1GIT-DIFF-FILES(1) Git Manual GIT-DIFF-FILES(1)
2
6 git-diff-files - Compares files in the working tree and the index
7
9 git diff-files [-q] [-0|-1|-2|-3|-c|--cc] [<common diff options>] [<path>...]
10
13 Compares the files in the working tree and the index. When paths are
14 specified, compares only those named paths. Otherwise all entries in
15 the index are compared. The output format is the same as for git
16 diff-index and git diff-tree.
17
19 -p, -u, --patch
20 Generate patch (see section on generating patches).
21 -s, --no-patch
23 Suppress diff output. Useful for commands like git show that show
24 the patch by default, or to cancel the effect of --patch.
25 -U<n>, --unified=<n>
27 Generate diffs with <n> lines of context instead of the usual
28 three. Implies -p.
29 --raw
31 Generate the diff in raw format. This is the default.
32 --patch-with-raw
34 Synonym for -p --raw.
35 --indent-heuristic
37 Enable the heuristic that shifts diff hunk boundaries to make
38 patches easier to read. This is the default.
39 --no-indent-heuristic
41 Disable the indent heuristic.
42 --minimal
44 Spend extra time to make sure the smallest possible diff is
45 produced.
46 --patience
48 Generate a diff using the "patience diff" algorithm.
49 --histogram
51 Generate a diff using the "histogram diff" algorithm.
52 --anchored=<text>
54 Generate a diff using the "anchored diff" algorithm.
55 This option may be specified more than once.
57 If a line exists in both the source and destination, exists only
59 once, and starts with this text, this algorithm attempts to prevent
60 it from appearing as a deletion or addition in the output. It uses
61 the "patience diff" algorithm internally.
62 --diff-algorithm={patience|minimal|histogram|myers}
64 Choose a diff algorithm. The variants are as follows:
65 default, myers
67 The basic greedy diff algorithm. Currently, this is the
68 default.
69 minimal
71 Spend extra time to make sure the smallest possible diff is
72 produced.
73 patience
75 Use "patience diff" algorithm when generating patches.
76 histogram
78 This algorithm extends the patience algorithm to "support
79 low-occurrence common elements".
80 For instance, if you configured the diff.algorithm variable to a
82 non-default value and want to use the default one, then you have to
83 use --diff-algorithm=default option.
84 --stat[=<width>[,<name-width>[,<count>]]]
86 Generate a diffstat. By default, as much space as necessary will be
87 used for the filename part, and the rest for the graph part.
88 Maximum width defaults to terminal width, or 80 columns if not
89 connected to a terminal, and can be overridden by <width>. The
90 width of the filename part can be limited by giving another width
91 <name-width> after a comma. The width of the graph part can be
92 limited by using --stat-graph-width=<width> (affects all commands
93 generating a stat graph) or by setting diff.statGraphWidth=<width>
94 (does not affect git format-patch). By giving a third parameter
95 <count>, you can limit the output to the first <count> lines,
96 followed by ... if there are more.
97 These parameters can also be set individually with
99 --stat-width=<width>, --stat-name-width=<name-width> and
100 --stat-count=<count>.
101 --compact-summary
103 Output a condensed summary of extended header information such as
104 file creations or deletions ("new" or "gone", optionally "+l" if
105 it’s a symlink) and mode changes ("+x" or "-x" for adding or
106 removing executable bit respectively) in diffstat. The information
107 is put between the filename part and the graph part. Implies
108 --stat.
109 --numstat
111 Similar to --stat, but shows number of added and deleted lines in
112 decimal notation and pathname without abbreviation, to make it more
113 machine friendly. For binary files, outputs two - instead of saying
114 0 0.
115 --shortstat
117 Output only the last line of the --stat format containing total
118 number of modified files, as well as number of added and deleted
119 lines.
120 --dirstat[=<param1,param2,...>]
122 Output the distribution of relative amount of changes for each
123 sub-directory. The behavior of --dirstat can be customized by
124 passing it a comma separated list of parameters. The defaults are
125 controlled by the diff.dirstat configuration variable (see git-
126 config(1)). The following parameters are available:
127 changes
129 Compute the dirstat numbers by counting the lines that have
130 been removed from the source, or added to the destination. This
131 ignores the amount of pure code movements within a file. In
132 other words, rearranging lines in a file is not counted as much
133 as other changes. This is the default behavior when no
134 parameter is given.
