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 --patch. Implies -p.
30 --output=<file>
32 Output to a specific file instead of stdout.
33 --output-indicator-new=<char>, --output-indicator-old=<char>,
35 --output-indicator-context=<char>
36 Specify the character used to indicate new, old or context lines in
37 the generated patch. Normally they are +, - and ' ' respectively.
38 --raw
40 Generate the diff in raw format. This is the default.
41 --patch-with-raw
43 Synonym for -p --raw.
44 --indent-heuristic
46 Enable the heuristic that shifts diff hunk boundaries to make
47 patches easier to read. This is the default.
48 --no-indent-heuristic
50 Disable the indent heuristic.
51 --minimal
53 Spend extra time to make sure the smallest possible diff is
54 produced.
55 --patience
57 Generate a diff using the "patience diff" algorithm.
58 --histogram
60 Generate a diff using the "histogram diff" algorithm.
61 --anchored=<text>
63 Generate a diff using the "anchored diff" algorithm.
64 This option may be specified more than once.
66 If a line exists in both the source and destination, exists only
68 once, and starts with this text, this algorithm attempts to prevent
69 it from appearing as a deletion or addition in the output. It uses
70 the "patience diff" algorithm internally.
71 --diff-algorithm={patience|minimal|histogram|myers}
73 Choose a diff algorithm. The variants are as follows:
74 default, myers
76 The basic greedy diff algorithm. Currently, this is the
77 default.
78 minimal
80 Spend extra time to make sure the smallest possible diff is
81 produced.
82 patience
84 Use "patience diff" algorithm when generating patches.
85 histogram
87 This algorithm extends the patience algorithm to "support
88 low-occurrence common elements".
89 For instance, if you configured the diff.algorithm variable to a
91 non-default value and want to use the default one, then you have to
92 use --diff-algorithm=default option.
93 --stat[=<width>[,<name-width>[,<count>]]]
95 Generate a diffstat. By default, as much space as necessary will be
96 used for the filename part, and the rest for the graph part.
97 Maximum width defaults to terminal width, or 80 columns if not
98 connected to a terminal, and can be overridden by <width>. The
99 width of the filename part can be limited by giving another width
100 <name-width> after a comma. The width of the graph part can be
101 limited by using --stat-graph-width=<width> (affects all commands
102 generating a stat graph) or by setting diff.statGraphWidth=<width>
103 (does not affect git format-patch). By giving a third parameter
104 <count>, you can limit the output to the first <count> lines,
105 followed by ... if there are more.
106 These parameters can also be set individually with
108 --stat-width=<width>, --stat-name-width=<name-width> and
109 --stat-count=<count>.
110 --compact-summary
112 Output a condensed summary of extended header information such as
113 file creations or deletions ("new" or "gone", optionally "+l" if
114 it’s a symlink) and mode changes ("+x" or "-x" for adding or
115 removing executable bit respectively) in diffstat. The information
116 is put between the filename part and the graph part. Implies
117 --stat.
118 --numstat
120 Similar to --stat, but shows number of added and deleted lines in
121 decimal notation and pathname without abbreviation, to make it more
122 machine friendly. For binary files, outputs two - instead of saying
123 0 0.
124 --shortstat
126 Output only the last line of the --stat format containing total
127 number of modified files, as well as number of added and deleted
128 lines.
129 -X[<param1,param2,...>], --dirstat[=<param1,param2,...>]
131 Output the distribution of relative amount of changes for each
132 sub-directory. The behavior of --dirstat can be customized by
133 passing it a comma separated list of parameters. The defaults are
134 controlled by the diff.dirstat configuration variable (see git-
135 config(1)). The following parameters are available:
136 changes
138 Compute the dirstat numbers by counting the lines that have
139 been removed from the source, or added to the destination. This
140 ignores the amount of pure code movements within a file. In
141 other words, rearranging lines in a file is not counted as much
142 as other changes. This is the default behavior when no
143 parameter is given.
144 lines
146 Compute the dirstat numbers by doing the regular line-based
147 diff analysis, and summing the removed/added line counts. (For
148 binary files, count 64-byte chunks instead, since binary files
149 have no natural concept of lines). This is a more expensive
150 --dirstat behavior than the changes behavior, but it does count
151 rearranged lines within a file as much as other changes. The
152 resulting output is consistent with what you get from the other
153 --*stat options.
154 files
156 Compute the dirstat numbers by counting the number of files
157 changed. Each changed file counts equally in the dirstat
158 analysis. This is the computationally cheapest --dirstat
159 behavior, since it does not have to look at the file contents
160 at all.
161 cumulative
163 Count changes in a child directory for the parent directory as
164 well. Note that when using cumulative, the sum of the
165 percentages reported may exceed 100%. The default
166 (non-cumulative) behavior can be specified with the
167 noncumulative parameter.
168 <limit>
170 An integer parameter specifies a cut-off percent (3% by
171 default). Directories contributing less than this percentage of
172 the changes are not shown in the output.
