Merge branch 'dl/subtree-limit-to-one-rev'
[git/git.git] / Documentation / rev-list-options.txt
1 Commit Limiting
2 ~~~~~~~~~~~~~~~
3
4 Besides specifying a range of commits that should be listed using the
5 special notations explained in the description, additional commit
6 limiting may be applied.
7
8 Using more options generally further limits the output (e.g.
9 `--since=<date1>` limits to commits newer than `<date1>`, and using it
10 with `--grep=<pattern>` further limits to commits whose log message
11 has a line that matches `<pattern>`), unless otherwise noted.
12
13 Note that these are applied before commit
14 ordering and formatting options, such as `--reverse`.
15
16 -<number>::
17 -n <number>::
18 --max-count=<number>::
19 Limit the number of commits to output.
20
21 --skip=<number>::
22 Skip 'number' commits before starting to show the commit output.
23
24 --since=<date>::
25 --after=<date>::
26 Show commits more recent than a specific date.
27
28 --until=<date>::
29 --before=<date>::
30 Show commits older than a specific date.
31
32 ifdef::git-rev-list[]
33 --max-age=<timestamp>::
34 --min-age=<timestamp>::
35 Limit the commits output to specified time range.
36 endif::git-rev-list[]
37
38 --author=<pattern>::
39 --committer=<pattern>::
40 Limit the commits output to ones with author/committer
41 header lines that match the specified pattern (regular
42 expression). With more than one `--author=<pattern>`,
43 commits whose author matches any of the given patterns are
44 chosen (similarly for multiple `--committer=<pattern>`).
45
46 --grep-reflog=<pattern>::
47 Limit the commits output to ones with reflog entries that
48 match the specified pattern (regular expression). With
49 more than one `--grep-reflog`, commits whose reflog message
50 matches any of the given patterns are chosen. It is an
51 error to use this option unless `--walk-reflogs` is in use.
52
53 --grep=<pattern>::
54 Limit the commits output to ones with log message that
55 matches the specified pattern (regular expression). With
56 more than one `--grep=<pattern>`, commits whose message
57 matches any of the given patterns are chosen (but see
58 `--all-match`).
59 ifndef::git-rev-list[]
60 +
61 When `--show-notes` is in effect, the message from the notes is
62 matched as if it were part of the log message.
63 endif::git-rev-list[]
64
65 --all-match::
66 Limit the commits output to ones that match all given `--grep`,
67 instead of ones that match at least one.
68
69 --invert-grep::
70 Limit the commits output to ones with log message that do not
71 match the pattern specified with `--grep=<pattern>`.
72
73 -i::
74 --regexp-ignore-case::
75 Match the regular expression limiting patterns without regard to letter
76 case.
77
78 --basic-regexp::
79 Consider the limiting patterns to be basic regular expressions;
80 this is the default.
81
82 -E::
83 --extended-regexp::
84 Consider the limiting patterns to be extended regular expressions
85 instead of the default basic regular expressions.
86
87 -F::
88 --fixed-strings::
89 Consider the limiting patterns to be fixed strings (don't interpret
90 pattern as a regular expression).
91
92 -P::
93 --perl-regexp::
94 Consider the limiting patterns to be Perl-compatible regular
95 expressions.
96 +
97 Support for these types of regular expressions is an optional
98 compile-time dependency. If Git wasn't compiled with support for them
99 providing this option will cause it to die.
100
101 --remove-empty::
102 Stop when a given path disappears from the tree.
103
104 --merges::
105 Print only merge commits. This is exactly the same as `--min-parents=2`.
106
107 --no-merges::
108 Do not print commits with more than one parent. This is
109 exactly the same as `--max-parents=1`.
110
111 --min-parents=<number>::
112 --max-parents=<number>::
113 --no-min-parents::
114 --no-max-parents::
115 Show only commits which have at least (or at most) that many parent
116 commits. In particular, `--max-parents=1` is the same as `--no-merges`,
117 `--min-parents=2` is the same as `--merges`. `--max-parents=0`
118 gives all root commits and `--min-parents=3` all octopus merges.
119 +
120 `--no-min-parents` and `--no-max-parents` reset these limits (to no limit)
121 again. Equivalent forms are `--min-parents=0` (any commit has 0 or more
122 parents) and `--max-parents=-1` (negative numbers denote no upper limit).
123
124 --first-parent::
125 Follow only the first parent commit upon seeing a merge
126 commit. This option can give a better overview when
127 viewing the evolution of a particular topic branch,
128 because merges into a topic branch tend to be only about
129 adjusting to updated upstream from time to time, and
130 this option allows you to ignore the individual commits
131 brought in to your history by such a merge. Cannot be
132 combined with --bisect.
