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
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 between two blob objects, or changes between two files on disk.
20 git diff [<options>] [--] [<path>...]
22 This form is to view the changes you made relative to the index
23 (staging area for the next commit). In other words, the differences
24 are what you could tell Git to further add to the index but you
25 still haven’t. You can stage these changes by using git-add(1).
26 git diff [<options>] --no-index [--] <path> <path>
28 This form is to compare the given two paths on the filesystem. You
29 can omit the --no-index option when running the command in a
30 working tree controlled by Git and at least one of the paths points
31 outside the working tree, or when running the command outside a
32 working tree controlled by Git.
33 git diff [<options>] --cached [<commit>] [--] [<path>...]
35 This form is to view the changes you staged for the next commit
36 relative to the named <commit>. Typically you would want comparison
37 with the latest commit, so if you do not give <commit>, it defaults
38 to HEAD. If HEAD does not exist (e.g. unborn branches) and <commit>
39 is not given, it shows all staged changes. --staged is a synonym of
40 --cached.
41 git diff [<options>] <commit> [--] [<path>...]
43 This form is to view the changes you have in your working tree
44 relative to the named <commit>. You can use HEAD to compare it with
45 the latest commit, or a branch name to compare with the tip of a
46 different branch.
47 git diff [<options>] <commit> <commit> [--] [<path>...]
49 This is to view the changes between two arbitrary <commit>.
50 git diff [<options>] <commit>..<commit> [--] [<path>...]
52 This is synonymous to the previous form. If <commit> on one side is
53 omitted, it will have the same effect as using HEAD instead.
54 git diff [<options>] <commit>...<commit> [--] [<path>...]
56 This form is to view the changes on the branch containing and up to
57 the second <commit>, starting at a common ancestor of both
58 <commit>. "git diff A...B" is equivalent to "git diff
59 $(git-merge-base A B) B". You can omit any one of <commit>, which
60 has the same effect as using HEAD instead.
61 Just in case if you are doing something exotic, it should be noted that
63 all of the <commit> in the above description, except in the last two
64 forms that use ".." notations, can be any <tree>.
65 For a more complete list of ways to spell <commit>, see "SPECIFYING
67 REVISIONS" section in gitrevisions(7). However, "diff" is about
68 comparing two endpoints, not ranges, and the range notations
69 ("<commit>..<commit>" and "<commit>...<commit>") do not mean a range as
70 defined in the "SPECIFYING RANGES" section in gitrevisions(7).
71 git diff [<options>] <blob> <blob>
73 This form is to view the differences between the raw contents of
74 two blob objects.
75
77 -p, -u, --patch
78 Generate patch (see section on generating patches). This is the
79 default.
80 -s, --no-patch
82 Suppress diff output. Useful for commands like git show that show
83 the patch by default, or to cancel the effect of --patch.
84 -U<n>, --unified=<n>
86 Generate diffs with <n> lines of context instead of the usual
87 three. Implies -p.
88 --raw
90 Generate the diff in raw format.
91 --patch-with-raw
93 Synonym for -p --raw.
94 --indent-heuristic
96 Enable the heuristic that shifts diff hunk boundaries to make
97 patches easier to read. This is the default.
98 --no-indent-heuristic
100 Disable the indent heuristic.
101 --minimal
103 Spend extra time to make sure the smallest possible diff is
104 produced.
105 --patience
107 Generate a diff using the "patience diff" algorithm.
108 --histogram
110 Generate a diff using the "histogram diff" algorithm.
111 --anchored=<text>
113 Generate a diff using the "anchored diff" algorithm.
114 This option may be specified more than once.
116 If a line exists in both the source and destination, exists only
118 once, and starts with this text, this algorithm attempts to prevent
119 it from appearing as a deletion or addition in the output. It uses
120 the "patience diff" algorithm internally.
121 --diff-algorithm={patience|minimal|histogram|myers}
123 Choose a diff algorithm. The variants are as follows:
124 default, myers
126 The basic greedy diff algorithm. Currently, this is the
127 default.
128 minimal
130 Spend extra time to make sure the smallest possible diff is
131 produced.
132 patience
134 Use "patience diff" algorithm when generating patches.
135 histogram
137 This algorithm extends the patience algorithm to "support
138 low-occurrence common elements".
139 For instance, if you configured the diff.algorithm variable to a
141 non-default value and want to use the default one, then you have to
142 use --diff-algorithm=default option.
143 --stat[=<width>[,<name-width>[,<count>]]]
145 Generate a diffstat. By default, as much space as necessary will be
146 used for the filename part, and the rest for the graph part.
