1GIT-DIFF-FILES(1) Git Manual GIT-DIFF-FILES(1)
2
6 git-diff-files - Compares files in the working tree and the index
7
9 git diff-files [-q] [-0|-1|-2|-3|-c|--cc] [<common diff options>]
10 [<path>...]
11
13 Compares the files in the working tree and the index. When paths are
14 specified, compares only those named paths. Otherwise all entries in
15 the index are compared. The output format is the same as for git
16 diff-index and git diff-tree.
17
19 -p, -u, --patch
20 Generate patch (see section on generating patches).
21 -U<n>, --unified=<n>
23 Generate diffs with <n> lines of context instead of the usual
24 three. Implies -p.
25 --raw
27 Generate the raw format. This is the default.
28 --patch-with-raw
30 Synonym for -p --raw.
31 --patience
33 Generate a diff using the "patience diff" algorithm.
34 --stat[=<width>[,<name-width>]]
36 Generate a diffstat. You can override the default output width for
37 80-column terminal by --stat=<width>. The width of the filename
38 part can be controlled by giving another width to it separated by a
39 comma.
40 --numstat
42 Similar to --stat, but shows number of added and deleted lines in
43 decimal notation and pathname without abbreviation, to make it more
44 machine friendly. For binary files, outputs two - instead of saying
45 0 0.
46 --shortstat
48 Output only the last line of the --stat format containing total
49 number of modified files, as well as number of added and deleted
50 lines.
51 --dirstat[=<limit>]
53 Output the distribution of relative amount of changes (number of
54 lines added or removed) for each sub-directory. Directories with
55 changes below a cut-off percent (3% by default) are not shown. The
56 cut-off percent can be set with --dirstat=<limit>. Changes in a
57 child directory are not counted for the parent directory, unless
58 --cumulative is used.
59 --dirstat-by-file[=<limit>]
61 Same as --dirstat, but counts changed files instead of lines.
62 --summary
64 Output a condensed summary of extended header information such as
65 creations, renames and mode changes.
66 --patch-with-stat
68 Synonym for -p --stat.
69 -z
71 When --raw, --numstat, --name-only or --name-status has been given,
72 do not munge pathnames and use NULs as output field terminators.
73 Without this option, each pathname output will have TAB, LF, double
75 quotes, and backslash characters replaced with \t, \n, \", and \\,
76 respectively, and the pathname will be enclosed in double quotes if
77 any of those replacements occurred.
78 --name-only
80 Show only names of changed files.
81 --name-status
83 Show only names and status of changed files. See the description of
84 the --diff-filter option on what the status letters mean.
85 --submodule[=<format>]
87 Chose the output format for submodule differences. <format> can be
88 one of short and log. short just shows pairs of commit names, this
89 format is used when this option is not given. log is the default
90 value for this option and lists the commits in that commit range
91 like the summary option of git-submodule(1) does.
92 --color[=<when>]
94 Show colored diff. The value must be always (the default), never,
95 or auto.
96 --no-color
98 Turn off colored diff, even when the configuration file gives the
99 default to color output. Same as --color=never.
100 --word-diff[=<mode>]
102 Show a word diff, using the <mode> to delimit changed words. By
103 default, words are delimited by whitespace; see --word-diff-regex
104 below. The <mode> defaults to plain, and must be one of:
105 color
107 Highlight changed words using only colors. Implies --color.
108 plain
110 Show words as [-removed-] and {added}. Makes no attempts to
111 escape the delimiters if they appear in the input, so the
112 output may be ambiguous.
113 porcelain
115 Use a special line-based format intended for script
116 consumption. Added/removed/unchanged runs are printed in the
117 usual unified diff format, starting with a +/-/` ` character at
118 the beginning of the line and extending to the end of the line.
119 Newlines in the input are represented by a tilde ~ on a line of
120 its own.
121 none
123 Disable word diff again.
124 Note that despite the name of the first mode, color is used to
126 highlight the changed parts in all modes if enabled.
