GIT-REBASE(1)

📦
version 2.56.0 · 28 versions
, in arch rolling

NAME 💅

Okay so this is git-rebase from GIT-REBASE(1) — literally “Reapply commits on top of another base tip” ✨. Bestie, it’s the glow-up tool for your commit history, and it ate.

SYNOPSIS 🎀

git rebase [-i | --interactive] [<options>] [--exec <cmd>]
        [--onto <newbase> | --keep-base] [<upstream> [<branch>]]
git rebase [-i | --interactive] [<options>] [--exec <cmd>] [--onto <newbase>]
        --root [<branch>]
git rebase (--continue|--skip|--abort|--quit|--edit-todo|--show-current-patch)

DESCRIPTION 🦋

Transplant a series of commits onto a different starting point. You can also use git rebase to reorder or combine commits — see INTERACTIVE MODE below for how to do that. Periodt.

For example, imagine that you have been working on the topic branch in this history, and you want to “catch up” to the work done on the master branch. It’s giving classic main-character plot twist.

             A---B---C topic
            /
       D---E---F---G master

You want to transplant the commits you made on topic since it diverged from master (i.e. A, B, and C), on top of the current master. You can do this by running git rebase master while the topic branch is checked out. If you want to rebase topic while on another branch, git rebase master topic is a shortcut for git checkout topic && git rebase master. Slay, honestly.

                     A'--B'--C' topic
                    /
       D---E---F---G master

If there is a merge conflict during this process, git rebase will stop at the first problematic commit and leave conflict markers. If this happens, you can do one of these things:

  1. Resolve the conflict. You can use git diff to find the markers (<<<<<<) and make edits to resolve the conflict. For each file you edit, you need to tell Git that the conflict has been resolved. You can mark the conflict as resolved with git add <filename>. After resolving all of the conflicts, you can continue the rebasing process with
git rebase --continue
  1. Stop the git rebase and return your branch to its original state with
git rebase --abort
  1. Skip the commit that caused the merge conflict with
git rebase --skip

If you don’t specify an <upstream> to rebase onto, the upstream configured in branch.<name>.remote and branch.<name>.merge options will be used (see git-config(1) for details) and the --fork-point option is assumed. If you are currently not on any branch or if the current branch does not have a configured upstream, the rebase will abort. No upstream, no rebase, no vibe.

Here is a simplified description of what git rebase <upstream> does:

  1. Make a list of all commits on your current branch since it branched off from <upstream> that do not have an equivalent commit in <upstream>.
  2. Check out <upstream> with the equivalent of git checkout --detach <upstream>.
  3. Replay the commits, one by one, in order. This is similar to running git cherry-pick <commit> for each commit. See REBASING MERGES for how merges are handled.
  4. Update your branch to point to the final commit with the equivalent of git checkout -B <branch>.
Note — When starting the rebase, ORIG_HEAD is set to point to the commit at the tip of the to-be-rebased branch. However, ORIG_HEAD is not guaranteed to still point to that commit at the end of the rebase if other commands that change ORIG_HEAD (like git reset) are used during the rebase. The previous branch tip, however, is accessible using the reflog of the current branch (i.e. @{1}, see gitrevisions(7)). 💖

TRANSPLANTING A TOPIC BRANCH WITH --ONTO 🧚

Here is how you would transplant a topic branch based on one branch to another, to pretend that you forked the topic branch from the latter branch, using rebase --onto. It’s lowkey giving “fake it till you make it,” but for history.

First let’s assume your topic is based on branch next. For example, a feature developed in topic depends on some functionality which is found in next.

       o---o---o---o---o  master
            \
             o---o---o---o---o  next
                              \
                               o---o---o  topic

We want to make topic forked from branch master; for example, because the functionality on which topic depends was merged into the more stable master branch. We want our tree to look like this:

       o---o---o---o---o  master
           |            \
           |             o'--o'--o'  topic
            \
             o---o---o---o---o  next

We can get this using the following command:

git rebase --onto master next topic

Another example of --onto option is to rebase part of a branch. If we have the following situation:

                               H---I---J topicB
                              /
                     E---F---G  topicA
                    /
       A---B---C---D  master

then the command

git rebase --onto master topicA topicB

would result in:

                    H'--I'--J'  topicB
                   /
                   | E---F---G  topicA
                   |/
       A---B---C---D  master

This is useful when topicB does not depend on topicA. Iconic separation, honestly.