147 Maximum width defaults to terminal width, or 80 columns if not
148 connected to a terminal, and can be overridden by <width>. The
149 width of the filename part can be limited by giving another width
150 <name-width> after a comma. The width of the graph part can be
151 limited by using --stat-graph-width=<width> (affects all commands
152 generating a stat graph) or by setting diff.statGraphWidth=<width>
153 (does not affect git format-patch). By giving a third parameter
154 <count>, you can limit the output to the first <count> lines,
155 followed by ... if there are more.
156 These parameters can also be set individually with
158 --stat-width=<width>, --stat-name-width=<name-width> and
159 --stat-count=<count>.
160 --compact-summary
162 Output a condensed summary of extended header information such as
163 file creations or deletions ("new" or "gone", optionally "+l" if
164 it’s a symlink) and mode changes ("+x" or "-x" for adding or
165 removing executable bit respectively) in diffstat. The information
166 is put between the filename part and the graph part. Implies
167 --stat.
168 --numstat
170 Similar to --stat, but shows number of added and deleted lines in
171 decimal notation and pathname without abbreviation, to make it more
172 machine friendly. For binary files, outputs two - instead of saying
173 0 0.
174 --shortstat
176 Output only the last line of the --stat format containing total
177 number of modified files, as well as number of added and deleted
178 lines.
179 --dirstat[=<param1,param2,...>]
181 Output the distribution of relative amount of changes for each
182 sub-directory. The behavior of --dirstat can be customized by
183 passing it a comma separated list of parameters. The defaults are
184 controlled by the diff.dirstat configuration variable (see git-
185 config(1)). The following parameters are available:
186 changes
188 Compute the dirstat numbers by counting the lines that have
189 been removed from the source, or added to the destination. This
190 ignores the amount of pure code movements within a file. In
191 other words, rearranging lines in a file is not counted as much
192 as other changes. This is the default behavior when no
193 parameter is given.
194 lines
196 Compute the dirstat numbers by doing the regular line-based
197 diff analysis, and summing the removed/added line counts. (For
198 binary files, count 64-byte chunks instead, since binary files
199 have no natural concept of lines). This is a more expensive
200 --dirstat behavior than the changes behavior, but it does count
201 rearranged lines within a file as much as other changes. The
202 resulting output is consistent with what you get from the other
203 --*stat options.
204 files
206 Compute the dirstat numbers by counting the number of files
207 changed. Each changed file counts equally in the dirstat
208 analysis. This is the computationally cheapest --dirstat
209 behavior, since it does not have to look at the file contents
210 at all.
211 cumulative
213 Count changes in a child directory for the parent directory as
214 well. Note that when using cumulative, the sum of the
215 percentages reported may exceed 100%. The default
216 (non-cumulative) behavior can be specified with the
217 noncumulative parameter.
218 <limit>
220 An integer parameter specifies a cut-off percent (3% by
221 default). Directories contributing less than this percentage of
222 the changes are not shown in the output.
223 Example: The following will count changed files, while ignoring
225 directories with less than 10% of the total amount of changed
226 files, and accumulating child directory counts in the parent
227 directories: --dirstat=files,10,cumulative.
228 --summary
230 Output a condensed summary of extended header information such as
231 creations, renames and mode changes.
232 --patch-with-stat
234 Synonym for -p --stat.
235 -z
237 When --raw, --numstat, --name-only or --name-status has been given,
238 do not munge pathnames and use NULs as output field terminators.
239 Without this option, pathnames with "unusual" characters are quoted
241 as explained for the configuration variable core.quotePath (see
242 git-config(1)).
243 --name-only
245 Show only names of changed files.
246 --name-status
248 Show only names and status of changed files. See the description of
249 the --diff-filter option on what the status letters mean.
250 --submodule[=<format>]
252 Specify how differences in submodules are shown. When specifying
253 --submodule=short the short format is used. This format just shows
254 the names of the commits at the beginning and end of the range.
255 When --submodule or --submodule=log is specified, the log format is
256 used. This format lists the commits in the range like git-
257 submodule(1) summary does. When --submodule=diff is specified, the
258 diff format is used. This format shows an inline diff of the
259 changes in the submodule contents between the commit range.
260 Defaults to diff.submodule or the short format if the config option
261 is unset.
262 --color[=<when>]
264 Show colored diff. --color (i.e. without =<when>) is the same as
265 --color=always. <when> can be one of always, never, or auto. It
266 can be changed by the color.ui and color.diff configuration
267 settings.
268 --no-color
270 Turn off colored diff. This can be used to override configuration
271 settings. It is the same as --color=never.
272 --color-moved[=<mode>]
274 Moved lines of code are colored differently. It can be changed by
275 the diff.colorMoved configuration setting. The <mode> defaults to
276 no if the option is not given and to zebra if the option with no
277 mode is given. The mode must be one of:
278 no
280 Moved lines are not highlighted.
281 default
283 Is a synonym for zebra. This may change to a more sensible mode
284 in the future.
