1GIT-DIFF-INDEX(1) Git Manual GIT-DIFF-INDEX(1)
2
6 git-diff-index - Compare a tree to the working tree or index
7
9 git diff-index [-m] [--cached] [<common diff options>] <tree-ish> [<path>...]
10
13 Compares the content and mode of the blobs found in a tree object with
14 the corresponding tracked files in the working tree, or with the
15 corresponding paths in the index. When <path> arguments are present,
16 compares only paths matching those patterns. Otherwise all tracked
17 files are compared.
18
20 -p, -u, --patch
21 Generate patch (see section on generating patches).
22 -s, --no-patch
24 Suppress diff output. Useful for commands like git show that show
25 the patch by default, or to cancel the effect of --patch.
26 -U<n>, --unified=<n>
28 Generate diffs with <n> lines of context instead of the usual
29 three. Implies -p.
30 --raw
32 Generate the diff in raw format. This is the default.
33 --patch-with-raw
35 Synonym for -p --raw.
36 --indent-heuristic
38 Enable the heuristic that shifts diff hunk boundaries to make
39 patches easier to read. This is the default.
40 --no-indent-heuristic
42 Disable the indent heuristic.
43 --minimal
45 Spend extra time to make sure the smallest possible diff is
46 produced.
47 --patience
49 Generate a diff using the "patience diff" algorithm.
50 --histogram
52 Generate a diff using the "histogram diff" algorithm.
53 --anchored=<text>
55 Generate a diff using the "anchored diff" algorithm.
56 This option may be specified more than once.
58 If a line exists in both the source and destination, exists only
60 once, and starts with this text, this algorithm attempts to prevent
61 it from appearing as a deletion or addition in the output. It uses
62 the "patience diff" algorithm internally.
63 --diff-algorithm={patience|minimal|histogram|myers}
65 Choose a diff algorithm. The variants are as follows:
66 default, myers
68 The basic greedy diff algorithm. Currently, this is the
69 default.
70 minimal
72 Spend extra time to make sure the smallest possible diff is
73 produced.
74 patience
76 Use "patience diff" algorithm when generating patches.
77 histogram
79 This algorithm extends the patience algorithm to "support
80 low-occurrence common elements".
81 For instance, if you configured the diff.algorithm variable to a
83 non-default value and want to use the default one, then you have to
84 use --diff-algorithm=default option.
85 --stat[=<width>[,<name-width>[,<count>]]]
87 Generate a diffstat. By default, as much space as necessary will be
88 used for the filename part, and the rest for the graph part.
89 Maximum width defaults to terminal width, or 80 columns if not
90 connected to a terminal, and can be overridden by <width>. The
91 width of the filename part can be limited by giving another width
92 <name-width> after a comma. The width of the graph part can be
93 limited by using --stat-graph-width=<width> (affects all commands
94 generating a stat graph) or by setting diff.statGraphWidth=<width>
95 (does not affect git format-patch). By giving a third parameter
96 <count>, you can limit the output to the first <count> lines,
97 followed by ... if there are more.
98 These parameters can also be set individually with
100 --stat-width=<width>, --stat-name-width=<name-width> and
101 --stat-count=<count>.
102 --compact-summary
104 Output a condensed summary of extended header information such as
105 file creations or deletions ("new" or "gone", optionally "+l" if
106 it’s a symlink) and mode changes ("+x" or "-x" for adding or
107 removing executable bit respectively) in diffstat. The information
108 is put between the filename part and the graph part. Implies
109 --stat.
110 --numstat
112 Similar to --stat, but shows number of added and deleted lines in
113 decimal notation and pathname without abbreviation, to make it more
114 machine friendly. For binary files, outputs two - instead of saying
115 0 0.
116 --shortstat
118 Output only the last line of the --stat format containing total
119 number of modified files, as well as number of added and deleted
120 lines.
121 --dirstat[=<param1,param2,...>]
123 Output the distribution of relative amount of changes for each
124 sub-directory. The behavior of --dirstat can be customized by
125 passing it a comma separated list of parameters. The defaults are
126 controlled by the diff.dirstat configuration variable (see git-
127 config(1)). The following parameters are available:
128 changes
130 Compute the dirstat numbers by counting the lines that have
131 been removed from the source, or added to the destination. This
132 ignores the amount of pure code movements within a file. In
133 other words, rearranging lines in a file is not counted as much
134 as other changes. This is the default behavior when no
135 parameter is given.
