1GIT-DIFF-TREE(1) Git Manual GIT-DIFF-TREE(1)
2
6 git-diff-tree - Compares the content and mode of blobs found via two
7 tree objects
8
10 git diff-tree [--stdin] [-m] [-s] [-v] [--no-commit-id] [--pretty]
11 [-t] [-r] [-c | --cc] [--combined-all-paths] [--root] [--merge-base]
12 [<common-diff-options>] <tree-ish> [<tree-ish>] [<path>...]
13
15 Compare the content and mode of blobs found via two tree objects.
16 If there is only one <tree-ish> given, the commit is compared with its
18 parents (see --stdin below).
19 Note that git diff-tree can use the tree encapsulated in a commit
21 object.
22
24 -p, -u, --patch
25 Generate patch (see the section called “GENERATING PATCH TEXT WITH
26 -P”).
27 -s, --no-patch
29 Suppress all output from the diff machinery. Useful for commands
30 like git show that show the patch by default to squelch their
31 output, or to cancel the effect of options like --patch, --stat
32 earlier on the command line in an alias.
33 -U<n>, --unified=<n>
35 Generate diffs with <n> lines of context instead of the usual
36 three. Implies --patch.
37 --output=<file>
39 Output to a specific file instead of stdout.
40 --output-indicator-new=<char>, --output-indicator-old=<char>,
42 --output-indicator-context=<char>
43 Specify the character used to indicate new, old or context lines in
44 the generated patch. Normally they are +, - and ' ' respectively.
45 --raw
47 Generate the diff in raw format. This is the default.
48 --patch-with-raw
50 Synonym for -p --raw.
51 --indent-heuristic
53 Enable the heuristic that shifts diff hunk boundaries to make
54 patches easier to read. This is the default.
55 --no-indent-heuristic
57 Disable the indent heuristic.
58 --minimal
60 Spend extra time to make sure the smallest possible diff is
61 produced.
62 --patience
64 Generate a diff using the "patience diff" algorithm.
65 --histogram
67 Generate a diff using the "histogram diff" algorithm.
68 --anchored=<text>
70 Generate a diff using the "anchored diff" algorithm.
71 This option may be specified more than once.
73 If a line exists in both the source and destination, exists only
75 once, and starts with this text, this algorithm attempts to prevent
76 it from appearing as a deletion or addition in the output. It uses
77 the "patience diff" algorithm internally.
78 --diff-algorithm={patience|minimal|histogram|myers}
80 Choose a diff algorithm. The variants are as follows:
81 default, myers
83 The basic greedy diff algorithm. Currently, this is the
84 default.
85 minimal
87 Spend extra time to make sure the smallest possible diff is
88 produced.
89 patience
91 Use "patience diff" algorithm when generating patches.
92 histogram
94 This algorithm extends the patience algorithm to "support
95 low-occurrence common elements".
96 For instance, if you configured the diff.algorithm variable to a
98 non-default value and want to use the default one, then you have to
99 use --diff-algorithm=default option.
100 --stat[=<width>[,<name-width>[,<count>]]]
102 Generate a diffstat. By default, as much space as necessary will be
103 used for the filename part, and the rest for the graph part.
104 Maximum width defaults to terminal width, or 80 columns if not
105 connected to a terminal, and can be overridden by <width>. The
106 width of the filename part can be limited by giving another width
107 <name-width> after a comma or by setting
108 diff.statNameWidth=<width>. The width of the graph part can be
109 limited by using --stat-graph-width=<width> or by setting
110 diff.statGraphWidth=<width>. Using --stat or --stat-graph-width
111 affects all commands generating a stat graph, while setting
112 diff.statNameWidth or diff.statGraphWidth does not affect git
113 format-patch. By giving a third parameter <count>, you can limit
114 the output to the first <count> lines, followed by ... if there
115 are more.
116 These parameters can also be set individually with
118 --stat-width=<width>, --stat-name-width=<name-width> and
119 --stat-count=<count>.
120 --compact-summary
122 Output a condensed summary of extended header information such as
123 file creations or deletions ("new" or "gone", optionally "+l" if
124 it’s a symlink) and mode changes ("+x" or "-x" for adding or
125 removing executable bit respectively) in diffstat. The information
126 is put between the filename part and the graph part. Implies
127 --stat.
128 --numstat
130 Similar to --stat, but shows number of added and deleted lines in
131 decimal notation and pathname without abbreviation, to make it more
132 machine friendly. For binary files, outputs two - instead of saying
133 0 0.
134 --shortstat
136 Output only the last line of the --stat format containing total
137 number of modified files, as well as number of added and deleted
138 lines.
139 -X[<param1,param2,...>], --dirstat[=<param1,param2,...>]
141 Output the distribution of relative amount of changes for each
142 sub-directory. The behavior of --dirstat can be customized by
143 passing it a comma separated list of parameters. The defaults are
144 controlled by the diff.dirstat configuration variable (see git-
145 config(1)). The following parameters are available:
146 changes
148 Compute the dirstat numbers by counting the lines that have
149 been removed from the source, or added to the destination. This
150 ignores the amount of pure code movements within a file. In
151 other words, rearranging lines in a file is not counted as much
152 as other changes. This is the default behavior when no
153 parameter is given.
154 lines
156 Compute the dirstat numbers by doing the regular line-based
157 diff analysis, and summing the removed/added line counts. (For
158 binary files, count 64-byte chunks instead, since binary files
159 have no natural concept of lines). This is a more expensive
160 --dirstat behavior than the changes behavior, but it does count
161 rearranged lines within a file as much as other changes. The
162 resulting output is consistent with what you get from the other
163 --*stat options.
164 files
166 Compute the dirstat numbers by counting the number of files
167 changed. Each changed file counts equally in the dirstat
168 analysis. This is the computationally cheapest --dirstat
169 behavior, since it does not have to look at the file contents
170 at all.
171 cumulative
173 Count changes in a child directory for the parent directory as
174 well. Note that when using cumulative, the sum of the
175 percentages reported may exceed 100%. The default
176 (non-cumulative) behavior can be specified with the
177 noncumulative parameter.
178 <limit>
180 An integer parameter specifies a cut-off percent (3% by
181 default). Directories contributing less than this percentage of
182 the changes are not shown in the output.
183 Example: The following will count changed files, while ignoring
185 directories with less than 10% of the total amount of changed
186 files, and accumulating child directory counts in the parent
187 directories: --dirstat=files,10,cumulative.
188 --cumulative
190 Synonym for --dirstat=cumulative
191 --dirstat-by-file[=<param1,param2>...]
193 Synonym for --dirstat=files,param1,param2...
194 --summary
196 Output a condensed summary of extended header information such as
197 creations, renames and mode changes.
198 --patch-with-stat
200 Synonym for -p --stat.
201 -z
203 When --raw, --numstat, --name-only or --name-status has been given,
204 do not munge pathnames and use NULs as output field terminators.
205 Without this option, pathnames with "unusual" characters are quoted
207 as explained for the configuration variable core.quotePath (see
208 git-config(1)).
209 --name-only
211 Show only names of changed files. The file names are often encoded
212 in UTF-8. For more information see the discussion about encoding in
213 the git-log(1) manual page.
214 --name-status
216 Show only names and status of changed files. See the description of
217 the --diff-filter option on what the status letters mean. Just like
218 --name-only the file names are often encoded in UTF-8.
219 --submodule[=<format>]
221 Specify how differences in submodules are shown. When specifying
222 --submodule=short the short format is used. This format just shows
223 the names of the commits at the beginning and end of the range.
224 When --submodule or --submodule=log is specified, the log format is
225 used. This format lists the commits in the range like git-
226 submodule(1) summary does. When --submodule=diff is specified, the
227 diff format is used. This format shows an inline diff of the
228 changes in the submodule contents between the commit range.