A range of commits could also be removed with rebase. If we have the following situation:

       E---F---G---H---I---J  topicA

then the command

git rebase --onto topicA~5 topicA~3 topicA

would result in the removal of commits F and G:

       E---H'---I'---J'  topicA

This is useful if F and G were flawed in some way, or should not be part of topicA. Note that the argument to --onto and the <upstream> parameter can be any valid commit-ish. Any. Valid. Commit-ish. 💅

MODE OPTIONS 🎀

The options in this section cannot be used with any other option — including not with each other. These are solo acts, bestie. No group chat.

  • --continue — Restart the rebasing process after having resolved a merge conflict.
  • --skip — Restart the rebasing process by skipping the current patch.
  • --abort — Abort the rebase operation and reset HEAD to the original branch. If <branch> was provided when the rebase operation was started, then HEAD will be reset to <branch>. Otherwise HEAD will be reset to where it was when the rebase operation was started.
  • --quit — Abort the rebase operation but HEAD is not reset back to the original branch. The index and working tree are also left unchanged as a result. If a temporary stash entry was created using --autostash, it will be saved to the stash list.
  • --edit-todo — Edit the todo list during an interactive rebase.
  • --show-current-patch — Show the current patch in an interactive rebase or when rebase is stopped because of conflicts. This is the equivalent of git show REBASE_HEAD.

OPTIONS 🌈

  • --onto <newbase> — Starting point at which to create the new commits. If the --onto option is not specified, the starting point is <upstream>. May be any valid commit, and not just an existing branch name.

As a special case, you may use "A...B" as a shortcut for the merge base of A and B if there is exactly one merge base. You can leave out at most one of A and B, in which case it defaults to HEAD.

See TRANSPLANTING A TOPIC BRANCH WITH --ONTO above for examples. 🫶

  • --keep-base — Set the starting point at which to create the new commits to the merge base of <upstream> and <branch>. Running git rebase --keep-base <upstream> <branch> is equivalent to running git rebase --reapply-cherry-picks --no-fork-point --onto <upstream>...<branch> <upstream> <branch>.

This option is useful in the case where one is developing a feature on top of an upstream branch. While the feature is being worked on, the upstream branch may advance and it may not be the best idea to keep rebasing on top of the upstream but to keep the base commit as-is. As the base commit is unchanged this option implies --reapply-cherry-picks to avoid losing commits.

Although both this option and --fork-point find the merge base between <upstream> and <branch>, this option uses the merge base as the starting point on which new commits will be created, whereas --fork-point uses the merge base to determine the set of commits which will be rebased.

See also INCOMPATIBLE OPTIONS below.

  • <upstream> — Upstream branch to compare against. May be any valid commit, not just an existing branch name. Defaults to the configured upstream for the current branch.
  • <branch> — Working branch; defaults to HEAD.
  • --apply — Use applying strategies to rebase (calling git-am internally). This option may become a no-op in the future once the merge backend handles everything the apply one does.

See also INCOMPATIBLE OPTIONS below.

  • --empty=(drop|keep|stop) — How to handle commits that are not empty to start and are not clean cherry-picks of any upstream commit, but which become empty after rebasing (because they contain a subset of already upstream changes):
    • drop — The commit will be dropped. This is the default behavior.
    • keep — The commit will be kept. This option is implied when --exec is specified unless -i/--interactive is also specified.
    • stop, ask — The rebase will halt when the commit is applied, allowing you to choose whether to drop it, edit files more, or just commit the empty changes. This option is implied when -i/--interactive is specified. ask is a deprecated synonym of stop.

Note that commits which start empty are kept (unless --no-keep-empty is specified), and commits which are clean cherry-picks (as determined by git log --cherry-mark ...) are detected and dropped as a preliminary step (unless --reapply-cherry-picks or --keep-base is passed).

See also INCOMPATIBLE OPTIONS below.

  • --no-keep-empty, --keep-empty — Do not keep commits that start empty before the rebase (i.e. that do not change anything from its parent) in the result. The default is to keep commits which start empty, since creating such commits requires passing the --allow-empty override flag to git commit, signifying that a user is very intentionally creating such a commit and thus wants to keep it.

Usage of this flag will probably be rare, since you can get rid of commits that start empty by just firing up an interactive rebase and removing the lines corresponding to the commits you don’t want. This flag exists as a convenient shortcut, such as for cases where external tools generate many empty commits and you want them all removed.

For commits which do not start empty but become empty after rebasing, see the --empty flag.

See also INCOMPATIBLE OPTIONS below. 🫶

  • --reapply-cherry-picks, --no-reapply-cherry-picks — bestie, this one's about those upstream commits that are already, like, spiritually in your branch 🍒. Instead of preemptively dropping every clean cherry-pick of an upstream commit, it re-applies them all. (And if those commits then turn out empty after rebasing — because they're just a subset of what upstream already has — how we treat them is a whole separate thing, controlled by the --empty flag.)