136 lines
138 Compute the dirstat numbers by doing the regular line-based
139 diff analysis, and summing the removed/added line counts. (For
140 binary files, count 64-byte chunks instead, since binary files
141 have no natural concept of lines). This is a more expensive
142 --dirstat behavior than the changes behavior, but it does count
143 rearranged lines within a file as much as other changes. The
144 resulting output is consistent with what you get from the other
145 --*stat options.
146 files
148 Compute the dirstat numbers by counting the number of files
149 changed. Each changed file counts equally in the dirstat
150 analysis. This is the computationally cheapest --dirstat
151 behavior, since it does not have to look at the file contents
152 at all.
153 cumulative
155 Count changes in a child directory for the parent directory as
156 well. Note that when using cumulative, the sum of the
157 percentages reported may exceed 100%. The default
158 (non-cumulative) behavior can be specified with the
159 noncumulative parameter.
160 <limit>
162 An integer parameter specifies a cut-off percent (3% by
163 default). Directories contributing less than this percentage of
164 the changes are not shown in the output.
165 Example: The following will count changed files, while ignoring
167 directories with less than 10% of the total amount of changed
168 files, and accumulating child directory counts in the parent
169 directories: --dirstat=files,10,cumulative.
170 --summary
172 Output a condensed summary of extended header information such as
173 creations, renames and mode changes.
174 --patch-with-stat
176 Synonym for -p --stat.
177 -z
179 When --raw, --numstat, --name-only or --name-status has been given,
180 do not munge pathnames and use NULs as output field terminators.
181 Without this option, pathnames with "unusual" characters are quoted
183 as explained for the configuration variable core.quotePath (see
184 git-config(1)).
185 --name-only
187 Show only names of changed files.
188 --name-status
190 Show only names and status of changed files. See the description of
191 the --diff-filter option on what the status letters mean.
192 --submodule[=<format>]
194 Specify how differences in submodules are shown. When specifying
195 --submodule=short the short format is used. This format just shows
196 the names of the commits at the beginning and end of the range.
197 When --submodule or --submodule=log is specified, the log format is
198 used. This format lists the commits in the range like git-
199 submodule(1) summary does. When --submodule=diff is specified, the
200 diff format is used. This format shows an inline diff of the
201 changes in the submodule contents between the commit range.
202 Defaults to diff.submodule or the short format if the config option
203 is unset.
204 --color[=<when>]
206 Show colored diff. --color (i.e. without =<when>) is the same as
207 --color=always. <when> can be one of always, never, or auto.
208 --no-color
210 Turn off colored diff. It is the same as --color=never.
211 --color-moved[=<mode>]
213 Moved lines of code are colored differently. The <mode> defaults to
214 no if the option is not given and to zebra if the option with no
215 mode is given. The mode must be one of:
216 no
218 Moved lines are not highlighted.
219 default
221 Is a synonym for zebra. This may change to a more sensible mode
222 in the future.
223 plain
225 Any line that is added in one location and was removed in
226 another location will be colored with color.diff.newMoved.
227 Similarly color.diff.oldMoved will be used for removed lines
228 that are added somewhere else in the diff. This mode picks up
229 any moved line, but it is not very useful in a review to
230 determine if a block of code was moved without permutation.
231 zebra
233 Blocks of moved text of at least 20 alphanumeric characters are
234 detected greedily. The detected blocks are painted using either
235 the color.diff.{old,new}Moved color or
236 color.diff.{old,new}MovedAlternative. The change between the
237 two colors indicates that a new block was detected.
238 dimmed_zebra
240 Similar to zebra, but additional dimming of uninteresting parts
241 of moved code is performed. The bordering lines of two adjacent
242 blocks are considered interesting, the rest is uninteresting.
243 --word-diff[=<mode>]
245 Show a word diff, using the <mode> to delimit changed words. By
246 default, words are delimited by whitespace; see --word-diff-regex
247 below. The <mode> defaults to plain, and must be one of:
248 color
250 Highlight changed words using only colors. Implies --color.
251 plain
253 Show words as [-removed-] and {+added+}. Makes no attempts to
254 escape the delimiters if they appear in the input, so the
255 output may be ambiguous.
256 porcelain
258 Use a special line-based format intended for script
259 consumption. Added/removed/unchanged runs are printed in the
260 usual unified diff format, starting with a +/-/` ` character at
261 the beginning of the line and extending to the end of the line.
262 Newlines in the input are represented by a tilde ~ on a line of
263 its own.
264 none
266 Disable word diff again.
267 Note that despite the name of the first mode, color is used to
269 highlight the changed parts in all modes if enabled.
270 --word-diff-regex=<regex>
272 Use <regex> to decide what a word is, instead of considering runs
273 of non-whitespace to be a word. Also implies --word-diff unless it
274 was already enabled.
