1GIT-REBASE(1) Git Manual GIT-REBASE(1)
2
6 git-rebase - Reapply commits on top of another base tip
7
9 git rebase [-i | --interactive] [<options>] [--exec <cmd>]
10 [--onto <newbase> | --keep-base] [<upstream> [<branch>]]
11 git rebase [-i | --interactive] [<options>] [--exec <cmd>] [--onto <newbase>]
12 --root [<branch>]
13 git rebase (--continue | --skip | --abort | --quit | --edit-todo | --show-current-patch)
14
16 If <branch> is specified, git rebase will perform an automatic git
17 switch <branch> before doing anything else. Otherwise it remains on the
18 current branch.
19 If <upstream> is not specified, the upstream configured in
21 branch.<name>.remote and branch.<name>.merge options will be used (see
22 git-config(1) for details) and the --fork-point option is assumed. If
23 you are currently not on any branch or if the current branch does not
24 have a configured upstream, the rebase will abort.
25 All changes made by commits in the current branch but that are not in
27 <upstream> are saved to a temporary area. This is the same set of
28 commits that would be shown by git log <upstream>..HEAD; or by git log
29 'fork_point'..HEAD, if --fork-point is active (see the description on
30 --fork-point below); or by git log HEAD, if the --root option is
31 specified.
32 The current branch is reset to <upstream> or <newbase> if the --onto
34 option was supplied. This has the exact same effect as git reset --hard
35 <upstream> (or <newbase>). ORIG_HEAD is set to point at the tip of the
36 branch before the reset.
37 The commits that were previously saved into the temporary area are then
39 reapplied to the current branch, one by one, in order. Note that any
40 commits in HEAD which introduce the same textual changes as a commit in
41 HEAD..<upstream> are omitted (i.e., a patch already accepted upstream
42 with a different commit message or timestamp will be skipped).
43 It is possible that a merge failure will prevent this process from
45 being completely automatic. You will have to resolve any such merge
46 failure and run git rebase --continue. Another option is to bypass the
47 commit that caused the merge failure with git rebase --skip. To check
48 out the original <branch> and remove the .git/rebase-apply working
49 files, use the command git rebase --abort instead.
50 Assume the following history exists and the current branch is "topic":
52 A---B---C topic
54 /
55 D---E---F---G master
56 From this point, the result of either of the following commands:
58 git rebase master
60 git rebase master topic
61 would be:
63 A'--B'--C' topic
65 /
66 D---E---F---G master
67 NOTE: The latter form is just a short-hand of git checkout topic
69 followed by git rebase master. When rebase exits topic will remain the
70 checked-out branch.
71 If the upstream branch already contains a change you have made (e.g.,
73 because you mailed a patch which was applied upstream), then that
74 commit will be skipped and warnings will be issued (if the merge
75 backend is used). For example, running git rebase master on the
76 following history (in which A' and A introduce the same set of changes,
77 but have different committer information):
78 A---B---C topic
80 /
81 D---E---A'---F master
82 will result in:
84 B'---C' topic
86 /
87 D---E---A'---F master
88 Here is how you would transplant a topic branch based on one branch to
90 another, to pretend that you forked the topic branch from the latter
91 branch, using rebase --onto.
92 First let’s assume your topic is based on branch next. For example, a
94 feature developed in topic depends on some functionality which is found
95 in next.
96 o---o---o---o---o master
98 \
99 o---o---o---o---o next
100 \
101 o---o---o topic
102 We want to make topic forked from branch master; for example, because
104 the functionality on which topic depends was merged into the more
105 stable master branch. We want our tree to look like this:
106 o---o---o---o---o master
108 | \
109 | o'--o'--o' topic
110 \
111 o---o---o---o---o next
112 We can get this using the following command:
114 git rebase --onto master next topic
116 Another example of --onto option is to rebase part of a branch. If we
118 have the following situation:
119 H---I---J topicB
121 /
122 E---F---G topicA
123 /
124 A---B---C---D master
125 then the command
127 git rebase --onto master topicA topicB
129 would result in:
131 H'--I'--J' topicB
133 /
134 | E---F---G topicA
135 |/
136 A---B---C---D master
137 This is useful when topicB does not depend on topicA.
139 A range of commits could also be removed with rebase. If we have the
141 following situation:
142 E---F---G---H---I---J topicA
144 then the command
146 git rebase --onto topicA~5 topicA~3 topicA
148 would result in the removal of commits F and G:
150 E---H'---I'---J' topicA
152 This is useful if F and G were flawed in some way, or should not be
154 part of topicA. Note that the argument to --onto and the <upstream>
155 parameter can be any valid commit-ish.
156 In case of conflict, git rebase will stop at the first problematic
158 commit and leave conflict markers in the tree. You can use git diff to
159 locate the markers (<<<<<<) and make edits to resolve the conflict. For
160 each file you edit, you need to tell Git that the conflict has been
161 resolved, typically this would be done with
162 git add <filename>
164 After resolving the conflict manually and updating the index with the
166 desired resolution, you can continue the rebasing process with
167 git rebase --continue
169 Alternatively, you can undo the git rebase with
171 git rebase --abort
173
175 --onto <newbase>
176 Starting point at which to create the new commits. If the --onto
177 option is not specified, the starting point is <upstream>. May be
178 any valid commit, and not just an existing branch name.
179 As a special case, you may use "A...B" as a shortcut for the merge
181 base of A and B if there is exactly one merge base. You can leave
182 out at most one of A and B, in which case it defaults to HEAD.
183 --keep-base
185 Set the starting point at which to create the new commits to the
186 merge base of <upstream> and <branch>. Running git rebase
187 --keep-base <upstream> <branch> is equivalent to running git rebase
188 --reapply-cherry-picks --no-fork-point --onto <upstream>...<branch>
189 <upstream> <branch>.
190 This option is useful in the case where one is developing a feature
192 on top of an upstream branch. While the feature is being worked on,
193 the upstream branch may advance and it may not be the best idea to
194 keep rebasing on top of the upstream but to keep the base commit
195 as-is. As the base commit is unchanged this option implies
196 --reapply-cherry-picks to avoid losing commits.
197 Although both this option and --fork-point find the merge base
199 between <upstream> and <branch>, this option uses the merge base as
200 the starting point on which new commits will be created, whereas
201 --fork-point uses the merge base to determine the set of commits
202 which will be rebased.