133
134 --not::
135 Reverses the meaning of the '{caret}' prefix (or lack thereof)
136 for all following revision specifiers, up to the next `--not`.
137
138 --all::
139 Pretend as if all the refs in `refs/`, along with `HEAD`, are
140 listed on the command line as '<commit>'.
141
142 --branches[=<pattern>]::
143 Pretend as if all the refs in `refs/heads` are listed
144 on the command line as '<commit>'. If '<pattern>' is given, limit
145 branches to ones matching given shell glob. If pattern lacks '?',
146 '{asterisk}', or '[', '/{asterisk}' at the end is implied.
147
148 --tags[=<pattern>]::
149 Pretend as if all the refs in `refs/tags` are listed
150 on the command line as '<commit>'. If '<pattern>' is given, limit
151 tags to ones matching given shell glob. If pattern lacks '?', '{asterisk}',
152 or '[', '/{asterisk}' at the end is implied.
153
154 --remotes[=<pattern>]::
155 Pretend as if all the refs in `refs/remotes` are listed
156 on the command line as '<commit>'. If '<pattern>' is given, limit
157 remote-tracking branches to ones matching given shell glob.
158 If pattern lacks '?', '{asterisk}', or '[', '/{asterisk}' at the end is implied.
159
160 --glob=<glob-pattern>::
161 Pretend as if all the refs matching shell glob '<glob-pattern>'
162 are listed on the command line as '<commit>'. Leading 'refs/',
163 is automatically prepended if missing. If pattern lacks '?', '{asterisk}',
164 or '[', '/{asterisk}' at the end is implied.
165
166 --exclude=<glob-pattern>::
167
168 Do not include refs matching '<glob-pattern>' that the next `--all`,
169 `--branches`, `--tags`, `--remotes`, or `--glob` would otherwise
170 consider. Repetitions of this option accumulate exclusion patterns
171 up to the next `--all`, `--branches`, `--tags`, `--remotes`, or
172 `--glob` option (other options or arguments do not clear
173 accumulated patterns).
174 +
175 The patterns given should not begin with `refs/heads`, `refs/tags`, or
176 `refs/remotes` when applied to `--branches`, `--tags`, or `--remotes`,
177 respectively, and they must begin with `refs/` when applied to `--glob`
178 or `--all`. If a trailing '/{asterisk}' is intended, it must be given
179 explicitly.
180
181 --reflog::
182 Pretend as if all objects mentioned by reflogs are listed on the
183 command line as `<commit>`.
184
185 --single-worktree::
186 By default, all working trees will be examined by the
187 following options when there are more than one (see
188 linkgit:git-worktree[1]): `--all`, `--reflog` and
189 `--indexed-objects`.
190 This option forces them to examine the current working tree
191 only.
192
193 --ignore-missing::
194 Upon seeing an invalid object name in the input, pretend as if
195 the bad input was not given.
196
197 ifndef::git-rev-list[]
198 --bisect::
199 Pretend as if the bad bisection ref `refs/bisect/bad`
200 was listed and as if it was followed by `--not` and the good
201 bisection refs `refs/bisect/good-*` on the command
202 line. Cannot be combined with --first-parent.
203 endif::git-rev-list[]
204
205 --stdin::
206 In addition to the '<commit>' listed on the command
207 line, read them from the standard input. If a `--` separator is
208 seen, stop reading commits and start reading paths to limit the
209 result.
210
211 ifdef::git-rev-list[]
212 --quiet::
213 Don't print anything to standard output. This form
214 is primarily meant to allow the caller to
215 test the exit status to see if a range of objects is fully
216 connected (or not). It is faster than redirecting stdout
217 to `/dev/null` as the output does not have to be formatted.
218 endif::git-rev-list[]
219
220 --cherry-mark::
221 Like `--cherry-pick` (see below) but mark equivalent commits
222 with `=` rather than omitting them, and inequivalent ones with `+`.
223
224 --cherry-pick::
225 Omit any commit that introduces the same change as
226 another commit on the ``other side'' when the set of
227 commits are limited with symmetric difference.
228 +
229 For example, if you have two branches, `A` and `B`, a usual way
230 to list all commits on only one side of them is with
231 `--left-right` (see the example below in the description of
232 the `--left-right` option). However, it shows the commits that were
233 cherry-picked from the other branch (for example, ``3rd on b'' may be
234 cherry-picked from branch A). With this option, such pairs of commits are
235 excluded from the output.
236
237 --left-only::
238 --right-only::
239 List only commits on the respective side of a symmetric difference,
240 i.e. only those which would be marked `<` resp. `>` by
241 `--left-right`.