Without --keep-base (or if you pass --no-reapply-cherry-picks), those commits get automatically dropped. And because that means reading, like, every upstream commit, it can get expensive in repos with a huge number of upstream commits that need reading 😭. When you're on the merge backend, you'll get a warning for each dropped commit (unless --quiet is chilling there), plus advice unless advice.skippedCherryPicks is set to false (peep git-config(1)).

Fun little twist: --reapply-cherry-picks actually lets rebase forgo reading all upstream commits, which can lowkey improve performance ✨.

See also INCOMPATIBLE OPTIONS below.

  • --allow-empty-message — straight-up a no-op, bestie 💀. Back in the day, rebasing commits with an empty message used to fail, and this option overrode that so they'd go through. Now? Commits with an empty message don't halt the rebase at all. So this flag is just here, vibing.

See also INCOMPATIBLE OPTIONS below.

  • -m, --merge — rebase using merging strategies. This is the default, main character energy 👑.

Heads up though: a rebase merge works by replaying each commit from the working branch on top of the <upstream> branch. Which means when a merge conflict pops off, the side reported as ours is the so-far rebased series (starting with <upstream>), and theirs is the working branch. So yeah — the sides are swapped 🙈. Don't let that confuse you, fr fr.

See also INCOMPATIBLE OPTIONS below.

  • -s <strategy>, --strategy=<strategy> — use the merge strategy you name instead of the default ort. This implies --merge.

And because git rebase replays each commit from the working branch on top of <upstream> using whatever strategy you gave it, using the ours strategy just empties out all the patches from <branch> — which, ngl, makes basically no sense 🫠.

See also INCOMPATIBLE OPTIONS below.

  • -X <strategy-option>, --strategy-option=<strategy-option> — pass that <strategy-option> straight through to the merge strategy 🎀. This implies --merge and, if you didn't specify a strategy, -s ort. And remember that reversed ours/theirs thing we mentioned under -m — it applies here too.

See also INCOMPATIBLE OPTIONS below.

  • --rerere-autoupdate, --no-rerere-autoupdate — so after the rerere mechanism reuses a recorded resolution on the current conflict to update the files in the working tree, this lets it also update the index with the result of the resolution. And --no-rerere-autoupdate is a genuinely good way to double-check what git-rerere(1) did and catch potential mismerges before you commit the result to the index with a separate git-add(1) — chef's kiss for the careful girlies 💅.
  • -S[<keyid>], --gpg-sign[=<keyid>], --no-gpg-sign — GPG-sign commits 🔐. The keyid argument is optional and defaults to the committer identity; if you do specify it, it must be stuck to the option with no space. And --no-gpg-sign is the useful one for countermanding both the commit.gpgSign configuration variable and any earlier --gpg-sign.
  • -q, --quiet — be quiet 🤫. Implies --no-stat.
  • -v, --verbose — be verbose 🗣️. Implies --stat.
  • --stat — show a diffstat of what changed upstream since the last rebase. This is also controlled by the config option rebase.stat.
  • -n, --no-stat — do not show a diffstat as part of the rebase process 🚫.
  • --no-verify — bypasses the pre-rebase hook. See also githooks(5).
  • --verify — allows the pre-rebase hook to run, which is the default. This one's here so you can override --no-verify. See also githooks(5).
  • -C<n> — make sure at least <n> lines of surrounding context match before and after each change ✅. When fewer lines of surrounding context exist, they all have to match. By default no context is ever ignored. Implies --apply.

See also INCOMPATIBLE OPTIONS below.

  • --no-ff, --force-rebase, -f — individually replay all the rebased commits instead of fast-forwarding over the unchanged ones. This guarantees the entire history of the rebased branch is composed of new commits, no cap 🔥.

You may find this helpful after reverting a topic branch merge, since this option recreates the topic branch with fresh commits so it can be remerged successfully without needing to "revert the reversion" (see the revert-a-faulty-merge How-To[1] for details).

  • --fork-point, --no-fork-point — use the reflog to find a better common ancestor between <upstream> and <branch> when calculating which commits <branch> introduced 🔍.

When --fork-point is active, fork_point gets used instead of <upstream> to calculate the set of commits to rebase, where fork_point is the result of the git merge-base --fork-point <upstream> <branch> command (see git-merge-base(1)). And if fork_point ends up being empty, <upstream> is used as a fallback.