203 See also INCOMPATIBLE OPTIONS below.
205 <upstream>
207 Upstream branch to compare against. May be any valid commit, not
208 just an existing branch name. Defaults to the configured upstream
209 for the current branch.
210 <branch>
212 Working branch; defaults to HEAD.
213 --continue
215 Restart the rebasing process after having resolved a merge
216 conflict.
217 --abort
219 Abort the rebase operation and reset HEAD to the original branch.
220 If <branch> was provided when the rebase operation was started,
221 then HEAD will be reset to <branch>. Otherwise HEAD will be reset
222 to where it was when the rebase operation was started.
223 --quit
225 Abort the rebase operation but HEAD is not reset back to the
226 original branch. The index and working tree are also left unchanged
227 as a result. If a temporary stash entry was created using
228 --autostash, it will be saved to the stash list.
229 --apply
231 Use applying strategies to rebase (calling git-am internally). This
232 option may become a no-op in the future once the merge backend
233 handles everything the apply one does.
234 See also INCOMPATIBLE OPTIONS below.
236 --empty={drop,keep,ask}
238 How to handle commits that are not empty to start and are not clean
239 cherry-picks of any upstream commit, but which become empty after
240 rebasing (because they contain a subset of already upstream
241 changes). With drop (the default), commits that become empty are
242 dropped. With keep, such commits are kept. With ask (implied by
243 --interactive), the rebase will halt when an empty commit is
244 applied allowing you to choose whether to drop it, edit files more,
245 or just commit the empty changes. Other options, like --exec, will
246 use the default of drop unless -i/--interactive is explicitly
247 specified.
248 Note that commits which start empty are kept (unless
250 --no-keep-empty is specified), and commits which are clean
251 cherry-picks (as determined by git log --cherry-mark ...) are
252 detected and dropped as a preliminary step (unless
253 --reapply-cherry-picks or --keep-base is passed).
254 See also INCOMPATIBLE OPTIONS below.
256 --no-keep-empty, --keep-empty
258 Do not keep commits that start empty before the rebase (i.e. that
259 do not change anything from its parent) in the result. The default
260 is to keep commits which start empty, since creating such commits
261 requires passing the --allow-empty override flag to git commit,
262 signifying that a user is very intentionally creating such a commit
263 and thus wants to keep it.
264 Usage of this flag will probably be rare, since you can get rid of
266 commits that start empty by just firing up an interactive rebase
267 and removing the lines corresponding to the commits you don’t want.
268 This flag exists as a convenient shortcut, such as for cases where
269 external tools generate many empty commits and you want them all
270 removed.
271 For commits which do not start empty but become empty after
273 rebasing, see the --empty flag.
274 See also INCOMPATIBLE OPTIONS below.
276 --reapply-cherry-picks, --no-reapply-cherry-picks
278 Reapply all clean cherry-picks of any upstream commit instead of
279 preemptively dropping them. (If these commits then become empty
280 after rebasing, because they contain a subset of already upstream
281 changes, the behavior towards them is controlled by the --empty
282 flag.)
283 In the absence of --keep-base (or if --no-reapply-cherry-picks is
285 given), these commits will be automatically dropped. Because this
286 necessitates reading all upstream commits, this can be expensive in
287 repositories with a large number of upstream commits that need to
288 be read. When using the merge backend, warnings will be issued for
289 each dropped commit (unless --quiet is given). Advice will also be
290 issued unless advice.skippedCherryPicks is set to false (see git-
291 config(1)).
292 --reapply-cherry-picks allows rebase to forgo reading all upstream
294 commits, potentially improving performance.
295 See also INCOMPATIBLE OPTIONS below.
297 --allow-empty-message
299 No-op. Rebasing commits with an empty message used to fail and this
300 option would override that behavior, allowing commits with empty
301 messages to be rebased. Now commits with an empty message do not
302 cause rebasing to halt.
303 See also INCOMPATIBLE OPTIONS below.
305 --skip
307 Restart the rebasing process by skipping the current patch.
308 --edit-todo
310 Edit the todo list during an interactive rebase.
311 --show-current-patch
313 Show the current patch in an interactive rebase or when rebase is
314 stopped because of conflicts. This is the equivalent of git show
315 REBASE_HEAD.
316 -m, --merge
318 Using merging strategies to rebase (default).
319 Note that a rebase merge works by replaying each commit from the
321 working branch on top of the <upstream> branch. Because of this,
322 when a merge conflict happens, the side reported as ours is the
323 so-far rebased series, starting with <upstream>, and theirs is the
324 working branch. In other words, the sides are swapped.
325 See also INCOMPATIBLE OPTIONS below.
327 -s <strategy>, --strategy=<strategy>
329 Use the given merge strategy, instead of the default ort. This
330 implies --merge.
331 Because git rebase replays each commit from the working branch on
333 top of the <upstream> branch using the given strategy, using the
334 ours strategy simply empties all patches from the <branch>, which
335 makes little sense.
336 See also INCOMPATIBLE OPTIONS below.
338 -X <strategy-option>, --strategy-option=<strategy-option>
340 Pass the <strategy-option> through to the merge strategy. This
341 implies --merge and, if no strategy has been specified, -s ort.
342 Note the reversal of ours and theirs as noted above for the -m
343 option.
344 See also INCOMPATIBLE OPTIONS below.
346 --rerere-autoupdate, --no-rerere-autoupdate
348 After the rerere mechanism reuses a recorded resolution on the
349 current conflict to update the files in the working tree, allow it
350 to also update the index with the result of resolution.
351 --no-rerere-autoupdate is a good way to double-check what rerere
352 did and catch potential mismerges, before committing the result to
353 the index with a separate git add.
354 -S[<keyid>], --gpg-sign[=<keyid>], --no-gpg-sign
356 GPG-sign commits. The keyid argument is optional and defaults to
357 the committer identity; if specified, it must be stuck to the
358 option without a space. --no-gpg-sign is useful to countermand
359 both commit.gpgSign configuration variable, and earlier --gpg-sign.
360 -q, --quiet
362 Be quiet. Implies --no-stat.
363 -v, --verbose
365 Be verbose. Implies --stat.
366 --stat
368 Show a diffstat of what changed upstream since the last rebase. The
369 diffstat is also controlled by the configuration option
370 rebase.stat.