135 lines
137 Compute the dirstat numbers by doing the regular line-based
138 diff analysis, and summing the removed/added line counts. (For
139 binary files, count 64-byte chunks instead, since binary files
140 have no natural concept of lines). This is a more expensive
141 --dirstat behavior than the changes behavior, but it does count
142 rearranged lines within a file as much as other changes. The
143 resulting output is consistent with what you get from the other
144 --*stat options.
145 files
147 Compute the dirstat numbers by counting the number of files
148 changed. Each changed file counts equally in the dirstat
149 analysis. This is the computationally cheapest --dirstat
150 behavior, since it does not have to look at the file contents
151 at all.
152 cumulative
154 Count changes in a child directory for the parent directory as
155 well. Note that when using cumulative, the sum of the
156 percentages reported may exceed 100%. The default
157 (non-cumulative) behavior can be specified with the
158 noncumulative parameter.
159 <limit>
161 An integer parameter specifies a cut-off percent (3% by
162 default). Directories contributing less than this percentage of
163 the changes are not shown in the output.
164 Example: The following will count changed files, while ignoring
166 directories with less than 10% of the total amount of changed
167 files, and accumulating child directory counts in the parent
168 directories: --dirstat=files,10,cumulative.
169 --summary
171 Output a condensed summary of extended header information such as
172 creations, renames and mode changes.
173 --patch-with-stat
175 Synonym for -p --stat.
176 -z
178 When --raw, --numstat, --name-only or --name-status has been given,
179 do not munge pathnames and use NULs as output field terminators.
180 Without this option, pathnames with "unusual" characters are quoted
182 as explained for the configuration variable core.quotePath (see
183 git-config(1)).
184 --name-only
186 Show only names of changed files.
187 --name-status
189 Show only names and status of changed files. See the description of
190 the --diff-filter option on what the status letters mean.
191 --submodule[=<format>]
193 Specify how differences in submodules are shown. When specifying
194 --submodule=short the short format is used. This format just shows
195 the names of the commits at the beginning and end of the range.
196 When --submodule or --submodule=log is specified, the log format is
197 used. This format lists the commits in the range like git-
198 submodule(1) summary does. When --submodule=diff is specified, the
199 diff format is used. This format shows an inline diff of the
200 changes in the submodule contents between the commit range.
201 Defaults to diff.submodule or the short format if the config option
202 is unset.
203 --color[=<when>]
205 Show colored diff. --color (i.e. without =<when>) is the same as
206 --color=always. <when> can be one of always, never, or auto.
207 --no-color
209 Turn off colored diff. It is the same as --color=never.
210 --color-moved[=<mode>]
212 Moved lines of code are colored differently. The <mode> defaults to
213 no if the option is not given and to zebra if the option with no
214 mode is given. The mode must be one of:
215 no
217 Moved lines are not highlighted.
218 default
220 Is a synonym for zebra. This may change to a more sensible mode
221 in the future.
222 plain
224 Any line that is added in one location and was removed in
225 another location will be colored with color.diff.newMoved.
226 Similarly color.diff.oldMoved will be used for removed lines
227 that are added somewhere else in the diff. This mode picks up
228 any moved line, but it is not very useful in a review to
229 determine if a block of code was moved without permutation.
230 blocks
232 Blocks of moved text of at least 20 alphanumeric characters are
233 detected greedily. The detected blocks are painted using either
234 the color.diff.{old,new}Moved color. Adjacent blocks cannot be
235 told apart.
236 zebra
238 Blocks of moved text are detected as in blocks mode. The blocks
239 are painted using either the color.diff.{old,new}Moved color or
240 color.diff.{old,new}MovedAlternative. The change between the
241 two colors indicates that a new block was detected.
242 dimmed-zebra
244 Similar to zebra, but additional dimming of uninteresting parts
245 of moved code is performed. The bordering lines of two adjacent
246 blocks are considered interesting, the rest is uninteresting.
247 dimmed_zebra is a deprecated synonym.
248 --color-moved-ws=<modes>
250 This configures how white spaces are ignored when performing the
251 move detection for --color-moved. These modes can be given as a
252 comma separated list:
253 ignore-space-at-eol
255 Ignore changes in whitespace at EOL.
256 ignore-space-change
258 Ignore changes in amount of whitespace. This ignores whitespace
259 at line end, and considers all other sequences of one or more
260 whitespace characters to be equivalent.