242 +
243 For example, `--cherry-pick --right-only A...B` omits those
244 commits from `B` which are in `A` or are patch-equivalent to a commit in
245 `A`. In other words, this lists the `+` commits from `git cherry A B`.
246 More precisely, `--cherry-pick --right-only --no-merges` gives the exact
247 list.
248
249 --cherry::
250 A synonym for `--right-only --cherry-mark --no-merges`; useful to
251 limit the output to the commits on our side and mark those that
252 have been applied to the other side of a forked history with
253 `git log --cherry upstream...mybranch`, similar to
254 `git cherry upstream mybranch`.
255
256 -g::
257 --walk-reflogs::
258 Instead of walking the commit ancestry chain, walk
259 reflog entries from the most recent one to older ones.
260 When this option is used you cannot specify commits to
261 exclude (that is, '{caret}commit', 'commit1..commit2',
262 and 'commit1\...commit2' notations cannot be used).
263 +
264 With `--pretty` format other than `oneline` (for obvious reasons),
265 this causes the output to have two extra lines of information
266 taken from the reflog. The reflog designator in the output may be shown
267 as `ref@{Nth}` (where `Nth` is the reverse-chronological index in the
268 reflog) or as `ref@{timestamp}` (with the timestamp for that entry),
269 depending on a few rules:
270 +
271 --
272 1. If the starting point is specified as `ref@{Nth}`, show the index
273 format.
274 +
275 2. If the starting point was specified as `ref@{now}`, show the
276 timestamp format.
277 +
278 3. If neither was used, but `--date` was given on the command line, show
279 the timestamp in the format requested by `--date`.
280 +
281 4. Otherwise, show the index format.
282 --
283 +
284 Under `--pretty=oneline`, the commit message is
285 prefixed with this information on the same line.
286 This option cannot be combined with `--reverse`.
287 See also linkgit:git-reflog[1].
288
289 --merge::
290 After a failed merge, show refs that touch files having a
291 conflict and don't exist on all heads to merge.
292
293 --boundary::
294 Output excluded boundary commits. Boundary commits are
295 prefixed with `-`.
296
297 ifdef::git-rev-list[]
298 --use-bitmap-index::
299
300 Try to speed up the traversal using the pack bitmap index (if
301 one is available). Note that when traversing with `--objects`,
302 trees and blobs will not have their associated path printed.
303
304 --progress=<header>::
305 Show progress reports on stderr as objects are considered. The
306 `<header>` text will be printed with each progress update.
307 endif::git-rev-list[]
308
309 History Simplification
310 ~~~~~~~~~~~~~~~~~~~~~~
311
312 Sometimes you are only interested in parts of the history, for example the
313 commits modifying a particular <path>. But there are two parts of
314 'History Simplification', one part is selecting the commits and the other
315 is how to do it, as there are various strategies to simplify the history.
316
317 The following options select the commits to be shown:
318
319 <paths>::
320 Commits modifying the given <paths> are selected.
321
322 --simplify-by-decoration::
323 Commits that are referred by some branch or tag are selected.
324
325 Note that extra commits can be shown to give a meaningful history.
326
327 The following options affect the way the simplification is performed:
328
329 Default mode::
330 Simplifies the history to the simplest history explaining the
331 final state of the tree. Simplest because it prunes some side
332 branches if the end result is the same (i.e. merging branches
333 with the same content)
334
335 --full-history::
336 Same as the default mode, but does not prune some history.
337
338 --dense::
339 Only the selected commits are shown, plus some to have a
340 meaningful history.
341
342 --sparse::
343 All commits in the simplified history are shown.
344
345 --simplify-merges::
346 Additional option to `--full-history` to remove some needless
347 merges from the resulting history, as there are no selected
348 commits contributing to this merge.
349
350 --ancestry-path::
351 When given a range of commits to display (e.g. 'commit1..commit2'
352 or 'commit2 {caret}commit1'), only display commits that exist
353 directly on the ancestry chain between the 'commit1' and
354 'commit2', i.e. commits that are both descendants of 'commit1',
355 and ancestors of 'commit2'.
356
357 A more detailed explanation follows.
358
359 Suppose you specified `foo` as the <paths>. We shall call commits
360 that modify `foo` !TREESAME, and the rest TREESAME. (In a diff
361 filtered for `foo`, they look different and equal, respectively.)
362
363 In the following, we will always refer to the same example history to
364 illustrate the differences between simplification settings. We assume
365 that you are filtering for a file `foo` in this commit graph:
366 -----------------------------------------------------------------------
367 .-A---M---N---O---P---Q
368 / / / / / /
369 I B C D E Y
370 \ / / / / /
371 `-------------' X
372 -----------------------------------------------------------------------
373 The horizontal line of history A---Q is taken to be the first parent of
374 each merge. The commits are:
375
376 * `I` is the initial commit, in which `foo` exists with contents
377 ``asdf'', and a file `quux` exists with contents ``quux''. Initial
378 commits are compared to an empty tree, so `I` is !TREESAME.