371 -n, --no-stat
373 Do not show a diffstat as part of the rebase process.
374 --no-verify
376 This option bypasses the pre-rebase hook. See also githooks(5).
377 --verify
379 Allows the pre-rebase hook to run, which is the default. This
380 option can be used to override --no-verify. See also githooks(5).
381 -C<n>
383 Ensure at least <n> lines of surrounding context match before and
384 after each change. When fewer lines of surrounding context exist
385 they all must match. By default no context is ever ignored. Implies
386 --apply.
387 See also INCOMPATIBLE OPTIONS below.
389 --no-ff, --force-rebase, -f
391 Individually replay all rebased commits instead of fast-forwarding
392 over the unchanged ones. This ensures that the entire history of
393 the rebased branch is composed of new commits.
394 You may find this helpful after reverting a topic branch merge, as
396 this option recreates the topic branch with fresh commits so it can
397 be remerged successfully without needing to "revert the reversion"
398 (see the revert-a-faulty-merge How-To[1] for details).
399 --fork-point, --no-fork-point
401 Use reflog to find a better common ancestor between <upstream> and
402 <branch> when calculating which commits have been introduced by
403 <branch>.
404 When --fork-point is active, fork_point will be used instead of
406 <upstream> to calculate the set of commits to rebase, where
407 fork_point is the result of git merge-base --fork-point <upstream>
408 <branch> command (see git-merge-base(1)). If fork_point ends up
409 being empty, the <upstream> will be used as a fallback.
410 If <upstream> or --keep-base is given on the command line, then the
412 default is --no-fork-point, otherwise the default is --fork-point.
413 See also rebase.forkpoint in git-config(1).
414 If your branch was based on <upstream> but <upstream> was rewound
416 and your branch contains commits which were dropped, this option
417 can be used with --keep-base in order to drop those commits from
418 your branch.
419 See also INCOMPATIBLE OPTIONS below.
421 --ignore-whitespace
423 Ignore whitespace differences when trying to reconcile differences.
424 Currently, each backend implements an approximation of this
425 behavior:
426 apply backend
428 When applying a patch, ignore changes in whitespace in context
429 lines. Unfortunately, this means that if the "old" lines being
430 replaced by the patch differ only in whitespace from the
431 existing file, you will get a merge conflict instead of a
432 successful patch application.
433 merge backend
435 Treat lines with only whitespace changes as unchanged when
436 merging. Unfortunately, this means that any patch hunks that
437 were intended to modify whitespace and nothing else will be
438 dropped, even if the other side had no changes that conflicted.
439 --whitespace=<option>
441 This flag is passed to the git apply program (see git-apply(1))
442 that applies the patch. Implies --apply.
443 See also INCOMPATIBLE OPTIONS below.
445 --committer-date-is-author-date
447 Instead of using the current time as the committer date, use the
448 author date of the commit being rebased as the committer date. This
449 option implies --force-rebase.
450 --ignore-date, --reset-author-date
452 Instead of using the author date of the original commit, use the
453 current time as the author date of the rebased commit. This option
454 implies --force-rebase.
455 See also INCOMPATIBLE OPTIONS below.
457 --signoff
459 Add a Signed-off-by trailer to all the rebased commits. Note that
460 if --interactive is given then only commits marked to be picked,
461 edited or reworded will have the trailer added.
462 See also INCOMPATIBLE OPTIONS below.
464 -i, --interactive
466 Make a list of the commits which are about to be rebased. Let the
467 user edit that list before rebasing. This mode can also be used to
468 split commits (see SPLITTING COMMITS below).
469 The commit list format can be changed by setting the configuration
471 option rebase.instructionFormat. A customized instruction format
472 will automatically have the long commit hash prepended to the
473 format.
474 See also INCOMPATIBLE OPTIONS below.
476 -r, --rebase-merges[=(rebase-cousins|no-rebase-cousins)]
478 By default, a rebase will simply drop merge commits from the todo
479 list, and put the rebased commits into a single, linear branch.
480 With --rebase-merges, the rebase will instead try to preserve the
481 branching structure within the commits that are to be rebased, by
482 recreating the merge commits. Any resolved merge conflicts or
483 manual amendments in these merge commits will have to be
484 resolved/re-applied manually.
485 By default, or when no-rebase-cousins was specified, commits which
487 do not have <upstream> as direct ancestor will keep their original
488 branch point, i.e. commits that would be excluded by git-log(1)'s
489 --ancestry-path option will keep their original ancestry by
490 default. If the rebase-cousins mode is turned on, such commits are
491 instead rebased onto <upstream> (or <onto>, if specified).
492 It is currently only possible to recreate the merge commits using
494 the ort merge strategy; different merge strategies can be used only
495 via explicit exec git merge -s <strategy> [...] commands.
496 See also REBASING MERGES and INCOMPATIBLE OPTIONS below.
498 -x <cmd>, --exec <cmd>
500 Append "exec <cmd>" after each line creating a commit in the final
501 history. <cmd> will be interpreted as one or more shell commands.
502 Any command that fails will interrupt the rebase, with exit code 1.
503 You may execute several commands by either using one instance of
505 --exec with several commands:
506 git rebase -i --exec "cmd1 && cmd2 && ..."
508 or by giving more than one --exec:
510 git rebase -i --exec "cmd1" --exec "cmd2" --exec ...
512 If --autosquash is used, exec lines will not be appended for the
514 intermediate commits, and will only appear at the end of each
515 squash/fixup series.
516 This uses the --interactive machinery internally, but it can be run
518 without an explicit --interactive.
519 See also INCOMPATIBLE OPTIONS below.
521 --root
523 Rebase all commits reachable from <branch>, instead of limiting
524 them with an <upstream>. This allows you to rebase the root
525 commit(s) on a branch. When used with --onto, it will skip changes
526 already contained in <newbase> (instead of <upstream>) whereas
527 without --onto it will operate on every change.
528 See also INCOMPATIBLE OPTIONS below.
530 --autosquash, --no-autosquash
532 When the commit log message begins with "squash! ..." or "fixup!
533 ..." or "amend! ...", and there is already a commit in the todo
534 list that matches the same ..., automatically modify the todo list
535 of rebase -i, so that the commit marked for squashing comes right
536 after the commit to be modified, and change the action of the moved
537 commit from pick to squash or fixup or fixup -C respectively. A
538 commit matches the ... if the commit subject matches, or if the
539 ... refers to the commit’s hash. As a fall-back, partial matches
540 of the commit subject work, too. The recommended way to create
541 fixup/amend/squash commits is by using the --fixup, --fixup=amend:
542 or --fixup=reword: and --squash options respectively of git-
543 commit(1).
