1 // Please don't remove this comment as asciidoc behaves badly when
2 // the first non-empty line is ifdef/ifndef. The symptom is that
3 // without this comment the <git-diff-core> attribute conditionally
4 // defined below ends up being defined unconditionally.
5 // Last checked with asciidoc 7.0.2.
7 ifndef::git-format-patch[]
13 endif::git-format-patch[]
15 ifdef::git-format-patch[]
18 Generate plain patches without any diffstats.
19 endif::git-format-patch[]
21 ifndef::git-format-patch[]
25 Generate patch (see section on generating patches).
32 Suppress diff output. Useful for commands like `git show` that
33 show the patch by default, or to cancel the effect of `--patch`.
34 endif::git-format-patch[]
38 Generate diffs with <n> lines of context instead of
40 ifndef::git-format-patch[]
42 endif::git-format-patch[]
45 Output to a specific file instead of stdout.
47 --output-indicator-new=<char>::
48 --output-indicator-old=<char>::
49 --output-indicator-context=<char>::
50 Specify the character used to indicate new, old or context
51 lines in the generated patch. Normally they are '+', '-' and
54 ifndef::git-format-patch[]
57 Generate the diff in raw format.
58 ifdef::git-diff-core[]
60 endif::git-diff-core[]
63 For each commit, show a summary of changes using the raw diff
64 format. See the "RAW OUTPUT FORMAT" section of
65 linkgit:git-diff[1]. This is different from showing the log
66 itself in raw format, which you can achieve with
69 endif::git-format-patch[]
71 ifndef::git-format-patch[]
73 Synonym for `-p --raw`.
74 endif::git-format-patch[]
78 Show the tree objects in the diff output.
82 Enable the heuristic that shifts diff hunk boundaries to make patches
83 easier to read. This is the default.
85 --no-indent-heuristic::
86 Disable the indent heuristic.
89 Spend extra time to make sure the smallest possible
93 Generate a diff using the "patience diff" algorithm.
96 Generate a diff using the "histogram diff" algorithm.
99 Generate a diff using the "anchored diff" algorithm.
101 This option may be specified more than once.
103 If a line exists in both the source and destination, exists only once,
104 and starts with this text, this algorithm attempts to prevent it from
105 appearing as a deletion or addition in the output. It uses the "patience
106 diff" algorithm internally.
108 --diff-algorithm={patience|minimal|histogram|myers}::
109 Choose a diff algorithm. The variants are as follows:
113 The basic greedy diff algorithm. Currently, this is the default.
115 Spend extra time to make sure the smallest possible diff is
118 Use "patience diff" algorithm when generating patches.
120 This algorithm extends the patience algorithm to "support
121 low-occurrence common elements".
124 For instance, if you configured the `diff.algorithm` variable to a
125 non-default value and want to use the default one, then you
126 have to use `--diff-algorithm=default` option.
128 --stat[=<width>[,<name-width>[,<count>]]]::
129 Generate a diffstat. By default, as much space as necessary
130 will be used for the filename part, and the rest for the graph
131 part. Maximum width defaults to terminal width, or 80 columns
132 if not connected to a terminal, and can be overridden by
133 `<width>`. The width of the filename part can be limited by
134 giving another width `<name-width>` after a comma. The width
135 of the graph part can be limited by using
136 `--stat-graph-width=<width>` (affects all commands generating
137 a stat graph) or by setting `diff.statGraphWidth=<width>`
138 (does not affect `git format-patch`).
139 By giving a third parameter `<count>`, you can limit the
140 output to the first `<count>` lines, followed by `...` if
143 These parameters can also be set individually with `--stat-width=<width>`,
144 `--stat-name-width=<name-width>` and `--stat-count=<count>`.
147 Output a condensed summary of extended header information such
148 as file creations or deletions ("new" or "gone", optionally "+l"
149 if it's a symlink) and mode changes ("+x" or "-x" for adding
150 or removing executable bit respectively) in diffstat. The
151 information is put between the filename part and the graph
152 part. Implies `--stat`.