173 Example: The following will count changed files, while ignoring
175 directories with less than 10% of the total amount of changed
176 files, and accumulating child directory counts in the parent
177 directories: --dirstat=files,10,cumulative.
178 --cumulative
180 Synonym for --dirstat=cumulative
181 --dirstat-by-file[=<param1,param2>...]
183 Synonym for --dirstat=files,param1,param2...
184 --summary
186 Output a condensed summary of extended header information such as
187 creations, renames and mode changes.
188 --patch-with-stat
190 Synonym for -p --stat.
191 -z
193 When --raw, --numstat, --name-only or --name-status has been given,
194 do not munge pathnames and use NULs as output field terminators.
195 Without this option, pathnames with "unusual" characters are quoted
197 as explained for the configuration variable core.quotePath (see
198 git-config(1)).
199 --name-only
201 Show only names of changed files.
202 --name-status
204 Show only names and status of changed files. See the description of
205 the --diff-filter option on what the status letters mean.
206 --submodule[=<format>]
208 Specify how differences in submodules are shown. When specifying
209 --submodule=short the short format is used. This format just shows
210 the names of the commits at the beginning and end of the range.
211 When --submodule or --submodule=log is specified, the log format is
212 used. This format lists the commits in the range like git-
213 submodule(1) summary does. When --submodule=diff is specified, the
214 diff format is used. This format shows an inline diff of the
215 changes in the submodule contents between the commit range.
216 Defaults to diff.submodule or the short format if the config option
217 is unset.
218 --color[=<when>]
220 Show colored diff. --color (i.e. without =<when>) is the same as
221 --color=always. <when> can be one of always, never, or auto.
222 --no-color
224 Turn off colored diff. It is the same as --color=never.
225 --color-moved[=<mode>]
227 Moved lines of code are colored differently. The <mode> defaults to
228 no if the option is not given and to zebra if the option with no
229 mode is given. The mode must be one of:
230 no
232 Moved lines are not highlighted.
233 default
235 Is a synonym for zebra. This may change to a more sensible mode
236 in the future.
237 plain
239 Any line that is added in one location and was removed in
240 another location will be colored with color.diff.newMoved.
241 Similarly color.diff.oldMoved will be used for removed lines
242 that are added somewhere else in the diff. This mode picks up
243 any moved line, but it is not very useful in a review to
244 determine if a block of code was moved without permutation.
245 blocks
247 Blocks of moved text of at least 20 alphanumeric characters are
248 detected greedily. The detected blocks are painted using either
249 the color.diff.{old,new}Moved color. Adjacent blocks cannot be
250 told apart.
251 zebra
253 Blocks of moved text are detected as in blocks mode. The blocks
254 are painted using either the color.diff.{old,new}Moved color or
255 color.diff.{old,new}MovedAlternative. The change between the
256 two colors indicates that a new block was detected.
257 dimmed-zebra
259 Similar to zebra, but additional dimming of uninteresting parts
260 of moved code is performed. The bordering lines of two adjacent
261 blocks are considered interesting, the rest is uninteresting.
262 dimmed_zebra is a deprecated synonym.
263 --no-color-moved
265 Turn off move detection. This can be used to override configuration
266 settings. It is the same as --color-moved=no.
267 --color-moved-ws=<modes>
269 This configures how whitespace is ignored when performing the move
270 detection for --color-moved. These modes can be given as a comma
271 separated list:
272 no
274 Do not ignore whitespace when performing move detection.
275 ignore-space-at-eol
277 Ignore changes in whitespace at EOL.
278 ignore-space-change
280 Ignore changes in amount of whitespace. This ignores whitespace
281 at line end, and considers all other sequences of one or more
282 whitespace characters to be equivalent.
283 ignore-all-space
285 Ignore whitespace when comparing lines. This ignores
286 differences even if one line has whitespace where the other
287 line has none.
288 allow-indentation-change
290 Initially ignore any whitespace in the move detection, then
291 group the moved code blocks only into a block if the change in
292 whitespace is the same per line. This is incompatible with the
293 other modes.
294 --no-color-moved-ws
296 Do not ignore whitespace when performing move detection. This can
297 be used to override configuration settings. It is the same as
298 --color-moved-ws=no.
299 --word-diff[=<mode>]
301 Show a word diff, using the <mode> to delimit changed words. By
302 default, words are delimited by whitespace; see --word-diff-regex
303 below. The <mode> defaults to plain, and must be one of:
304 color
306 Highlight changed words using only colors. Implies --color.
307 plain
309 Show words as [-removed-] and {+added+}. Makes no attempts to
310 escape the delimiters if they appear in the input, so the
311 output may be ambiguous.
312 porcelain
314 Use a special line-based format intended for script
315 consumption. Added/removed/unchanged runs are printed in the
316 usual unified diff format, starting with a +/-/` ` character at
317 the beginning of the line and extending to the end of the line.