544 If the --autosquash option is enabled by default using the
546 configuration variable rebase.autoSquash, this option can be used
547 to override and disable this setting.
548 See also INCOMPATIBLE OPTIONS below.
550 --autostash, --no-autostash
552 Automatically create a temporary stash entry before the operation
553 begins, and apply it after the operation ends. This means that you
554 can run rebase on a dirty worktree. However, use with care: the
555 final stash application after a successful rebase might result in
556 non-trivial conflicts.
557 --reschedule-failed-exec, --no-reschedule-failed-exec
559 Automatically reschedule exec commands that failed. This only makes
560 sense in interactive mode (or when an --exec option was provided).
561 Even though this option applies once a rebase is started, it’s set
563 for the whole rebase at the start based on either the
564 rebase.rescheduleFailedExec configuration (see git-config(1) or
565 "CONFIGURATION" below) or whether this option is provided.
566 Otherwise an explicit --no-reschedule-failed-exec at the start
567 would be overridden by the presence of
568 rebase.rescheduleFailedExec=true configuration.
569 --update-refs, --no-update-refs
571 Automatically force-update any branches that point to commits that
572 are being rebased. Any branches that are checked out in a worktree
573 are not updated in this way.
574 If the configuration variable rebase.updateRefs is set, then this
576 option can be used to override and disable this setting.
577
579 The following options:
580 • --apply
582 • --whitespace
584 • -C
586 are incompatible with the following options:
588 • --merge
590 • --strategy
592 • --strategy-option
594 • --allow-empty-message
596 • --[no-]autosquash
598 • --rebase-merges
600 • --interactive
602 • --exec
604 • --no-keep-empty
606 • --empty=
608 • --reapply-cherry-picks
610 • --edit-todo
612 • --update-refs
614 • --root when used in combination with --onto
616 In addition, the following pairs of options are incompatible:
618 • --keep-base and --onto
620 • --keep-base and --root
622 • --fork-point and --root
624
626 git rebase has two primary backends: apply and merge. (The apply
627 backend used to be known as the am backend, but the name led to
628 confusion as it looks like a verb instead of a noun. Also, the merge
629 backend used to be known as the interactive backend, but it is now used
630 for non-interactive cases as well. Both were renamed based on
631 lower-level functionality that underpinned each.) There are some subtle
632 differences in how these two backends behave:
633 Empty commits
635 The apply backend unfortunately drops intentionally empty commits, i.e.
636 commits that started empty, though these are rare in practice. It also
637 drops commits that become empty and has no option for controlling this
638 behavior.
639 The merge backend keeps intentionally empty commits by default (though
641 with -i they are marked as empty in the todo list editor, or they can
642 be dropped automatically with --no-keep-empty).
643 Similar to the apply backend, by default the merge backend drops
645 commits that become empty unless -i/--interactive is specified (in
646 which case it stops and asks the user what to do). The merge backend
647 also has an --empty={drop,keep,ask} option for changing the behavior of
648 handling commits that become empty.
649 Directory rename detection
651 Due to the lack of accurate tree information (arising from constructing
652 fake ancestors with the limited information available in patches),
653 directory rename detection is disabled in the apply backend. Disabled
654 directory rename detection means that if one side of history renames a
655 directory and the other adds new files to the old directory, then the
656 new files will be left behind in the old directory without any warning
657 at the time of rebasing that you may want to move these files into the
658 new directory.
659 Directory rename detection works with the merge backend to provide you
661 warnings in such cases.
662 Context
664 The apply backend works by creating a sequence of patches (by calling
665 format-patch internally), and then applying the patches in sequence
666 (calling am internally). Patches are composed of multiple hunks, each
667 with line numbers, a context region, and the actual changes. The line
668 numbers have to be taken with some fuzz, since the other side will
669 likely have inserted or deleted lines earlier in the file. The context
670 region is meant to help find how to adjust the line numbers in order to
671 apply the changes to the right lines. However, if multiple areas of the
672 code have the same surrounding lines of context, the wrong one can be
673 picked. There are real-world cases where this has caused commits to be
674 reapplied incorrectly with no conflicts reported. Setting diff.context
675 to a larger value may prevent such types of problems, but increases the
676 chance of spurious conflicts (since it will require more lines of
677 matching context to apply).
678 The merge backend works with a full copy of each relevant file,
680 insulating it from these types of problems.
681 Labelling of conflicts markers
683 When there are content conflicts, the merge machinery tries to annotate
684 each side’s conflict markers with the commits where the content came
685 from. Since the apply backend drops the original information about the
686 rebased commits and their parents (and instead generates new fake
687 commits based off limited information in the generated patches), those
688 commits cannot be identified; instead it has to fall back to a commit
689 summary. Also, when merge.conflictStyle is set to diff3 or zdiff3, the
690 apply backend will use "constructed merge base" to label the content
691 from the merge base, and thus provide no information about the merge
692 base commit whatsoever.
693 The merge backend works with the full commits on both sides of history
695 and thus has no such limitations.
696 Hooks
698 The apply backend has not traditionally called the post-commit hook,
699 while the merge backend has. Both have called the post-checkout hook,
700 though the merge backend has squelched its output. Further, both
701 backends only call the post-checkout hook with the starting point
702 commit of the rebase, not the intermediate commits nor the final
703 commit. In each case, the calling of these hooks was by accident of
704 implementation rather than by design (both backends were originally
705 implemented as shell scripts and happened to invoke other commands like
706 git checkout or git commit that would call the hooks). Both backends
707 should have the same behavior, though it is not entirely clear which,
708 if any, is correct. We will likely make rebase stop calling either of
709 these hooks in the future.
710 Interruptability
712 The apply backend has safety problems with an ill-timed interrupt; if
713 the user presses Ctrl-C at the wrong time to try to abort the rebase,
714 the rebase can enter a state where it cannot be aborted with a
715 subsequent git rebase --abort. The merge backend does not appear to
716 suffer from the same shortcoming. (See
717 https://lore.kernel.org/git/20200207132152.GC2868@szeder.dev/ for
718 details.)