229 Defaults to diff.submodule or the short format if the config option
230 is unset.
231 --color[=<when>]
233 Show colored diff. --color (i.e. without =<when>) is the same as
234 --color=always. <when> can be one of always, never, or auto.
235 --no-color
237 Turn off colored diff. It is the same as --color=never.
238 --color-moved[=<mode>]
240 Moved lines of code are colored differently. The <mode> defaults to
241 no if the option is not given and to zebra if the option with no
242 mode is given. The mode must be one of:
243 no
245 Moved lines are not highlighted.
246 default
248 Is a synonym for zebra. This may change to a more sensible mode
249 in the future.
250 plain
252 Any line that is added in one location and was removed in
253 another location will be colored with color.diff.newMoved.
254 Similarly color.diff.oldMoved will be used for removed lines
255 that are added somewhere else in the diff. This mode picks up
256 any moved line, but it is not very useful in a review to
257 determine if a block of code was moved without permutation.
258 blocks
260 Blocks of moved text of at least 20 alphanumeric characters are
261 detected greedily. The detected blocks are painted using either
262 the color.diff.{old,new}Moved color. Adjacent blocks cannot be
263 told apart.
264 zebra
266 Blocks of moved text are detected as in blocks mode. The blocks
267 are painted using either the color.diff.{old,new}Moved color or
268 color.diff.{old,new}MovedAlternative. The change between the
269 two colors indicates that a new block was detected.
270 dimmed-zebra
272 Similar to zebra, but additional dimming of uninteresting parts
273 of moved code is performed. The bordering lines of two adjacent
274 blocks are considered interesting, the rest is uninteresting.
275 dimmed_zebra is a deprecated synonym.
276 --no-color-moved
278 Turn off move detection. This can be used to override configuration
279 settings. It is the same as --color-moved=no.
280 --color-moved-ws=<modes>
282 This configures how whitespace is ignored when performing the move
283 detection for --color-moved. These modes can be given as a comma
284 separated list:
285 no
287 Do not ignore whitespace when performing move detection.
288 ignore-space-at-eol
290 Ignore changes in whitespace at EOL.
291 ignore-space-change
293 Ignore changes in amount of whitespace. This ignores whitespace
294 at line end, and considers all other sequences of one or more
295 whitespace characters to be equivalent.
296 ignore-all-space
298 Ignore whitespace when comparing lines. This ignores
299 differences even if one line has whitespace where the other
300 line has none.
301 allow-indentation-change
303 Initially ignore any whitespace in the move detection, then
304 group the moved code blocks only into a block if the change in
305 whitespace is the same per line. This is incompatible with the
306 other modes.
307 --no-color-moved-ws
309 Do not ignore whitespace when performing move detection. This can
310 be used to override configuration settings. It is the same as
311 --color-moved-ws=no.
312 --word-diff[=<mode>]
314 Show a word diff, using the <mode> to delimit changed words. By
315 default, words are delimited by whitespace; see --word-diff-regex
316 below. The <mode> defaults to plain, and must be one of:
317 color
319 Highlight changed words using only colors. Implies --color.
320 plain
322 Show words as [-removed-] and {+added+}. Makes no attempts to
323 escape the delimiters if they appear in the input, so the
324 output may be ambiguous.
325 porcelain
327 Use a special line-based format intended for script
328 consumption. Added/removed/unchanged runs are printed in the
329 usual unified diff format, starting with a +/-/` ` character at
330 the beginning of the line and extending to the end of the line.
331 Newlines in the input are represented by a tilde ~ on a line of
332 its own.
333 none
335 Disable word diff again.
336 Note that despite the name of the first mode, color is used to
338 highlight the changed parts in all modes if enabled.
339 --word-diff-regex=<regex>
341 Use <regex> to decide what a word is, instead of considering runs
342 of non-whitespace to be a word. Also implies --word-diff unless it
343 was already enabled.
344 Every non-overlapping match of the <regex> is considered a word.
346 Anything between these matches is considered whitespace and
347 ignored(!) for the purposes of finding differences. You may want to
348 append |[^[:space:]] to your regular expression to make sure that
349 it matches all non-whitespace characters. A match that contains a
350 newline is silently truncated(!) at the newline.
351 For example, --word-diff-regex=. will treat each character as a
353 word and, correspondingly, show differences character by character.
354 The regex can also be set via a diff driver or configuration
356 option, see gitattributes(5) or git-config(1). Giving it explicitly
357 overrides any diff driver or configuration setting. Diff drivers
358 override configuration settings.
359 --color-words[=<regex>]
361 Equivalent to --word-diff=color plus (if a regex was specified)
362 --word-diff-regex=<regex>.
363 --no-renames
365 Turn off rename detection, even when the configuration file gives
366 the default to do so.
367 --[no-]rename-empty
369 Whether to use empty blobs as rename source.
370 --check
372 Warn if changes introduce conflict markers or whitespace errors.
373 What are considered whitespace errors is controlled by
374 core.whitespace configuration. By default, trailing whitespaces
375 (including lines that consist solely of whitespaces) and a space
376 character that is immediately followed by a tab character inside
377 the initial indent of the line are considered whitespace errors.
378 Exits with non-zero status if problems are found. Not compatible
379 with --exit-code.
380 --ws-error-highlight=<kind>
382 Highlight whitespace errors in the context, old or new lines of the
383 diff. Multiple values are separated by comma, none resets previous
384 values, default reset the list to new and all is a shorthand for
385 old,new,context. When this option is not given, and the
386 configuration variable diff.wsErrorHighlight is not set, only
387 whitespace errors in new lines are highlighted. The whitespace
388 errors are colored with color.diff.whitespace.
389 --full-index
391 Instead of the first handful of characters, show the full pre- and
392 post-image blob object names on the "index" line when generating
393 patch format output.
394 --binary
396 In addition to --full-index, output a binary diff that can be
397 applied with git-apply. Implies --patch.
398 --abbrev[=<n>]
400 Instead of showing the full 40-byte hexadecimal object name in
401 diff-raw format output and diff-tree header lines, show the
402 shortest prefix that is at least <n> hexdigits long that uniquely
403 refers the object. In diff-patch output format, --full-index takes
404 higher precedence, i.e. if --full-index is specified, full blob
405 names will be shown regardless of --abbrev. Non default number of
406 digits can be specified with --abbrev=<n>.
407 -B[<n>][/<m>], --break-rewrites[=[<n>][/<m>]]
409 Break complete rewrite changes into pairs of delete and create.
410 This serves two purposes:
411 It affects the way a change that amounts to a total rewrite of a
413 file not as a series of deletion and insertion mixed together with
414 a very few lines that happen to match textually as the context, but
415 as a single deletion of everything old followed by a single
416 insertion of everything new, and the number m controls this aspect
417 of the -B option (defaults to 60%). -B/70% specifies that less
418 than 30% of the original should remain in the result for Git to
419 consider it a total rewrite (i.e. otherwise the resulting patch
420 will be a series of deletion and insertion mixed together with
421 context lines).
422 When used with -M, a totally-rewritten file is also considered as
424 the source of a rename (usually -M only considers a file that
425 disappeared as the source of a rename), and the number n controls
426 this aspect of the -B option (defaults to 50%). -B20% specifies
427 that a change with addition and deletion compared to 20% or more of
428 the file’s size are eligible for being picked up as a possible
429 source of a rename to another file.