318 Newlines in the input are represented by a tilde ~ on a line of
319 its own.
320 none
322 Disable word diff again.
323 Note that despite the name of the first mode, color is used to
325 highlight the changed parts in all modes if enabled.
326 --word-diff-regex=<regex>
328 Use <regex> to decide what a word is, instead of considering runs
329 of non-whitespace to be a word. Also implies --word-diff unless it
330 was already enabled.
331 Every non-overlapping match of the <regex> is considered a word.
333 Anything between these matches is considered whitespace and
334 ignored(!) for the purposes of finding differences. You may want to
335 append |[^[:space:]] to your regular expression to make sure that
336 it matches all non-whitespace characters. A match that contains a
337 newline is silently truncated(!) at the newline.
338 For example, --word-diff-regex=. will treat each character as a
340 word and, correspondingly, show differences character by character.
341 The regex can also be set via a diff driver or configuration
343 option, see gitattributes(5) or git-config(1). Giving it explicitly
344 overrides any diff driver or configuration setting. Diff drivers
345 override configuration settings.
346 --color-words[=<regex>]
348 Equivalent to --word-diff=color plus (if a regex was specified)
349 --word-diff-regex=<regex>.
350 --no-renames
352 Turn off rename detection, even when the configuration file gives
353 the default to do so.
354 --[no-]rename-empty
356 Whether to use empty blobs as rename source.
357 --check
359 Warn if changes introduce conflict markers or whitespace errors.
360 What are considered whitespace errors is controlled by
361 core.whitespace configuration. By default, trailing whitespaces
362 (including lines that consist solely of whitespaces) and a space
363 character that is immediately followed by a tab character inside
364 the initial indent of the line are considered whitespace errors.
365 Exits with non-zero status if problems are found. Not compatible
366 with --exit-code.
367 --ws-error-highlight=<kind>
369 Highlight whitespace errors in the context, old or new lines of the
370 diff. Multiple values are separated by comma, none resets previous
371 values, default reset the list to new and all is a shorthand for
372 old,new,context. When this option is not given, and the
373 configuration variable diff.wsErrorHighlight is not set, only
374 whitespace errors in new lines are highlighted. The whitespace
375 errors are colored with color.diff.whitespace.
376 --full-index
378 Instead of the first handful of characters, show the full pre- and
379 post-image blob object names on the "index" line when generating
380 patch format output.
381 --binary
383 In addition to --full-index, output a binary diff that can be
384 applied with git-apply. Implies --patch.
385 --abbrev[=<n>]
387 Instead of showing the full 40-byte hexadecimal object name in
388 diff-raw format output and diff-tree header lines, show only a
389 partial prefix. This is independent of the --full-index option
390 above, which controls the diff-patch output format. Non default
391 number of digits can be specified with --abbrev=<n>.
392 -B[<n>][/<m>], --break-rewrites[=[<n>][/<m>]]
394 Break complete rewrite changes into pairs of delete and create.
395 This serves two purposes:
396 It affects the way a change that amounts to a total rewrite of a
398 file not as a series of deletion and insertion mixed together with
399 a very few lines that happen to match textually as the context, but
400 as a single deletion of everything old followed by a single
401 insertion of everything new, and the number m controls this aspect
402 of the -B option (defaults to 60%). -B/70% specifies that less
403 than 30% of the original should remain in the result for Git to
404 consider it a total rewrite (i.e. otherwise the resulting patch
405 will be a series of deletion and insertion mixed together with
406 context lines).
407 When used with -M, a totally-rewritten file is also considered as
409 the source of a rename (usually -M only considers a file that
410 disappeared as the source of a rename), and the number n controls
411 this aspect of the -B option (defaults to 50%). -B20% specifies
412 that a change with addition and deletion compared to 20% or more of
413 the file’s size are eligible for being picked up as a possible
414 source of a rename to another file.
415 -M[<n>], --find-renames[=<n>]
417 Detect renames. If n is specified, it is a threshold on the
418 similarity index (i.e. amount of addition/deletions compared to the
419 file’s size). For example, -M90% means Git should consider a
420 delete/add pair to be a rename if more than 90% of the file hasn’t
421 changed. Without a % sign, the number is to be read as a fraction,
422 with a decimal point before it. I.e., -M5 becomes 0.5, and is thus
423 the same as -M50%. Similarly, -M05 is the same as -M5%. To limit
424 detection to exact renames, use -M100%. The default similarity
425 index is 50%.
426 -C[<n>], --find-copies[=<n>]
428 Detect copies as well as renames. See also --find-copies-harder. If
429 n is specified, it has the same meaning as for -M<n>.
430 --find-copies-harder
432 For performance reasons, by default, -C option finds copies only if
433 the original file of the copy was modified in the same changeset.
434 This flag makes the command inspect unmodified files as candidates
435 for the source of copy. This is a very expensive operation for
436 large projects, so use it with caution. Giving more than one -C
437 option has the same effect.