285 plain
287 Any line that is added in one location and was removed in
288 another location will be colored with color.diff.newMoved.
289 Similarly color.diff.oldMoved will be used for removed lines
290 that are added somewhere else in the diff. This mode picks up
291 any moved line, but it is not very useful in a review to
292 determine if a block of code was moved without permutation.
293 zebra
295 Blocks of moved text of at least 20 alphanumeric characters are
296 detected greedily. The detected blocks are painted using either
297 the color.diff.{old,new}Moved color or
298 color.diff.{old,new}MovedAlternative. The change between the
299 two colors indicates that a new block was detected.
300 dimmed_zebra
302 Similar to zebra, but additional dimming of uninteresting parts
303 of moved code is performed. The bordering lines of two adjacent
304 blocks are considered interesting, the rest is uninteresting.
305 --word-diff[=<mode>]
307 Show a word diff, using the <mode> to delimit changed words. By
308 default, words are delimited by whitespace; see --word-diff-regex
309 below. The <mode> defaults to plain, and must be one of:
310 color
312 Highlight changed words using only colors. Implies --color.
313 plain
315 Show words as [-removed-] and {+added+}. Makes no attempts to
316 escape the delimiters if they appear in the input, so the
317 output may be ambiguous.
318 porcelain
320 Use a special line-based format intended for script
321 consumption. Added/removed/unchanged runs are printed in the
322 usual unified diff format, starting with a +/-/` ` character at
323 the beginning of the line and extending to the end of the line.
324 Newlines in the input are represented by a tilde ~ on a line of
325 its own.
326 none
328 Disable word diff again.
329 Note that despite the name of the first mode, color is used to
331 highlight the changed parts in all modes if enabled.
332 --word-diff-regex=<regex>
334 Use <regex> to decide what a word is, instead of considering runs
335 of non-whitespace to be a word. Also implies --word-diff unless it
336 was already enabled.
337 Every non-overlapping match of the <regex> is considered a word.
339 Anything between these matches is considered whitespace and
340 ignored(!) for the purposes of finding differences. You may want to
341 append |[^[:space:]] to your regular expression to make sure that
342 it matches all non-whitespace characters. A match that contains a
343 newline is silently truncated(!) at the newline.
344 For example, --word-diff-regex=. will treat each character as a
346 word and, correspondingly, show differences character by character.
347 The regex can also be set via a diff driver or configuration
349 option, see gitattributes(5) or git-config(1). Giving it explicitly
350 overrides any diff driver or configuration setting. Diff drivers
351 override configuration settings.
352 --color-words[=<regex>]
354 Equivalent to --word-diff=color plus (if a regex was specified)
355 --word-diff-regex=<regex>.
356 --no-renames
358 Turn off rename detection, even when the configuration file gives
359 the default to do so.
360 --check
362 Warn if changes introduce conflict markers or whitespace errors.
363 What are considered whitespace errors is controlled by
364 core.whitespace configuration. By default, trailing whitespaces
365 (including lines that consist solely of whitespaces) and a space
366 character that is immediately followed by a tab character inside
367 the initial indent of the line are considered whitespace errors.
368 Exits with non-zero status if problems are found. Not compatible
369 with --exit-code.
370 --ws-error-highlight=<kind>
372 Highlight whitespace errors in the context, old or new lines of the
373 diff. Multiple values are separated by comma, none resets previous
374 values, default reset the list to new and all is a shorthand for
375 old,new,context. When this option is not given, and the
376 configuration variable diff.wsErrorHighlight is not set, only
377 whitespace errors in new lines are highlighted. The whitespace
378 errors are colored with color.diff.whitespace.
379 --full-index
381 Instead of the first handful of characters, show the full pre- and
382 post-image blob object names on the "index" line when generating
383 patch format output.
384 --binary
386 In addition to --full-index, output a binary diff that can be
387 applied with git-apply.
388 --abbrev[=<n>]
390 Instead of showing the full 40-byte hexadecimal object name in
391 diff-raw format output and diff-tree header lines, show only a
392 partial prefix. This is independent of the --full-index option
393 above, which controls the diff-patch output format. Non default
394 number of digits can be specified with --abbrev=<n>.
395 -B[<n>][/<m>], --break-rewrites[=[<n>][/<m>]]
397 Break complete rewrite changes into pairs of delete and create.
398 This serves two purposes:
399 It affects the way a change that amounts to a total rewrite of a
401 file not as a series of deletion and insertion mixed together with
402 a very few lines that happen to match textually as the context, but
403 as a single deletion of everything old followed by a single
404 insertion of everything new, and the number m controls this aspect
405 of the -B option (defaults to 60%). -B/70% specifies that less
406 than 30% of the original should remain in the result for Git to
407 consider it a total rewrite (i.e. otherwise the resulting patch
408 will be a series of deletion and insertion mixed together with
409 context lines).