275 Every non-overlapping match of the <regex> is considered a word.
277 Anything between these matches is considered whitespace and
278 ignored(!) for the purposes of finding differences. You may want to
279 append |[^[:space:]] to your regular expression to make sure that
280 it matches all non-whitespace characters. A match that contains a
281 newline is silently truncated(!) at the newline.
282 For example, --word-diff-regex=. will treat each character as a
284 word and, correspondingly, show differences character by character.
285 The regex can also be set via a diff driver or configuration
287 option, see gitattributes(5) or git-config(1). Giving it explicitly
288 overrides any diff driver or configuration setting. Diff drivers
289 override configuration settings.
290 --color-words[=<regex>]
292 Equivalent to --word-diff=color plus (if a regex was specified)
293 --word-diff-regex=<regex>.
294 --no-renames
296 Turn off rename detection, even when the configuration file gives
297 the default to do so.
298 --check
300 Warn if changes introduce conflict markers or whitespace errors.
301 What are considered whitespace errors is controlled by
302 core.whitespace configuration. By default, trailing whitespaces
303 (including lines that consist solely of whitespaces) and a space
304 character that is immediately followed by a tab character inside
305 the initial indent of the line are considered whitespace errors.
306 Exits with non-zero status if problems are found. Not compatible
307 with --exit-code.
308 --ws-error-highlight=<kind>
310 Highlight whitespace errors in the context, old or new lines of the
311 diff. Multiple values are separated by comma, none resets previous
312 values, default reset the list to new and all is a shorthand for
313 old,new,context. When this option is not given, and the
314 configuration variable diff.wsErrorHighlight is not set, only
315 whitespace errors in new lines are highlighted. The whitespace
316 errors are colored with color.diff.whitespace.
317 --full-index
319 Instead of the first handful of characters, show the full pre- and
320 post-image blob object names on the "index" line when generating
321 patch format output.
322 --binary
324 In addition to --full-index, output a binary diff that can be
325 applied with git-apply.
326 --abbrev[=<n>]
328 Instead of showing the full 40-byte hexadecimal object name in
329 diff-raw format output and diff-tree header lines, show only a
330 partial prefix. This is independent of the --full-index option
331 above, which controls the diff-patch output format. Non default
332 number of digits can be specified with --abbrev=<n>.
333 -B[<n>][/<m>], --break-rewrites[=[<n>][/<m>]]
335 Break complete rewrite changes into pairs of delete and create.
336 This serves two purposes:
337 It affects the way a change that amounts to a total rewrite of a
339 file not as a series of deletion and insertion mixed together with
340 a very few lines that happen to match textually as the context, but
341 as a single deletion of everything old followed by a single
342 insertion of everything new, and the number m controls this aspect
343 of the -B option (defaults to 60%). -B/70% specifies that less
344 than 30% of the original should remain in the result for Git to
345 consider it a total rewrite (i.e. otherwise the resulting patch
346 will be a series of deletion and insertion mixed together with
347 context lines).
348 When used with -M, a totally-rewritten file is also considered as
350 the source of a rename (usually -M only considers a file that
351 disappeared as the source of a rename), and the number n controls
352 this aspect of the -B option (defaults to 50%). -B20% specifies
353 that a change with addition and deletion compared to 20% or more of
354 the file’s size are eligible for being picked up as a possible
355 source of a rename to another file.
356 -M[<n>], --find-renames[=<n>]
358 Detect renames. If n is specified, it is a threshold on the
359 similarity index (i.e. amount of addition/deletions compared to the
360 file’s size). For example, -M90% means Git should consider a
361 delete/add pair to be a rename if more than 90% of the file hasn’t
362 changed. Without a % sign, the number is to be read as a fraction,
363 with a decimal point before it. I.e., -M5 becomes 0.5, and is thus
364 the same as -M50%. Similarly, -M05 is the same as -M5%. To limit
365 detection to exact renames, use -M100%. The default similarity
366 index is 50%.
367 -C[<n>], --find-copies[=<n>]
369 Detect copies as well as renames. See also --find-copies-harder. If
370 n is specified, it has the same meaning as for -M<n>.
371 --find-copies-harder
373 For performance reasons, by default, -C option finds copies only if
374 the original file of the copy was modified in the same changeset.
375 This flag makes the command inspect unmodified files as candidates
376 for the source of copy. This is a very expensive operation for
377 large projects, so use it with caution. Giving more than one -C
378 option has the same effect.