155 Similar to `--stat`, but shows number of added and
156 deleted lines in decimal notation and pathname without
157 abbreviation, to make it more machine friendly. For
158 binary files, outputs two `-` instead of saying
162 Output only the last line of the `--stat` format containing total
163 number of modified files, as well as number of added and deleted
166 -X[<param1,param2,...>]::
167 --dirstat[=<param1,param2,...>]::
168 Output the distribution of relative amount of changes for each
169 sub-directory. The behavior of `--dirstat` can be customized by
170 passing it a comma separated list of parameters.
171 The defaults are controlled by the `diff.dirstat` configuration
172 variable (see linkgit:git-config[1]).
173 The following parameters are available:
177 Compute the dirstat numbers by counting the lines that have been
178 removed from the source, or added to the destination. This ignores
179 the amount of pure code movements within a file. In other words,
180 rearranging lines in a file is not counted as much as other changes.
181 This is the default behavior when no parameter is given.
183 Compute the dirstat numbers by doing the regular line-based diff
184 analysis, and summing the removed/added line counts. (For binary
185 files, count 64-byte chunks instead, since binary files have no
186 natural concept of lines). This is a more expensive `--dirstat`
187 behavior than the `changes` behavior, but it does count rearranged
188 lines within a file as much as other changes. The resulting output
189 is consistent with what you get from the other `--*stat` options.
191 Compute the dirstat numbers by counting the number of files changed.
192 Each changed file counts equally in the dirstat analysis. This is
193 the computationally cheapest `--dirstat` behavior, since it does
194 not have to look at the file contents at all.
196 Count changes in a child directory for the parent directory as well.
197 Note that when using `cumulative`, the sum of the percentages
198 reported may exceed 100%. The default (non-cumulative) behavior can
199 be specified with the `noncumulative` parameter.
201 An integer parameter specifies a cut-off percent (3% by default).
202 Directories contributing less than this percentage of the changes
203 are not shown in the output.
206 Example: The following will count changed files, while ignoring
207 directories with less than 10% of the total amount of changed files,
208 and accumulating child directory counts in the parent directories:
209 `--dirstat=files,10,cumulative`.
212 Synonym for --dirstat=cumulative
214 --dirstat-by-file[=<param1,param2>...]::
215 Synonym for --dirstat=files,param1,param2...
218 Output a condensed summary of extended header information
219 such as creations, renames and mode changes.
221 ifndef::git-format-patch[]
223 Synonym for `-p --stat`.
224 endif::git-format-patch[]
226 ifndef::git-format-patch[]
230 Separate the commits with NULs instead of with new newlines.
232 Also, when `--raw` or `--numstat` has been given, do not munge
233 pathnames and use NULs as output field terminators.
236 When `--raw`, `--numstat`, `--name-only` or `--name-status` has been
237 given, do not munge pathnames and use NULs as output field terminators.
240 Without this option, pathnames with "unusual" characters are quoted as
241 explained for the configuration variable `core.quotePath` (see
242 linkgit:git-config[1]).
245 Show only names of changed files.
248 Show only names and status of changed files. See the description
249 of the `--diff-filter` option on what the status letters mean.
251 --submodule[=<format>]::
252 Specify how differences in submodules are shown. When specifying
253 `--submodule=short` the 'short' format is used. This format just
254 shows the names of the commits at the beginning and end of the range.
255 When `--submodule` or `--submodule=log` is specified, the 'log'
256 format is used. This format lists the commits in the range like
257 linkgit:git-submodule[1] `summary` does. When `--submodule=diff`
258 is specified, the 'diff' format is used. This format shows an
259 inline diff of the changes in the submodule contents between the
260 commit range. Defaults to `diff.submodule` or the 'short' format
261 if the config option is unset.
265 `--color` (i.e. without '=<when>') is the same as `--color=always`.
266 '<when>' can be one of `always`, `never`, or `auto`.
268 It can be changed by the `color.ui` and `color.diff`
269 configuration settings.
273 Turn off colored diff.
275 This can be used to override configuration settings.
277 It is the same as `--color=never`.
279 --color-moved[=<mode>]::
280 Moved lines of code are colored differently.
282 It can be changed by the `diff.colorMoved` configuration setting.
284 The <mode> defaults to 'no' if the option is not given
285 and to 'zebra' if the option with no mode is given.
286 The mode must be one of:
290 Moved lines are not highlighted.
292 Is a synonym for `zebra`. This may change to a more sensible mode
295 Any line that is added in one location and was removed
296 in another location will be colored with 'color.diff.newMoved'.