410 When used with -M, a totally-rewritten file is also considered as
412 the source of a rename (usually -M only considers a file that
413 disappeared as the source of a rename), and the number n controls
414 this aspect of the -B option (defaults to 50%). -B20% specifies
415 that a change with addition and deletion compared to 20% or more of
416 the file’s size are eligible for being picked up as a possible
417 source of a rename to another file.
418 -M[<n>], --find-renames[=<n>]
420 Detect renames. If n is specified, it is a threshold on the
421 similarity index (i.e. amount of addition/deletions compared to the
422 file’s size). For example, -M90% means Git should consider a
423 delete/add pair to be a rename if more than 90% of the file hasn’t
424 changed. Without a % sign, the number is to be read as a fraction,
425 with a decimal point before it. I.e., -M5 becomes 0.5, and is thus
426 the same as -M50%. Similarly, -M05 is the same as -M5%. To limit
427 detection to exact renames, use -M100%. The default similarity
428 index is 50%.
429 -C[<n>], --find-copies[=<n>]
431 Detect copies as well as renames. See also --find-copies-harder. If
432 n is specified, it has the same meaning as for -M<n>.
433 --find-copies-harder
435 For performance reasons, by default, -C option finds copies only if
436 the original file of the copy was modified in the same changeset.
437 This flag makes the command inspect unmodified files as candidates
438 for the source of copy. This is a very expensive operation for
439 large projects, so use it with caution. Giving more than one -C
440 option has the same effect.
441 -D, --irreversible-delete
443 Omit the preimage for deletes, i.e. print only the header but not
444 the diff between the preimage and /dev/null. The resulting patch is
445 not meant to be applied with patch or git apply; this is solely for
446 people who want to just concentrate on reviewing the text after the
447 change. In addition, the output obviously lacks enough information
448 to apply such a patch in reverse, even manually, hence the name of
449 the option.
450 When used together with -B, omit also the preimage in the deletion
452 part of a delete/create pair.
453 -l<num>
455 The -M and -C options require O(n^2) processing time where n is the
456 number of potential rename/copy targets. This option prevents
457 rename/copy detection from running if the number of rename/copy
458 targets exceeds the specified number.
459 --diff-filter=[(A|C|D|M|R|T|U|X|B)...[*]]
461 Select only files that are Added (A), Copied (C), Deleted (D),
462 Modified (M), Renamed (R), have their type (i.e. regular file,
463 symlink, submodule, ...) changed (T), are Unmerged (U), are Unknown
464 (X), or have had their pairing Broken (B). Any combination of the
465 filter characters (including none) can be used. When *
466 (All-or-none) is added to the combination, all paths are selected
467 if there is any file that matches other criteria in the comparison;
468 if there is no file that matches other criteria, nothing is
469 selected.
470 Also, these upper-case letters can be downcased to exclude. E.g.
472 --diff-filter=ad excludes added and deleted paths.
473 Note that not all diffs can feature all types. For instance, diffs
475 from the index to the working tree can never have Added entries
476 (because the set of paths included in the diff is limited by what
477 is in the index). Similarly, copied and renamed entries cannot
478 appear if detection for those types is disabled.
479 -S<string>
481 Look for differences that change the number of occurrences of the
482 specified string (i.e. addition/deletion) in a file. Intended for
483 the scripter’s use.
484 It is useful when you’re looking for an exact block of code (like a
486 struct), and want to know the history of that block since it first
487 came into being: use the feature iteratively to feed the
488 interesting block in the preimage back into -S, and keep going
489 until you get the very first version of the block.
490 -G<regex>
492 Look for differences whose patch text contains added/removed lines
493 that match <regex>.
494 To illustrate the difference between -S<regex> --pickaxe-regex and
496 -G<regex>, consider a commit with the following diff in the same
497 file:
498 + return !regexec(regexp, two->ptr, 1, ®match, 0);
500 ...
501 - hit = !regexec(regexp, mf2.ptr, 1, ®match, 0);
502 While git log -G"regexec\(regexp" will show this commit, git log
504 -S"regexec\(regexp" --pickaxe-regex will not (because the number of
505 occurrences of that string did not change).
506 See the pickaxe entry in gitdiffcore(7) for more information.
508 --find-object=<object-id>
510 Look for differences that change the number of occurrences of the
511 specified object. Similar to -S, just the argument is different in
512 that it doesn’t search for a specific string but for a specific
513 object id.
514 The object can be a blob or a submodule commit. It implies the -t
516 option in git-log to also find trees.
517 --pickaxe-all
519 When -S or -G finds a change, show all the changes in that
520 changeset, not just the files that contain the change in <string>.
521 --pickaxe-regex
523 Treat the <string> given to -S as an extended POSIX regular
524 expression to match.
