4. no optional "score" number
851 5. tab-separated pathname(s) of the file
853 For -c and --cc, only the destination or final path is shown even if
855 the file was renamed on any side of history. With --combined-all-paths,
856 the name of the path in each parent is shown followed by the name of
857 the path in the merge commit.
858 Examples for -c and --cc without --combined-all-paths:
860 ::100644 100644 100644 fabadb8 cc95eb0 4866510 MM desc.c
862 ::100755 100755 100755 52b7a2d 6d1ac04 d2ac7d7 RM bar.sh
863 ::100644 100644 100644 e07d6c5 9042e82 ee91881 RR phooey.c
864 Examples when --combined-all-paths added to either -c or --cc:
866 ::100644 100644 100644 fabadb8 cc95eb0 4866510 MM desc.c desc.c desc.c
868 ::100755 100755 100755 52b7a2d 6d1ac04 d2ac7d7 RM foo.sh bar.sh bar.sh
869 ::100644 100644 100644 e07d6c5 9042e82 ee91881 RR fooey.c fuey.c phooey.c
870 Note that combined diff lists only files which were modified from all
872 parents.
873
875 Running git-diff(1), git-log(1), git-show(1), git-diff-index(1), git-
876 diff-tree(1), or git-diff-files(1) with the -p option produces patch
877 text. You can customize the creation of patch text via the
878 GIT_EXTERNAL_DIFF and the GIT_DIFF_OPTS environment variables.
879 What the -p option produces is slightly different from the traditional
881 diff format:
882 1. It is preceded with a "git diff" header that looks like this:
884 diff --git a/file1 b/file2
886 The a/ and b/ filenames are the same unless rename/copy is
888 involved. Especially, even for a creation or a deletion, /dev/null
889 is not used in place of the a/ or b/ filenames.
890 When rename/copy is involved, file1 and file2 show the name of the
892 source file of the rename/copy and the name of the file that
893 rename/copy produces, respectively.
894 2. It is followed by one or more extended header lines:
896 old mode <mode>
898 new mode <mode>
899 deleted file mode <mode>
900 new file mode <mode>
901 copy from <path>
902 copy to <path>
903 rename from <path>
904 rename to <path>
905 similarity index <number>
906 dissimilarity index <number>
907 index <hash>..<hash> <mode>
908 File modes are printed as 6-digit octal numbers including the file
910 type and file permission bits.
911 Path names in extended headers do not include the a/ and b/
913 prefixes.
914 The similarity index is the percentage of unchanged lines, and the
916 dissimilarity index is the percentage of changed lines. It is a
917 rounded down integer, followed by a percent sign. The similarity
918 index value of 100% is thus reserved for two equal files, while
919 100% dissimilarity means that no line from the old file made it
920 into the new one.
921 The index line includes the blob object names before and after the
923 change. The <mode> is included if the file mode does not change;
924 otherwise, separate lines indicate the old and the new mode.
925 3. Pathnames with "unusual" characters are quoted as explained for the
927 configuration variable core.quotePath (see git-config(1)).
928 4. All the file1 files in the output refer to files before the commit,
930 and all the file2 files refer to files after the commit. It is
931 incorrect to apply each change to each file sequentially. For
932 example, this patch will swap a and b:
933 diff --git a/a b/b
935 rename from a
936 rename to b
937 diff --git a/b b/a
938 rename from b
939 rename to a
940
942 Any diff-generating command can take the -c or --cc option to produce a
943 combined diff when showing a merge. This is the default format when
944 showing merges with git-diff(1) or git-show(1). Note also that you can
945 give the -m option to any of these commands to force generation of
946 diffs with individual parents of a merge.