127 --word-diff-regex=<regex>
129 Use <regex> to decide what a word is, instead of considering runs
130 of non-whitespace to be a word. Also implies --word-diff unless it
131 was already enabled.
132 Every non-overlapping match of the <regex> is considered a word.
134 Anything between these matches is considered whitespace and
135 ignored(!) for the purposes of finding differences. You may want to
136 append |[^[:space:]] to your regular expression to make sure that
137 it matches all non-whitespace characters. A match that contains a
138 newline is silently truncated(!) at the newline.
139 The regex can also be set via a diff driver or configuration
141 option, see gitattributes(1) or git-config(1). Giving it explicitly
142 overrides any diff driver or configuration setting. Diff drivers
143 override configuration settings.
144 --color-words[=<regex>]
146 Equivalent to --word-diff=color plus (if a regex was specified)
147 --word-diff-regex=<regex>.
148 --no-renames
150 Turn off rename detection, even when the configuration file gives
151 the default to do so.
152 --check
154 Warn if changes introduce trailing whitespace or an indent that
155 uses a space before a tab. Exits with non-zero status if problems
156 are found. Not compatible with --exit-code.
157 --full-index
159 Instead of the first handful of characters, show the full pre- and
160 post-image blob object names on the "index" line when generating
161 patch format output.
162 --binary
164 In addition to --full-index, output a binary diff that can be
165 applied with git-apply.
166 --abbrev[=<n>]
168 Instead of showing the full 40-byte hexadecimal object name in
169 diff-raw format output and diff-tree header lines, show only a
170 partial prefix. This is independent of the --full-index option
171 above, which controls the diff-patch output format. Non default
172 number of digits can be specified with --abbrev=<n>.
173 -B[<n>][/<m>], --break-rewrites[=[<n>][/<m>]]
175 Break complete rewrite changes into pairs of delete and create.
176 This serves two purposes:
177 It affects the way a change that amounts to a total rewrite of a
179 file not as a series of deletion and insertion mixed together with
180 a very few lines that happen to match textually as the context, but
181 as a single deletion of everything old followed by a single
182 insertion of everything new, and the number m controls this aspect
183 of the -B option (defaults to 60%). -B/70% specifies that less
184 than 30% of the original should remain in the result for git to
185 consider it a total rewrite (i.e. otherwise the resulting patch
186 will be a series of deletion and insertion mixed together with
187 context lines).
188 When used with -M, a totally-rewritten file is also considered as
190 the source of a rename (usually -M only considers a file that
191 disappeared as the source of a rename), and the number n controls
192 this aspect of the -B option (defaults to 50%). -B20% specifies
193 that a change with addition and deletion compared to 20% or more of
194 the file’s size are eligible for being picked up as a possible
195 source of a rename to another file.
196 -M[<n>], --find-renames[=<n>]
198 Detect renames. If n is specified, it is a is a threshold on the
199 similarity index (i.e. amount of addition/deletions compared to the
200 file’s size). For example, -M90% means git should consider a
201 delete/add pair to be a rename if more than 90% of the file hasn’t
202 changed.
203 -C[<n>], --find-copies[=<n>]
205 Detect copies as well as renames. See also --find-copies-harder. If
206 n is specified, it has the same meaning as for -M<n>.
207 --find-copies-harder
209 For performance reasons, by default, -C option finds copies only if
210 the original file of the copy was modified in the same changeset.
211 This flag makes the command inspect unmodified files as candidates
212 for the source of copy. This is a very expensive operation for
213 large projects, so use it with caution. Giving more than one -C
214 option has the same effect.
215 -l<num>
217 The -M and -C options require O(n^2) processing time where n is the
218 number of potential rename/copy targets. This option prevents
219 rename/copy detection from running if the number of rename/copy
220 targets exceeds the specified number.