525 -O<orderfile>
527 Control the order in which files appear in the output. This
528 overrides the diff.orderFile configuration variable (see git-
529 config(1)). To cancel diff.orderFile, use -O/dev/null.
530 The output order is determined by the order of glob patterns in
532 <orderfile>. All files with pathnames that match the first pattern
533 are output first, all files with pathnames that match the second
534 pattern (but not the first) are output next, and so on. All files
535 with pathnames that do not match any pattern are output last, as if
536 there was an implicit match-all pattern at the end of the file. If
537 multiple pathnames have the same rank (they match the same pattern
538 but no earlier patterns), their output order relative to each other
539 is the normal order.
540 <orderfile> is parsed as follows:
542 · Blank lines are ignored, so they can be used as separators for
544 readability.
545 · Lines starting with a hash ("#") are ignored, so they can be
547 used for comments. Add a backslash ("\") to the beginning of
548 the pattern if it starts with a hash.
549 · Each other line contains a single pattern.
551 Patterns have the same syntax and semantics as patterns used for
553 fnmatch(3) without the FNM_PATHNAME flag, except a pathname also
554 matches a pattern if removing any number of the final pathname
555 components matches the pattern. For example, the pattern "foo*bar"
556 matches "fooasdfbar" and "foo/bar/baz/asdf" but not "foobarx".
557 -R
559 Swap two inputs; that is, show differences from index or on-disk
560 file to tree contents.
561 --relative[=<path>]
563 When run from a subdirectory of the project, it can be told to
564 exclude changes outside the directory and show pathnames relative
565 to it with this option. When you are not in a subdirectory (e.g. in
566 a bare repository), you can name which subdirectory to make the
567 output relative to by giving a <path> as an argument.
568 -a, --text
570 Treat all files as text.
571 --ignore-cr-at-eol
573 Ignore carriage-return at the end of line when doing a comparison.
574 --ignore-space-at-eol
576 Ignore changes in whitespace at EOL.
577 -b, --ignore-space-change
579 Ignore changes in amount of whitespace. This ignores whitespace at
580 line end, and considers all other sequences of one or more
581 whitespace characters to be equivalent.
582 -w, --ignore-all-space
584 Ignore whitespace when comparing lines. This ignores differences
585 even if one line has whitespace where the other line has none.
586 --ignore-blank-lines
588 Ignore changes whose lines are all blank.
589 --inter-hunk-context=<lines>
591 Show the context between diff hunks, up to the specified number of
592 lines, thereby fusing hunks that are close to each other. Defaults
593 to diff.interHunkContext or 0 if the config option is unset.
594 -W, --function-context
596 Show whole surrounding functions of changes.
597 --exit-code
599 Make the program exit with codes similar to diff(1). That is, it
600 exits with 1 if there were differences and 0 means no differences.
601 --quiet
603 Disable all output of the program. Implies --exit-code.
604 --ext-diff
606 Allow an external diff helper to be executed. If you set an
607 external diff driver with gitattributes(5), you need to use this
608 option with git-log(1) and friends.
609 --no-ext-diff
611 Disallow external diff drivers.
612 --textconv, --no-textconv
614 Allow (or disallow) external text conversion filters to be run when
615 comparing binary files. See gitattributes(5) for details. Because
616 textconv filters are typically a one-way conversion, the resulting
617 diff is suitable for human consumption, but cannot be applied. For
618 this reason, textconv filters are enabled by default only for git-
619 diff(1) and git-log(1), but not for git-format-patch(1) or diff
620 plumbing commands.
621 --ignore-submodules[=<when>]
623 Ignore changes to submodules in the diff generation. <when> can be
624 either "none", "untracked", "dirty" or "all", which is the default.
625 Using "none" will consider the submodule modified when it either
626 contains untracked or modified files or its HEAD differs from the
627 commit recorded in the superproject and can be used to override any
628 settings of the ignore option in git-config(1) or gitmodules(5).
629 When "untracked" is used submodules are not considered dirty when
630 they only contain untracked content (but they are still scanned for
631 modified content). Using "dirty" ignores all changes to the work
632 tree of submodules, only changes to the commits stored in the
633 superproject are shown (this was the behavior until 1.7.0). Using
634 "all" hides all changes to submodules.
635 --src-prefix=<prefix>
637 Show the given source prefix instead of "a/".
638 --dst-prefix=<prefix>
640 Show the given destination prefix instead of "b/".
641 --no-prefix
643 Do not show any source or destination prefix.
644 --line-prefix=<prefix>
646 Prepend an additional prefix to every line of output.
647 --ita-invisible-in-index
649 By default entries added by "git add -N" appear as an existing
650 empty file in "git diff" and a new file in "git diff --cached".