719 Commit Rewording
721 When a conflict occurs while rebasing, rebase stops and asks the user
722 to resolve. Since the user may need to make notable changes while
723 resolving conflicts, after conflicts are resolved and the user has run
724 git rebase --continue, the rebase should open an editor and ask the
725 user to update the commit message. The merge backend does this, while
726 the apply backend blindly applies the original commit message.
727 Miscellaneous differences
729 There are a few more behavioral differences that most folks would
730 probably consider inconsequential but which are mentioned for
731 completeness:
732 • Reflog: The two backends will use different wording when describing
734 the changes made in the reflog, though both will make use of the
735 word "rebase".
736 • Progress, informational, and error messages: The two backends
738 provide slightly different progress and informational messages.
739 Also, the apply backend writes error messages (such as "Your files
740 would be overwritten...") to stdout, while the merge backend writes
741 them to stderr.
742 • State directories: The two backends keep their state in different
744 directories under .git/
745
747 The merge mechanism (git merge and git pull commands) allows the
748 backend merge strategies to be chosen with -s option. Some strategies
749 can also take their own options, which can be passed by giving
750 -X<option> arguments to git merge and/or git pull.
751 ort
753 This is the default merge strategy when pulling or merging one
754 branch. This strategy can only resolve two heads using a 3-way
755 merge algorithm. When there is more than one common ancestor that
756 can be used for 3-way merge, it creates a merged tree of the common
757 ancestors and uses that as the reference tree for the 3-way merge.
758 This has been reported to result in fewer merge conflicts without
759 causing mismerges by tests done on actual merge commits taken from
760 Linux 2.6 kernel development history. Additionally this strategy
761 can detect and handle merges involving renames. It does not make
762 use of detected copies. The name for this algorithm is an acronym
763 ("Ostensibly Recursive’s Twin") and came from the fact that it was
764 written as a replacement for the previous default algorithm,
765 recursive.
766 The ort strategy can take the following options:
768 ours
770 This option forces conflicting hunks to be auto-resolved
771 cleanly by favoring our version. Changes from the other tree
772 that do not conflict with our side are reflected in the merge
773 result. For a binary file, the entire contents are taken from
774 our side.
775 This should not be confused with the ours merge strategy, which
777 does not even look at what the other tree contains at all. It
778 discards everything the other tree did, declaring our history
779 contains all that happened in it.
780 theirs
782 This is the opposite of ours; note that, unlike ours, there is
783 no theirs merge strategy to confuse this merge option with.
784 ignore-space-change, ignore-all-space, ignore-space-at-eol,
786 ignore-cr-at-eol
787 Treats lines with the indicated type of whitespace change as
788 unchanged for the sake of a three-way merge. Whitespace changes
789 mixed with other changes to a line are not ignored. See also
790 git-diff(1) -b, -w, --ignore-space-at-eol, and
791 --ignore-cr-at-eol.
792 • If their version only introduces whitespace changes to a
794 line, our version is used;
795 • If our version introduces whitespace changes but their
797 version includes a substantial change, their version is
798 used;
799 • Otherwise, the merge proceeds in the usual way.
801 renormalize
803 This runs a virtual check-out and check-in of all three stages
804 of a file when resolving a three-way merge. This option is
805 meant to be used when merging branches with different clean
806 filters or end-of-line normalization rules. See "Merging
807 branches with differing checkin/checkout attributes" in
808 gitattributes(5) for details.
809 no-renormalize
811 Disables the renormalize option. This overrides the
812 merge.renormalize configuration variable.
813 find-renames[=<n>]
815 Turn on rename detection, optionally setting the similarity
816 threshold. This is the default. This overrides the
817 merge.renames configuration variable. See also git-diff(1)
818 --find-renames.
819 rename-threshold=<n>
821 Deprecated synonym for find-renames=<n>.
822 subtree[=<path>]
824 This option is a more advanced form of subtree strategy, where
825 the strategy makes a guess on how two trees must be shifted to
826 match with each other when merging. Instead, the specified path
827 is prefixed (or stripped from the beginning) to make the shape
828 of two trees to match.
829 recursive
831 This can only resolve two heads using a 3-way merge algorithm. When
832 there is more than one common ancestor that can be used for 3-way
833 merge, it creates a merged tree of the common ancestors and uses
834 that as the reference tree for the 3-way merge. This has been
835 reported to result in fewer merge conflicts without causing
836 mismerges by tests done on actual merge commits taken from Linux
837 2.6 kernel development history. Additionally this can detect and
838 handle merges involving renames. It does not make use of detected
839 copies. This was the default strategy for resolving two heads from
840 Git v0.99.9k until v2.33.0.
841 The recursive strategy takes the same options as ort. However,
843 there are three additional options that ort ignores (not documented
844 above) that are potentially useful with the recursive strategy:
845 patience
847 Deprecated synonym for diff-algorithm=patience.
848 diff-algorithm=[patience|minimal|histogram|myers]
850 Use a different diff algorithm while merging, which can help
851 avoid mismerges that occur due to unimportant matching lines
852 (such as braces from distinct functions). See also git-diff(1)
853 --diff-algorithm. Note that ort specifically uses
854 diff-algorithm=histogram, while recursive defaults to the
855 diff.algorithm config setting.
856 no-renames
858 Turn off rename detection. This overrides the merge.renames
859 configuration variable. See also git-diff(1) --no-renames.
860 resolve
862 This can only resolve two heads (i.e. the current branch and
863 another branch you pulled from) using a 3-way merge algorithm. It
864 tries to carefully detect criss-cross merge ambiguities. It does
865 not handle renames.
866 octopus
868 This resolves cases with more than two heads, but refuses to do a
869 complex merge that needs manual resolution. It is primarily meant
870 to be used for bundling topic branch heads together. This is the
871 default merge strategy when pulling or merging more than one
872 branch.
873 ours
875 This resolves any number of heads, but the resulting tree of the
876 merge is always that of the current branch head, effectively
877 ignoring all changes from all other branches. It is meant to be
878 used to supersede old development history of side branches. Note
879 that this is different from the -Xours option to the recursive
880 merge strategy.