947 A "combined diff" format looks like this:
949 diff --combined describe.c
951 index fabadb8,cc95eb0..4866510
952 --- a/describe.c
953 +++ b/describe.c
954 @@@ -98,20 -98,12 +98,20 @@@
955 return (a_date > b_date) ? -1 : (a_date == b_date) ? 0 : 1;
956 }
957 - static void describe(char *arg)
959 -static void describe(struct commit *cmit, int last_one)
960 ++static void describe(char *arg, int last_one)
961 {
962 + unsigned char sha1[20];
963 + struct commit *cmit;
964 struct commit_list *list;
965 static int initialized = 0;
966 struct commit_name *n;
967 + if (get_sha1(arg, sha1) < 0)
969 + usage(describe_usage);
970 + cmit = lookup_commit_reference(sha1);
971 + if (!cmit)
972 + usage(describe_usage);
973 +
974 if (!initialized) {
975 initialized = 1;
976 for_each_ref(get_name);
977 1. It is preceded with a "git diff" header, that looks like this (when
979 the -c option is used):
980 diff --combined file
982 or like this (when the --cc option is used):
984 diff --cc file
986 2. It is followed by one or more extended header lines (this example
988 shows a merge with two parents):
989 index <hash>,<hash>..<hash>
991 mode <mode>,<mode>..<mode>
992 new file mode <mode>
993 deleted file mode <mode>,<mode>
994 The mode <mode>,<mode>..<mode> line appears only if at least one of
996 the <mode> is different from the rest. Extended headers with
997 information about detected contents movement (renames and copying
998 detection) are designed to work with diff of two <tree-ish> and are
999 not used by combined diff format.
1000 3. It is followed by two-line from-file/to-file header
1002 --- a/file
1004 +++ b/file
1005 Similar to two-line header for traditional unified diff format,
1007 /dev/null is used to signal created or deleted files.
1008 However, if the --combined-all-paths option is provided, instead of
1010 a two-line from-file/to-file you get a N+1 line from-file/to-file
1011 header, where N is the number of parents in the merge commit
1012 --- a/file
1014 --- a/file
1015 --- a/file
1016 +++ b/file
1017 This extended format can be useful if rename or copy detection is
1019 active, to allow you to see the original name of the file in
1020 different parents.
1021 4. Chunk header format is modified to prevent people from accidentally
1023 feeding it to patch -p1. Combined diff format was created for
1024 review of merge commit changes, and was not meant to be applied.
1025 The change is similar to the change in the extended index header:
1026 @@@ <from-file-range> <from-file-range> <to-file-range> @@@
1028 There are (number of parents + 1) @ characters in the chunk header
1030 for combined diff format.
1031 Unlike the traditional unified diff format, which shows two files A and
1033 B with a single column that has - (minus — appears in A but removed in
1034 B), + (plus — missing in A but added to B), or " " (space — unchanged)
1035 prefix, this format compares two or more files file1, file2,... with
1036 one file X, and shows how X differs from each of fileN. One column for
1037 each of fileN is prepended to the output line to note how X’s line is
1038 different from it.
1039 A - character in the column N means that the line appears in fileN but
1041 it does not appear in the result. A + character in the column N means
1042 that the line appears in the result, and fileN does not have that line
1043 (in other words, the line was added, from the point of view of that
1044 parent).
1045 In the above example output, the function signature was changed from
1047 both files (hence two - removals from both file1 and file2, plus ++ to
1048 mean one line that was added does not appear in either file1 or file2).
1049 Also eight other lines are the same from file1 but do not appear in
1050 file2 (hence prefixed with +).
1051 When shown by git diff-tree -c, it compares the parents of a merge
1053 commit with the merge result (i.e. file1..fileN are the parents). When
1054 shown by git diff-files -c, it compares the two unresolved merge
1055 parents with the working tree file (i.e. file1 is stage 2 aka "our
1056 version", file2 is stage 3 aka "their version").
1057
1059 The --summary option describes newly added, deleted, renamed and copied
1060 files. The --stat option adds diffstat(1) graph to the output. These
1061 options can be combined with other options, such as -p, and are meant
1062 for human consumption.