221 --diff-filter=[(A|C|D|M|R|T|U|X|B)...[*]]
223 Select only files that are Added (A), Copied (C), Deleted (D),
224 Modified (M), Renamed (R), have their type (i.e. regular file,
225 symlink, submodule, ...) changed (T), are Unmerged (U), are Unknown
226 (X), or have had their pairing Broken (B). Any combination of the
227 filter characters (including none) can be used. When *
228 (All-or-none) is added to the combination, all paths are selected
229 if there is any file that matches other criteria in the comparison;
230 if there is no file that matches other criteria, nothing is
231 selected.
232 -S<string>
234 Look for differences that introduce or remove an instance of
235 <string>. Note that this is different than the string simply
236 appearing in diff output; see the pickaxe entry in gitdiffcore(7)
237 for more details.
238 -G<regex>
240 Look for differences whose added or removed line matches the given
241 <regex>.
242 --pickaxe-all
244 When -S or -G finds a change, show all the changes in that
245 changeset, not just the files that contain the change in <string>.
246 --pickaxe-regex
248 Make the <string> not a plain string but an extended POSIX regex to
249 match.
250 -O<orderfile>
252 Output the patch in the order specified in the <orderfile>, which
253 has one shell glob pattern per line.
254 -R
256 Swap two inputs; that is, show differences from index or on-disk
257 file to tree contents.
258 --relative[=<path>]
260 When run from a subdirectory of the project, it can be told to
261 exclude changes outside the directory and show pathnames relative
262 to it with this option. When you are not in a subdirectory (e.g. in
263 a bare repository), you can name which subdirectory to make the
264 output relative to by giving a <path> as an argument.
265 -a, --text
267 Treat all files as text.
268 --ignore-space-at-eol
270 Ignore changes in whitespace at EOL.
271 -b, --ignore-space-change
273 Ignore changes in amount of whitespace. This ignores whitespace at
274 line end, and considers all other sequences of one or more
275 whitespace characters to be equivalent.
276 -w, --ignore-all-space
278 Ignore whitespace when comparing lines. This ignores differences
279 even if one line has whitespace where the other line has none.
280 --inter-hunk-context=<lines>
282 Show the context between diff hunks, up to the specified number of
283 lines, thereby fusing hunks that are close to each other.
284 --exit-code
286 Make the program exit with codes similar to diff(1). That is, it
287 exits with 1 if there were differences and 0 means no differences.
288 --quiet
290 Disable all output of the program. Implies --exit-code.
291 --ext-diff
293 Allow an external diff helper to be executed. If you set an
294 external diff driver with gitattributes(5), you need to use this
295 option with git-log(1) and friends.
296 --no-ext-diff
298 Disallow external diff drivers.
299 --ignore-submodules[=<when>]
301 Ignore changes to submodules in the diff generation. <when> can be
302 either "none", "untracked", "dirty" or "all", which is the default
303 Using "none" will consider the submodule modified when it either
304 contains untracked or modified files or its HEAD differs from the
305 commit recorded in the superproject and can be used to override any
306 settings of the ignore option in git-config(1) or gitmodules(5).
307 When "untracked" is used submodules are not considered dirty when
308 they only contain untracked content (but they are still scanned for
309 modified content). Using "dirty" ignores all changes to the work
310 tree of submodules, only changes to the commits stored in the
311 superproject are shown (this was the behavior until 1.7.0). Using
312 "all" hides all changes to submodules.
313 --src-prefix=<prefix>
315 Show the given source prefix instead of "a/".
316 --dst-prefix=<prefix>
318 Show the given destination prefix instead of "b/".
319 --no-prefix
321 Do not show any source or destination prefix.
322 For more detailed explanation on these common options, see also
324 gitdiffcore(7).
325 -1 --base, -2 --ours, -3 --theirs, -0
327 Diff against the "base" version, "our branch" or "their branch"
328 respectively. With these options, diffs for merged entries are not
329 shown.
330 The default is to diff against our branch (-2) and the cleanly
332 resolved paths. The option -0 can be given to omit diff output for
333 unmerged entries and just show "Unmerged".