379
380 * In `A`, `foo` contains just ``foo''.
381
382 * `B` contains the same change as `A`. Its merge `M` is trivial and
383 hence TREESAME to all parents.
384
385 * `C` does not change `foo`, but its merge `N` changes it to ``foobar'',
386 so it is not TREESAME to any parent.
387
388 * `D` sets `foo` to ``baz''. Its merge `O` combines the strings from
389 `N` and `D` to ``foobarbaz''; i.e., it is not TREESAME to any parent.
390
391 * `E` changes `quux` to ``xyzzy'', and its merge `P` combines the
392 strings to ``quux xyzzy''. `P` is TREESAME to `O`, but not to `E`.
393
394 * `X` is an independent root commit that added a new file `side`, and `Y`
395 modified it. `Y` is TREESAME to `X`. Its merge `Q` added `side` to `P`, and
396 `Q` is TREESAME to `P`, but not to `Y`.
397
398 `rev-list` walks backwards through history, including or excluding
399 commits based on whether `--full-history` and/or parent rewriting
400 (via `--parents` or `--children`) are used. The following settings
401 are available.
402
403 Default mode::
404 Commits are included if they are not TREESAME to any parent
405 (though this can be changed, see `--sparse` below). If the
406 commit was a merge, and it was TREESAME to one parent, follow
407 only that parent. (Even if there are several TREESAME
408 parents, follow only one of them.) Otherwise, follow all
409 parents.
410 +
411 This results in:
412 +
413 -----------------------------------------------------------------------
414 .-A---N---O
415 / / /
416 I---------D
417 -----------------------------------------------------------------------
418 +
419 Note how the rule to only follow the TREESAME parent, if one is
420 available, removed `B` from consideration entirely. `C` was
421 considered via `N`, but is TREESAME. Root commits are compared to an
422 empty tree, so `I` is !TREESAME.
423 +
424 Parent/child relations are only visible with `--parents`, but that does
425 not affect the commits selected in default mode, so we have shown the
426 parent lines.
427
428 --full-history without parent rewriting::
429 This mode differs from the default in one point: always follow
430 all parents of a merge, even if it is TREESAME to one of them.
431 Even if more than one side of the merge has commits that are
432 included, this does not imply that the merge itself is! In
433 the example, we get
434 +
435 -----------------------------------------------------------------------
436 I A B N D O P Q
437 -----------------------------------------------------------------------
438 +
439 `M` was excluded because it is TREESAME to both parents. `E`,
440 `C` and `B` were all walked, but only `B` was !TREESAME, so the others
441 do not appear.
442 +
443 Note that without parent rewriting, it is not really possible to talk
444 about the parent/child relationships between the commits, so we show
445 them disconnected.
446
447 --full-history with parent rewriting::
448 Ordinary commits are only included if they are !TREESAME
449 (though this can be changed, see `--sparse` below).
450 +
451 Merges are always included. However, their parent list is rewritten:
452 Along each parent, prune away commits that are not included
453 themselves. This results in
454 +
455 -----------------------------------------------------------------------
456 .-A---M---N---O---P---Q
457 / / / / /
458 I B / D /
459 \ / / / /
460 `-------------'
461 -----------------------------------------------------------------------
462 +
463 Compare to `--full-history` without rewriting above. Note that `E`
464 was pruned away because it is TREESAME, but the parent list of P was
465 rewritten to contain `E`'s parent `I`. The same happened for `C` and
466 `N`, and `X`, `Y` and `Q`.
467
468 In addition to the above settings, you can change whether TREESAME
469 affects inclusion:
470
471 --dense::
472 Commits that are walked are included if they are not TREESAME
473 to any parent.
474
475 --sparse::
476 All commits that are walked are included.
477 +
478 Note that without `--full-history`, this still simplifies merges: if
479 one of the parents is TREESAME, we follow only that one, so the other
480 sides of the merge are never walked.
481
482 --simplify-merges::
483 First, build a history graph in the same way that
484 `--full-history` with parent rewriting does (see above).
485 +
486 Then simplify each commit `C` to its replacement `C'` in the final
487 history according to the following rules:
488 +
489 --
490 * Set `C'` to `C`.
491 +
492 * Replace each parent `P` of `C'` with its simplification `P'`. In
493 the process, drop parents that are ancestors of other parents or that are
494 root commits TREESAME to an empty tree, and remove duplicates, but take care
495 to never drop all parents that we are TREESAME to.