379 -D, --irreversible-delete
381 Omit the preimage for deletes, i.e. print only the header but not
382 the diff between the preimage and /dev/null. The resulting patch is
383 not meant to be applied with patch or git apply; this is solely for
384 people who want to just concentrate on reviewing the text after the
385 change. In addition, the output obviously lacks enough information
386 to apply such a patch in reverse, even manually, hence the name of
387 the option.
388 When used together with -B, omit also the preimage in the deletion
390 part of a delete/create pair.
391 -l<num>
393 The -M and -C options require O(n^2) processing time where n is the
394 number of potential rename/copy targets. This option prevents
395 rename/copy detection from running if the number of rename/copy
396 targets exceeds the specified number.
397 --diff-filter=[(A|C|D|M|R|T|U|X|B)...[*]]
399 Select only files that are Added (A), Copied (C), Deleted (D),
400 Modified (M), Renamed (R), have their type (i.e. regular file,
401 symlink, submodule, ...) changed (T), are Unmerged (U), are Unknown
402 (X), or have had their pairing Broken (B). Any combination of the
403 filter characters (including none) can be used. When *
404 (All-or-none) is added to the combination, all paths are selected
405 if there is any file that matches other criteria in the comparison;
406 if there is no file that matches other criteria, nothing is
407 selected.
408 Also, these upper-case letters can be downcased to exclude. E.g.
410 --diff-filter=ad excludes added and deleted paths.
411 Note that not all diffs can feature all types. For instance, diffs
413 from the index to the working tree can never have Added entries
414 (because the set of paths included in the diff is limited by what
415 is in the index). Similarly, copied and renamed entries cannot
416 appear if detection for those types is disabled.
417 -S<string>
419 Look for differences that change the number of occurrences of the
420 specified string (i.e. addition/deletion) in a file. Intended for
421 the scripter’s use.
422 It is useful when you’re looking for an exact block of code (like a
424 struct), and want to know the history of that block since it first
425 came into being: use the feature iteratively to feed the
426 interesting block in the preimage back into -S, and keep going
427 until you get the very first version of the block.
428 -G<regex>
430 Look for differences whose patch text contains added/removed lines
431 that match <regex>.
432 To illustrate the difference between -S<regex> --pickaxe-regex and
434 -G<regex>, consider a commit with the following diff in the same
435 file:
436 + return !regexec(regexp, two->ptr, 1, ®match, 0);
438 ...
439 - hit = !regexec(regexp, mf2.ptr, 1, ®match, 0);
440 While git log -G"regexec\(regexp" will show this commit, git log
442 -S"regexec\(regexp" --pickaxe-regex will not (because the number of
443 occurrences of that string did not change).
444 See the pickaxe entry in gitdiffcore(7) for more information.
446 --find-object=<object-id>
448 Look for differences that change the number of occurrences of the
449 specified object. Similar to -S, just the argument is different in
450 that it doesn’t search for a specific string but for a specific
451 object id.
452 The object can be a blob or a submodule commit. It implies the -t
454 option in git-log to also find trees.
455 --pickaxe-all
457 When -S or -G finds a change, show all the changes in that
458 changeset, not just the files that contain the change in <string>.
459 --pickaxe-regex
461 Treat the <string> given to -S as an extended POSIX regular
462 expression to match.
463 -O<orderfile>
465 Control the order in which files appear in the output. This
466 overrides the diff.orderFile configuration variable (see git-
467 config(1)). To cancel diff.orderFile, use -O/dev/null.
468 The output order is determined by the order of glob patterns in
470 <orderfile>. All files with pathnames that match the first pattern
471 are output first, all files with pathnames that match the second
472 pattern (but not the first) are output next, and so on. All files
473 with pathnames that do not match any pattern are output last, as if
474 there was an implicit match-all pattern at the end of the file. If
475 multiple pathnames have the same rank (they match the same pattern
476 but no earlier patterns), their output order relative to each other
477 is the normal order.
478 <orderfile> is parsed as follows:
480 · Blank lines are ignored, so they can be used as separators for
482 readability.
483 · Lines starting with a hash ("#") are ignored, so they can be
485 used for comments. Add a backslash ("\") to the beginning of
486 the pattern if it starts with a hash.
487 · Each other line contains a single pattern.
489 Patterns have the same syntax and semantics as patterns used for
491 fnmatch(3) without the FNM_PATHNAME flag, except a pathname also
492 matches a pattern if removing any number of the final pathname
493 components matches the pattern. For example, the pattern "foo*bar"
494 matches "fooasdfbar" and "foo/bar/baz/asdf" but not "foobarx".
495 -R
497 Swap two inputs; that is, show differences from index or on-disk
498 file to tree contents.