334 -c, --cc
336 This compares stage 2 (our branch), stage 3 (their branch) and the
337 working tree file and outputs a combined diff, similar to the way
338 diff-tree shows a merge commit with these flags.
339 -q
341 Remain silent even on nonexistent files
342
344 The raw output format from "git-diff-index", "git-diff-tree",
345 "git-diff-files" and "git diff --raw" are very similar.
346 These commands all compare two sets of things; what is compared
348 differs:
349 git-diff-index <tree-ish>
351 compares the <tree-ish> and the files on the filesystem.
352 git-diff-index --cached <tree-ish>
354 compares the <tree-ish> and the index.
355 git-diff-tree [-r] <tree-ish-1> <tree-ish-2> [<pattern>...]
357 compares the trees named by the two arguments.
358 git-diff-files [<pattern>...]
360 compares the index and the files on the filesystem.
361 The "git-diff-tree" command begins its output by printing the hash of
363 what is being compared. After that, all the commands print one output
364 line per changed file.
365 An output line is formatted this way:
367 in-place edit :100644 100644 bcd1234... 0123456... M file0
369 copy-edit :100644 100644 abcd123... 1234567... C68 file1 file2
370 rename-edit :100644 100644 abcd123... 1234567... R86 file1 file3
371 create :000000 100644 0000000... 1234567... A file4
372 delete :100644 000000 1234567... 0000000... D file5
373 unmerged :000000 000000 0000000... 0000000... U file6
374 That is, from the left to the right:
377 1. a colon.
379 2. mode for "src"; 000000 if creation or unmerged.
381 3. a space.
383 4. mode for "dst"; 000000 if deletion or unmerged.
385 5. a space.
387 6. sha1 for "src"; 0{40} if creation or unmerged.
389 7. a space.
391 8. sha1 for "dst"; 0{40} if creation, unmerged or "look at work tree".
393 9. a space.
395 10. status, followed by optional "score" number.
397 11. a tab or a NUL when -z option is used.
399 12. path for "src"
401 13. a tab or a NUL when -z option is used; only exists for C or R.
403 14. path for "dst"; only exists for C or R.
405 15. an LF or a NUL when -z option is used, to terminate the record.
407 Possible status letters are:
409 · A: addition of a file
411 · C: copy of a file into a new one
413 · D: deletion of a file
415 · M: modification of the contents or mode of a file
417 · R: renaming of a file
419 · T: change in the type of the file
421 · U: file is unmerged (you must complete the merge before it can be
423 committed)
424 · X: "unknown" change type (most probably a bug, please report it)
426 Status letters C and R are always followed by a score (denoting the
428 percentage of similarity between the source and target of the move or
429 copy), and are the only ones to be so.
430 <sha1> is shown as all 0’s if a file is new on the filesystem and it is
432 out of sync with the index.
433 Example:
435 :100644 100644 5be4a4...... 000000...... M file.c
437 When -z option is not used, TAB, LF, and backslash characters in
440 pathnames are represented as \t, \n, and \\, respectively.
441
443 "git-diff-tree", "git-diff-files" and "git-diff --raw" can take -c or
444 --cc option to generate diff output also for merge commits. The output
445 differs from the format described above in the following way:
446 1. there is a colon for each parent
448 2. there are more "src" modes and "src" sha1
450 3. status is concatenated status characters for each parent
452 4. no optional "score" number
454 5. single path, only for "dst"
456 Example:
458 ::100644 100644 100644 fabadb8... cc95eb0... 4866510... MM describe.c
460 Note that combined diff lists only files which were modified from all
463 parents.
464
466 When "git-diff-index", "git-diff-tree", or "git-diff-files" are run
467 with a -p option, "git diff" without the --raw option, or "git log"
468 with the "-p" option, they do not produce the output described above;
469 instead they produce a patch file. You can customize the creation of
470 such patches via the GIT_EXTERNAL_DIFF and the GIT_DIFF_OPTS
471 environment variables.