496 +
497 * If after this parent rewriting, `C'` is a root or merge commit (has
498 zero or >1 parents), a boundary commit, or !TREESAME, it remains.
499 Otherwise, it is replaced with its only parent.
500 --
501 +
502 The effect of this is best shown by way of comparing to
503 `--full-history` with parent rewriting. The example turns into:
504 +
505 -----------------------------------------------------------------------
506 .-A---M---N---O
507 / / /
508 I B D
509 \ / /
510 `---------'
511 -----------------------------------------------------------------------
512 +
513 Note the major differences in `N`, `P`, and `Q` over `--full-history`:
514 +
515 --
516 * `N`'s parent list had `I` removed, because it is an ancestor of the
517 other parent `M`. Still, `N` remained because it is !TREESAME.
518 +
519 * `P`'s parent list similarly had `I` removed. `P` was then
520 removed completely, because it had one parent and is TREESAME.
521 +
522 * `Q`'s parent list had `Y` simplified to `X`. `X` was then removed, because it
523 was a TREESAME root. `Q` was then removed completely, because it had one
524 parent and is TREESAME.
525 --
526
527 Finally, there is a fifth simplification mode available:
528
529 --ancestry-path::
530 Limit the displayed commits to those directly on the ancestry
531 chain between the ``from'' and ``to'' commits in the given commit
532 range. I.e. only display commits that are ancestor of the ``to''
533 commit and descendants of the ``from'' commit.
534 +
535 As an example use case, consider the following commit history:
536 +
537 -----------------------------------------------------------------------
538 D---E-------F
539 / \ \
540 B---C---G---H---I---J
541 / \
542 A-------K---------------L--M
543 -----------------------------------------------------------------------
544 +
545 A regular 'D..M' computes the set of commits that are ancestors of `M`,
546 but excludes the ones that are ancestors of `D`. This is useful to see
547 what happened to the history leading to `M` since `D`, in the sense
548 that ``what does `M` have that did not exist in `D`''. The result in this
549 example would be all the commits, except `A` and `B` (and `D` itself,
550 of course).
551 +
552 When we want to find out what commits in `M` are contaminated with the
553 bug introduced by `D` and need fixing, however, we might want to view
554 only the subset of 'D..M' that are actually descendants of `D`, i.e.
555 excluding `C` and `K`. This is exactly what the `--ancestry-path`
556 option does. Applied to the 'D..M' range, it results in:
557 +
558 -----------------------------------------------------------------------
559 E-------F
560 \ \
561 G---H---I---J
562 \
563 L--M
564 -----------------------------------------------------------------------
565
566 The `--simplify-by-decoration` option allows you to view only the
567 big picture of the topology of the history, by omitting commits
568 that are not referenced by tags. Commits are marked as !TREESAME
569 (in other words, kept after history simplification rules described
570 above) if (1) they are referenced by tags, or (2) they change the
571 contents of the paths given on the command line. All other
572 commits are marked as TREESAME (subject to be simplified away).
573
574 ifdef::git-rev-list[]
575 Bisection Helpers
576 ~~~~~~~~~~~~~~~~~
577
578 --bisect::
579 Limit output to the one commit object which is roughly halfway between
580 included and excluded commits. Note that the bad bisection ref
581 `refs/bisect/bad` is added to the included commits (if it
582 exists) and the good bisection refs `refs/bisect/good-*` are
583 added to the excluded commits (if they exist). Thus, supposing there
584 are no refs in `refs/bisect/`, if
585 +
586 -----------------------------------------------------------------------
587 $ git rev-list --bisect foo ^bar ^baz
588 -----------------------------------------------------------------------
589 +
590 outputs 'midpoint', the output of the two commands
591 +
592 -----------------------------------------------------------------------
593 $ git rev-list foo ^midpoint
594 $ git rev-list midpoint ^bar ^baz
595 -----------------------------------------------------------------------
596 +
597 would be of roughly the same length. Finding the change which
598 introduces a regression is thus reduced to a binary search: repeatedly
599 generate and test new 'midpoint's until the commit chain is of length
600 one. Cannot be combined with --first-parent.
601
602 --bisect-vars::
603 This calculates the same as `--bisect`, except that refs in
604 `refs/bisect/` are not used, and except that this outputs
605 text ready to be eval'ed by the shell. These lines will assign the
606 name of the midpoint revision to the variable `bisect_rev`, and the
607 expected number of commits to be tested after `bisect_rev` is tested
608 to `bisect_nr`, the expected number of commits to be tested if
609 `bisect_rev` turns out to be good to `bisect_good`, the expected
610 number of commits to be tested if `bisect_rev` turns out to be bad to
611 `bisect_bad`, and the number of commits we are bisecting right now to
612 `bisect_all`.