499 --relative[=<path>]
501 When run from a subdirectory of the project, it can be told to
502 exclude changes outside the directory and show pathnames relative
503 to it with this option. When you are not in a subdirectory (e.g. in
504 a bare repository), you can name which subdirectory to make the
505 output relative to by giving a <path> as an argument.
506 -a, --text
508 Treat all files as text.
509 --ignore-cr-at-eol
511 Ignore carriage-return at the end of line when doing a comparison.
512 --ignore-space-at-eol
514 Ignore changes in whitespace at EOL.
515 -b, --ignore-space-change
517 Ignore changes in amount of whitespace. This ignores whitespace at
518 line end, and considers all other sequences of one or more
519 whitespace characters to be equivalent.
520 -w, --ignore-all-space
522 Ignore whitespace when comparing lines. This ignores differences
523 even if one line has whitespace where the other line has none.
524 --ignore-blank-lines
526 Ignore changes whose lines are all blank.
527 --inter-hunk-context=<lines>
529 Show the context between diff hunks, up to the specified number of
530 lines, thereby fusing hunks that are close to each other. Defaults
531 to diff.interHunkContext or 0 if the config option is unset.
532 -W, --function-context
534 Show whole surrounding functions of changes.
535 --exit-code
537 Make the program exit with codes similar to diff(1). That is, it
538 exits with 1 if there were differences and 0 means no differences.
539 --quiet
541 Disable all output of the program. Implies --exit-code.
542 --ext-diff
544 Allow an external diff helper to be executed. If you set an
545 external diff driver with gitattributes(5), you need to use this
546 option with git-log(1) and friends.
547 --no-ext-diff
549 Disallow external diff drivers.
550 --textconv, --no-textconv
552 Allow (or disallow) external text conversion filters to be run when
553 comparing binary files. See gitattributes(5) for details. Because
554 textconv filters are typically a one-way conversion, the resulting
555 diff is suitable for human consumption, but cannot be applied. For
556 this reason, textconv filters are enabled by default only for git-
557 diff(1) and git-log(1), but not for git-format-patch(1) or diff
558 plumbing commands.
559 --ignore-submodules[=<when>]
561 Ignore changes to submodules in the diff generation. <when> can be
562 either "none", "untracked", "dirty" or "all", which is the default.
563 Using "none" will consider the submodule modified when it either
564 contains untracked or modified files or its HEAD differs from the
565 commit recorded in the superproject and can be used to override any
566 settings of the ignore option in git-config(1) or gitmodules(5).
567 When "untracked" is used submodules are not considered dirty when
568 they only contain untracked content (but they are still scanned for
569 modified content). Using "dirty" ignores all changes to the work
570 tree of submodules, only changes to the commits stored in the
571 superproject are shown (this was the behavior until 1.7.0). Using
572 "all" hides all changes to submodules.
573 --src-prefix=<prefix>
575 Show the given source prefix instead of "a/".
576 --dst-prefix=<prefix>
578 Show the given destination prefix instead of "b/".
579 --no-prefix
581 Do not show any source or destination prefix.
582 --line-prefix=<prefix>
584 Prepend an additional prefix to every line of output.
585 --ita-invisible-in-index
587 By default entries added by "git add -N" appear as an existing
588 empty file in "git diff" and a new file in "git diff --cached".
589 This option makes the entry appear as a new file in "git diff" and
590 non-existent in "git diff --cached". This option could be reverted
591 with --ita-visible-in-index. Both options are experimental and
592 could be removed in future.
593 For more detailed explanation on these common options, see also
595 gitdiffcore(7).
596 <tree-ish>
598 The id of a tree object to diff against.
599 --cached
601 do not consider the on-disk file at all
602 -m
604 By default, files recorded in the index but not checked out are
605 reported as deleted. This flag makes git diff-index say that all
606 non-checked-out files are up to date.
607
609 The raw output format from "git-diff-index", "git-diff-tree",
610 "git-diff-files" and "git diff --raw" are very similar.
611 These commands all compare two sets of things; what is compared
613 differs:
614 git-diff-index <tree-ish>
616 compares the <tree-ish> and the files on the filesystem.
617 git-diff-index --cached <tree-ish>
619 compares the <tree-ish> and the index.
620 git-diff-tree [-r] <tree-ish-1> <tree-ish-2> [<pattern>...]
622 compares the trees named by the two arguments.
623 git-diff-files [<pattern>...]
625 compares the index and the files on the filesystem.
626 The "git-diff-tree" command begins its output by printing the hash of
628 what is being compared. After that, all the commands print one output
629 line per changed file.