472 What the -p option produces is slightly different from the traditional
474 diff format:
475 1. It is preceded with a "git diff" header that looks like this:
477 diff --git a/file1 b/file2
479 The a/ and b/ filenames are the same unless rename/copy is
481 involved. Especially, even for a creation or a deletion, /dev/null
482 is not used in place of the a/ or b/ filenames.
483 When rename/copy is involved, file1 and file2 show the name of the
485 source file of the rename/copy and the name of the file that
486 rename/copy produces, respectively.
487 2. It is followed by one or more extended header lines:
489 old mode <mode>
491 new mode <mode>
492 deleted file mode <mode>
493 new file mode <mode>
494 copy from <path>
495 copy to <path>
496 rename from <path>
497 rename to <path>
498 similarity index <number>
499 dissimilarity index <number>
500 index <hash>..<hash> <mode>
501 File modes are printed as 6-digit octal numbers including the file
503 type and file permission bits.
504 Path names in extended headers do not include the a/ and b/
506 prefixes.
507 The similarity index is the percentage of unchanged lines, and the
509 dissimilarity index is the percentage of changed lines. It is a
510 rounded down integer, followed by a percent sign. The similarity
511 index value of 100% is thus reserved for two equal files, while
512 100% dissimilarity means that no line from the old file made it
513 into the new one.
514 The index line includes the SHA-1 checksum before and after the
516 change. The <mode> is included if the file mode does not change;
517 otherwise, separate lines indicate the old and the new mode.
518 3. TAB, LF, double quote and backslash characters in pathnames are
520 represented as \t, \n, \" and \\, respectively. If there is need
521 for such substitution then the whole pathname is put in double
522 quotes.
523 4. All the file1 files in the output refer to files before the commit,
525 and all the file2 files refer to files after the commit. It is
526 incorrect to apply each change to each file sequentially. For
527 example, this patch will swap a and b:
528 diff --git a/a b/b
530 rename from a
531 rename to b
532 diff --git a/b b/a
533 rename from b
534 rename to a
535
537 Any diff-generating command can take the ‘-c` or --cc option to produce
538 a combined diff when showing a merge. This is the default format when
539 showing merges with git-diff(1) or git-show(1). Note also that you can
540 give the `-m’ option to any of these commands to force generation of
541 diffs with individual parents of a merge.
542 A combined diff format looks like this:
544 diff --combined describe.c
546 index fabadb8,cc95eb0..4866510
547 --- a/describe.c
548 +++ b/describe.c
549 @@@ -98,20 -98,12 +98,20 @@@
550 return (a_date > b_date) ? -1 : (a_date == b_date) ? 0 : 1;
551 }
552 - static void describe(char *arg)
554 -static void describe(struct commit *cmit, int last_one)
555 ++static void describe(char *arg, int last_one)
556 {
557 + unsigned char sha1[20];
558 + struct commit *cmit;
559 struct commit_list *list;
560 static int initialized = 0;
561 struct commit_name *n;
562 + if (get_sha1(arg, sha1) < 0)
564 + usage(describe_usage);
565 + cmit = lookup_commit_reference(sha1);
566 + if (!cmit)
567 + usage(describe_usage);
568 +
569 if (!initialized) {
570 initialized = 1;
571 for_each_ref(get_name);
572 1. It is preceded with a "git diff" header, that looks like this (when
576 -c option is used):
577 diff --combined file
579 or like this (when --cc option is used):
581 diff --cc file
583 2. It is followed by one or more extended header lines (this example
585 shows a merge with two parents):
586 index <hash>,<hash>..<hash>
588 mode <mode>,<mode>..<mode>
589 new file mode <mode>
590 deleted file mode <mode>,<mode>
591 The mode <mode>,<mode>..<mode> line appears only if at least one of
593 the <mode> is different from the rest. Extended headers with
594 information about detected contents movement (renames and copying
595 detection) are designed to work with diff of two <tree-ish> and are
596 not used by combined diff format.