430 -M[<n>], --find-renames[=<n>]
432 Detect renames. If n is specified, it is a threshold on the
433 similarity index (i.e. amount of addition/deletions compared to the
434 file’s size). For example, -M90% means Git should consider a
435 delete/add pair to be a rename if more than 90% of the file hasn’t
436 changed. Without a % sign, the number is to be read as a fraction,
437 with a decimal point before it. I.e., -M5 becomes 0.5, and is thus
438 the same as -M50%. Similarly, -M05 is the same as -M5%. To limit
439 detection to exact renames, use -M100%. The default similarity
440 index is 50%.
441 -C[<n>], --find-copies[=<n>]
443 Detect copies as well as renames. See also --find-copies-harder. If
444 n is specified, it has the same meaning as for -M<n>.
445 --find-copies-harder
447 For performance reasons, by default, -C option finds copies only if
448 the original file of the copy was modified in the same changeset.
449 This flag makes the command inspect unmodified files as candidates
450 for the source of copy. This is a very expensive operation for
451 large projects, so use it with caution. Giving more than one -C
452 option has the same effect.
453 -D, --irreversible-delete
455 Omit the preimage for deletes, i.e. print only the header but not
456 the diff between the preimage and /dev/null. The resulting patch is
457 not meant to be applied with patch or git apply; this is solely for
458 people who want to just concentrate on reviewing the text after the
459 change. In addition, the output obviously lacks enough information
460 to apply such a patch in reverse, even manually, hence the name of
461 the option.
462 When used together with -B, omit also the preimage in the deletion
464 part of a delete/create pair.
465 -l<num>
467 The -M and -C options involve some preliminary steps that can
468 detect subsets of renames/copies cheaply, followed by an exhaustive
469 fallback portion that compares all remaining unpaired destinations
470 to all relevant sources. (For renames, only remaining unpaired
471 sources are relevant; for copies, all original sources are
472 relevant.) For N sources and destinations, this exhaustive check is
473 O(N^2). This option prevents the exhaustive portion of rename/copy
474 detection from running if the number of source/destination files
475 involved exceeds the specified number. Defaults to
476 diff.renameLimit. Note that a value of 0 is treated as unlimited.
477 --diff-filter=[(A|C|D|M|R|T|U|X|B)...[*]]
479 Select only files that are Added (A), Copied (C), Deleted (D),
480 Modified (M), Renamed (R), have their type (i.e. regular file,
481 symlink, submodule, ...) changed (T), are Unmerged (U), are Unknown
482 (X), or have had their pairing Broken (B). Any combination of the
483 filter characters (including none) can be used. When *
484 (All-or-none) is added to the combination, all paths are selected
485 if there is any file that matches other criteria in the comparison;
486 if there is no file that matches other criteria, nothing is
487 selected.
488 Also, these upper-case letters can be downcased to exclude. E.g.
490 --diff-filter=ad excludes added and deleted paths.
491 Note that not all diffs can feature all types. For instance, copied
493 and renamed entries cannot appear if detection for those types is
494 disabled.
495 -S<string>
497 Look for differences that change the number of occurrences of the
498 specified string (i.e. addition/deletion) in a file. Intended for
499 the scripter’s use.
500 It is useful when you’re looking for an exact block of code (like a
502 struct), and want to know the history of that block since it first
503 came into being: use the feature iteratively to feed the
504 interesting block in the preimage back into -S, and keep going
505 until you get the very first version of the block.
506 Binary files are searched as well.
508 -G<regex>
510 Look for differences whose patch text contains added/removed lines
511 that match <regex>.
512 To illustrate the difference between -S<regex> --pickaxe-regex and
514 -G<regex>, consider a commit with the following diff in the same
515 file:
516 + return frotz(nitfol, two->ptr, 1, 0);
518 ...
519 - hit = frotz(nitfol, mf2.ptr, 1, 0);
520 While git log -G"frotz\(nitfol" will show this commit, git log
522 -S"frotz\(nitfol" --pickaxe-regex will not (because the number of
523 occurrences of that string did not change).
524 Unless --text is supplied patches of binary files without a
526 textconv filter will be ignored.
527 See the pickaxe entry in gitdiffcore(7) for more information.
529 --find-object=<object-id>
531 Look for differences that change the number of occurrences of the
532 specified object. Similar to -S, just the argument is different in
533 that it doesn’t search for a specific string but for a specific
534 object id.
535 The object can be a blob or a submodule commit. It implies the -t
537 option in git-log to also find trees.
538 --pickaxe-all
540 When -S or -G finds a change, show all the changes in that
541 changeset, not just the files that contain the change in <string>.
542 --pickaxe-regex
544 Treat the <string> given to -S as an extended POSIX regular
545 expression to match.
546 -O<orderfile>
548 Control the order in which files appear in the output. This
549 overrides the diff.orderFile configuration variable (see git-
550 config(1)). To cancel diff.orderFile, use -O/dev/null.
551 The output order is determined by the order of glob patterns in
553 <orderfile>. All files with pathnames that match the first pattern
554 are output first, all files with pathnames that match the second
555 pattern (but not the first) are output next, and so on. All files
556 with pathnames that do not match any pattern are output last, as if
557 there was an implicit match-all pattern at the end of the file. If
558 multiple pathnames have the same rank (they match the same pattern
559 but no earlier patterns), their output order relative to each other
560 is the normal order.
561 <orderfile> is parsed as follows:
563 • Blank lines are ignored, so they can be used as separators for
565 readability.
566 • Lines starting with a hash ("#") are ignored, so they can be
568 used for comments. Add a backslash ("\") to the beginning of
569 the pattern if it starts with a hash.
570 • Each other line contains a single pattern.
572 Patterns have the same syntax and semantics as patterns used for
574 fnmatch(3) without the FNM_PATHNAME flag, except a pathname also
575 matches a pattern if removing any number of the final pathname
576 components matches the pattern. For example, the pattern "foo*bar"
577 matches "fooasdfbar" and "foo/bar/baz/asdf" but not "foobarx".
578 --skip-to=<file>, --rotate-to=<file>
580 Discard the files before the named <file> from the output (i.e.
581 skip to), or move them to the end of the output (i.e. rotate to).
582 These options were invented primarily for the use of the git
583 difftool command, and may not be very useful otherwise.
584 -R
586 Swap two inputs; that is, show differences from index or on-disk
587 file to tree contents.
588 --relative[=<path>], --no-relative
590 When run from a subdirectory of the project, it can be told to
591 exclude changes outside the directory and show pathnames relative
592 to it with this option. When you are not in a subdirectory (e.g. in
593 a bare repository), you can name which subdirectory to make the
594 output relative to by giving a <path> as an argument.
595 --no-relative can be used to countermand both diff.relative config
596 option and previous --relative.
597 -a, --text
599 Treat all files as text.
600 --ignore-cr-at-eol
602 Ignore carriage-return at the end of line when doing a comparison.
603 --ignore-space-at-eol
605 Ignore changes in whitespace at EOL.
606 -b, --ignore-space-change
608 Ignore changes in amount of whitespace. This ignores whitespace at
609 line end, and considers all other sequences of one or more
610 whitespace characters to be equivalent.
611 -w, --ignore-all-space
613 Ignore whitespace when comparing lines. This ignores differences
614 even if one line has whitespace where the other line has none.
615 --ignore-blank-lines
617 Ignore changes whose lines are all blank.
618 -I<regex>, --ignore-matching-lines=<regex>
620 Ignore changes whose all lines match <regex>. This option may be
621 specified more than once.
622 --inter-hunk-context=<lines>
624 Show the context between diff hunks, up to the specified number of
625 lines, thereby fusing hunks that are close to each other. Defaults
626 to diff.interHunkContext or 0 if the config option is unset.
627 -W, --function-context
629 Show whole function as context lines for each change. The function
630 names are determined in the same way as git diff works out patch
631 hunk headers (see Defining a custom hunk-header in
632 gitattributes(5)).