613
614 --bisect-all::
615 This outputs all the commit objects between the included and excluded
616 commits, ordered by their distance to the included and excluded
617 commits. Refs in `refs/bisect/` are not used. The farthest
618 from them is displayed first. (This is the only one displayed by
619 `--bisect`.)
620 +
621 This is useful because it makes it easy to choose a good commit to
622 test when you want to avoid to test some of them for some reason (they
623 may not compile for example).
624 +
625 This option can be used along with `--bisect-vars`, in this case,
626 after all the sorted commit objects, there will be the same text as if
627 `--bisect-vars` had been used alone.
628 endif::git-rev-list[]
629
630
631 Commit Ordering
632 ~~~~~~~~~~~~~~~
633
634 By default, the commits are shown in reverse chronological order.
635
636 --date-order::
637 Show no parents before all of its children are shown, but
638 otherwise show commits in the commit timestamp order.
639
640 --author-date-order::
641 Show no parents before all of its children are shown, but
642 otherwise show commits in the author timestamp order.
643
644 --topo-order::
645 Show no parents before all of its children are shown, and
646 avoid showing commits on multiple lines of history
647 intermixed.
648 +
649 For example, in a commit history like this:
650 +
651 ----------------------------------------------------------------
652
653 ---1----2----4----7
654 \ \
655 3----5----6----8---
656
657 ----------------------------------------------------------------
658 +
659 where the numbers denote the order of commit timestamps, `git
660 rev-list` and friends with `--date-order` show the commits in the
661 timestamp order: 8 7 6 5 4 3 2 1.
662 +
663 With `--topo-order`, they would show 8 6 5 3 7 4 2 1 (or 8 7 4 2 6 5
664 3 1); some older commits are shown before newer ones in order to
665 avoid showing the commits from two parallel development track mixed
666 together.
667
668 --reverse::
669 Output the commits chosen to be shown (see Commit Limiting
670 section above) in reverse order. Cannot be combined with
671 `--walk-reflogs`.
672
673 Object Traversal
674 ~~~~~~~~~~~~~~~~
675
676 These options are mostly targeted for packing of Git repositories.
677
678 ifdef::git-rev-list[]
679 --objects::
680 Print the object IDs of any object referenced by the listed
681 commits. `--objects foo ^bar` thus means ``send me
682 all object IDs which I need to download if I have the commit
683 object _bar_ but not _foo_''.
684
685 --in-commit-order::
686 Print tree and blob ids in order of the commits. The tree
687 and blob ids are printed after they are first referenced
688 by a commit.
689
690 --objects-edge::
691 Similar to `--objects`, but also print the IDs of excluded
692 commits prefixed with a ``-'' character. This is used by
693 linkgit:git-pack-objects[1] to build a ``thin'' pack, which records
694 objects in deltified form based on objects contained in these
695 excluded commits to reduce network traffic.
696
697 --objects-edge-aggressive::
698 Similar to `--objects-edge`, but it tries harder to find excluded
699 commits at the cost of increased time. This is used instead of
700 `--objects-edge` to build ``thin'' packs for shallow repositories.
701
702 --indexed-objects::
703 Pretend as if all trees and blobs used by the index are listed
704 on the command line. Note that you probably want to use
705 `--objects`, too.
706
707 --unpacked::
708 Only useful with `--objects`; print the object IDs that are not
709 in packs.
710
711 --filter=<filter-spec>::
712 Only useful with one of the `--objects*`; omits objects (usually
713 blobs) from the list of printed objects. The '<filter-spec>'
714 may be one of the following:
715 +
716 The form '--filter=blob:none' omits all blobs.
717 +
718 The form '--filter=blob:limit=<n>[kmg]' omits blobs larger than n bytes
719 or units. n may be zero. The suffixes k, m, and g can be used to name
720 units in KiB, MiB, or GiB. For example, 'blob:limit=1k' is the same
721 as 'blob:limit=1024'.
722 +
723 The form '--filter=sparse:oid=<blob-ish>' uses a sparse-checkout
724 specification contained in the blob (or blob-expression) '<blob-ish>'
725 to omit blobs that would not be not required for a sparse checkout on
726 the requested refs.
727 +
728 The form '--filter=sparse:path=<path>' similarly uses a sparse-checkout
729 specification contained in <path>.
730 +
731 The form '--filter=tree:<depth>' omits all blobs and trees whose depth
732 from the root tree is >= <depth> (minimum depth if an object is located
733 at multiple depths in the commits traversed). <depth>=0 will not include
734 any trees or blobs unless included explicitly in the command-line (or
735 standard input when --stdin is used). <depth>=1 will include only the
736 tree and blobs which are referenced directly by a commit reachable from
737 <commit> or an explicitly-given object. <depth>=2 is like <depth>=1
738 while also including trees and blobs one more level removed from an
739 explicitly-given commit or tree.