297 Similarly 'color.diff.oldMoved' will be used for removed lines
298 that are added somewhere else in the diff. This mode picks up any
299 moved line, but it is not very useful in a review to determine
300 if a block of code was moved without permutation.
302 Blocks of moved text of at least 20 alphanumeric characters
303 are detected greedily. The detected blocks are
304 painted using either the 'color.diff.{old,new}Moved' color.
305 Adjacent blocks cannot be told apart.
307 Blocks of moved text are detected as in 'blocks' mode. The blocks
308 are painted using either the 'color.diff.{old,new}Moved' color or
309 'color.diff.{old,new}MovedAlternative'. The change between
310 the two colors indicates that a new block was detected.
312 Similar to 'zebra', but additional dimming of uninteresting parts
313 of moved code is performed. The bordering lines of two adjacent
314 blocks are considered interesting, the rest is uninteresting.
315 `dimmed_zebra` is a deprecated synonym.
319 Turn off move detection. This can be used to override configuration
320 settings. It is the same as `--color-moved=no`.
322 --color-moved-ws=<modes>::
323 This configures how whitespace is ignored when performing the
324 move detection for `--color-moved`.
326 It can be set by the `diff.colorMovedWS` configuration setting.
328 These modes can be given as a comma separated list:
332 Do not ignore whitespace when performing move detection.
333 ignore-space-at-eol::
334 Ignore changes in whitespace at EOL.
335 ignore-space-change::
336 Ignore changes in amount of whitespace. This ignores whitespace
337 at line end, and considers all other sequences of one or
338 more whitespace characters to be equivalent.
340 Ignore whitespace when comparing lines. This ignores differences
341 even if one line has whitespace where the other line has none.
342 allow-indentation-change::
343 Initially ignore any whitespace in the move detection, then
344 group the moved code blocks only into a block if the change in
345 whitespace is the same per line. This is incompatible with the
349 --no-color-moved-ws::
350 Do not ignore whitespace when performing move detection. This can be
351 used to override configuration settings. It is the same as
352 `--color-moved-ws=no`.
354 --word-diff[=<mode>]::
355 Show a word diff, using the <mode> to delimit changed words.
356 By default, words are delimited by whitespace; see
357 `--word-diff-regex` below. The <mode> defaults to 'plain', and
362 Highlight changed words using only colors. Implies `--color`.
364 Show words as `[-removed-]` and `{+added+}`. Makes no
365 attempts to escape the delimiters if they appear in the input,
366 so the output may be ambiguous.
368 Use a special line-based format intended for script
369 consumption. Added/removed/unchanged runs are printed in the
370 usual unified diff format, starting with a `+`/`-`/` `
371 character at the beginning of the line and extending to the
372 end of the line. Newlines in the input are represented by a
373 tilde `~` on a line of its own.
375 Disable word diff again.
378 Note that despite the name of the first mode, color is used to
379 highlight the changed parts in all modes if enabled.
381 --word-diff-regex=<regex>::
382 Use <regex> to decide what a word is, instead of considering
383 runs of non-whitespace to be a word. Also implies
384 `--word-diff` unless it was already enabled.
386 Every non-overlapping match of the
387 <regex> is considered a word. Anything between these matches is
388 considered whitespace and ignored(!) for the purposes of finding
389 differences. You may want to append `|[^[:space:]]` to your regular
390 expression to make sure that it matches all non-whitespace characters.
391 A match that contains a newline is silently truncated(!) at the
394 For example, `--word-diff-regex=.` will treat each character as a word
395 and, correspondingly, show differences character by character.
397 The regex can also be set via a diff driver or configuration option, see
398 linkgit:gitattributes[5] or linkgit:git-config[1]. Giving it explicitly
399 overrides any diff driver or configuration setting. Diff drivers
400 override configuration settings.
402 --color-words[=<regex>]::
403 Equivalent to `--word-diff=color` plus (if a regex was
404 specified) `--word-diff-regex=<regex>`.
405 endif::git-format-patch[]
408 Turn off rename detection, even when the configuration
409 file gives the default to do so.