633 --exit-code
635 Make the program exit with codes similar to diff(1). That is, it
636 exits with 1 if there were differences and 0 means no differences.
637 --quiet
639 Disable all output of the program. Implies --exit-code.
640 --ext-diff
642 Allow an external diff helper to be executed. If you set an
643 external diff driver with gitattributes(5), you need to use this
644 option with git-log(1) and friends.
645 --no-ext-diff
647 Disallow external diff drivers.
648 --textconv, --no-textconv
650 Allow (or disallow) external text conversion filters to be run when
651 comparing binary files. See gitattributes(5) for details. Because
652 textconv filters are typically a one-way conversion, the resulting
653 diff is suitable for human consumption, but cannot be applied. For
654 this reason, textconv filters are enabled by default only for git-
655 diff(1) and git-log(1), but not for git-format-patch(1) or diff
656 plumbing commands.
657 --ignore-submodules[=<when>]
659 Ignore changes to submodules in the diff generation. <when> can be
660 either "none", "untracked", "dirty" or "all", which is the default.
661 Using "none" will consider the submodule modified when it either
662 contains untracked or modified files or its HEAD differs from the
663 commit recorded in the superproject and can be used to override any
664 settings of the ignore option in git-config(1) or gitmodules(5).
665 When "untracked" is used submodules are not considered dirty when
666 they only contain untracked content (but they are still scanned for
667 modified content). Using "dirty" ignores all changes to the work
668 tree of submodules, only changes to the commits stored in the
669 superproject are shown (this was the behavior until 1.7.0). Using
670 "all" hides all changes to submodules.
671 --src-prefix=<prefix>
673 Show the given source prefix instead of "a/".
674 --dst-prefix=<prefix>
676 Show the given destination prefix instead of "b/".
677 --no-prefix
679 Do not show any source or destination prefix.
680 --default-prefix
682 Use the default source and destination prefixes ("a/" and "b/").
683 This is usually the default already, but may be used to override
684 config such as diff.noprefix.
685 --line-prefix=<prefix>
687 Prepend an additional prefix to every line of output.
688 --ita-invisible-in-index
690 By default entries added by "git add -N" appear as an existing
691 empty file in "git diff" and a new file in "git diff --cached".
692 This option makes the entry appear as a new file in "git diff" and
693 non-existent in "git diff --cached". This option could be reverted
694 with --ita-visible-in-index. Both options are experimental and
695 could be removed in future.
696 For more detailed explanation on these common options, see also
698 gitdiffcore(7).
699 <tree-ish>
701 The id of a tree object.
702 <path>...
704 If provided, the results are limited to a subset of files matching
705 one of the provided pathspecs.
706 -r
708 Recurse into sub-trees.
709 -t
711 Show tree entry itself as well as subtrees. Implies -r.
712 --root
714 When --root is specified the initial commit will be shown as a big
715 creation event. This is equivalent to a diff against the NULL tree.
716 --merge-base
718 Instead of comparing the <tree-ish>s directly, use the merge base
719 between the two <tree-ish>s as the "before" side. There must be two
720 <tree-ish>s given and they must both be commits.
721 --stdin
723 When --stdin is specified, the command does not take <tree-ish>
724 arguments from the command line. Instead, it reads lines containing
725 either two <tree>, one <commit>, or a list of <commit> from its
726 standard input. (Use a single space as separator.)
727 When two trees are given, it compares the first tree with the
729 second. When a single commit is given, it compares the commit with
730 its parents. The remaining commits, when given, are used as if they
731 are parents of the first commit.
732 When comparing two trees, the ID of both trees (separated by a
734 space and terminated by a newline) is printed before the
735 difference. When comparing commits, the ID of the first (or only)
736 commit, followed by a newline, is printed.
737 The following flags further affect the behavior when comparing
739 commits (but not trees).
740 -m
742 By default, git diff-tree --stdin does not show differences for
743 merge commits. With this flag, it shows differences to that commit
744 from all of its parents. See also -c.
745 -s
747 By default, git diff-tree --stdin shows differences, either in
748 machine-readable form (without -p) or in patch form (with -p). This
749 output can be suppressed. It is only useful with the -v flag.
750 -v
752 This flag causes git diff-tree --stdin to also show the commit
753 message before the differences.
754 --pretty[=<format>], --format=<format>
756 Pretty-print the contents of the commit logs in a given format,
757 where <format> can be one of oneline, short, medium, full, fuller,
758 reference, email, raw, format:<string> and tformat:<string>. When
759 <format> is none of the above, and has %placeholder in it, it acts
760 as if --pretty=tformat:<format> were given.
761 See the "PRETTY FORMATS" section for some additional details for
763 each format. When =<format> part is omitted, it defaults to medium.
764 Note: you can specify the default pretty format in the repository
766 configuration (see git-config(1)).
767 --abbrev-commit
769 Instead of showing the full 40-byte hexadecimal commit object name,
770 show a prefix that names the object uniquely. "--abbrev=<n>" (which
771 also modifies diff output, if it is displayed) option can be used
772 to specify the minimum length of the prefix.
773 This should make "--pretty=oneline" a whole lot more readable for
775 people using 80-column terminals.
776 --no-abbrev-commit
778 Show the full 40-byte hexadecimal commit object name. This negates
779 --abbrev-commit, either explicit or implied by other options such
780 as "--oneline". It also overrides the log.abbrevCommit variable.
781 --oneline
783 This is a shorthand for "--pretty=oneline --abbrev-commit" used
784 together.
785 --encoding=<encoding>
787 Commit objects record the character encoding used for the log
788 message in their encoding header; this option can be used to tell
789 the command to re-code the commit log message in the encoding
790 preferred by the user. For non plumbing commands this defaults to
791 UTF-8. Note that if an object claims to be encoded in X and we are
792 outputting in X, we will output the object verbatim; this means
793 that invalid sequences in the original commit may be copied to the
794 output. Likewise, if iconv(3) fails to convert the commit, we will
795 quietly output the original object verbatim.
796 --expand-tabs=<n>, --expand-tabs, --no-expand-tabs
798 Perform a tab expansion (replace each tab with enough spaces to
799 fill to the next display column that is a multiple of <n>) in the
800 log message before showing it in the output. --expand-tabs is a
801 short-hand for --expand-tabs=8, and --no-expand-tabs is a
802 short-hand for --expand-tabs=0, which disables tab expansion.
803 By default, tabs are expanded in pretty formats that indent the log
805 message by 4 spaces (i.e. medium, which is the default, full, and
806 fuller).
807 --notes[=<ref>]
809 Show the notes (see git-notes(1)) that annotate the commit, when
810 showing the commit log message. This is the default for git log,
811 git show and git whatchanged commands when there is no --pretty,
812 --format, or --oneline option given on the command line.
813 By default, the notes shown are from the notes refs listed in the
815 core.notesRef and notes.displayRef variables (or corresponding
816 environment overrides). See git-config(1) for more details.
817 With an optional <ref> argument, use the ref to find the notes to
819 display. The ref can specify the full refname when it begins with
820 refs/notes/; when it begins with notes/, refs/ and otherwise
821 refs/notes/ is prefixed to form the full name of the ref.
822 Multiple --notes options can be combined to control which notes are
824 being displayed. Examples: "--notes=foo" will show only notes from
825 "refs/notes/foo"; "--notes=foo --notes" will show both notes from
826 "refs/notes/foo" and from the default notes ref(s).
827 --no-notes
829 Do not show notes. This negates the above --notes option, by
830 resetting the list of notes refs from which notes are shown.
831 Options are parsed in the order given on the command line, so e.g.
832 "--notes --notes=foo --no-notes --notes=bar" will only show notes
833 from "refs/notes/bar".
834 --show-notes-by-default
836 Show the default notes unless options for displaying specific notes
837 are given.