881 subtree
883 This is a modified ort strategy. When merging trees A and B, if B
884 corresponds to a subtree of A, B is first adjusted to match the
885 tree structure of A, instead of reading the trees at the same
886 level. This adjustment is also done to the common ancestor tree.
887 With the strategies that use 3-way merge (including the default, ort),
889 if a change is made on both branches, but later reverted on one of the
890 branches, that change will be present in the merged result; some people
891 find this behavior confusing. It occurs because only the heads and the
892 merge base are considered when performing a merge, not the individual
893 commits. The merge algorithm therefore considers the reverted change as
894 no change at all, and substitutes the changed version instead.
895
897 You should understand the implications of using git rebase on a
898 repository that you share. See also RECOVERING FROM UPSTREAM REBASE
899 below.
900 When the rebase is run, it will first execute a pre-rebase hook if one
902 exists. You can use this hook to do sanity checks and reject the rebase
903 if it isn’t appropriate. Please see the template pre-rebase hook script
904 for an example.
905 Upon completion, <branch> will be the current branch.
907
909 Rebasing interactively means that you have a chance to edit the commits
910 which are rebased. You can reorder the commits, and you can remove them
911 (weeding out bad or otherwise unwanted patches).
912 The interactive mode is meant for this type of workflow:
914 1. have a wonderful idea
916 2. hack on the code
918 3. prepare a series for submission
920 4. submit
922 where point 2. consists of several instances of
924 a) regular use
926 1. finish something worthy of a commit
928 2. commit
930 b) independent fixup
932 1. realize that something does not work
934 2. fix that
936 3. commit it
938 Sometimes the thing fixed in b.2. cannot be amended to the not-quite
940 perfect commit it fixes, because that commit is buried deeply in a
941 patch series. That is exactly what interactive rebase is for: use it
942 after plenty of "a"s and "b"s, by rearranging and editing commits, and
943 squashing multiple commits into one.
944 Start it with the last commit you want to retain as-is:
946 git rebase -i <after-this-commit>
948 An editor will be fired up with all the commits in your current branch
950 (ignoring merge commits), which come after the given commit. You can
951 reorder the commits in this list to your heart’s content, and you can
952 remove them. The list looks more or less like this:
953 pick deadbee The oneline of this commit
955 pick fa1afe1 The oneline of the next commit
956 ...
957 The oneline descriptions are purely for your pleasure; git rebase will
959 not look at them but at the commit names ("deadbee" and "fa1afe1" in
960 this example), so do not delete or edit the names.
961 By replacing the command "pick" with the command "edit", you can tell
963 git rebase to stop after applying that commit, so that you can edit the
964 files and/or the commit message, amend the commit, and continue
965 rebasing.
966 To interrupt the rebase (just like an "edit" command would do, but
968 without cherry-picking any commit first), use the "break" command.
969 If you just want to edit the commit message for a commit, replace the
971 command "pick" with the command "reword".
972 To drop a commit, replace the command "pick" with "drop", or just
974 delete the matching line.
975 If you want to fold two or more commits into one, replace the command
977 "pick" for the second and subsequent commits with "squash" or "fixup".
978 If the commits had different authors, the folded commit will be
979 attributed to the author of the first commit. The suggested commit
980 message for the folded commit is the concatenation of the first
981 commit’s message with those identified by "squash" commands, omitting
982 the messages of commits identified by "fixup" commands, unless "fixup
983 -c" is used. In that case the suggested commit message is only the
984 message of the "fixup -c" commit, and an editor is opened allowing you
985 to edit the message. The contents (patch) of the "fixup -c" commit are
986 still incorporated into the folded commit. If there is more than one
987 "fixup -c" commit, the message from the final one is used. You can also
988 use "fixup -C" to get the same behavior as "fixup -c" except without
989 opening an editor.
990 git rebase will stop when "pick" has been replaced with "edit" or when
992 a command fails due to merge errors. When you are done editing and/or
993 resolving conflicts you can continue with git rebase --continue.
994 For example, if you want to reorder the last 5 commits, such that what
996 was HEAD~4 becomes the new HEAD. To achieve that, you would call git
997 rebase like this:
998 $ git rebase -i HEAD~5
1000 And move the first patch to the end of the list.
1002 You might want to recreate merge commits, e.g. if you have a history
1004 like this:
1005 X
1007 \
1008 A---M---B
1009 /
1010 ---o---O---P---Q
1011 Suppose you want to rebase the side branch starting at "A" to "Q". Make
1013 sure that the current HEAD is "B", and call
1014 $ git rebase -i -r --onto Q O
1016 Reordering and editing commits usually creates untested intermediate
1018 steps. You may want to check that your history editing did not break
1019 anything by running a test, or at least recompiling at intermediate
1020 points in history by using the "exec" command (shortcut "x"). You may
1021 do so by creating a todo list like this one:
1022 pick deadbee Implement feature XXX
1024 fixup f1a5c00 Fix to feature XXX
1025 exec make
1026 pick c0ffeee The oneline of the next commit
1027 edit deadbab The oneline of the commit after
1028 exec cd subdir; make test
1029 ...
1030 The interactive rebase will stop when a command fails (i.e. exits with
1032 non-0 status) to give you an opportunity to fix the problem. You can
1033 continue with git rebase --continue.
1034 The "exec" command launches the command in a shell (the one specified
1036 in $SHELL, or the default shell if $SHELL is not set), so you can use
1037 shell features (like "cd", ">", ";" ...). The command is run from the
1038 root of the working tree.
1039 $ git rebase -i --exec "make test"
1041 This command lets you check that intermediate commits are compilable.
1043 The todo list becomes like that:
1044 pick 5928aea one
1046 exec make test
1047 pick 04d0fda two
1048 exec make test
1049 pick ba46169 three
1050 exec make test
1051 pick f4593f9 four
1052 exec make test
1053
1055 In interactive mode, you can mark commits with the action "edit".
1056 However, this does not necessarily mean that git rebase expects the
1057 result of this edit to be exactly one commit. Indeed, you can undo the
1058 commit, or you can add other commits. This can be used to split a
1059 commit into two:
1060 • Start an interactive rebase with git rebase -i <commit>^, where
1062 <commit> is the commit you want to split. In fact, any commit range
1063 will do, as long as it contains that commit.
1064 • Mark the commit you want to split with the action "edit".
1066 • When it comes to editing that commit, execute git reset HEAD^. The
1068 effect is that the HEAD is rewound by one, and the index follows
1069 suit. However, the working tree stays the same.