1063 When showing a change that involves a rename or a copy, --stat output
1065 formats the pathnames compactly by combining common prefix and suffix
1066 of the pathnames. For example, a change that moves arch/i386/Makefile
1067 to arch/x86/Makefile while modifying 4 lines will be shown like this:
1068 arch/{i386 => x86}/Makefile | 4 +--
1070 The --numstat option gives the diffstat(1) information but is designed
1072 for easier machine consumption. An entry in --numstat output looks like
1073 this:
1074 1 2 README
1076 3 1 arch/{i386 => x86}/Makefile
1077 That is, from left to right:
1079 1. the number of added lines;
1081 2. a tab;
1083 3. the number of deleted lines;
1085 4. a tab;
1087 5. pathname (possibly with rename/copy information);
1089 6. a newline.
1091 When -z output option is in effect, the output is formatted this way:
1093 1 2 README NUL
1095 3 1 NUL arch/i386/Makefile NUL arch/x86/Makefile NUL
1096 That is:
1098 1. the number of added lines;
1100 2. a tab;
1102 3. the number of deleted lines;
1104 4. a tab;
1106 5. a NUL (only exists if renamed/copied);
1108 6. pathname in preimage;
1110 7. a NUL (only exists if renamed/copied);
1112 8. pathname in postimage (only exists if renamed/copied);
1114 9. a NUL.
1116 The extra NUL before the preimage path in renamed case is to allow
1118 scripts that read the output to tell if the current record being read
1119 is a single-path record or a rename/copy record without reading ahead.
1120 After reading added and deleted lines, reading up to NUL would yield
1121 the pathname, but if that is NUL, the record will show two paths.
1122
1124 Various ways to check your working tree
1125 $ git diff [1m(1)
1127 $ git diff --cached [1m(2)
1128 $ git diff HEAD [1m(3)
1129 1. Changes in the working tree not yet staged for the next
1131 commit.
1132 2. Changes between the index and your last commit; what you
1133 would be committing if you run "git commit" without "-a"
1134 option.
1135 3. Changes in the working tree since your last commit; what
1136 you would be committing if you run "git commit -a"
1137 Comparing with arbitrary commits
1139 $ git diff test [1m(1)
1141 $ git diff HEAD -- ./test [1m(2)
1142 $ git diff HEAD^ HEAD [1m(3)
1143 1. Instead of using the tip of the current branch, compare
1145 with the tip of "test" branch.
1146 2. Instead of comparing with the tip of "test" branch,
1147 compare with the tip of the current branch, but limit the
1148 comparison to the file "test".
1149 3. Compare the version before the last commit and the last
1150 commit.
1151 Comparing branches
1153 $ git diff topic master [1m(1)
1155 $ git diff topic..master [1m(2)
1156 $ git diff topic...master [1m(3)
1157 1. Changes between the tips of the topic and the master
1159 branches.
1160 2. Same as above.
1161 3. Changes that occurred on the master branch since when the
1162 topic branch was started off it.
1163 Limiting the diff output
1165 $ git diff --diff-filter=MRC [1m(1)
1167 $ git diff --name-status [1m(2)
1168 $ git diff arch/i386 include/asm-i386 [1m(3)
1169 1. Show only modification, rename, and copy, but not addition
1171 or deletion.
1172 2. Show only names and the nature of change, but not actual
1173 diff output.
1174 3. Limit diff output to named subtrees.
1175 Munging the diff output
1177 $ git diff --find-copies-harder -B -C [1m(1)
1179 $ git diff -R [1m(2)
1180 1. Spend extra cycles to find renames, copies and complete
1182 rewrites (very expensive).
1183 2. Output diff in reverse.
1184
1186 diff(1), git-difftool(1), git-log(1), gitdiffcore(7), git-format-
1187 patch(1), git-apply(1)
1188
1190 Part of the git(1) suite
1191Git 2.26.2 2020-04-20 GIT-DIFF(1)