438 -D, --irreversible-delete
440 Omit the preimage for deletes, i.e. print only the header but not
441 the diff between the preimage and /dev/null. The resulting patch is
442 not meant to be applied with patch or git apply; this is solely for
443 people who want to just concentrate on reviewing the text after the
444 change. In addition, the output obviously lacks enough information
445 to apply such a patch in reverse, even manually, hence the name of
446 the option.
447 When used together with -B, omit also the preimage in the deletion
449 part of a delete/create pair.
450 -l<num>
452 The -M and -C options require O(n^2) processing time where n is the
453 number of potential rename/copy targets. This option prevents
454 rename/copy detection from running if the number of rename/copy
455 targets exceeds the specified number.
456 --diff-filter=[(A|C|D|M|R|T|U|X|B)...[*]]
458 Select only files that are Added (A), Copied (C), Deleted (D),
459 Modified (M), Renamed (R), have their type (i.e. regular file,
460 symlink, submodule, ...) changed (T), are Unmerged (U), are Unknown
461 (X), or have had their pairing Broken (B). Any combination of the
462 filter characters (including none) can be used. When *
463 (All-or-none) is added to the combination, all paths are selected
464 if there is any file that matches other criteria in the comparison;
465 if there is no file that matches other criteria, nothing is
466 selected.
467 Also, these upper-case letters can be downcased to exclude. E.g.
469 --diff-filter=ad excludes added and deleted paths.
470 Note that not all diffs can feature all types. For instance, diffs
472 from the index to the working tree can never have Added entries
473 (because the set of paths included in the diff is limited by what
474 is in the index). Similarly, copied and renamed entries cannot
475 appear if detection for those types is disabled.
476 -S<string>
478 Look for differences that change the number of occurrences of the
479 specified string (i.e. addition/deletion) in a file. Intended for
480 the scripter’s use.
481 It is useful when you’re looking for an exact block of code (like a
483 struct), and want to know the history of that block since it first
484 came into being: use the feature iteratively to feed the
485 interesting block in the preimage back into -S, and keep going
486 until you get the very first version of the block.
487 Binary files are searched as well.
489 -G<regex>
491 Look for differences whose patch text contains added/removed lines
492 that match <regex>.
493 To illustrate the difference between -S<regex> --pickaxe-regex and
495 -G<regex>, consider a commit with the following diff in the same
496 file:
497 + return !regexec(regexp, two->ptr, 1, ®match, 0);
499 ...
500 - hit = !regexec(regexp, mf2.ptr, 1, ®match, 0);
501 While git log -G"regexec\(regexp" will show this commit, git log
503 -S"regexec\(regexp" --pickaxe-regex will not (because the number of
504 occurrences of that string did not change).
505 Unless --text is supplied patches of binary files without a
507 textconv filter will be ignored.
508 See the pickaxe entry in gitdiffcore(7) for more information.
510 --find-object=<object-id>
512 Look for differences that change the number of occurrences of the
513 specified object. Similar to -S, just the argument is different in
514 that it doesn’t search for a specific string but for a specific
515 object id.
516 The object can be a blob or a submodule commit. It implies the -t
518 option in git-log to also find trees.
519 --pickaxe-all
521 When -S or -G finds a change, show all the changes in that
522 changeset, not just the files that contain the change in <string>.
523 --pickaxe-regex
525 Treat the <string> given to -S as an extended POSIX regular
526 expression to match.
527 -O<orderfile>
529 Control the order in which files appear in the output. This
530 overrides the diff.orderFile configuration variable (see git-
531 config(1)). To cancel diff.orderFile, use -O/dev/null.
532 The output order is determined by the order of glob patterns in
534 <orderfile>. All files with pathnames that match the first pattern
535 are output first, all files with pathnames that match the second
536 pattern (but not the first) are output next, and so on. All files
537 with pathnames that do not match any pattern are output last, as if
538 there was an implicit match-all pattern at the end of the file. If
539 multiple pathnames have the same rank (they match the same pattern
540 but no earlier patterns), their output order relative to each other
541 is the normal order.
542 <orderfile> is parsed as follows:
544 · Blank lines are ignored, so they can be used as separators for
546 readability.
547 · Lines starting with a hash ("#") are ignored, so they can be
549 used for comments. Add a backslash ("\") to the beginning of
550 the pattern if it starts with a hash.
551 · Each other line contains a single pattern.
553 Patterns have the same syntax and semantics as patterns used for
555 fnmatch(3) without the FNM_PATHNAME flag, except a pathname also
556 matches a pattern if removing any number of the final pathname
557 components matches the pattern. For example, the pattern "foo*bar"
558 matches "fooasdfbar" and "foo/bar/baz/asdf" but not "foobarx".
559 -R
561 Swap two inputs; that is, show differences from index or on-disk
562 file to tree contents.