630 An output line is formatted this way:
632 in-place edit :100644 100644 bcd1234... 0123456... M file0
634 copy-edit :100644 100644 abcd123... 1234567... C68 file1 file2
635 rename-edit :100644 100644 abcd123... 1234567... R86 file1 file3
636 create :000000 100644 0000000... 1234567... A file4
637 delete :100644 000000 1234567... 0000000... D file5
638 unmerged :000000 000000 0000000... 0000000... U file6
639 That is, from the left to the right:
642 1. a colon.
644 2. mode for "src"; 000000 if creation or unmerged.
646 3. a space.
648 4. mode for "dst"; 000000 if deletion or unmerged.
650 5. a space.
652 6. sha1 for "src"; 0{40} if creation or unmerged.
654 7. a space.
656 8. sha1 for "dst"; 0{40} if creation, unmerged or "look at work tree".
658 9. a space.
660 10. status, followed by optional "score" number.
662 11. a tab or a NUL when -z option is used.
664 12. path for "src"
666 13. a tab or a NUL when -z option is used; only exists for C or R.
668 14. path for "dst"; only exists for C or R.
670 15. an LF or a NUL when -z option is used, to terminate the record.
672 Possible status letters are:
674 · A: addition of a file
676 · C: copy of a file into a new one
678 · D: deletion of a file
680 · M: modification of the contents or mode of a file
682 · R: renaming of a file
684 · T: change in the type of the file
686 · U: file is unmerged (you must complete the merge before it can be
688 committed)
689 · X: "unknown" change type (most probably a bug, please report it)
691 Status letters C and R are always followed by a score (denoting the
693 percentage of similarity between the source and target of the move or
694 copy). Status letter M may be followed by a score (denoting the
695 percentage of dissimilarity) for file rewrites.
696 <sha1> is shown as all 0’s if a file is new on the filesystem and it is
698 out of sync with the index.
699 Example:
701 :100644 100644 5be4a4...... 000000...... M file.c
703 Without the -z option, pathnames with "unusual" characters are quoted
706 as explained for the configuration variable core.quotePath (see git-
707 config(1)). Using -z the filename is output verbatim and the line is
708 terminated by a NUL byte.
709
711 "git-diff-tree", "git-diff-files" and "git-diff --raw" can take -c or
712 --cc option to generate diff output also for merge commits. The output
713 differs from the format described above in the following way:
714 1. there is a colon for each parent
716 2. there are more "src" modes and "src" sha1
718 3. status is concatenated status characters for each parent
720 4. no optional "score" number
722 5. single path, only for "dst"
724 Example:
726 ::100644 100644 100644 fabadb8... cc95eb0... 4866510... MM describe.c
728 Note that combined diff lists only files which were modified from all
731 parents.
732
734 When "git-diff-index", "git-diff-tree", or "git-diff-files" are run
735 with a -p option, "git diff" without the --raw option, or "git log"
736 with the "-p" option, they do not produce the output described above;
737 instead they produce a patch file. You can customize the creation of
738 such patches via the GIT_EXTERNAL_DIFF and the GIT_DIFF_OPTS
739 environment variables.
740 What the -p option produces is slightly different from the traditional
742 diff format:
743 1. It is preceded with a "git diff" header that looks like this:
745 diff --git a/file1 b/file2
747 The a/ and b/ filenames are the same unless rename/copy is
749 involved. Especially, even for a creation or a deletion, /dev/null
750 is not used in place of the a/ or b/ filenames.
751 When rename/copy is involved, file1 and file2 show the name of the
753 source file of the rename/copy and the name of the file that
754 rename/copy produces, respectively.
755 2. It is followed by one or more extended header lines:
757 old mode <mode>
759 new mode <mode>
760 deleted file mode <mode>
761 new file mode <mode>
762 copy from <path>
763 copy to <path>
764 rename from <path>
765 rename to <path>
766 similarity index <number>
767 dissimilarity index <number>
768 index <hash>..<hash> <mode>
769 File modes are printed as 6-digit octal numbers including the file
771 type and file permission bits.
772 Path names in extended headers do not include the a/ and b/
774 prefixes.
775 The similarity index is the percentage of unchanged lines, and the
777 dissimilarity index is the percentage of changed lines. It is a
778 rounded down integer, followed by a percent sign. The similarity
779 index value of 100% is thus reserved for two equal files, while
780 100% dissimilarity means that no line from the old file made it
781 into the new one.
782 The index line includes the SHA-1 checksum before and after the
784 change. The <mode> is included if the file mode does not change;
785 otherwise, separate lines indicate the old and the new mode.
786 3. Pathnames with "unusual" characters are quoted as explained for the
788 configuration variable core.quotePath (see git-config(1)).