651 This option makes the entry appear as a new file in "git diff" and
652 non-existent in "git diff --cached". This option could be reverted
653 with --ita-visible-in-index. Both options are experimental and
654 could be removed in future.
655 For more detailed explanation on these common options, see also
657 gitdiffcore(7).
658 -1 --base, -2 --ours, -3 --theirs
660 Compare the working tree with the "base" version (stage #1), "our
661 branch" (stage #2) or "their branch" (stage #3). The index contains
662 these stages only for unmerged entries i.e. while resolving
663 conflicts. See git-read-tree(1) section "3-Way Merge" for detailed
664 information.
665 -0
667 Omit diff output for unmerged entries and just show "Unmerged". Can
668 be used only when comparing the working tree with the index.
669 <path>...
671 The <paths> parameters, when given, are used to limit the diff to
672 the named paths (you can give directory names and get diff for all
673 files under them).
674
676 The raw output format from "git-diff-index", "git-diff-tree",
677 "git-diff-files" and "git diff --raw" are very similar.
678 These commands all compare two sets of things; what is compared
680 differs:
681 git-diff-index <tree-ish>
683 compares the <tree-ish> and the files on the filesystem.
684 git-diff-index --cached <tree-ish>
686 compares the <tree-ish> and the index.
687 git-diff-tree [-r] <tree-ish-1> <tree-ish-2> [<pattern>...]
689 compares the trees named by the two arguments.
690 git-diff-files [<pattern>...]
692 compares the index and the files on the filesystem.
693 The "git-diff-tree" command begins its output by printing the hash of
695 what is being compared. After that, all the commands print one output
696 line per changed file.
697 An output line is formatted this way:
699 in-place edit :100644 100644 bcd1234... 0123456... M file0
701 copy-edit :100644 100644 abcd123... 1234567... C68 file1 file2
702 rename-edit :100644 100644 abcd123... 1234567... R86 file1 file3
703 create :000000 100644 0000000... 1234567... A file4
704 delete :100644 000000 1234567... 0000000... D file5
705 unmerged :000000 000000 0000000... 0000000... U file6
706 That is, from the left to the right:
709 1. a colon.
711 2. mode for "src"; 000000 if creation or unmerged.
713 3. a space.
715 4. mode for "dst"; 000000 if deletion or unmerged.
717 5. a space.
719 6. sha1 for "src"; 0{40} if creation or unmerged.
721 7. a space.
723 8. sha1 for "dst"; 0{40} if creation, unmerged or "look at work tree".
725 9. a space.
727 10. status, followed by optional "score" number.
729 11. a tab or a NUL when -z option is used.
731 12. path for "src"
733 13. a tab or a NUL when -z option is used; only exists for C or R.
735 14. path for "dst"; only exists for C or R.
737 15. an LF or a NUL when -z option is used, to terminate the record.
739 Possible status letters are:
741 · A: addition of a file
743 · C: copy of a file into a new one
745 · D: deletion of a file
747 · M: modification of the contents or mode of a file
749 · R: renaming of a file
751 · T: change in the type of the file
753 · U: file is unmerged (you must complete the merge before it can be
755 committed)
756 · X: "unknown" change type (most probably a bug, please report it)
758 Status letters C and R are always followed by a score (denoting the
760 percentage of similarity between the source and target of the move or
761 copy). Status letter M may be followed by a score (denoting the
762 percentage of dissimilarity) for file rewrites.
763 <sha1> is shown as all 0’s if a file is new on the filesystem and it is
765 out of sync with the index.
766 Example:
768 :100644 100644 5be4a4...... 000000...... M file.c
770 Without the -z option, pathnames with "unusual" characters are quoted
773 as explained for the configuration variable core.quotePath (see git-
774 config(1)). Using -z the filename is output verbatim and the line is
775 terminated by a NUL byte.
776
778 "git-diff-tree", "git-diff-files" and "git-diff --raw" can take -c or
779 --cc option to generate diff output also for merge commits. The output
780 differs from the format described above in the following way:
781 1. there is a colon for each parent
783 2. there are more "src" modes and "src" sha1
785 3. status is concatenated status characters for each parent
787 4. no optional "score" number
789 5. single path, only for "dst"
791 Example:
793 ::100644 100644 100644 fabadb8... cc95eb0... 4866510... MM describe.c
795 Note that combined diff lists only files which were modified from all
798 parents.
799
801 When "git-diff-index", "git-diff-tree", or "git-diff-files" are run
802 with a -p option, "git diff" without the --raw option, or "git log"
803 with the "-p" option, they do not produce the output described above;
804 instead they produce a patch file. You can customize the creation of
805 such patches via the GIT_EXTERNAL_DIFF and the GIT_DIFF_OPTS
806 environment variables.