563 --relative[=<path>]
565 When run from a subdirectory of the project, it can be told to
566 exclude changes outside the directory and show pathnames relative
567 to it with this option. When you are not in a subdirectory (e.g. in
568 a bare repository), you can name which subdirectory to make the
569 output relative to by giving a <path> as an argument.
570 -a, --text
572 Treat all files as text.
573 --ignore-cr-at-eol
575 Ignore carriage-return at the end of line when doing a comparison.
576 --ignore-space-at-eol
578 Ignore changes in whitespace at EOL.
579 -b, --ignore-space-change
581 Ignore changes in amount of whitespace. This ignores whitespace at
582 line end, and considers all other sequences of one or more
583 whitespace characters to be equivalent.
584 -w, --ignore-all-space
586 Ignore whitespace when comparing lines. This ignores differences
587 even if one line has whitespace where the other line has none.
588 --ignore-blank-lines
590 Ignore changes whose lines are all blank.
591 --inter-hunk-context=<lines>
593 Show the context between diff hunks, up to the specified number of
594 lines, thereby fusing hunks that are close to each other. Defaults
595 to diff.interHunkContext or 0 if the config option is unset.
596 -W, --function-context
598 Show whole surrounding functions of changes.
599 --exit-code
601 Make the program exit with codes similar to diff(1). That is, it
602 exits with 1 if there were differences and 0 means no differences.
603 --quiet
605 Disable all output of the program. Implies --exit-code.
606 --ext-diff
608 Allow an external diff helper to be executed. If you set an
609 external diff driver with gitattributes(5), you need to use this
610 option with git-log(1) and friends.
611 --no-ext-diff
613 Disallow external diff drivers.
614 --textconv, --no-textconv
616 Allow (or disallow) external text conversion filters to be run when
617 comparing binary files. See gitattributes(5) for details. Because
618 textconv filters are typically a one-way conversion, the resulting
619 diff is suitable for human consumption, but cannot be applied. For
620 this reason, textconv filters are enabled by default only for git-
621 diff(1) and git-log(1), but not for git-format-patch(1) or diff
622 plumbing commands.
623 --ignore-submodules[=<when>]
625 Ignore changes to submodules in the diff generation. <when> can be
626 either "none", "untracked", "dirty" or "all", which is the default.
627 Using "none" will consider the submodule modified when it either
628 contains untracked or modified files or its HEAD differs from the
629 commit recorded in the superproject and can be used to override any
630 settings of the ignore option in git-config(1) or gitmodules(5).
631 When "untracked" is used submodules are not considered dirty when
632 they only contain untracked content (but they are still scanned for
633 modified content). Using "dirty" ignores all changes to the work
634 tree of submodules, only changes to the commits stored in the
635 superproject are shown (this was the behavior until 1.7.0). Using
636 "all" hides all changes to submodules.
637 --src-prefix=<prefix>
639 Show the given source prefix instead of "a/".
640 --dst-prefix=<prefix>
642 Show the given destination prefix instead of "b/".
643 --no-prefix
645 Do not show any source or destination prefix.
646 --line-prefix=<prefix>
648 Prepend an additional prefix to every line of output.
649 --ita-invisible-in-index
651 By default entries added by "git add -N" appear as an existing
652 empty file in "git diff" and a new file in "git diff --cached".
653 This option makes the entry appear as a new file in "git diff" and
654 non-existent in "git diff --cached". This option could be reverted
655 with --ita-visible-in-index. Both options are experimental and
656 could be removed in future.
657 For more detailed explanation on these common options, see also
659 gitdiffcore(7).
660 <tree-ish>
662 The id of a tree object to diff against.
663 --cached
665 do not consider the on-disk file at all
666 -m
668 By default, files recorded in the index but not checked out are
669 reported as deleted. This flag makes git diff-index say that all
670 non-checked-out files are up to date.
671
673 The raw output format from "git-diff-index", "git-diff-tree",
674 "git-diff-files" and "git diff --raw" are very similar.
675 These commands all compare two sets of things; what is compared
677 differs:
678 git-diff-index <tree-ish>
680 compares the <tree-ish> and the files on the filesystem.
681 git-diff-index --cached <tree-ish>
683 compares the <tree-ish> and the index.
684 git-diff-tree [-r] <tree-ish-1> <tree-ish-2> [<pattern>...]
686 compares the trees named by the two arguments.
687 git-diff-files [<pattern>...]
689 compares the index and the files on the filesystem.
690 The "git-diff-tree" command begins its output by printing the hash of
692 what is being compared. After that, all the commands print one output
693 line per changed file.
694 An output line is formatted this way:
696 in-place edit :100644 100644 bcd1234 0123456 M file0
698 copy-edit :100644 100644 abcd123 1234567 C68 file1 file2
699 rename-edit :100644 100644 abcd123 1234567 R86 file1 file3
700 create :000000 100644 0000000 1234567 A file4
701 delete :100644 000000 1234567 0000000 D file5
702 unmerged :000000 000000 0000000 0000000 U file6
703 That is, from the left to the right:
706 1. a colon.