261 ignore-all-space
263 Ignore whitespace when comparing lines. This ignores
264 differences even if one line has whitespace where the other
265 line has none.
266 allow-indentation-change
268 Initially ignore any white spaces in the move detection, then
269 group the moved code blocks only into a block if the change in
270 whitespace is the same per line. This is incompatible with the
271 other modes.
272 --word-diff[=<mode>]
274 Show a word diff, using the <mode> to delimit changed words. By
275 default, words are delimited by whitespace; see --word-diff-regex
276 below. The <mode> defaults to plain, and must be one of:
277 color
279 Highlight changed words using only colors. Implies --color.
280 plain
282 Show words as [-removed-] and {+added+}. Makes no attempts to
283 escape the delimiters if they appear in the input, so the
284 output may be ambiguous.
285 porcelain
287 Use a special line-based format intended for script
288 consumption. Added/removed/unchanged runs are printed in the
289 usual unified diff format, starting with a +/-/` ` character at
290 the beginning of the line and extending to the end of the line.
291 Newlines in the input are represented by a tilde ~ on a line of
292 its own.
293 none
295 Disable word diff again.
296 Note that despite the name of the first mode, color is used to
298 highlight the changed parts in all modes if enabled.
299 --word-diff-regex=<regex>
301 Use <regex> to decide what a word is, instead of considering runs
302 of non-whitespace to be a word. Also implies --word-diff unless it
303 was already enabled.
304 Every non-overlapping match of the <regex> is considered a word.
306 Anything between these matches is considered whitespace and
307 ignored(!) for the purposes of finding differences. You may want to
308 append |[^[:space:]] to your regular expression to make sure that
309 it matches all non-whitespace characters. A match that contains a
310 newline is silently truncated(!) at the newline.
311 For example, --word-diff-regex=. will treat each character as a
313 word and, correspondingly, show differences character by character.
314 The regex can also be set via a diff driver or configuration
316 option, see gitattributes(5) or git-config(1). Giving it explicitly
317 overrides any diff driver or configuration setting. Diff drivers
318 override configuration settings.
319 --color-words[=<regex>]
321 Equivalent to --word-diff=color plus (if a regex was specified)
322 --word-diff-regex=<regex>.
323 --no-renames
325 Turn off rename detection, even when the configuration file gives
326 the default to do so.
327 --check
329 Warn if changes introduce conflict markers or whitespace errors.
330 What are considered whitespace errors is controlled by
331 core.whitespace configuration. By default, trailing whitespaces
332 (including lines that consist solely of whitespaces) and a space
333 character that is immediately followed by a tab character inside
334 the initial indent of the line are considered whitespace errors.
335 Exits with non-zero status if problems are found. Not compatible
336 with --exit-code.
337 --ws-error-highlight=<kind>
339 Highlight whitespace errors in the context, old or new lines of the
340 diff. Multiple values are separated by comma, none resets previous
341 values, default reset the list to new and all is a shorthand for
342 old,new,context. When this option is not given, and the
343 configuration variable diff.wsErrorHighlight is not set, only
344 whitespace errors in new lines are highlighted. The whitespace
345 errors are colored with color.diff.whitespace.
346 --full-index
348 Instead of the first handful of characters, show the full pre- and
349 post-image blob object names on the "index" line when generating
350 patch format output.
351 --binary
353 In addition to --full-index, output a binary diff that can be
354 applied with git-apply.
355 --abbrev[=<n>]
357 Instead of showing the full 40-byte hexadecimal object name in
358 diff-raw format output and diff-tree header lines, show only a
359 partial prefix. This is independent of the --full-index option
360 above, which controls the diff-patch output format. Non default
361 number of digits can be specified with --abbrev=<n>.
362 -B[<n>][/<m>], --break-rewrites[=[<n>][/<m>]]
364 Break complete rewrite changes into pairs of delete and create.
365 This serves two purposes:
366 It affects the way a change that amounts to a total rewrite of a
368 file not as a series of deletion and insertion mixed together with
369 a very few lines that happen to match textually as the context, but
370 as a single deletion of everything old followed by a single
371 insertion of everything new, and the number m controls this aspect
372 of the -B option (defaults to 60%). -B/70% specifies that less
373 than 30% of the original should remain in the result for Git to
374 consider it a total rewrite (i.e. otherwise the resulting patch
375 will be a series of deletion and insertion mixed together with
376 context lines).