597 3. It is followed by two-line from-file/to-file header
599 --- a/file
601 +++ b/file
602 Similar to two-line header for traditional unified diff format,
604 /dev/null is used to signal created or deleted files.
605 4. Chunk header format is modified to prevent people from accidentally
607 feeding it to patch -p1. Combined diff format was created for
608 review of merge commit changes, and was not meant for apply. The
609 change is similar to the change in the extended index header:
610 @@@ <from-file-range> <from-file-range> <to-file-range> @@@
612 There are (number of parents + 1) @ characters in the chunk header
614 for combined diff format.
615 Unlike the traditional unified diff format, which shows two files A and
617 B with a single column that has - (minus — appears in A but removed in
618 B), + (plus — missing in A but added to B), or " " (space — unchanged)
619 prefix, this format compares two or more files file1, file2,... with
620 one file X, and shows how X differs from each of fileN. One column for
621 each of fileN is prepended to the output line to note how X’s line is
622 different from it.
623 A - character in the column N means that the line appears in fileN but
625 it does not appear in the result. A + character in the column N means
626 that the line appears in the result, and fileN does not have that line
627 (in other words, the line was added, from the point of view of that
628 parent).
629 In the above example output, the function signature was changed from
631 both files (hence two - removals from both file1 and file2, plus ++ to
632 mean one line that was added does not appear in either file1 nor
633 file2). Also eight other lines are the same from file1 but do not
634 appear in file2 (hence prefixed with +).
635 When shown by git diff-tree -c, it compares the parents of a merge
637 commit with the merge result (i.e. file1..fileN are the parents). When
638 shown by git diff-files -c, it compares the two unresolved merge
639 parents with the working tree file (i.e. file1 is stage 2 aka "our
640 version", file2 is stage 3 aka "their version").
641
643 The --summary option describes newly added, deleted, renamed and copied
644 files. The --stat option adds diffstat(1) graph to the output. These
645 options can be combined with other options, such as -p, and are meant
646 for human consumption.
647 When showing a change that involves a rename or a copy, --stat output
649 formats the pathnames compactly by combining common prefix and suffix
650 of the pathnames. For example, a change that moves arch/i386/Makefile
651 to arch/x86/Makefile while modifying 4 lines will be shown like this:
652 arch/{i386 => x86}/Makefile | 4 +--
654 The --numstat option gives the diffstat(1) information but is designed
657 for easier machine consumption. An entry in --numstat output looks like
658 this:
659 1 2 README
661 3 1 arch/{i386 => x86}/Makefile
662 That is, from left to right:
665 1. the number of added lines;
667 2. a tab;
669 3. the number of deleted lines;
671 4. a tab;
673 5. pathname (possibly with rename/copy information);
675 6. a newline.
677 When -z output option is in effect, the output is formatted this way:
679 1 2 README NUL
681 3 1 NUL arch/i386/Makefile NUL arch/x86/Makefile NUL
682 That is:
685 1. the number of added lines;
687 2. a tab;
689 3. the number of deleted lines;
691 4. a tab;
693 5. a NUL (only exists if renamed/copied);
695 6. pathname in preimage;
697 7. a NUL (only exists if renamed/copied);
699 8. pathname in postimage (only exists if renamed/copied);
701 9. a NUL.
703 The extra NUL before the preimage path in renamed case is to allow
705 scripts that read the output to tell if the current record being read
706 is a single-path record or a rename/copy record without reading ahead.
707 After reading added and deleted lines, reading up to NUL would yield
708 the pathname, but if that is NUL, the record will show two paths.
709
711 Written by Linus Torvalds <torvalds@osdl.org[1]>
712
714 Documentation by David Greaves, Junio C Hamano and the git-list
715 <git@vger.kernel.org[2]>.
716
718 Part of the git(1) suite
719
721 1. torvalds@osdl.org
722 mailto:torvalds@osdl.org
723 2. git@vger.kernel.org
725 mailto:git@vger.kernel.org
726Git 1.7.4.4 04/11/2011 GIT-DIFF-FILES(1)