411 --[no-]rename-empty::
412 Whether to use empty blobs as rename source.
414 ifndef::git-format-patch[]
416 Warn if changes introduce conflict markers or whitespace errors.
417 What are considered whitespace errors is controlled by `core.whitespace`
418 configuration. By default, trailing whitespaces (including
419 lines that consist solely of whitespaces) and a space character
420 that is immediately followed by a tab character inside the
421 initial indent of the line are considered whitespace errors.
422 Exits with non-zero status if problems are found. Not compatible
425 --ws-error-highlight=<kind>::
426 Highlight whitespace errors in the `context`, `old` or `new`
427 lines of the diff. Multiple values are separated by comma,
428 `none` resets previous values, `default` reset the list to
429 `new` and `all` is a shorthand for `old,new,context`. When
430 this option is not given, and the configuration variable
431 `diff.wsErrorHighlight` is not set, only whitespace errors in
432 `new` lines are highlighted. The whitespace errors are colored
433 with `color.diff.whitespace`.
435 endif::git-format-patch[]
438 Instead of the first handful of characters, show the full
439 pre- and post-image blob object names on the "index"
440 line when generating patch format output.
443 In addition to `--full-index`, output a binary diff that
444 can be applied with `git-apply`.
445 ifndef::git-format-patch[]
447 endif::git-format-patch[]
450 Instead of showing the full 40-byte hexadecimal object
451 name in diff-raw format output and diff-tree header
452 lines, show the shortest prefix that is at least '<n>'
453 hexdigits long that uniquely refers the object.
454 In diff-patch output format, `--full-index` takes higher
455 precedence, i.e. if `--full-index` is specified, full blob
456 names will be shown regardless of `--abbrev`.
457 Non default number of digits can be specified with `--abbrev=<n>`.
460 --break-rewrites[=[<n>][/<m>]]::
461 Break complete rewrite changes into pairs of delete and
462 create. This serves two purposes:
464 It affects the way a change that amounts to a total rewrite of a file
465 not as a series of deletion and insertion mixed together with a very
466 few lines that happen to match textually as the context, but as a
467 single deletion of everything old followed by a single insertion of
468 everything new, and the number `m` controls this aspect of the -B
469 option (defaults to 60%). `-B/70%` specifies that less than 30% of the
470 original should remain in the result for Git to consider it a total
471 rewrite (i.e. otherwise the resulting patch will be a series of
472 deletion and insertion mixed together with context lines).
474 When used with -M, a totally-rewritten file is also considered as the
475 source of a rename (usually -M only considers a file that disappeared
476 as the source of a rename), and the number `n` controls this aspect of
477 the -B option (defaults to 50%). `-B20%` specifies that a change with
478 addition and deletion compared to 20% or more of the file's size are
479 eligible for being picked up as a possible source of a rename to
483 --find-renames[=<n>]::
488 If generating diffs, detect and report renames for each commit.
489 For following files across renames while traversing history, see
492 If `n` is specified, it is a threshold on the similarity
493 index (i.e. amount of addition/deletions compared to the
494 file's size). For example, `-M90%` means Git should consider a
495 delete/add pair to be a rename if more than 90% of the file
496 hasn't changed. Without a `%` sign, the number is to be read as
497 a fraction, with a decimal point before it. I.e., `-M5` becomes
498 0.5, and is thus the same as `-M50%`. Similarly, `-M05` is
499 the same as `-M5%`. To limit detection to exact renames, use
500 `-M100%`. The default similarity index is 50%.
503 --find-copies[=<n>]::
504 Detect copies as well as renames. See also `--find-copies-harder`.
505 If `n` is specified, it has the same meaning as for `-M<n>`.
507 --find-copies-harder::
508 For performance reasons, by default, `-C` option finds copies only
509 if the original file of the copy was modified in the same
510 changeset. This flag makes the command
511 inspect unmodified files as candidates for the source of
512 copy. This is a very expensive operation for large
513 projects, so use it with caution. Giving more than one
514 `-C` option has the same effect.
517 --irreversible-delete::
518 Omit the preimage for deletes, i.e. print only the header but not
519 the diff between the preimage and `/dev/null`. The resulting patch
520 is not meant to be applied with `patch` or `git apply`; this is
521 solely for people who want to just concentrate on reviewing the
522 text after the change. In addition, the output obviously lacks
523 enough information to apply such a patch in reverse, even manually,
524 hence the name of the option.