377 When used with -M, a totally-rewritten file is also considered as
379 the source of a rename (usually -M only considers a file that
380 disappeared as the source of a rename), and the number n controls
381 this aspect of the -B option (defaults to 50%). -B20% specifies
382 that a change with addition and deletion compared to 20% or more of
383 the file’s size are eligible for being picked up as a possible
384 source of a rename to another file.
385 -M[<n>], --find-renames[=<n>]
387 Detect renames. If n is specified, it is a threshold on the
388 similarity index (i.e. amount of addition/deletions compared to the
389 file’s size). For example, -M90% means Git should consider a
390 delete/add pair to be a rename if more than 90% of the file hasn’t
391 changed. Without a % sign, the number is to be read as a fraction,
392 with a decimal point before it. I.e., -M5 becomes 0.5, and is thus
393 the same as -M50%. Similarly, -M05 is the same as -M5%. To limit
394 detection to exact renames, use -M100%. The default similarity
395 index is 50%.
396 -C[<n>], --find-copies[=<n>]
398 Detect copies as well as renames. See also --find-copies-harder. If
399 n is specified, it has the same meaning as for -M<n>.
400 --find-copies-harder
402 For performance reasons, by default, -C option finds copies only if
403 the original file of the copy was modified in the same changeset.
404 This flag makes the command inspect unmodified files as candidates
405 for the source of copy. This is a very expensive operation for
406 large projects, so use it with caution. Giving more than one -C
407 option has the same effect.
408 -D, --irreversible-delete
410 Omit the preimage for deletes, i.e. print only the header but not
411 the diff between the preimage and /dev/null. The resulting patch is
412 not meant to be applied with patch or git apply; this is solely for
413 people who want to just concentrate on reviewing the text after the
414 change. In addition, the output obviously lacks enough information
415 to apply such a patch in reverse, even manually, hence the name of
416 the option.
417 When used together with -B, omit also the preimage in the deletion
419 part of a delete/create pair.
420 -l<num>
422 The -M and -C options require O(n^2) processing time where n is the
423 number of potential rename/copy targets. This option prevents
424 rename/copy detection from running if the number of rename/copy
425 targets exceeds the specified number.
426 --diff-filter=[(A|C|D|M|R|T|U|X|B)...[*]]
428 Select only files that are Added (A), Copied (C), Deleted (D),
429 Modified (M), Renamed (R), have their type (i.e. regular file,
430 symlink, submodule, ...) changed (T), are Unmerged (U), are Unknown
431 (X), or have had their pairing Broken (B). Any combination of the
432 filter characters (including none) can be used. When *
433 (All-or-none) is added to the combination, all paths are selected
434 if there is any file that matches other criteria in the comparison;
435 if there is no file that matches other criteria, nothing is
436 selected.
437 Also, these upper-case letters can be downcased to exclude. E.g.
439 --diff-filter=ad excludes added and deleted paths.
440 Note that not all diffs can feature all types. For instance, diffs
442 from the index to the working tree can never have Added entries
443 (because the set of paths included in the diff is limited by what
444 is in the index). Similarly, copied and renamed entries cannot
445 appear if detection for those types is disabled.
446 -S<string>
448 Look for differences that change the number of occurrences of the
449 specified string (i.e. addition/deletion) in a file. Intended for
450 the scripter’s use.
451 It is useful when you’re looking for an exact block of code (like a
453 struct), and want to know the history of that block since it first
454 came into being: use the feature iteratively to feed the
455 interesting block in the preimage back into -S, and keep going
456 until you get the very first version of the block.
457 -G<regex>
459 Look for differences whose patch text contains added/removed lines
460 that match <regex>.
461 To illustrate the difference between -S<regex> --pickaxe-regex and
463 -G<regex>, consider a commit with the following diff in the same
464 file:
465 + return !regexec(regexp, two->ptr, 1, ®match, 0);
467 ...
468 - hit = !regexec(regexp, mf2.ptr, 1, ®match, 0);
469 While git log -G"regexec\(regexp" will show this commit, git log
471 -S"regexec\(regexp" --pickaxe-regex will not (because the number of
472 occurrences of that string did not change).
473 See the pickaxe entry in gitdiffcore(7) for more information.
475 --find-object=<object-id>
477 Look for differences that change the number of occurrences of the
478 specified object. Similar to -S, just the argument is different in
479 that it doesn’t search for a specific string but for a specific
480 object id.