740
741 --no-filter::
742 Turn off any previous `--filter=` argument.
743
744 --filter-print-omitted::
745 Only useful with `--filter=`; prints a list of the objects omitted
746 by the filter. Object IDs are prefixed with a ``~'' character.
747
748 --missing=<missing-action>::
749 A debug option to help with future "partial clone" development.
750 This option specifies how missing objects are handled.
751 +
752 The form '--missing=error' requests that rev-list stop with an error if
753 a missing object is encountered. This is the default action.
754 +
755 The form '--missing=allow-any' will allow object traversal to continue
756 if a missing object is encountered. Missing objects will silently be
757 omitted from the results.
758 +
759 The form '--missing=allow-promisor' is like 'allow-any', but will only
760 allow object traversal to continue for EXPECTED promisor missing objects.
761 Unexpected missing objects will raise an error.
762 +
763 The form '--missing=print' is like 'allow-any', but will also print a
764 list of the missing objects. Object IDs are prefixed with a ``?'' character.
765
766 --exclude-promisor-objects::
767 (For internal use only.) Prefilter object traversal at
768 promisor boundary. This is used with partial clone. This is
769 stronger than `--missing=allow-promisor` because it limits the
770 traversal, rather than just silencing errors about missing
771 objects.
772 endif::git-rev-list[]
773
774 --no-walk[=(sorted|unsorted)]::
775 Only show the given commits, but do not traverse their ancestors.
776 This has no effect if a range is specified. If the argument
777 `unsorted` is given, the commits are shown in the order they were
778 given on the command line. Otherwise (if `sorted` or no argument
779 was given), the commits are shown in reverse chronological order
780 by commit time.
781 Cannot be combined with `--graph`.
782
783 --do-walk::
784 Overrides a previous `--no-walk`.
785
786 Commit Formatting
787 ~~~~~~~~~~~~~~~~~
788
789 ifdef::git-rev-list[]
790 Using these options, linkgit:git-rev-list[1] will act similar to the
791 more specialized family of commit log tools: linkgit:git-log[1],
792 linkgit:git-show[1], and linkgit:git-whatchanged[1]
793 endif::git-rev-list[]
794
795 include::pretty-options.txt[]
796
797 --relative-date::
798 Synonym for `--date=relative`.
799
800 --date=<format>::
801 Only takes effect for dates shown in human-readable format, such
802 as when using `--pretty`. `log.date` config variable sets a default
803 value for the log command's `--date` option. By default, dates
804 are shown in the original time zone (either committer's or
805 author's). If `-local` is appended to the format (e.g.,
806 `iso-local`), the user's local time zone is used instead.
807 +
808 `--date=relative` shows dates relative to the current time,
809 e.g. ``2 hours ago''. The `-local` option has no effect for
810 `--date=relative`.
811 +
812 `--date=local` is an alias for `--date=default-local`.
813 +
814 `--date=iso` (or `--date=iso8601`) shows timestamps in a ISO 8601-like format.
815 The differences to the strict ISO 8601 format are:
816
817 - a space instead of the `T` date/time delimiter
818 - a space between time and time zone
819 - no colon between hours and minutes of the time zone
820
821 +
822 `--date=iso-strict` (or `--date=iso8601-strict`) shows timestamps in strict
823 ISO 8601 format.
824 +
825 `--date=rfc` (or `--date=rfc2822`) shows timestamps in RFC 2822
826 format, often found in email messages.
827 +
828 `--date=short` shows only the date, but not the time, in `YYYY-MM-DD` format.
829 +
830 `--date=raw` shows the date as seconds since the epoch (1970-01-01
831 00:00:00 UTC), followed by a space, and then the timezone as an offset
832 from UTC (a `+` or `-` with four digits; the first two are hours, and
833 the second two are minutes). I.e., as if the timestamp were formatted
834 with `strftime("%s %z")`).
835 Note that the `-local` option does not affect the seconds-since-epoch
836 value (which is always measured in UTC), but does switch the accompanying
837 timezone value.
838 +
839 `--date=human` shows the timezone if the timezone does not match the
840 current time-zone, and doesn't print the whole date if that matches
841 (ie skip printing year for dates that are "this year", but also skip
842 the whole date itself if it's in the last few days and we can just say
843 what weekday it was). For older dates the hour and minute is also
844 omitted.
845 +
846 `--date=unix` shows the date as a Unix epoch timestamp (seconds since
847 1970). As with `--raw`, this is always in UTC and therefore `-local`
848 has no effect.