838 --show-notes[=<ref>], --[no-]standard-notes
840 These options are deprecated. Use the above --notes/--no-notes
841 options instead.
842 --show-signature
844 Check the validity of a signed commit object by passing the
845 signature to gpg --verify and show the output.
846 --no-commit-id
848 git diff-tree outputs a line with the commit ID when applicable.
849 This flag suppressed the commit ID output.
850 -c
852 This flag changes the way a merge commit is displayed (which means
853 it is useful only when the command is given one <tree-ish>, or
854 --stdin). It shows the differences from each of the parents to the
855 merge result simultaneously instead of showing pairwise diff
856 between a parent and the result one at a time (which is what the -m
857 option does). Furthermore, it lists only files which were modified
858 from all parents.
859 --cc
861 This flag changes the way a merge commit patch is displayed, in a
862 similar way to the -c option. It implies the -c and -p options and
863 further compresses the patch output by omitting uninteresting hunks
864 whose contents in the parents have only two variants and the merge
865 result picks one of them without modification. When all hunks are
866 uninteresting, the commit itself and the commit log message are not
867 shown, just like in any other "empty diff" case.
868 --combined-all-paths
870 This flag causes combined diffs (used for merge commits) to list
871 the name of the file from all parents. It thus only has effect when
872 -c or --cc are specified, and is likely only useful if filename
873 changes are detected (i.e. when either rename or copy detection
874 have been requested).
875 --always
877 Show the commit itself and the commit log message even if the diff
878 itself is empty.
879
881 If the commit is a merge, and if the pretty-format is not oneline,
882 email or raw, an additional line is inserted before the Author: line.
883 This line begins with "Merge: " and the hashes of ancestral commits are
884 printed, separated by spaces. Note that the listed commits may not
885 necessarily be the list of the direct parent commits if you have
886 limited your view of history: for example, if you are only interested
887 in changes related to a certain directory or file.
888 There are several built-in formats, and you can define additional
890 formats by setting a pretty.<name> config option to either another
891 format name, or a format: string, as described below (see git-
892 config(1)). Here are the details of the built-in formats:
893 • oneline
895 <hash> <title-line>
897 This is designed to be as compact as possible.
899 • short
901 commit <hash>
903 Author: <author>
904 <title-line>
906 • medium
908 commit <hash>
910 Author: <author>
911 Date: <author-date>
912 <title-line>
914 <full-commit-message>
916 • full
918 commit <hash>
920 Author: <author>
921 Commit: <committer>
922 <title-line>
924 <full-commit-message>
926 • fuller
928 commit <hash>
930 Author: <author>
931 AuthorDate: <author-date>
932 Commit: <committer>
933 CommitDate: <committer-date>
934 <title-line>
936 <full-commit-message>
938 • reference
940 <abbrev-hash> (<title-line>, <short-author-date>)
942 This format is used to refer to another commit in a commit message
944 and is the same as --pretty='format:%C(auto)%h (%s, %ad)'. By
945 default, the date is formatted with --date=short unless another
946 --date option is explicitly specified. As with any format: with
947 format placeholders, its output is not affected by other options
948 like --decorate and --walk-reflogs.
949 • email
951 From <hash> <date>
953 From: <author>
954 Date: <author-date>
955 Subject: [PATCH] <title-line>
956 <full-commit-message>
958 • mboxrd
960 Like email, but lines in the commit message starting with "From "
962 (preceded by zero or more ">") are quoted with ">" so they aren’t
963 confused as starting a new commit.
964 • raw
966 The raw format shows the entire commit exactly as stored in the
968 commit object. Notably, the hashes are displayed in full,
969 regardless of whether --abbrev or --no-abbrev are used, and parents
970 information show the true parent commits, without taking grafts or
971 history simplification into account. Note that this format affects
972 the way commits are displayed, but not the way the diff is shown
973 e.g. with git log --raw. To get full object names in a raw diff
974 format, use --no-abbrev.
975 • format:<format-string>
977 The format:<format-string> format allows you to specify which
979 information you want to show. It works a little bit like printf
980 format, with the notable exception that you get a newline with %n
981 instead of \n.
982 E.g, format:"The author of %h was %an, %ar%nThe title was >>%s<<%n"
984 would show something like this:
985 The author of fe6e0ee was Junio C Hamano, 23 hours ago
987 The title was >>t4119: test autocomputing -p<n> for traditional diff input.<<
988 The placeholders are:
990 • Placeholders that expand to a single literal character:
992 %n
994 newline
995 %%
997 a raw %
998 %x00
1000 %x followed by two hexadecimal digits is replaced with a
1001 byte with the hexadecimal digits' value (we will call this
1002 "literal formatting code" in the rest of this document).
1003 • Placeholders that affect formatting of later placeholders:
1005 %Cred
1007 switch color to red
1008 %Cgreen
1010 switch color to green
1011 %Cblue
1013 switch color to blue
1014 %Creset
1016 reset color
1017 %C(...)
1019 color specification, as described under Values in the
1020 "CONFIGURATION FILE" section of git-config(1). By default,
1021 colors are shown only when enabled for log output (by
1022 color.diff, color.ui, or --color, and respecting the auto
1023 settings of the former if we are going to a terminal).
1024 %C(auto,...) is accepted as a historical synonym for the
1025 default (e.g., %C(auto,red)). Specifying %C(always,...)
1026 will show the colors even when color is not otherwise
1027 enabled (though consider just using --color=always to
1028 enable color for the whole output, including this format
1029 and anything else git might color). auto alone (i.e.
1030 %C(auto)) will turn on auto coloring on the next
1031 placeholders until the color is switched again.
1032 %m
1034 left (<), right (>) or boundary (-) mark
1035 %w([<w>[,<i1>[,<i2>]]])
1037 switch line wrapping, like the -w option of git-
1038 shortlog(1).
1039 %<( <N> [,trunc|ltrunc|mtrunc])
1041 make the next placeholder take at least N column widths,
1042 padding spaces on the right if necessary. Optionally
1043 truncate (with ellipsis ..) at the left (ltrunc) ..ft, the
1044 middle (mtrunc) mi..le, or the end (trunc) rig.., if the
1045 output is longer than N columns. Note 1: that truncating
1046 only works correctly with N >= 2. Note 2: spaces around the
1047 N and M (see below) values are optional. Note 3: Emojis and
1048 other wide characters will take two display columns, which
1049 may over-run column boundaries. Note 4: decomposed
1050 character combining marks may be misplaced at padding
1051 boundaries.
1052 %<|( <M> )
1054 make the next placeholder take at least until Mth display
1055 column, padding spaces on the right if necessary. Use
1056 negative M values for column positions measured from the
1057 right hand edge of the terminal window.