481 The object can be a blob or a submodule commit. It implies the -t
483 option in git-log to also find trees.
484 --pickaxe-all
486 When -S or -G finds a change, show all the changes in that
487 changeset, not just the files that contain the change in <string>.
488 --pickaxe-regex
490 Treat the <string> given to -S as an extended POSIX regular
491 expression to match.
492 -O<orderfile>
494 Control the order in which files appear in the output. This
495 overrides the diff.orderFile configuration variable (see git-
496 config(1)). To cancel diff.orderFile, use -O/dev/null.
497 The output order is determined by the order of glob patterns in
499 <orderfile>. All files with pathnames that match the first pattern
500 are output first, all files with pathnames that match the second
501 pattern (but not the first) are output next, and so on. All files
502 with pathnames that do not match any pattern are output last, as if
503 there was an implicit match-all pattern at the end of the file. If
504 multiple pathnames have the same rank (they match the same pattern
505 but no earlier patterns), their output order relative to each other
506 is the normal order.
507 <orderfile> is parsed as follows:
509 · Blank lines are ignored, so they can be used as separators for
511 readability.
512 · Lines starting with a hash ("#") are ignored, so they can be
514 used for comments. Add a backslash ("\") to the beginning of
515 the pattern if it starts with a hash.
516 · Each other line contains a single pattern.
518 Patterns have the same syntax and semantics as patterns used for
520 fnmatch(3) without the FNM_PATHNAME flag, except a pathname also
521 matches a pattern if removing any number of the final pathname
522 components matches the pattern. For example, the pattern "foo*bar"
523 matches "fooasdfbar" and "foo/bar/baz/asdf" but not "foobarx".
524 -R
526 Swap two inputs; that is, show differences from index or on-disk
527 file to tree contents.
528 --relative[=<path>]
530 When run from a subdirectory of the project, it can be told to
531 exclude changes outside the directory and show pathnames relative
532 to it with this option. When you are not in a subdirectory (e.g. in
533 a bare repository), you can name which subdirectory to make the
534 output relative to by giving a <path> as an argument.
535 -a, --text
537 Treat all files as text.
538 --ignore-cr-at-eol
540 Ignore carriage-return at the end of line when doing a comparison.
541 --ignore-space-at-eol
543 Ignore changes in whitespace at EOL.
544 -b, --ignore-space-change
546 Ignore changes in amount of whitespace. This ignores whitespace at
547 line end, and considers all other sequences of one or more
548 whitespace characters to be equivalent.
549 -w, --ignore-all-space
551 Ignore whitespace when comparing lines. This ignores differences
552 even if one line has whitespace where the other line has none.
553 --ignore-blank-lines
555 Ignore changes whose lines are all blank.
556 --inter-hunk-context=<lines>
558 Show the context between diff hunks, up to the specified number of
559 lines, thereby fusing hunks that are close to each other. Defaults
560 to diff.interHunkContext or 0 if the config option is unset.
561 -W, --function-context
563 Show whole surrounding functions of changes.
564 --exit-code
566 Make the program exit with codes similar to diff(1). That is, it
567 exits with 1 if there were differences and 0 means no differences.
568 --quiet
570 Disable all output of the program. Implies --exit-code.
571 --ext-diff
573 Allow an external diff helper to be executed. If you set an
574 external diff driver with gitattributes(5), you need to use this
575 option with git-log(1) and friends.
576 --no-ext-diff
578 Disallow external diff drivers.
579 --textconv, --no-textconv
581 Allow (or disallow) external text conversion filters to be run when
582 comparing binary files. See gitattributes(5) for details. Because
583 textconv filters are typically a one-way conversion, the resulting
584 diff is suitable for human consumption, but cannot be applied. For
585 this reason, textconv filters are enabled by default only for git-
586 diff(1) and git-log(1), but not for git-format-patch(1) or diff
587 plumbing commands.
588 --ignore-submodules[=<when>]
590 Ignore changes to submodules in the diff generation. <when> can be
591 either "none", "untracked", "dirty" or "all", which is the default.
592 Using "none" will consider the submodule modified when it either
593 contains untracked or modified files or its HEAD differs from the
594 commit recorded in the superproject and can be used to override any
595 settings of the ignore option in git-config(1) or gitmodules(5).