708 2. mode for "src"; 000000 if creation or unmerged.
710 3. a space.
712 4. mode for "dst"; 000000 if deletion or unmerged.
714 5. a space.
716 6. sha1 for "src"; 0{40} if creation or unmerged.
718 7. a space.
720 8. sha1 for "dst"; 0{40} if creation, unmerged or "look at work tree".
722 9. a space.
724 10. status, followed by optional "score" number.
726 11. a tab or a NUL when -z option is used.
728 12. path for "src"
730 13. a tab or a NUL when -z option is used; only exists for C or R.
732 14. path for "dst"; only exists for C or R.
734 15. an LF or a NUL when -z option is used, to terminate the record.
736 Possible status letters are:
738 · A: addition of a file
740 · C: copy of a file into a new one
742 · D: deletion of a file
744 · M: modification of the contents or mode of a file
746 · R: renaming of a file
748 · T: change in the type of the file
750 · U: file is unmerged (you must complete the merge before it can be
752 committed)
753 · X: "unknown" change type (most probably a bug, please report it)
755 Status letters C and R are always followed by a score (denoting the
757 percentage of similarity between the source and target of the move or
758 copy). Status letter M may be followed by a score (denoting the
759 percentage of dissimilarity) for file rewrites.
760 <sha1> is shown as all 0’s if a file is new on the filesystem and it is
762 out of sync with the index.
763 Example:
765 :100644 100644 5be4a4a 0000000 M file.c
767 Without the -z option, pathnames with "unusual" characters are quoted
770 as explained for the configuration variable core.quotePath (see git-
771 config(1)). Using -z the filename is output verbatim and the line is
772 terminated by a NUL byte.
773
775 "git-diff-tree", "git-diff-files" and "git-diff --raw" can take -c or
776 --cc option to generate diff output also for merge commits. The output
777 differs from the format described above in the following way:
778 1. there is a colon for each parent
780 2. there are more "src" modes and "src" sha1
782 3. status is concatenated status characters for each parent
784 4. no optional "score" number
786 5. tab-separated pathname(s) of the file
788 For -c and --cc, only the destination or final path is shown even if
790 the file was renamed on any side of history. With --combined-all-paths,
791 the name of the path in each parent is shown followed by the name of
792 the path in the merge commit.
793 Examples for -c and --cc without --combined-all-paths:
795 ::100644 100644 100644 fabadb8 cc95eb0 4866510 MM desc.c
797 ::100755 100755 100755 52b7a2d 6d1ac04 d2ac7d7 RM bar.sh
798 ::100644 100644 100644 e07d6c5 9042e82 ee91881 RR phooey.c
799 Examples when --combined-all-paths added to either -c or --cc:
802 ::100644 100644 100644 fabadb8 cc95eb0 4866510 MM desc.c desc.c desc.c
804 ::100755 100755 100755 52b7a2d 6d1ac04 d2ac7d7 RM foo.sh bar.sh bar.sh
805 ::100644 100644 100644 e07d6c5 9042e82 ee91881 RR fooey.c fuey.c phooey.c
806 Note that combined diff lists only files which were modified from all
809 parents.
810
812 Running git-diff(1), git-log(1), git-show(1), git-diff-index(1), git-
813 diff-tree(1), or git-diff-files(1) with the -p option produces patch
814 text. You can customize the creation of patch text via the
815 GIT_EXTERNAL_DIFF and the GIT_DIFF_OPTS environment variables.
816 What the -p option produces is slightly different from the traditional
818 diff format:
819 1. It is preceded with a "git diff" header that looks like this:
821 diff --git a/file1 b/file2
823 The a/ and b/ filenames are the same unless rename/copy is
825 involved. Especially, even for a creation or a deletion, /dev/null
826 is not used in place of the a/ or b/ filenames.
827 When rename/copy is involved, file1 and file2 show the name of the
829 source file of the rename/copy and the name of the file that
830 rename/copy produces, respectively.
831 2. It is followed by one or more extended header lines:
833 old mode <mode>
835 new mode <mode>
836 deleted file mode <mode>
837 new file mode <mode>
838 copy from <path>
839 copy to <path>
840 rename from <path>
841 rename to <path>
842 similarity index <number>
843 dissimilarity index <number>
844 index <hash>..<hash> <mode>
845 File modes are printed as 6-digit octal numbers including the file
847 type and file permission bits.
848 Path names in extended headers do not include the a/ and b/
850 prefixes.
851 The similarity index is the percentage of unchanged lines, and the
853 dissimilarity index is the percentage of changed lines. It is a
854 rounded down integer, followed by a percent sign. The similarity
855 index value of 100% is thus reserved for two equal files, while
856 100% dissimilarity means that no line from the old file made it
857 into the new one.
858 The index line includes the blob object names before and after the
860 change. The <mode> is included if the file mode does not change;
861 otherwise, separate lines indicate the old and the new mode.