1070 • Now add the changes to the index that you want to have in the first
1072 commit. You can use git add (possibly interactively) or git gui (or
1073 both) to do that.
1074 • Commit the now-current index with whatever commit message is
1076 appropriate now.
1077 • Repeat the last two steps until your working tree is clean.
1079 • Continue the rebase with git rebase --continue.
1081 If you are not absolutely sure that the intermediate revisions are
1083 consistent (they compile, pass the testsuite, etc.) you should use git
1084 stash to stash away the not-yet-committed changes after each commit,
1085 test, and amend the commit if fixes are necessary.
1086
1088 Rebasing (or any other form of rewriting) a branch that others have
1089 based work on is a bad idea: anyone downstream of it is forced to
1090 manually fix their history. This section explains how to do the fix
1091 from the downstream’s point of view. The real fix, however, would be to
1092 avoid rebasing the upstream in the first place.
1093 To illustrate, suppose you are in a situation where someone develops a
1095 subsystem branch, and you are working on a topic that is dependent on
1096 this subsystem. You might end up with a history like the following:
1097 o---o---o---o---o---o---o---o master
1099 \
1100 o---o---o---o---o subsystem
1101 \
1102 *---*---* topic
1103 If subsystem is rebased against master, the following happens:
1105 o---o---o---o---o---o---o---o master
1107 \ \
1108 o---o---o---o---o o'--o'--o'--o'--o' subsystem
1109 \
1110 *---*---* topic
1111 If you now continue development as usual, and eventually merge topic to
1113 subsystem, the commits from subsystem will remain duplicated forever:
1114 o---o---o---o---o---o---o---o master
1116 \ \
1117 o---o---o---o---o o'--o'--o'--o'--o'--M subsystem
1118 \ /
1119 *---*---*-..........-*--* topic
1120 Such duplicates are generally frowned upon because they clutter up
1122 history, making it harder to follow. To clean things up, you need to
1123 transplant the commits on topic to the new subsystem tip, i.e., rebase
1124 topic. This becomes a ripple effect: anyone downstream from topic is
1125 forced to rebase too, and so on!
1126 There are two kinds of fixes, discussed in the following subsections:
1128 Easy case: The changes are literally the same.
1130 This happens if the subsystem rebase was a simple rebase and had no
1131 conflicts.
1132 Hard case: The changes are not the same.
1134 This happens if the subsystem rebase had conflicts, or used
1135 --interactive to omit, edit, squash, or fixup commits; or if the
1136 upstream used one of commit --amend, reset, or a full history
1137 rewriting command like filter-repo[2].
1138 The easy case
1140 Only works if the changes (patch IDs based on the diff contents) on
1141 subsystem are literally the same before and after the rebase subsystem
1142 did.
1143 In that case, the fix is easy because git rebase knows to skip changes
1145 that are already present in the new upstream (unless
1146 --reapply-cherry-picks is given). So if you say (assuming you’re on
1147 topic)
1148 $ git rebase subsystem
1150 you will end up with the fixed history
1152 o---o---o---o---o---o---o---o master
1154 \
1155 o'--o'--o'--o'--o' subsystem
1156 \
1157 *---*---* topic
1158 The hard case
1160 Things get more complicated if the subsystem changes do not exactly
1161 correspond to the ones before the rebase.
1162 Note
1164 While an "easy case recovery" sometimes appears to be successful
1165 even in the hard case, it may have unintended consequences. For
1166 example, a commit that was removed via git rebase --interactive
1167 will be resurrected!
1168 The idea is to manually tell git rebase "where the old subsystem ended
1170 and your topic began", that is, what the old merge base between them
1171 was. You will have to find a way to name the last commit of the old
1172 subsystem, for example:
1173 • With the subsystem reflog: after git fetch, the old tip of
1175 subsystem is at subsystem@{1}. Subsequent fetches will increase the
1176 number. (See git-reflog(1).)
1177 • Relative to the tip of topic: knowing that your topic has three
1179 commits, the old tip of subsystem must be topic~3.
1180 You can then transplant the old subsystem..topic to the new tip by
1182 saying (for the reflog case, and assuming you are on topic already):
1183 $ git rebase --onto subsystem subsystem@{1}
1185 The ripple effect of a "hard case" recovery is especially bad: everyone
1187 downstream from topic will now have to perform a "hard case" recovery
1188 too!
1189
1191 The interactive rebase command was originally designed to handle
1192 individual patch series. As such, it makes sense to exclude merge
1193 commits from the todo list, as the developer may have merged the
1194 then-current master while working on the branch, only to rebase all the
1195 commits onto master eventually (skipping the merge commits).
1196 However, there are legitimate reasons why a developer may want to
1198 recreate merge commits: to keep the branch structure (or "commit
1199 topology") when working on multiple, inter-related branches.
1200 In the following example, the developer works on a topic branch that
1202 refactors the way buttons are defined, and on another topic branch that
1203 uses that refactoring to implement a "Report a bug" button. The output
1204 of git log --graph --format=%s -5 may look like this:
1205 * Merge branch 'report-a-bug'
1207 |\
1208 | * Add the feedback button
1209 * | Merge branch 'refactor-button'
1210 |\ \
1211 | |/
1212 | * Use the Button class for all buttons
1213 | * Extract a generic Button class from the DownloadButton one
1214 The developer might want to rebase those commits to a newer master
1216 while keeping the branch topology, for example when the first topic
1217 branch is expected to be integrated into master much earlier than the
1218 second one, say, to resolve merge conflicts with changes to the
1219 DownloadButton class that made it into master.
1220 This rebase can be performed using the --rebase-merges option. It will
1222 generate a todo list looking like this:
1223 label onto
1225 # Branch: refactor-button
1227 reset onto
1228 pick 123456 Extract a generic Button class from the DownloadButton one
1229 pick 654321 Use the Button class for all buttons
1230 label refactor-button
1231 # Branch: report-a-bug
1233 reset refactor-button # Use the Button class for all buttons
1234 pick abcdef Add the feedback button
1235 label report-a-bug
1236 reset onto
1238 merge -C a1b2c3 refactor-button # Merge 'refactor-button'
1239 merge -C 6f5e4d report-a-bug # Merge 'report-a-bug'
1240 In contrast to a regular interactive rebase, there are label, reset and
1242 merge commands in addition to pick ones.