596 When "untracked" is used submodules are not considered dirty when
597 they only contain untracked content (but they are still scanned for
598 modified content). Using "dirty" ignores all changes to the work
599 tree of submodules, only changes to the commits stored in the
600 superproject are shown (this was the behavior until 1.7.0). Using
601 "all" hides all changes to submodules.
602 --src-prefix=<prefix>
604 Show the given source prefix instead of "a/".
605 --dst-prefix=<prefix>
607 Show the given destination prefix instead of "b/".
608 --no-prefix
610 Do not show any source or destination prefix.
611 --line-prefix=<prefix>
613 Prepend an additional prefix to every line of output.
614 --ita-invisible-in-index
616 By default entries added by "git add -N" appear as an existing
617 empty file in "git diff" and a new file in "git diff --cached".
618 This option makes the entry appear as a new file in "git diff" and
619 non-existent in "git diff --cached". This option could be reverted
620 with --ita-visible-in-index. Both options are experimental and
621 could be removed in future.
622 For more detailed explanation on these common options, see also
624 gitdiffcore(7).
625 -1 --base, -2 --ours, -3 --theirs, -0
627 Diff against the "base" version, "our branch" or "their branch"
628 respectively. With these options, diffs for merged entries are not
629 shown.
630 The default is to diff against our branch (-2) and the cleanly
632 resolved paths. The option -0 can be given to omit diff output for
633 unmerged entries and just show "Unmerged".
634 -c, --cc
636 This compares stage 2 (our branch), stage 3 (their branch) and the
637 working tree file and outputs a combined diff, similar to the way
638 diff-tree shows a merge commit with these flags.
639 -q
641 Remain silent even on nonexistent files
642
644 The raw output format from "git-diff-index", "git-diff-tree",
645 "git-diff-files" and "git diff --raw" are very similar.
646 These commands all compare two sets of things; what is compared
648 differs:
649 git-diff-index <tree-ish>
651 compares the <tree-ish> and the files on the filesystem.
652 git-diff-index --cached <tree-ish>
654 compares the <tree-ish> and the index.
655 git-diff-tree [-r] <tree-ish-1> <tree-ish-2> [<pattern>...]
657 compares the trees named by the two arguments.
658 git-diff-files [<pattern>...]
660 compares the index and the files on the filesystem.
661 The "git-diff-tree" command begins its output by printing the hash of
663 what is being compared. After that, all the commands print one output
664 line per changed file.
665 An output line is formatted this way:
667 in-place edit :100644 100644 bcd1234 0123456 M file0
669 copy-edit :100644 100644 abcd123 1234567 C68 file1 file2
670 rename-edit :100644 100644 abcd123 1234567 R86 file1 file3
671 create :000000 100644 0000000 1234567 A file4
672 delete :100644 000000 1234567 0000000 D file5
673 unmerged :000000 000000 0000000 0000000 U file6
674 That is, from the left to the right:
677 1. a colon.
679 2. mode for "src"; 000000 if creation or unmerged.
681 3. a space.
683 4. mode for "dst"; 000000 if deletion or unmerged.
685 5. a space.
687 6. sha1 for "src"; 0{40} if creation or unmerged.
689 7. a space.
691 8. sha1 for "dst"; 0{40} if creation, unmerged or "look at work tree".
693 9. a space.
695 10. status, followed by optional "score" number.
697 11. a tab or a NUL when -z option is used.
699 12. path for "src"
701 13. a tab or a NUL when -z option is used; only exists for C or R.
703 14. path for "dst"; only exists for C or R.
705 15. an LF or a NUL when -z option is used, to terminate the record.
707 Possible status letters are:
709 · A: addition of a file
711 · C: copy of a file into a new one
713 · D: deletion of a file
715 · M: modification of the contents or mode of a file
717 · R: renaming of a file
719 · T: change in the type of the file
721 · U: file is unmerged (you must complete the merge before it can be
723 committed)
724 · X: "unknown" change type (most probably a bug, please report it)
726 Status letters C and R are always followed by a score (denoting the
728 percentage of similarity between the source and target of the move or
729 copy). Status letter M may be followed by a score (denoting the
730 percentage of dissimilarity) for file rewrites.
731 <sha1> is shown as all 0’s if a file is new on the filesystem and it is
733 out of sync with the index.