1058 %>( <N> ), %>|( <M> )
1060 similar to %<( <N> ), %<|( <M> ) respectively, but padding
1061 spaces on the left
1062 %>>( <N> ), %>>|( <M> )
1064 similar to %>( <N> ), %>|( <M> ) respectively, except that
1065 if the next placeholder takes more spaces than given and
1066 there are spaces on its left, use those spaces
1067 %><( <N> ), %><|( <M> )
1069 similar to %<( <N> ), %<|( <M> ) respectively, but padding
1070 both sides (i.e. the text is centered)
1071 • Placeholders that expand to information extracted from the
1073 commit:
1074 %H
1076 commit hash
1077 %h
1079 abbreviated commit hash
1080 %T
1082 tree hash
1083 %t
1085 abbreviated tree hash
1086 %P
1088 parent hashes
1089 %p
1091 abbreviated parent hashes
1092 %an
1094 author name
1095 %aN
1097 author name (respecting .mailmap, see git-shortlog(1) or
1098 git-blame(1))
1099 %ae
1101 author email
1102 %aE
1104 author email (respecting .mailmap, see git-shortlog(1) or
1105 git-blame(1))
1106 %al
1108 author email local-part (the part before the @ sign)
1109 %aL
1111 author local-part (see %al) respecting .mailmap, see git-
1112 shortlog(1) or git-blame(1))
1113 %ad
1115 author date (format respects --date= option)
1116 %aD
1118 author date, RFC2822 style
1119 %ar
1121 author date, relative
1122 %at
1124 author date, UNIX timestamp
1125 %ai
1127 author date, ISO 8601-like format
1128 %aI
1130 author date, strict ISO 8601 format
1131 %as
1133 author date, short format (YYYY-MM-DD)
1134 %ah
1136 author date, human style (like the --date=human option of
1137 git-rev-list(1))
1138 %cn
1140 committer name
1141 %cN
1143 committer name (respecting .mailmap, see git-shortlog(1) or
1144 git-blame(1))
1145 %ce
1147 committer email
1148 %cE
1150 committer email (respecting .mailmap, see git-shortlog(1)
1151 or git-blame(1))
1152 %cl
1154 committer email local-part (the part before the @ sign)
1155 %cL
1157 committer local-part (see %cl) respecting .mailmap, see
1158 git-shortlog(1) or git-blame(1))
1159 %cd
1161 committer date (format respects --date= option)
1162 %cD
1164 committer date, RFC2822 style
1165 %cr
1167 committer date, relative
1168 %ct
1170 committer date, UNIX timestamp
1171 %ci
1173 committer date, ISO 8601-like format
1174 %cI
1176 committer date, strict ISO 8601 format
1177 %cs
1179 committer date, short format (YYYY-MM-DD)
1180 %ch
1182 committer date, human style (like the --date=human option
1183 of git-rev-list(1))
1184 %d
1186 ref names, like the --decorate option of git-log(1)
1187 %D
1189 ref names without the " (", ")" wrapping.
1190 %(decorate[:<options>])
1192 ref names with custom decorations. The decorate string may
1193 be followed by a colon and zero or more comma-separated
1194 options. Option values may contain literal formatting
1195 codes. These must be used for commas (%x2C) and closing
1196 parentheses (%x29), due to their role in the option syntax.
1197 • prefix=<value>: Shown before the list of ref names.
1199 Defaults to " (".
1200 • suffix=<value>: Shown after the list of ref names.
1202 Defaults to ")".
1203 • separator=<value>: Shown between ref names. Defaults to
1205 ", ".
1206 • pointer=<value>: Shown between HEAD and the branch it
1208 points to, if any. Defaults to " -> ".
1209 • tag=<value>: Shown before tag names. Defaults to
1211 "tag: ".
1212 For example, to produce decorations with no wrapping or tag
1214 annotations, and spaces as separators:
1215 + %(decorate:prefix=,suffix=,tag=,separator= )
1217 %(describe[:<options>])
1219 human-readable name, like git-describe(1); empty string for
1220 undescribable commits. The describe string may be followed by a
1221 colon and zero or more comma-separated options. Descriptions
1222 can be inconsistent when tags are added or removed at the same
1223 time.
1224 • tags[=<bool-value>]: Instead of only considering annotated
1226 tags, consider lightweight tags as well.
1227 • abbrev=<number>: Instead of using the default number of
1229 hexadecimal digits (which will vary according to the number
1230 of objects in the repository with a default of 7) of the
1231 abbreviated object name, use <number> digits, or as many
1232 digits as needed to form a unique object name.
1233 • match=<pattern>: Only consider tags matching the given
1235 glob(7) pattern, excluding the "refs/tags/" prefix.
1236 • exclude=<pattern>: Do not consider tags matching the given
1238 glob(7) pattern, excluding the "refs/tags/" prefix.
1239 %S
1241 ref name given on the command line by which the commit was
1242 reached (like git log --source), only works with git log
1243 %e
1245 encoding
1246 %s
1248 subject
1249 %f
1251 sanitized subject line, suitable for a filename
1252 %b
1254 body
1255 %B
1257 raw body (unwrapped subject and body)
1258 %N
1260 commit notes
1261 %GG
1263 raw verification message from GPG for a signed commit
1264 %G?
1266 show "G" for a good (valid) signature, "B" for a bad signature,
1267 "U" for a good signature with unknown validity, "X" for a good
1268 signature that has expired, "Y" for a good signature made by an
1269 expired key, "R" for a good signature made by a revoked key,
1270 "E" if the signature cannot be checked (e.g. missing key) and
1271 "N" for no signature
1272 %GS
1274 show the name of the signer for a signed commit
1275 %GK
1277 show the key used to sign a signed commit
1278 %GF
1280 show the fingerprint of the key used to sign a signed commit
1281 %GP
1283 show the fingerprint of the primary key whose subkey was used
1284 to sign a signed commit
1285 %GT
1287 show the trust level for the key used to sign a signed commit
1288 %gD
1290 reflog selector, e.g., refs/stash@{1} or refs/stash@{2 minutes
1291 ago}; the format follows the rules described for the -g option.
1292 The portion before the @ is the refname as given on the command
1293 line (so git log -g refs/heads/master would yield
1294 refs/heads/master@{0}).
1295 %gd
1297 shortened reflog selector; same as %gD, but the refname portion
1298 is shortened for human readability (so refs/heads/master
1299 becomes just master).
1300 %gn
1302 reflog identity name
1303 %gN
1305 reflog identity name (respecting .mailmap, see git-shortlog(1)
1306 or git-blame(1))
1307 %ge
1309 reflog identity email
1310 %gE
1312 reflog identity email (respecting .mailmap, see git-shortlog(1)
1313 or git-blame(1))
1314 %gs
1316 reflog subject
1317 %(trailers[:<options>])
1319 display the trailers of the body as interpreted by git-
1320 interpret-trailers(1). The trailers string may be followed by a
1321 colon and zero or more comma-separated options. If any option
1322 is provided multiple times, the last occurrence wins.
1323 • key=<key>: only show trailers with specified <key>.
1325 Matching is done case-insensitively and trailing colon is
1326 optional. If option is given multiple times trailer lines
1327 matching any of the keys are shown. This option
1328 automatically enables the only option so that non-trailer
1329 lines in the trailer block are hidden. If that is not
1330 desired it can be disabled with only=false. E.g.,
1331 %(trailers:key=Reviewed-by) shows trailer lines with key
1332 Reviewed-by.
1333 • only[=<bool>]: select whether non-trailer lines from the
1335 trailer block should be included.
1336 • separator=<sep>: specify a separator inserted between
1338 trailer lines. When this option is not given each trailer
1339 line is terminated with a line feed character. The string
1340 <sep> may contain the literal formatting codes described
1341 above. To use comma as separator one must use %x2C as it
1342 would otherwise be parsed as next option. E.g.,
1343 %(trailers:key=Ticket,separator=%x2C ) shows all trailer
1344 lines whose key is "Ticket" separated by a comma and a
1345 space.
1346 • unfold[=<bool>]: make it behave as if interpret-trailer’s
1348 --unfold option was given. E.g.,
1349 %(trailers:only,unfold=true) unfolds and shows all trailer
1350 lines.
1351 • keyonly[=<bool>]: only show the key part of the trailer.
1353 • valueonly[=<bool>]: only show the value part of the
1355 trailer.
1356 • key_value_separator=<sep>: specify a separator inserted
1358 between trailer lines. When this option is not given each
1359 trailer key-value pair is separated by ": ". Otherwise it
1360 shares the same semantics as separator=<sep> above.