862 3. Pathnames with "unusual" characters are quoted as explained for the
864 configuration variable core.quotePath (see git-config(1)).
865 4. All the file1 files in the output refer to files before the commit,
867 and all the file2 files refer to files after the commit. It is
868 incorrect to apply each change to each file sequentially. For
869 example, this patch will swap a and b:
870 diff --git a/a b/b
872 rename from a
873 rename to b
874 diff --git a/b b/a
875 rename from b
876 rename to a
877
879 Any diff-generating command can take the -c or --cc option to produce a
880 combined diff when showing a merge. This is the default format when
881 showing merges with git-diff(1) or git-show(1). Note also that you can
882 give the -m option to any of these commands to force generation of
883 diffs with individual parents of a merge.
884 A "combined diff" format looks like this:
886 diff --combined describe.c
888 index fabadb8,cc95eb0..4866510
889 --- a/describe.c
890 +++ b/describe.c
891 @@@ -98,20 -98,12 +98,20 @@@
892 return (a_date > b_date) ? -1 : (a_date == b_date) ? 0 : 1;
893 }
894 - static void describe(char *arg)
896 -static void describe(struct commit *cmit, int last_one)
897 ++static void describe(char *arg, int last_one)
898 {
899 + unsigned char sha1[20];
900 + struct commit *cmit;
901 struct commit_list *list;
902 static int initialized = 0;
903 struct commit_name *n;
904 + if (get_sha1(arg, sha1) < 0)
906 + usage(describe_usage);
907 + cmit = lookup_commit_reference(sha1);
908 + if (!cmit)
909 + usage(describe_usage);
910 +
911 if (!initialized) {
912 initialized = 1;
913 for_each_ref(get_name);
914 1. It is preceded with a "git diff" header, that looks like this (when
918 the -c option is used):
919 diff --combined file
921 or like this (when the --cc option is used):
923 diff --cc file
925 2. It is followed by one or more extended header lines (this example
927 shows a merge with two parents):
928 index <hash>,<hash>..<hash>
930 mode <mode>,<mode>..<mode>
931 new file mode <mode>
932 deleted file mode <mode>,<mode>
933 The mode <mode>,<mode>..<mode> line appears only if at least one of
935 the <mode> is different from the rest. Extended headers with
936 information about detected contents movement (renames and copying
937 detection) are designed to work with diff of two <tree-ish> and are
938 not used by combined diff format.
939 3. It is followed by two-line from-file/to-file header
941 --- a/file
943 +++ b/file
944 Similar to two-line header for traditional unified diff format,
946 /dev/null is used to signal created or deleted files.
947 However, if the --combined-all-paths option is provided, instead of
949 a two-line from-file/to-file you get a N+1 line from-file/to-file
950 header, where N is the number of parents in the merge commit
951 --- a/file
953 --- a/file
954 --- a/file
955 +++ b/file
956 This extended format can be useful if rename or copy detection is
958 active, to allow you to see the original name of the file in
959 different parents.
960 4. Chunk header format is modified to prevent people from accidentally
962 feeding it to patch -p1. Combined diff format was created for
963 review of merge commit changes, and was not meant to be applied.
964 The change is similar to the change in the extended index header:
965 @@@ <from-file-range> <from-file-range> <to-file-range> @@@
967 There are (number of parents + 1) @ characters in the chunk header
969 for combined diff format.
970 Unlike the traditional unified diff format, which shows two files A and
972 B with a single column that has - (minus — appears in A but removed in
973 B), + (plus — missing in A but added to B), or " " (space — unchanged)
974 prefix, this format compares two or more files file1, file2,... with
975 one file X, and shows how X differs from each of fileN. One column for
976 each of fileN is prepended to the output line to note how X’s line is
977 different from it.
978 A - character in the column N means that the line appears in fileN but
980 it does not appear in the result. A + character in the column N means
981 that the line appears in the result, and fileN does not have that line
982 (in other words, the line was added, from the point of view of that
983 parent).
984 In the above example output, the function signature was changed from
986 both files (hence two - removals from both file1 and file2, plus ++ to
987 mean one line that was added does not appear in either file1 or file2).
988 Also eight other lines are the same from file1 but do not appear in
989 file2 (hence prefixed with +).
990 When shown by git diff-tree -c, it compares the parents of a merge
992 commit with the merge result (i.e. file1..fileN are the parents). When
993 shown by git diff-files -c, it compares the two unresolved merge
994 parents with the working tree file (i.e. file1 is stage 2 aka "our
995 version", file2 is stage 3 aka "their version").
996
998 The --summary option describes newly added, deleted, renamed and copied
999 files. The --stat option adds diffstat(1) graph to the output. These
1000 options can be combined with other options, such as -p, and are meant
1001 for human consumption.