526 When used together with `-B`, omit also the preimage in the deletion part
527 of a delete/create pair.
530 The `-M` and `-C` options require O(n^2) processing time where n
531 is the number of potential rename/copy targets. This
532 option prevents rename/copy detection from running if
533 the number of rename/copy targets exceeds the specified
536 ifndef::git-format-patch[]
537 --diff-filter=[(A|C|D|M|R|T|U|X|B)...[*]]::
538 Select only files that are Added (`A`), Copied (`C`),
539 Deleted (`D`), Modified (`M`), Renamed (`R`), have their
540 type (i.e. regular file, symlink, submodule, ...) changed (`T`),
541 are Unmerged (`U`), are
542 Unknown (`X`), or have had their pairing Broken (`B`).
543 Any combination of the filter characters (including none) can be used.
544 When `*` (All-or-none) is added to the combination, all
545 paths are selected if there is any file that matches
546 other criteria in the comparison; if there is no file
547 that matches other criteria, nothing is selected.
549 Also, these upper-case letters can be downcased to exclude. E.g.
550 `--diff-filter=ad` excludes added and deleted paths.
552 Note that not all diffs can feature all types. For instance, diffs
553 from the index to the working tree can never have Added entries
554 (because the set of paths included in the diff is limited by what is in
555 the index). Similarly, copied and renamed entries cannot appear if
556 detection for those types is disabled.
559 Look for differences that change the number of occurrences of
560 the specified string (i.e. addition/deletion) in a file.
561 Intended for the scripter's use.
563 It is useful when you're looking for an exact block of code (like a
564 struct), and want to know the history of that block since it first
565 came into being: use the feature iteratively to feed the interesting
566 block in the preimage back into `-S`, and keep going until you get the
567 very first version of the block.
569 Binary files are searched as well.
572 Look for differences whose patch text contains added/removed
573 lines that match <regex>.
575 To illustrate the difference between `-S<regex> --pickaxe-regex` and
576 `-G<regex>`, consider a commit with the following diff in the same
580 + return frotz(nitfol, two->ptr, 1, 0);
582 - hit = frotz(nitfol, mf2.ptr, 1, 0);
585 While `git log -G"frotz\(nitfol"` will show this commit, `git log
586 -S"frotz\(nitfol" --pickaxe-regex` will not (because the number of
587 occurrences of that string did not change).
589 Unless `--text` is supplied patches of binary files without a textconv
590 filter will be ignored.
592 See the 'pickaxe' entry in linkgit:gitdiffcore[7] for more
595 --find-object=<object-id>::
596 Look for differences that change the number of occurrences of
597 the specified object. Similar to `-S`, just the argument is different
598 in that it doesn't search for a specific string but for a specific
601 The object can be a blob or a submodule commit. It implies the `-t` option in
602 `git-log` to also find trees.
605 When `-S` or `-G` finds a change, show all the changes in that
606 changeset, not just the files that contain the change
610 Treat the <string> given to `-S` as an extended POSIX regular
613 endif::git-format-patch[]
616 Control the order in which files appear in the output.
617 This overrides the `diff.orderFile` configuration variable
618 (see linkgit:git-config[1]). To cancel `diff.orderFile`,
621 The output order is determined by the order of glob patterns in
623 All files with pathnames that match the first pattern are output
624 first, all files with pathnames that match the second pattern (but not
625 the first) are output next, and so on.
626 All files with pathnames that do not match any pattern are output
627 last, as if there was an implicit match-all pattern at the end of the
629 If multiple pathnames have the same rank (they match the same pattern
630 but no earlier patterns), their output order relative to each other is
633 <orderfile> is parsed as follows:
636 - Blank lines are ignored, so they can be used as separators for
639 - Lines starting with a hash ("`#`") are ignored, so they can be used
640 for comments. Add a backslash ("`\`") to the beginning of the
641 pattern if it starts with a hash.
643 - Each other line contains a single pattern.
646 Patterns have the same syntax and semantics as patterns used for
647 fnmatch(3) without the FNM_PATHNAME flag, except a pathname also
648 matches a pattern if removing any number of the final pathname
649 components matches the pattern. For example, the pattern "`foo*bar`"
650 matches "`fooasdfbar`" and "`foo/bar/baz/asdf`" but not "`foobarx`".