734 Example:
736 :100644 100644 5be4a4a 0000000 M file.c
738 Without the -z option, pathnames with "unusual" characters are quoted
741 as explained for the configuration variable core.quotePath (see git-
742 config(1)). Using -z the filename is output verbatim and the line is
743 terminated by a NUL byte.
744
746 "git-diff-tree", "git-diff-files" and "git-diff --raw" can take -c or
747 --cc option to generate diff output also for merge commits. The output
748 differs from the format described above in the following way:
749 1. there is a colon for each parent
751 2. there are more "src" modes and "src" sha1
753 3. status is concatenated status characters for each parent
755 4. no optional "score" number
757 5. single path, only for "dst"
759 Example:
761 ::100644 100644 100644 fabadb8 cc95eb0 4866510 MM describe.c
763 Note that combined diff lists only files which were modified from all
766 parents.
767
769 When "git-diff-index", "git-diff-tree", or "git-diff-files" are run
770 with a -p option, "git diff" without the --raw option, or "git log"
771 with the "-p" option, they do not produce the output described above;
772 instead they produce a patch file. You can customize the creation of
773 such patches via the GIT_EXTERNAL_DIFF and the GIT_DIFF_OPTS
774 environment variables.
775 What the -p option produces is slightly different from the traditional
777 diff format:
778 1. It is preceded with a "git diff" header that looks like this:
780 diff --git a/file1 b/file2
782 The a/ and b/ filenames are the same unless rename/copy is
784 involved. Especially, even for a creation or a deletion, /dev/null
785 is not used in place of the a/ or b/ filenames.
786 When rename/copy is involved, file1 and file2 show the name of the
788 source file of the rename/copy and the name of the file that
789 rename/copy produces, respectively.
790 2. It is followed by one or more extended header lines:
792 old mode <mode>
794 new mode <mode>
795 deleted file mode <mode>
796 new file mode <mode>
797 copy from <path>
798 copy to <path>
799 rename from <path>
800 rename to <path>
801 similarity index <number>
802 dissimilarity index <number>
803 index <hash>..<hash> <mode>
804 File modes are printed as 6-digit octal numbers including the file
806 type and file permission bits.
807 Path names in extended headers do not include the a/ and b/
809 prefixes.
810 The similarity index is the percentage of unchanged lines, and the
812 dissimilarity index is the percentage of changed lines. It is a
813 rounded down integer, followed by a percent sign. The similarity
814 index value of 100% is thus reserved for two equal files, while
815 100% dissimilarity means that no line from the old file made it
816 into the new one.
817 The index line includes the SHA-1 checksum before and after the
819 change. The <mode> is included if the file mode does not change;
820 otherwise, separate lines indicate the old and the new mode.
821 3. Pathnames with "unusual" characters are quoted as explained for the
823 configuration variable core.quotePath (see git-config(1)).
824 4. All the file1 files in the output refer to files before the commit,
826 and all the file2 files refer to files after the commit. It is
827 incorrect to apply each change to each file sequentially. For
828 example, this patch will swap a and b:
829 diff --git a/a b/b
831 rename from a
832 rename to b
833 diff --git a/b b/a
834 rename from b
835 rename to a
836
838 Any diff-generating command can take the -c or --cc option to produce a
839 combined diff when showing a merge. This is the default format when
840 showing merges with git-diff(1) or git-show(1). Note also that you can
841 give the -m option to any of these commands to force generation of
842 diffs with individual parents of a merge.
843 A combined diff format looks like this:
845 diff --combined describe.c
847 index fabadb8,cc95eb0..4866510
848 --- a/describe.c
849 +++ b/describe.c
850 @@@ -98,20 -98,12 +98,20 @@@
851 return (a_date > b_date) ? -1 : (a_date == b_date) ? 0 : 1;
852 }
853 - static void describe(char *arg)
855 -static void describe(struct commit *cmit, int last_one)
856 ++static void describe(char *arg, int last_one)
857 {
858 + unsigned char sha1[20];
859 + struct commit *cmit;
860 struct commit_list *list;
861 static int initialized = 0;
862 struct commit_name *n;
863 + if (get_sha1(arg, sha1) < 0)
865 + usage(describe_usage);
866 + cmit = lookup_commit_reference(sha1);
867 + if (!cmit)
868 + usage(describe_usage);
869 +
870 if (!initialized) {
871 initialized = 1;
872 for_each_ref(get_name);
873 1. It is preceded with a "git diff" header, that looks like this (when
877 -c option is used):
878 diff --combined file
880 or like this (when --cc option is used):
882 diff --cc file
884 2. It is followed by one or more extended header lines (this example
886 shows a merge with two parents):
887 index <hash>,<hash>..<hash>
889 mode <mode>,<mode>..<mode>
890 new file mode <mode>
891 deleted file mode <mode>,<mode>
892 The mode <mode>,<mode>..<mode> line appears only if at least one of
894 the <mode> is different from the rest. Extended headers with
895 information about detected contents movement (renames and copying
896 detection) are designed to work with diff of two <tree-ish> and are
897 not used by combined diff format.