807 What the -p option produces is slightly different from the traditional
809 diff format:
810 1. It is preceded with a "git diff" header that looks like this:
812 diff --git a/file1 b/file2
814 The a/ and b/ filenames are the same unless rename/copy is
816 involved. Especially, even for a creation or a deletion, /dev/null
817 is not used in place of the a/ or b/ filenames.
818 When rename/copy is involved, file1 and file2 show the name of the
820 source file of the rename/copy and the name of the file that
821 rename/copy produces, respectively.
822 2. It is followed by one or more extended header lines:
824 old mode <mode>
826 new mode <mode>
827 deleted file mode <mode>
828 new file mode <mode>
829 copy from <path>
830 copy to <path>
831 rename from <path>
832 rename to <path>
833 similarity index <number>
834 dissimilarity index <number>
835 index <hash>..<hash> <mode>
836 File modes are printed as 6-digit octal numbers including the file
838 type and file permission bits.
839 Path names in extended headers do not include the a/ and b/
841 prefixes.
842 The similarity index is the percentage of unchanged lines, and the
844 dissimilarity index is the percentage of changed lines. It is a
845 rounded down integer, followed by a percent sign. The similarity
846 index value of 100% is thus reserved for two equal files, while
847 100% dissimilarity means that no line from the old file made it
848 into the new one.
849 The index line includes the SHA-1 checksum before and after the
851 change. The <mode> is included if the file mode does not change;
852 otherwise, separate lines indicate the old and the new mode.
853 3. Pathnames with "unusual" characters are quoted as explained for the
855 configuration variable core.quotePath (see git-config(1)).
856 4. All the file1 files in the output refer to files before the commit,
858 and all the file2 files refer to files after the commit. It is
859 incorrect to apply each change to each file sequentially. For
860 example, this patch will swap a and b:
861 diff --git a/a b/b
863 rename from a
864 rename to b
865 diff --git a/b b/a
866 rename from b
867 rename to a
868
870 Any diff-generating command can take the -c or --cc option to produce a
871 combined diff when showing a merge. This is the default format when
872 showing merges with git-diff(1) or git-show(1). Note also that you can
873 give the -m option to any of these commands to force generation of
874 diffs with individual parents of a merge.
875 A combined diff format looks like this:
877 diff --combined describe.c
879 index fabadb8,cc95eb0..4866510
880 --- a/describe.c
881 +++ b/describe.c
882 @@@ -98,20 -98,12 +98,20 @@@
883 return (a_date > b_date) ? -1 : (a_date == b_date) ? 0 : 1;
884 }
885 - static void describe(char *arg)
887 -static void describe(struct commit *cmit, int last_one)
888 ++static void describe(char *arg, int last_one)
889 {
890 + unsigned char sha1[20];
891 + struct commit *cmit;
892 struct commit_list *list;
893 static int initialized = 0;
894 struct commit_name *n;
895 + if (get_sha1(arg, sha1) < 0)
897 + usage(describe_usage);
898 + cmit = lookup_commit_reference(sha1);
899 + if (!cmit)
900 + usage(describe_usage);
901 +
902 if (!initialized) {
903 initialized = 1;
904 for_each_ref(get_name);
905 1. It is preceded with a "git diff" header, that looks like this (when
909 -c option is used):
910 diff --combined file
912 or like this (when --cc option is used):
914 diff --cc file
916 2. It is followed by one or more extended header lines (this example
918 shows a merge with two parents):
919 index <hash>,<hash>..<hash>
921 mode <mode>,<mode>..<mode>
922 new file mode <mode>
923 deleted file mode <mode>,<mode>
924 The mode <mode>,<mode>..<mode> line appears only if at least one of
926 the <mode> is different from the rest. Extended headers with
927 information about detected contents movement (renames and copying
928 detection) are designed to work with diff of two <tree-ish> and are
929 not used by combined diff format.
930 3. It is followed by two-line from-file/to-file header
932 --- a/file
934 +++ b/file
935 Similar to two-line header for traditional unified diff format,
937 /dev/null is used to signal created or deleted files.
938 4. Chunk header format is modified to prevent people from accidentally
940 feeding it to patch -p1. Combined diff format was created for
941 review of merge commit changes, and was not meant for apply. The
942 change is similar to the change in the extended index header:
943 @@@ <from-file-range> <from-file-range> <to-file-range> @@@
945 There are (number of parents + 1) @ characters in the chunk header
947 for combined diff format.
948 Unlike the traditional unified diff format, which shows two files A and
950 B with a single column that has - (minus — appears in A but removed in
951 B), + (plus — missing in A but added to B), or " " (space — unchanged)
952 prefix, this format compares two or more files file1, file2,... with
953 one file X, and shows how X differs from each of fileN. One column for
954 each of fileN is prepended to the output line to note how X’s line is
955 different from it.