1002 When showing a change that involves a rename or a copy, --stat output
1004 formats the pathnames compactly by combining common prefix and suffix
1005 of the pathnames. For example, a change that moves arch/i386/Makefile
1006 to arch/x86/Makefile while modifying 4 lines will be shown like this:
1007 arch/{i386 => x86}/Makefile | 4 +--
1009 The --numstat option gives the diffstat(1) information but is designed
1012 for easier machine consumption. An entry in --numstat output looks like
1013 this:
1014 1 2 README
1016 3 1 arch/{i386 => x86}/Makefile
1017 That is, from left to right:
1020 1. the number of added lines;
1022 2. a tab;
1024 3. the number of deleted lines;
1026 4. a tab;
1028 5. pathname (possibly with rename/copy information);
1030 6. a newline.
1032 When -z output option is in effect, the output is formatted this way:
1034 1 2 README NUL
1036 3 1 NUL arch/i386/Makefile NUL arch/x86/Makefile NUL
1037 That is:
1040 1. the number of added lines;
1042 2. a tab;
1044 3. the number of deleted lines;
1046 4. a tab;
1048 5. a NUL (only exists if renamed/copied);
1050 6. pathname in preimage;
1052 7. a NUL (only exists if renamed/copied);
1054 8. pathname in postimage (only exists if renamed/copied);
1056 9. a NUL.
1058 The extra NUL before the preimage path in renamed case is to allow
1060 scripts that read the output to tell if the current record being read
1061 is a single-path record or a rename/copy record without reading ahead.
1062 After reading added and deleted lines, reading up to NUL would yield
1063 the pathname, but if that is NUL, the record will show two paths.
1064
1066 You can choose whether you want to trust the index file entirely (using
1067 the --cached flag) or ask the diff logic to show any files that don’t
1068 match the stat state as being "tentatively changed". Both of these
1069 operations are very useful indeed.
1070
1072 If --cached is specified, it allows you to ask:
1073 show me the differences between HEAD and the current index
1075 contents (the ones I'd write using 'git write-tree')
1076 For example, let’s say that you have worked on your working directory,
1078 updated some files in the index and are ready to commit. You want to
1079 see exactly what you are going to commit, without having to write a new
1080 tree object and compare it that way, and to do that, you just do
1081 git diff-index --cached HEAD
1083 Example: let’s say I had renamed commit.c to git-commit.c, and I had
1085 done an update-index to make that effective in the index file. git
1086 diff-files wouldn’t show anything at all, since the index file matches
1087 my working directory. But doing a git diff-index does:
1088 torvalds@ppc970:~/git> git diff-index --cached HEAD
1090 -100644 blob 4161aecc6700a2eb579e842af0b7f22b98443f74 commit.c
1091 +100644 blob 4161aecc6700a2eb579e842af0b7f22b98443f74 git-commit.c
1092 You can see easily that the above is a rename.
1094 In fact, git diff-index --cached should always be entirely equivalent
1096 to actually doing a git write-tree and comparing that. Except this one
1097 is much nicer for the case where you just want to check where you are.
1098 So doing a git diff-index --cached is basically very useful when you
1100 are asking yourself "what have I already marked for being committed,
1101 and what’s the difference to a previous tree".
1102
1104 The "non-cached" mode takes a different approach, and is potentially
1105 the more useful of the two in that what it does can’t be emulated with
1106 a git write-tree + git diff-tree. Thus that’s the default mode. The
1107 non-cached version asks the question:
1108 show me the differences between HEAD and the currently checked out
1110 tree - index contents _and_ files that aren't up to date
1111 which is obviously a very useful question too, since that tells you
1113 what you could commit. Again, the output matches the git diff-tree -r
1114 output to a tee, but with a twist.
1115 The twist is that if some file doesn’t match the index, we don’t have a
1117 backing store thing for it, and we use the magic "all-zero" sha1 to
1118 show that. So let’s say that you have edited kernel/sched.c, but have
1119 not actually done a git update-index on it yet - there is no "object"
1120 associated with the new state, and you get:
1121 torvalds@ppc970:~/v2.6/linux> git diff-index --abbrev HEAD
1123 :100644 100664 7476bb... 000000... kernel/sched.c
1124 i.e., it shows that the tree has changed, and that kernel/sched.c is
1126 not up to date and may contain new stuff. The all-zero sha1 means that
1127 to get the real diff, you need to look at the object in the working
1128 directory directly rather than do an object-to-object diff.
1129 Note
1131 As with other commands of this type, git diff-index does not
1132 actually look at the contents of the file at all. So maybe
1133 kernel/sched.c hasn’t actually changed, and it’s just that you
1134 touched it. In either case, it’s a note that you need to git
1135 update-index it to make the index be in sync.
1136 Note
1138 You can have a mixture of files show up as "has been updated" and
1139 "is still dirty in the working directory" together. You can always
1140 tell which file is in which state, since the "has been updated"
1141 ones show a valid sha1, and the "not in sync with the index" ones
1142 will always have the special all-zero sha1.
1143
1145 Part of the git(1) suite
1146Git 2.24.1 12/10/2019 GIT-DIFF-INDEX(1)