652 ifndef::git-format-patch[]
654 Swap two inputs; that is, show differences from index or
655 on-disk file to tree contents.
656 endif::git-format-patch[]
658 --relative[=<path>]::
660 When run from a subdirectory of the project, it can be
661 told to exclude changes outside the directory and show
662 pathnames relative to it with this option. When you are
663 not in a subdirectory (e.g. in a bare repository), you
664 can name which subdirectory to make the output relative
665 to by giving a <path> as an argument.
666 `--no-relative` can be used to countermand both `diff.relative` config
667 option and previous `--relative`.
671 Treat all files as text.
674 Ignore carriage-return at the end of line when doing a comparison.
676 --ignore-space-at-eol::
677 Ignore changes in whitespace at EOL.
680 --ignore-space-change::
681 Ignore changes in amount of whitespace. This ignores whitespace
682 at line end, and considers all other sequences of one or
683 more whitespace characters to be equivalent.
687 Ignore whitespace when comparing lines. This ignores
688 differences even if one line has whitespace where the other
691 --ignore-blank-lines::
692 Ignore changes whose lines are all blank.
695 --ignore-matching-lines=<regex>::
696 Ignore changes whose all lines match <regex>. This option may
697 be specified more than once.
699 --inter-hunk-context=<lines>::
700 Show the context between diff hunks, up to the specified number
701 of lines, thereby fusing hunks that are close to each other.
702 Defaults to `diff.interHunkContext` or 0 if the config option
707 Show whole function as context lines for each change.
708 The function names are determined in the same way as
709 `git diff` works out patch hunk headers (see 'Defining a
710 custom hunk-header' in linkgit:gitattributes[5]).
712 ifndef::git-format-patch[]
715 Make the program exit with codes similar to diff(1).
716 That is, it exits with 1 if there were differences and
717 0 means no differences.
720 Disable all output of the program. Implies `--exit-code`.
722 endif::git-format-patch[]
725 Allow an external diff helper to be executed. If you set an
726 external diff driver with linkgit:gitattributes[5], you need
727 to use this option with linkgit:git-log[1] and friends.
730 Disallow external diff drivers.
734 Allow (or disallow) external text conversion filters to be run
735 when comparing binary files. See linkgit:gitattributes[5] for
736 details. Because textconv filters are typically a one-way
737 conversion, the resulting diff is suitable for human
738 consumption, but cannot be applied. For this reason, textconv
739 filters are enabled by default only for linkgit:git-diff[1] and
740 linkgit:git-log[1], but not for linkgit:git-format-patch[1] or
741 diff plumbing commands.
743 --ignore-submodules[=<when>]::
744 Ignore changes to submodules in the diff generation. <when> can be
745 either "none", "untracked", "dirty" or "all", which is the default.
746 Using "none" will consider the submodule modified when it either contains
747 untracked or modified files or its HEAD differs from the commit recorded
748 in the superproject and can be used to override any settings of the
749 'ignore' option in linkgit:git-config[1] or linkgit:gitmodules[5]. When
750 "untracked" is used submodules are not considered dirty when they only
751 contain untracked content (but they are still scanned for modified
752 content). Using "dirty" ignores all changes to the work tree of submodules,
753 only changes to the commits stored in the superproject are shown (this was
754 the behavior until 1.7.0). Using "all" hides all changes to submodules.
756 --src-prefix=<prefix>::
757 Show the given source prefix instead of "a/".
759 --dst-prefix=<prefix>::
760 Show the given destination prefix instead of "b/".
763 Do not show any source or destination prefix.
765 --line-prefix=<prefix>::
766 Prepend an additional prefix to every line of output.
768 --ita-invisible-in-index::
769 By default entries added by "git add -N" appear as an existing
770 empty file in "git diff" and a new file in "git diff --cached".
771 This option makes the entry appear as a new file in "git diff"
772 and non-existent in "git diff --cached". This option could be
773 reverted with `--ita-visible-in-index`. Both options are
774 experimental and could be removed in future.
776 For more detailed explanation on these common options, see also
777 linkgit:gitdiffcore[7].