956 A - character in the column N means that the line appears in fileN but
958 it does not appear in the result. A + character in the column N means
959 that the line appears in the result, and fileN does not have that line
960 (in other words, the line was added, from the point of view of that
961 parent).
962 In the above example output, the function signature was changed from
964 both files (hence two - removals from both file1 and file2, plus ++ to
965 mean one line that was added does not appear in either file1 or file2).
966 Also eight other lines are the same from file1 but do not appear in
967 file2 (hence prefixed with +).
968 When shown by git diff-tree -c, it compares the parents of a merge
970 commit with the merge result (i.e. file1..fileN are the parents). When
971 shown by git diff-files -c, it compares the two unresolved merge
972 parents with the working tree file (i.e. file1 is stage 2 aka "our
973 version", file2 is stage 3 aka "their version").
974
976 The --summary option describes newly added, deleted, renamed and copied
977 files. The --stat option adds diffstat(1) graph to the output. These
978 options can be combined with other options, such as -p, and are meant
979 for human consumption.
980 When showing a change that involves a rename or a copy, --stat output
982 formats the pathnames compactly by combining common prefix and suffix
983 of the pathnames. For example, a change that moves arch/i386/Makefile
984 to arch/x86/Makefile while modifying 4 lines will be shown like this:
985 arch/{i386 => x86}/Makefile | 4 +--
987 The --numstat option gives the diffstat(1) information but is designed
990 for easier machine consumption. An entry in --numstat output looks like
991 this:
992 1 2 README
994 3 1 arch/{i386 => x86}/Makefile
995 That is, from left to right:
998 1. the number of added lines;
1000 2. a tab;
1002 3. the number of deleted lines;
1004 4. a tab;
1006 5. pathname (possibly with rename/copy information);
1008 6. a newline.
1010 When -z output option is in effect, the output is formatted this way:
1012 1 2 README NUL
1014 3 1 NUL arch/i386/Makefile NUL arch/x86/Makefile NUL
1015 That is:
1018 1. the number of added lines;
1020 2. a tab;
1022 3. the number of deleted lines;
1024 4. a tab;
1026 5. a NUL (only exists if renamed/copied);
1028 6. pathname in preimage;
1030 7. a NUL (only exists if renamed/copied);
1032 8. pathname in postimage (only exists if renamed/copied);
1034 9. a NUL.
1036 The extra NUL before the preimage path in renamed case is to allow
1038 scripts that read the output to tell if the current record being read
1039 is a single-path record or a rename/copy record without reading ahead.
1040 After reading added and deleted lines, reading up to NUL would yield
1041 the pathname, but if that is NUL, the record will show two paths.
1042
1044 Various ways to check your working tree
1045 $ git diff [1m(1)
1047 $ git diff --cached [1m(2)
1048 $ git diff HEAD [1m(3)
1049 1. Changes in the working tree not yet staged for the next commit.
1051 2. Changes between the index and your last commit; what you would
1052 be committing if you run "git commit" without "-a" option.
1053 3. Changes in the working tree since your last commit; what you
1054 would be committing if you run "git commit -a"
1055 Comparing with arbitrary commits
1057 $ git diff test [1m(1)
1059 $ git diff HEAD -- ./test [1m(2)
1060 $ git diff HEAD^ HEAD [1m(3)
1061 1. Instead of using the tip of the current branch, compare with the
1063 tip of "test" branch.
1064 2. Instead of comparing with the tip of "test" branch, compare with
1065 the tip of the current branch, but limit the comparison to the file
1066 "test".
1067 3. Compare the version before the last commit and the last commit.
1068 Comparing branches
1070 $ git diff topic master [1m(1)
1072 $ git diff topic..master [1m(2)
1073 $ git diff topic...master [1m(3)
1074 1. Changes between the tips of the topic and the master branches.
1076 2. Same as above.
1077 3. Changes that occurred on the master branch since when the topic
1078 branch was started off it.
1079 Limiting the diff output
1081 $ git diff --diff-filter=MRC [1m(1)
1083 $ git diff --name-status [1m(2)
1084 $ git diff arch/i386 include/asm-i386 [1m(3)
1085 1. Show only modification, rename, and copy, but not addition or
1087 deletion.
1088 2. Show only names and the nature of change, but not actual diff
1089 output.
1090 3. Limit diff output to named subtrees.
1091 Munging the diff output
1093 $ git diff --find-copies-harder -B -C [1m(1)
1095 $ git diff -R [1m(2)
1096 1. Spend extra cycles to find renames, copies and complete rewrites
1098 (very expensive).
1099 2. Output diff in reverse.
1100
1102 diff(1), git-difftool(1), git-log(1), gitdiffcore(7), git-format-
1103 patch(1), git-apply(1)
1104
1106 Part of the git(1) suite
1107Git 2.18.1 05/14/2019 GIT-DIFF(1)