789 4. All the file1 files in the output refer to files before the commit,
791 and all the file2 files refer to files after the commit. It is
792 incorrect to apply each change to each file sequentially. For
793 example, this patch will swap a and b:
794 diff --git a/a b/b
796 rename from a
797 rename to b
798 diff --git a/b b/a
799 rename from b
800 rename to a
801
803 Any diff-generating command can take the -c or --cc option to produce a
804 combined diff when showing a merge. This is the default format when
805 showing merges with git-diff(1) or git-show(1). Note also that you can
806 give the -m option to any of these commands to force generation of
807 diffs with individual parents of a merge.
808 A combined diff format looks like this:
810 diff --combined describe.c
812 index fabadb8,cc95eb0..4866510
813 --- a/describe.c
814 +++ b/describe.c
815 @@@ -98,20 -98,12 +98,20 @@@
816 return (a_date > b_date) ? -1 : (a_date == b_date) ? 0 : 1;
817 }
818 - static void describe(char *arg)
820 -static void describe(struct commit *cmit, int last_one)
821 ++static void describe(char *arg, int last_one)
822 {
823 + unsigned char sha1[20];
824 + struct commit *cmit;
825 struct commit_list *list;
826 static int initialized = 0;
827 struct commit_name *n;
828 + if (get_sha1(arg, sha1) < 0)
830 + usage(describe_usage);
831 + cmit = lookup_commit_reference(sha1);
832 + if (!cmit)
833 + usage(describe_usage);
834 +
835 if (!initialized) {
836 initialized = 1;
837 for_each_ref(get_name);
838 1. It is preceded with a "git diff" header, that looks like this (when
842 -c option is used):
843 diff --combined file
845 or like this (when --cc option is used):
847 diff --cc file
849 2. It is followed by one or more extended header lines (this example
851 shows a merge with two parents):
852 index <hash>,<hash>..<hash>
854 mode <mode>,<mode>..<mode>
855 new file mode <mode>
856 deleted file mode <mode>,<mode>
857 The mode <mode>,<mode>..<mode> line appears only if at least one of
859 the <mode> is different from the rest. Extended headers with
860 information about detected contents movement (renames and copying
861 detection) are designed to work with diff of two <tree-ish> and are
862 not used by combined diff format.
863 3. It is followed by two-line from-file/to-file header
865 --- a/file
867 +++ b/file
868 Similar to two-line header for traditional unified diff format,
870 /dev/null is used to signal created or deleted files.
871 4. Chunk header format is modified to prevent people from accidentally
873 feeding it to patch -p1. Combined diff format was created for
874 review of merge commit changes, and was not meant for apply. The
875 change is similar to the change in the extended index header:
876 @@@ <from-file-range> <from-file-range> <to-file-range> @@@
878 There are (number of parents + 1) @ characters in the chunk header
880 for combined diff format.
881 Unlike the traditional unified diff format, which shows two files A and
883 B with a single column that has - (minus — appears in A but removed in
884 B), + (plus — missing in A but added to B), or " " (space — unchanged)
885 prefix, this format compares two or more files file1, file2,... with
886 one file X, and shows how X differs from each of fileN. One column for
887 each of fileN is prepended to the output line to note how X’s line is
888 different from it.
889 A - character in the column N means that the line appears in fileN but
891 it does not appear in the result. A + character in the column N means
892 that the line appears in the result, and fileN does not have that line
893 (in other words, the line was added, from the point of view of that
894 parent).
895 In the above example output, the function signature was changed from
897 both files (hence two - removals from both file1 and file2, plus ++ to
898 mean one line that was added does not appear in either file1 or file2).
899 Also eight other lines are the same from file1 but do not appear in
900 file2 (hence prefixed with +).
901 When shown by git diff-tree -c, it compares the parents of a merge
903 commit with the merge result (i.e. file1..fileN are the parents). When
904 shown by git diff-files -c, it compares the two unresolved merge
905 parents with the working tree file (i.e. file1 is stage 2 aka "our
906 version", file2 is stage 3 aka "their version").
907
909 The --summary option describes newly added, deleted, renamed and copied
910 files. The --stat option adds diffstat(1) graph to the output. These
911 options can be combined with other options, such as -p, and are meant
912 for human consumption.
913 When showing a change that involves a rename or a copy, --stat output
915 formats the pathnames compactly by combining common prefix and suffix
916 of the pathnames. For example, a change that moves arch/i386/Makefile
917 to arch/x86/Makefile while modifying 4 lines will be shown like this:
918 arch/{i386 => x86}/Makefile | 4 +--
920 The --numstat option gives the diffstat(1) information but is designed
923 for easier machine consumption. An entry in --numstat output looks like
924 this:
925 1 2 README
927 3 1 arch/{i386 => x86}/Makefile
928 That is, from left to right:
931 1. the number of added lines;
933 2. a tab;
935 3. the number of deleted lines;
937 4. a tab;
939 5. pathname (possibly with rename/copy information);
941 6. a newline.