1243 The label command associates a label with the current HEAD when that
1245 command is executed. These labels are created as worktree-local refs
1246 (refs/rewritten/<label>) that will be deleted when the rebase finishes.
1247 That way, rebase operations in multiple worktrees linked to the same
1248 repository do not interfere with one another. If the label command
1249 fails, it is rescheduled immediately, with a helpful message how to
1250 proceed.
1251 The reset command resets the HEAD, index and worktree to the specified
1253 revision. It is similar to an exec git reset --hard <label>, but
1254 refuses to overwrite untracked files. If the reset command fails, it is
1255 rescheduled immediately, with a helpful message how to edit the todo
1256 list (this typically happens when a reset command was inserted into the
1257 todo list manually and contains a typo).
1258 The merge command will merge the specified revision(s) into whatever is
1260 HEAD at that time. With -C <original-commit>, the commit message of the
1261 specified merge commit will be used. When the -C is changed to a
1262 lower-case -c, the message will be opened in an editor after a
1263 successful merge so that the user can edit the message.
1264 If a merge command fails for any reason other than merge conflicts
1266 (i.e. when the merge operation did not even start), it is rescheduled
1267 immediately.
1268 By default, the merge command will use the ort merge strategy for
1270 regular merges, and octopus for octopus merges. One can specify a
1271 default strategy for all merges using the --strategy argument when
1272 invoking rebase, or can override specific merges in the interactive
1273 list of commands by using an exec command to call git merge explicitly
1274 with a --strategy argument. Note that when calling git merge explicitly
1275 like this, you can make use of the fact that the labels are
1276 worktree-local refs (the ref refs/rewritten/onto would correspond to
1277 the label onto, for example) in order to refer to the branches you want
1278 to merge.
1279 Note: the first command (label onto) labels the revision onto which the
1281 commits are rebased; The name onto is just a convention, as a nod to
1282 the --onto option.
1283 It is also possible to introduce completely new merge commits from
1285 scratch by adding a command of the form merge <merge-head>. This form
1286 will generate a tentative commit message and always open an editor to
1287 let the user edit it. This can be useful e.g. when a topic branch turns
1288 out to address more than a single concern and wants to be split into
1289 two or even more topic branches. Consider this todo list:
1290 pick 192837 Switch from GNU Makefiles to CMake
1292 pick 5a6c7e Document the switch to CMake
1293 pick 918273 Fix detection of OpenSSL in CMake
1294 pick afbecd http: add support for TLS v1.3
1295 pick fdbaec Fix detection of cURL in CMake on Windows
1296 The one commit in this list that is not related to CMake may very well
1298 have been motivated by working on fixing all those bugs introduced by
1299 switching to CMake, but it addresses a different concern. To split this
1300 branch into two topic branches, the todo list could be edited like
1301 this:
1302 label onto
1304 pick afbecd http: add support for TLS v1.3
1306 label tlsv1.3
1307 reset onto
1309 pick 192837 Switch from GNU Makefiles to CMake
1310 pick 918273 Fix detection of OpenSSL in CMake
1311 pick fdbaec Fix detection of cURL in CMake on Windows
1312 pick 5a6c7e Document the switch to CMake
1313 label cmake
1314 reset onto
1316 merge tlsv1.3
1317 merge cmake
1318
1320 Everything below this line in this section is selectively included from
1321 the git-config(1) documentation. The content is the same as what’s
1322 found there:
1323 rebase.backend
1325 Default backend to use for rebasing. Possible choices are apply or
1326 merge. In the future, if the merge backend gains all remaining
1327 capabilities of the apply backend, this setting may become unused.
1328 rebase.stat
1330 Whether to show a diffstat of what changed upstream since the last
1331 rebase. False by default.
1332 rebase.autoSquash
1334 If set to true enable --autosquash option by default.
1335 rebase.autoStash
1337 When set to true, automatically create a temporary stash entry
1338 before the operation begins, and apply it after the operation ends.
1339 This means that you can run rebase on a dirty worktree. However,
1340 use with care: the final stash application after a successful
1341 rebase might result in non-trivial conflicts. This option can be
1342 overridden by the --no-autostash and --autostash options of git-
1343 rebase(1). Defaults to false.
1344 rebase.updateRefs
1346 If set to true enable --update-refs option by default.
1347 rebase.missingCommitsCheck
1349 If set to "warn", git rebase -i will print a warning if some
1350 commits are removed (e.g. a line was deleted), however the rebase
1351 will still proceed. If set to "error", it will print the previous
1352 warning and stop the rebase, git rebase --edit-todo can then be
1353 used to correct the error. If set to "ignore", no checking is done.
1354 To drop a commit without warning or error, use the drop command in
1355 the todo list. Defaults to "ignore".
1356 rebase.instructionFormat
1358 A format string, as specified in git-log(1), to be used for the
1359 todo list during an interactive rebase. The format will
1360 automatically have the long commit hash prepended to the format.
1361 rebase.abbreviateCommands
1363 If set to true, git rebase will use abbreviated command names in
1364 the todo list resulting in something like this:
1365 p deadbee The oneline of the commit
1367 p fa1afe1 The oneline of the next commit
1368 ...
1369 instead of:
1371 pick deadbee The oneline of the commit
1373 pick fa1afe1 The oneline of the next commit
1374 ...
1375 Defaults to false.
1377 rebase.rescheduleFailedExec
1379 Automatically reschedule exec commands that failed. This only makes
1380 sense in interactive mode (or when an --exec option was provided).
1381 This is the same as specifying the --reschedule-failed-exec option.
1382 rebase.forkPoint
1384 If set to false set --no-fork-point option by default.
1385 sequence.editor
1387 Text editor used by git rebase -i for editing the rebase
1388 instruction file. The value is meant to be interpreted by the shell
1389 when it is used. It can be overridden by the GIT_SEQUENCE_EDITOR
1390 environment variable. When not configured the default commit
1391 message editor is used instead.
1392
1394 Part of the git(1) suite
1395
1397 1. revert-a-faulty-merge How-To
1398 file:///usr/share/doc/git/howto/revert-a-faulty-merge.html
1399 2. filter-repo
1401 https://github.com/newren/git-filter-repo
1402Git 2.39.1 2023-01-13 GIT-REBASE(1)