898 3. It is followed by two-line from-file/to-file header
900 --- a/file
902 +++ b/file
903 Similar to two-line header for traditional unified diff format,
905 /dev/null is used to signal created or deleted files.
906 4. Chunk header format is modified to prevent people from accidentally
908 feeding it to patch -p1. Combined diff format was created for
909 review of merge commit changes, and was not meant for apply. The
910 change is similar to the change in the extended index header:
911 @@@ <from-file-range> <from-file-range> <to-file-range> @@@
913 There are (number of parents + 1) @ characters in the chunk header
915 for combined diff format.
916 Unlike the traditional unified diff format, which shows two files A and
918 B with a single column that has - (minus — appears in A but removed in
919 B), + (plus — missing in A but added to B), or " " (space — unchanged)
920 prefix, this format compares two or more files file1, file2,... with
921 one file X, and shows how X differs from each of fileN. One column for
922 each of fileN is prepended to the output line to note how X’s line is
923 different from it.
924 A - character in the column N means that the line appears in fileN but
926 it does not appear in the result. A + character in the column N means
927 that the line appears in the result, and fileN does not have that line
928 (in other words, the line was added, from the point of view of that
929 parent).
930 In the above example output, the function signature was changed from
932 both files (hence two - removals from both file1 and file2, plus ++ to
933 mean one line that was added does not appear in either file1 or file2).
934 Also eight other lines are the same from file1 but do not appear in
935 file2 (hence prefixed with +).
936 When shown by git diff-tree -c, it compares the parents of a merge
938 commit with the merge result (i.e. file1..fileN are the parents). When
939 shown by git diff-files -c, it compares the two unresolved merge
940 parents with the working tree file (i.e. file1 is stage 2 aka "our
941 version", file2 is stage 3 aka "their version").
942
944 The --summary option describes newly added, deleted, renamed and copied
945 files. The --stat option adds diffstat(1) graph to the output. These
946 options can be combined with other options, such as -p, and are meant
947 for human consumption.
948 When showing a change that involves a rename or a copy, --stat output
950 formats the pathnames compactly by combining common prefix and suffix
951 of the pathnames. For example, a change that moves arch/i386/Makefile
952 to arch/x86/Makefile while modifying 4 lines will be shown like this:
953 arch/{i386 => x86}/Makefile | 4 +--
955 The --numstat option gives the diffstat(1) information but is designed
958 for easier machine consumption. An entry in --numstat output looks like
959 this:
960 1 2 README
962 3 1 arch/{i386 => x86}/Makefile
963 That is, from left to right:
966 1. the number of added lines;
968 2. a tab;
970 3. the number of deleted lines;
972 4. a tab;
974 5. pathname (possibly with rename/copy information);
976 6. a newline.
978 When -z output option is in effect, the output is formatted this way:
980 1 2 README NUL
982 3 1 NUL arch/i386/Makefile NUL arch/x86/Makefile NUL
983 That is:
986 1. the number of added lines;
988 2. a tab;
990 3. the number of deleted lines;
992 4. a tab;
994 5. a NUL (only exists if renamed/copied);
996 6. pathname in preimage;
998 7. a NUL (only exists if renamed/copied);
1000 8. pathname in postimage (only exists if renamed/copied);
1002 9. a NUL.
1004 The extra NUL before the preimage path in renamed case is to allow
1006 scripts that read the output to tell if the current record being read
1007 is a single-path record or a rename/copy record without reading ahead.
1008 After reading added and deleted lines, reading up to NUL would yield
1009 the pathname, but if that is NUL, the record will show two paths.
1010
1012 Part of the git(1) suite
1013Git 2.20.1 12/15/2018 GIT-DIFF-FILES(1)