943 When -z output option is in effect, the output is formatted this way:
945 1 2 README NUL
947 3 1 NUL arch/i386/Makefile NUL arch/x86/Makefile NUL
948 That is:
951 1. the number of added lines;
953 2. a tab;
955 3. the number of deleted lines;
957 4. a tab;
959 5. a NUL (only exists if renamed/copied);
961 6. pathname in preimage;
963 7. a NUL (only exists if renamed/copied);
965 8. pathname in postimage (only exists if renamed/copied);
967 9. a NUL.
969 The extra NUL before the preimage path in renamed case is to allow
971 scripts that read the output to tell if the current record being read
972 is a single-path record or a rename/copy record without reading ahead.
973 After reading added and deleted lines, reading up to NUL would yield
974 the pathname, but if that is NUL, the record will show two paths.
975
977 You can choose whether you want to trust the index file entirely (using
978 the --cached flag) or ask the diff logic to show any files that don’t
979 match the stat state as being "tentatively changed". Both of these
980 operations are very useful indeed.
981
983 If --cached is specified, it allows you to ask:
984 show me the differences between HEAD and the current index
986 contents (the ones I'd write using 'git write-tree')
987 For example, let’s say that you have worked on your working directory,
989 updated some files in the index and are ready to commit. You want to
990 see exactly what you are going to commit, without having to write a new
991 tree object and compare it that way, and to do that, you just do
992 git diff-index --cached HEAD
994 Example: let’s say I had renamed commit.c to git-commit.c, and I had
996 done an update-index to make that effective in the index file. git
997 diff-files wouldn’t show anything at all, since the index file matches
998 my working directory. But doing a git diff-index does:
999 torvalds@ppc970:~/git> git diff-index --cached HEAD
1001 -100644 blob 4161aecc6700a2eb579e842af0b7f22b98443f74 commit.c
1002 +100644 blob 4161aecc6700a2eb579e842af0b7f22b98443f74 git-commit.c
1003 You can see easily that the above is a rename.
1005 In fact, git diff-index --cached should always be entirely equivalent
1007 to actually doing a git write-tree and comparing that. Except this one
1008 is much nicer for the case where you just want to check where you are.
1009 So doing a git diff-index --cached is basically very useful when you
1011 are asking yourself "what have I already marked for being committed,
1012 and what’s the difference to a previous tree".
1013
1015 The "non-cached" mode takes a different approach, and is potentially
1016 the more useful of the two in that what it does can’t be emulated with
1017 a git write-tree + git diff-tree. Thus that’s the default mode. The
1018 non-cached version asks the question:
1019 show me the differences between HEAD and the currently checked out
1021 tree - index contents _and_ files that aren't up to date
1022 which is obviously a very useful question too, since that tells you
1024 what you could commit. Again, the output matches the git diff-tree -r
1025 output to a tee, but with a twist.
1026 The twist is that if some file doesn’t match the index, we don’t have a
1028 backing store thing for it, and we use the magic "all-zero" sha1 to
1029 show that. So let’s say that you have edited kernel/sched.c, but have
1030 not actually done a git update-index on it yet - there is no "object"
1031 associated with the new state, and you get:
1032 torvalds@ppc970:~/v2.6/linux> git diff-index --abbrev HEAD
1034 :100644 100664 7476bb... 000000... kernel/sched.c
1035 i.e., it shows that the tree has changed, and that kernel/sched.c is
1037 not up to date and may contain new stuff. The all-zero sha1 means that
1038 to get the real diff, you need to look at the object in the working
1039 directory directly rather than do an object-to-object diff.
1040 Note
1042 As with other commands of this type, git diff-index does not
1043 actually look at the contents of the file at all. So maybe
1044 kernel/sched.c hasn’t actually changed, and it’s just that you
1045 touched it. In either case, it’s a note that you need to git
1046 update-index it to make the index be in sync.
1047 Note
1049 You can have a mixture of files show up as "has been updated" and
1050 "is still dirty in the working directory" together. You can always
1051 tell which file is in which state, since the "has been updated"
1052 ones show a valid sha1, and the "not in sync with the index" ones
1053 will always have the special all-zero sha1.
1054
1056 Part of the git(1) suite
1057Git 2.18.1 05/14/2019 GIT-DIFF-INDEX(1)