1361 Note
1363 Some placeholders may depend on other options given to the revision
1364 traversal engine. For example, the %g* reflog options will insert
1365 an empty string unless we are traversing reflog entries (e.g., by
1366 git log -g). The %d and %D placeholders will use the "short"
1367 decoration format if --decorate was not already provided on the
1368 command line.
1369 The boolean options accept an optional value [=<bool-value>]. The
1371 values true, false, on, off etc. are all accepted. See the "boolean"
1372 sub-section in "EXAMPLES" in git-config(1). If a boolean option is
1373 given with no value, it’s enabled.
1374 If you add a + (plus sign) after % of a placeholder, a line-feed is
1376 inserted immediately before the expansion if and only if the
1377 placeholder expands to a non-empty string.
1378 If you add a - (minus sign) after % of a placeholder, all consecutive
1380 line-feeds immediately preceding the expansion are deleted if and only
1381 if the placeholder expands to an empty string.
1382 If you add a ` ` (space) after % of a placeholder, a space is inserted
1384 immediately before the expansion if and only if the placeholder expands
1385 to a non-empty string.
1386 • tformat:
1388 The tformat: format works exactly like format:, except that it
1390 provides "terminator" semantics instead of "separator" semantics.
1391 In other words, each commit has the message terminator character
1392 (usually a newline) appended, rather than a separator placed
1393 between entries. This means that the final entry of a single-line
1394 format will be properly terminated with a new line, just as the
1395 "oneline" format does. For example:
1396 $ git log -2 --pretty=format:%h 4da45bef \
1398 | perl -pe '$_ .= " -- NO NEWLINE\n" unless /\n/'
1399 4da45be
1400 7134973 -- NO NEWLINE
1401 $ git log -2 --pretty=tformat:%h 4da45bef \
1403 | perl -pe '$_ .= " -- NO NEWLINE\n" unless /\n/'
1404 4da45be
1405 7134973
1406 In addition, any unrecognized string that has a % in it is
1408 interpreted as if it has tformat: in front of it. For example,
1409 these two are equivalent:
1410 $ git log -2 --pretty=tformat:%h 4da45bef
1412 $ git log -2 --pretty=%h 4da45bef
1413
1415 The raw output format from "git-diff-index", "git-diff-tree",
1416 "git-diff-files" and "git diff --raw" are very similar.
1417 These commands all compare two sets of things; what is compared
1419 differs:
1420 git-diff-index <tree-ish>
1422 compares the <tree-ish> and the files on the filesystem.
1423 git-diff-index --cached <tree-ish>
1425 compares the <tree-ish> and the index.
1426 git-diff-tree [-r] <tree-ish-1> <tree-ish-2> [<pattern>...]
1428 compares the trees named by the two arguments.
1429 git-diff-files [<pattern>...]
1431 compares the index and the files on the filesystem.
1432 The "git-diff-tree" command begins its output by printing the hash of
1434 what is being compared. After that, all the commands print one output
1435 line per changed file.
1436 An output line is formatted this way:
1438 in-place edit :100644 100644 bcd1234 0123456 M file0
1440 copy-edit :100644 100644 abcd123 1234567 C68 file1 file2
1441 rename-edit :100644 100644 abcd123 1234567 R86 file1 file3
1442 create :000000 100644 0000000 1234567 A file4
1443 delete :100644 000000 1234567 0000000 D file5
1444 unmerged :000000 000000 0000000 0000000 U file6
1445 That is, from the left to the right:
1447 1. a colon.
1449 2. mode for "src"; 000000 if creation or unmerged.
1451 3. a space.
1453 4. mode for "dst"; 000000 if deletion or unmerged.
1455 5. a space.
1457 6. sha1 for "src"; 0{40} if creation or unmerged.
1459 7. a space.
1461 8. sha1 for "dst"; 0{40} if deletion, unmerged or "work tree out of
1463 sync with the index".
1464 9. a space.
1466 10. status, followed by optional "score" number.
1468 11. a tab or a NUL when -z option is used.
1470 12. path for "src"
1472 13. a tab or a NUL when -z option is used; only exists for C or R.
1474 14. path for "dst"; only exists for C or R.
1476 15. an LF or a NUL when -z option is used, to terminate the record.
1478 Possible status letters are:
1480 • A: addition of a file
1482 • C: copy of a file into a new one
1484 • D: deletion of a file
1486 • M: modification of the contents or mode of a file
1488 • R: renaming of a file
1490 • T: change in the type of the file (regular file, symbolic link or
1492 submodule)
1493 • U: file is unmerged (you must complete the merge before it can be
1495 committed)
1496 • X: "unknown" change type (most probably a bug, please report it)
1498 Status letters C and R are always followed by a score (denoting the
1500 percentage of similarity between the source and target of the move or
1501 copy). Status letter M may be followed by a score (denoting the
1502 percentage of dissimilarity) for file rewrites.
1503 The sha1 for "dst" is shown as all 0’s if a file on the filesystem is
1505 out of sync with the index.
1506 Example:
1508 :100644 100644 5be4a4a 0000000 M file.c
1510 Without the -z option, pathnames with "unusual" characters are quoted
1512 as explained for the configuration variable core.quotePath (see git-
1513 config(1)). Using -z the filename is output verbatim and the line is
1514 terminated by a NUL byte.
1515
1517 "git-diff-tree", "git-diff-files" and "git-diff --raw" can take -c or
1518 --cc option to generate diff output also for merge commits. The output
1519 differs from the format described above in the following way:
1520 1. there is a colon for each parent
1522 2. there are more "src" modes and "src" sha1
1524 3. status is concatenated status characters for each parent
1526 4. no optional "score" number
1528 5. tab-separated pathname(s) of the file
1530 For -c and --cc, only the destination or final path is shown even if
1532 the file was renamed on any side of history. With --combined-all-paths,
1533 the name of the path in each parent is shown followed by the name of
1534 the path in the merge commit.
1535 Examples for -c and --cc without --combined-all-paths:
1537 ::100644 100644 100644 fabadb8 cc95eb0 4866510 MM desc.c
1539 ::100755 100755 100755 52b7a2d 6d1ac04 d2ac7d7 RM bar.sh
1540 ::100644 100644 100644 e07d6c5 9042e82 ee91881 RR phooey.c
1541 Examples when --combined-all-paths added to either -c or --cc:
1543 ::100644 100644 100644 fabadb8 cc95eb0 4866510 MM desc.c desc.c desc.c
1545 ::100755 100755 100755 52b7a2d 6d1ac04 d2ac7d7 RM foo.sh bar.sh bar.sh
1546 ::100644 100644 100644 e07d6c5 9042e82 ee91881 RR fooey.c fuey.c phooey.c
1547 Note that combined diff lists only files which were modified from all
1549 parents.
1550
1552 Running git-diff(1), git-log(1), git-show(1), git-diff-index(1), git-
1553 diff-tree(1), or git-diff-files(1) with the -p option produces patch
1554 text. You can customize the creation of patch text via the
1555 GIT_EXTERNAL_DIFF and the GIT_DIFF_OPTS environment variables (see
1556 git(1)), and the diff attribute (see gitattributes(5)).
1557 What the -p option produces is slightly different from the traditional
1559 diff format:
1560 1. It is preceded by a "git diff" header that looks like this:
1562 diff --git a/file1 b/file2
1564 The a/ and b/ filenames are the same unless rename/copy is
1566 involved. Especially, even for a creation or a deletion, /dev/null
1567 is not used in place of the a/ or b/ filenames.
1568 When a rename/copy is involved, file1 and file2 show the name of
1570 the source file of the rename/copy and the name of the file that
1571 the rename/copy produces, respectively.