849 +
850 `--date=format:...` feeds the format `...` to your system `strftime`,
851 except for %z and %Z, which are handled internally.
852 Use `--date=format:%c` to show the date in your system locale's
853 preferred format. See the `strftime` manual for a complete list of
854 format placeholders. When using `-local`, the correct syntax is
855 `--date=format-local:...`.
856 +
857 `--date=default` is the default format, and is similar to
858 `--date=rfc2822`, with a few exceptions:
859
860 - there is no comma after the day-of-week
861
862 - the time zone is omitted when the local time zone is used
863
864 ifdef::git-rev-list[]
865 --header::
866 Print the contents of the commit in raw-format; each record is
867 separated with a NUL character.
868 endif::git-rev-list[]
869
870 --parents::
871 Print also the parents of the commit (in the form "commit parent...").
872 Also enables parent rewriting, see 'History Simplification' above.
873
874 --children::
875 Print also the children of the commit (in the form "commit child...").
876 Also enables parent rewriting, see 'History Simplification' above.
877
878 ifdef::git-rev-list[]
879 --timestamp::
880 Print the raw commit timestamp.
881 endif::git-rev-list[]
882
883 --left-right::
884 Mark which side of a symmetric difference a commit is reachable from.
885 Commits from the left side are prefixed with `<` and those from
886 the right with `>`. If combined with `--boundary`, those
887 commits are prefixed with `-`.
888 +
889 For example, if you have this topology:
890 +
891 -----------------------------------------------------------------------
892 y---b---b branch B
893 / \ /
894 / .
895 / / \
896 o---x---a---a branch A
897 -----------------------------------------------------------------------
898 +
899 you would get an output like this:
900 +
901 -----------------------------------------------------------------------
902 $ git rev-list --left-right --boundary --pretty=oneline A...B
903
904 >bbbbbbb... 3rd on b
905 >bbbbbbb... 2nd on b
906 <aaaaaaa... 3rd on a
907 <aaaaaaa... 2nd on a
908 -yyyyyyy... 1st on b
909 -xxxxxxx... 1st on a
910 -----------------------------------------------------------------------
911
912 --graph::
913 Draw a text-based graphical representation of the commit history
914 on the left hand side of the output. This may cause extra lines
915 to be printed in between commits, in order for the graph history
916 to be drawn properly.
917 Cannot be combined with `--no-walk`.
918 +
919 This enables parent rewriting, see 'History Simplification' above.
920 +
921 This implies the `--topo-order` option by default, but the
922 `--date-order` option may also be specified.
923
924 --show-linear-break[=<barrier>]::
925 When --graph is not used, all history branches are flattened
926 which can make it hard to see that the two consecutive commits
927 do not belong to a linear branch. This option puts a barrier
928 in between them in that case. If `<barrier>` is specified, it
929 is the string that will be shown instead of the default one.
930
931 ifdef::git-rev-list[]
932 --count::
933 Print a number stating how many commits would have been
934 listed, and suppress all other output. When used together
935 with `--left-right`, instead print the counts for left and
936 right commits, separated by a tab. When used together with
937 `--cherry-mark`, omit patch equivalent commits from these
938 counts and print the count for equivalent commits separated
939 by a tab.
940 endif::git-rev-list[]
941
942 ifndef::git-rev-list[]
943 Diff Formatting
944 ~~~~~~~~~~~~~~~
945
946 Listed below are options that control the formatting of diff output.
947 Some of them are specific to linkgit:git-rev-list[1], however other diff
948 options may be given. See linkgit:git-diff-files[1] for more options.
949
950 -c::
951 With this option, diff output for a merge commit
952 shows the differences from each of the parents to the merge result
953 simultaneously instead of showing pairwise diff between a parent
954 and the result one at a time. Furthermore, it lists only files
955 which were modified from all parents.
956
957 --cc::
958 This flag implies the `-c` option and further compresses the
959 patch output by omitting uninteresting hunks whose contents in
960 the parents have only two variants and the merge result picks
961 one of them without modification.
962
963 --combined-all-paths::
964 This flag causes combined diffs (used for merge commits) to
965 list the name of the file from all parents. It thus only has
966 effect when -c or --cc are specified, and is likely only
967 useful if filename changes are detected (i.e. when either
968 rename or copy detection have been requested).
969
970 -m::
971 This flag makes the merge commits show the full diff like
972 regular commits; for each merge parent, a separate log entry
973 and diff is generated. An exception is that only diff against
974 the first parent is shown when `--first-parent` option is given;
975 in that case, the output represents the changes the merge
976 brought _into_ the then-current branch.
977
978 -r::
979 Show recursive diffs.
980
981 -t::
982 Show the tree objects in the diff output. This implies `-r`.
983 endif::git-rev-list[]