1572 2. It is followed by one or more extended header lines:
1574 old mode <mode>
1576 new mode <mode>
1577 deleted file mode <mode>
1578 new file mode <mode>
1579 copy from <path>
1580 copy to <path>
1581 rename from <path>
1582 rename to <path>
1583 similarity index <number>
1584 dissimilarity index <number>
1585 index <hash>..<hash> <mode>
1586 File modes are printed as 6-digit octal numbers including the file
1588 type and file permission bits.
1589 Path names in extended headers do not include the a/ and b/
1591 prefixes.
1592 The similarity index is the percentage of unchanged lines, and the
1594 dissimilarity index is the percentage of changed lines. It is a
1595 rounded down integer, followed by a percent sign. The similarity
1596 index value of 100% is thus reserved for two equal files, while
1597 100% dissimilarity means that no line from the old file made it
1598 into the new one.
1599 The index line includes the blob object names before and after the
1601 change. The <mode> is included if the file mode does not change;
1602 otherwise, separate lines indicate the old and the new mode.
1603 3. Pathnames with "unusual" characters are quoted as explained for the
1605 configuration variable core.quotePath (see git-config(1)).
1606 4. All the file1 files in the output refer to files before the commit,
1608 and all the file2 files refer to files after the commit. It is
1609 incorrect to apply each change to each file sequentially. For
1610 example, this patch will swap a and b:
1611 diff --git a/a b/b
1613 rename from a
1614 rename to b
1615 diff --git a/b b/a
1616 rename from b
1617 rename to a
1618 5. Hunk headers mention the name of the function to which the hunk
1620 applies. See "Defining a custom hunk-header" in gitattributes(5)
1621 for details of how to tailor this to specific languages.
1622
1624 Any diff-generating command can take the -c or --cc option to produce a
1625 combined diff when showing a merge. This is the default format when
1626 showing merges with git-diff(1) or git-show(1). Note also that you can
1627 give suitable --diff-merges option to any of these commands to force
1628 generation of diffs in a specific format.
1629 A "combined diff" format looks like this:
1631 diff --combined describe.c
1633 index fabadb8,cc95eb0..4866510
1634 --- a/describe.c
1635 +++ b/describe.c
1636 @@@ -98,20 -98,12 +98,20 @@@
1637 return (a_date > b_date) ? -1 : (a_date == b_date) ? 0 : 1;
1638 }
1639 - static void describe(char *arg)
1641 -static void describe(struct commit *cmit, int last_one)
1642 ++static void describe(char *arg, int last_one)
1643 {
1644 + unsigned char sha1[20];
1645 + struct commit *cmit;
1646 struct commit_list *list;
1647 static int initialized = 0;
1648 struct commit_name *n;
1649 + if (get_sha1(arg, sha1) < 0)
1651 + usage(describe_usage);
1652 + cmit = lookup_commit_reference(sha1);
1653 + if (!cmit)
1654 + usage(describe_usage);
1655 +
1656 if (!initialized) {
1657 initialized = 1;
1658 for_each_ref(get_name);
1659 1. It is preceded by a "git diff" header, that looks like this (when
1661 the -c option is used):
1662 diff --combined file
1664 or like this (when the --cc option is used):
1666 diff --cc file
1668 2. It is followed by one or more extended header lines (this example
1670 shows a merge with two parents):
1671 index <hash>,<hash>..<hash>
1673 mode <mode>,<mode>..<mode>
1674 new file mode <mode>
1675 deleted file mode <mode>,<mode>
1676 The mode <mode>,<mode>..<mode> line appears only if at least one of
1678 the <mode> is different from the rest. Extended headers with
1679 information about detected content movement (renames and copying
1680 detection) are designed to work with the diff of two <tree-ish> and
1681 are not used by combined diff format.
1682 3. It is followed by a two-line from-file/to-file header:
1684 --- a/file
1686 +++ b/file
1687 Similar to the two-line header for the traditional unified diff
1689 format, /dev/null is used to signal created or deleted files.
1690 However, if the --combined-all-paths option is provided, instead of
1692 a two-line from-file/to-file, you get an N+1 line from-file/to-file
1693 header, where N is the number of parents in the merge commit:
1694 --- a/file
1696 --- a/file
1697 --- a/file
1698 +++ b/file
1699 This extended format can be useful if rename or copy detection is
1701 active, to allow you to see the original name of the file in
1702 different parents.
1703 4. Chunk header format is modified to prevent people from accidentally
1705 feeding it to patch -p1. Combined diff format was created for
1706 review of merge commit changes, and was not meant to be applied.
1707 The change is similar to the change in the extended index header:
1708 @@@ <from-file-range> <from-file-range> <to-file-range> @@@
1710 There are (number of parents + 1) @ characters in the chunk header
1712 for combined diff format.
1713 Unlike the traditional unified diff format, which shows two files A and
1715 B with a single column that has - (minus — appears in A but removed in
1716 B), + (plus — missing in A but added to B), or " " (space — unchanged)
1717 prefix, this format compares two or more files file1, file2,... with
1718 one file X, and shows how X differs from each of fileN. One column for
1719 each of fileN is prepended to the output line to note how X’s line is
1720 different from it.
1721 A - character in the column N means that the line appears in fileN but
1723 it does not appear in the result. A + character in the column N means
1724 that the line appears in the result, and fileN does not have that line
1725 (in other words, the line was added, from the point of view of that
1726 parent).
1727 In the above example output, the function signature was changed from
1729 both files (hence two - removals from both file1 and file2, plus ++ to
1730 mean one line that was added does not appear in either file1 or file2).
1731 Also, eight other lines are the same from file1 but do not appear in
1732 file2 (hence prefixed with +).
1733 When shown by git diff-tree -c, it compares the parents of a merge
1735 commit with the merge result (i.e. file1..fileN are the parents). When
1736 shown by git diff-files -c, it compares the two unresolved merge
1737 parents with the working tree file (i.e. file1 is stage 2 aka "our
1738 version", file2 is stage 3 aka "their version").
1739
1741 The --summary option describes newly added, deleted, renamed and copied
1742 files. The --stat option adds diffstat(1) graph to the output. These
1743 options can be combined with other options, such as -p, and are meant
1744 for human consumption.
1745 When showing a change that involves a rename or a copy, --stat output
1747 formats the pathnames compactly by combining common prefix and suffix
1748 of the pathnames. For example, a change that moves arch/i386/Makefile
1749 to arch/x86/Makefile while modifying 4 lines will be shown like this:
1750 arch/{i386 => x86}/Makefile | 4 +--
1752 The --numstat option gives the diffstat(1) information but is designed
1754 for easier machine consumption. An entry in --numstat output looks like
1755 this:
1756 1 2 README
1758 3 1 arch/{i386 => x86}/Makefile
1759 That is, from left to right:
1761 1. the number of added lines;
1763 2. a tab;
1765 3. the number of deleted lines;
1767 4. a tab;
1769 5. pathname (possibly with rename/copy information);
1771 6. a newline.
1773 When -z output option is in effect, the output is formatted this way:
1775 1 2 README NUL
1777 3 1 NUL arch/i386/Makefile NUL arch/x86/Makefile NUL
1778 That is:
1780 1. the number of added lines;
1782 2. a tab;
1784 3. the number of deleted lines;
1786 4. a tab;
1788 5. a NUL (only exists if renamed/copied);
1790 6. pathname in preimage;
1792 7. a NUL (only exists if renamed/copied);
1794 8. pathname in postimage (only exists if renamed/copied);
1796 9. a NUL.
1798 The extra NUL before the preimage path in renamed case is to allow
1800 scripts that read the output to tell if the current record being read
1801 is a single-path record or a rename/copy record without reading ahead.
1802 After reading added and deleted lines, reading up to NUL would yield
1803 the pathname, but if that is NUL, the record will show two paths.
1804
1806 Part of the git(1) suite
1807Git 2.43.0 11/20/2023 GIT-DIFF-TREE(1)