If <upstream> or --keep-base is given on the command line, then the default is --no-fork-point; otherwise the default is --fork-point. See also rebase.forkpoint in git-config(1).

And if your branch was based on <upstream> but <upstream> got rewound and your branch contains commits which were dropped, you can use this option with --keep-base to drop those commits from your branch 🦋.

See also INCOMPATIBLE OPTIONS below.

  • --ignore-whitespace — ignore whitespace differences when trying to reconcile differences. Currently, each backend implements its own little approximation of this behavior:
    • apply backend — when applying a patch, ignore changes in whitespace in context lines. Unfortunately, this means if the "old" lines being replaced by the patch differ only in whitespace from the existing file, you'll get a merge conflict instead of a successful patch application 😭.
    • merge backend — treat lines with only whitespace changes as unchanged when merging. Unfortunately, this means any patch hunks that were intended to modify whitespace and nothing else will be dropped, even if the other side had no changes that conflicted 💀.
  • --whitespace=<option> — this flag gets passed to the git apply program (see git-apply(1)) that applies the patch. Implies --apply.

See also INCOMPATIBLE OPTIONS below.

  • --committer-date-is-author-date — instead of using the current time as the committer date, use the author date of the commit being rebased as the committer date 🕰️. This option implies --force-rebase.
Warning — the history walking machinery assumes that commits have non-decreasing commit timestamps. You should genuinely consider whether you actually need this option. And then, if you do, only use it to override the committer date when rebasing commits on top of a base whose commit is older (in terms of the commit date) than the oldest commit you are applying (in terms of the author date). Like, be careful, bestie 🙏.
  • --ignore-date, --reset-author-date — instead of using the author date of the original commit, use the current time as the author date of the rebased commit. This option implies --force-rebase.

See also INCOMPATIBLE OPTIONS below.

  • --signoff — add a Signed-off-by trailer to all the rebased commits ✍️. Note that if --interactive is given, then only commits marked to be picked, edited or reworded will have the trailer added.

See also INCOMPATIBLE OPTIONS below.

  • --trailer=<trailer> — append the given trailer to every rebased commit message, processed via git-interpret-trailers(1) 🎀. This option implies --force-rebase.

See also INCOMPATIBLE OPTIONS below.

  • -i, --interactive — makes a list of the commits which are about to be rebased and lets you edit that list before rebasing ✨. This mode can also be used to split commits (see SPLITTING COMMITS below).

The commit list format can be changed by setting the configuration option rebase.instructionFormat. A customized instruction format will automatically have the commit hash prepended to the format.

See also INCOMPATIBLE OPTIONS below.

  • -r, --rebase-merges[=(rebase-cousins|no-rebase-cousins)], --no-rebase-merges — by default, a rebase just drops merge commits from the todo list and puts the rebased commits into a single, linear branch 🧵. With --rebase-merges, the rebase instead tries to preserve the branching structure within the commits that are to be rebased by recreating the merge commits. Any resolved merge conflicts or manual amendments in these merge commits will have to be resolved/re-applied manually, jsyk. And --no-rebase-merges can be used to countermand both the rebase.rebaseMerges config option and a previous --rebase-merges.

When rebasing merges, there are two modes: rebase-cousins and no-rebase-cousins. If you don't specify the mode, it defaults to no-rebase-cousins. In no-rebase-cousins mode, commits which don't have <upstream> as a direct ancestor keep their original branch point — i.e. commits that would be excluded by git-log(1)'s --ancestry-path option keep their original ancestry by default. In rebase-cousins mode, those commits are instead rebased onto <upstream> (or <onto>, if specified) 🌈.

It is currently only possible to recreate the merge commits using the ort merge strategy; different merge strategies can be used only via explicit exec git merge -s <strategy> [...] commands.

See also REBASING MERGES and INCOMPATIBLE OPTIONS below.

  • -x <cmd>, --exec <cmd> — appends exec <cmd> after each line creating a commit in the final history 🚀. <cmd> will be interpreted as one or more shell commands. Any command that fails will interrupt the rebase, with exit code 1.

You may execute several commands by either using one instance of --exec with several commands:

git rebase -i --exec "cmd1 && cmd2 && ..."

or by giving more than one --exec:

git rebase -i --exec "cmd1" --exec "cmd2" --exec ...

If --autosquash is used, exec lines will not be appended for the intermediate commits, and will only appear at the end of each squash/fixup series 👀.

This uses the --interactive machinery internally, but it can be run without an explicit --interactive.

See also INCOMPATIBLE OPTIONS below.

  • --root — rebase all commits reachable from <branch>, instead of limiting them with an <upstream>. This is the glow-up that lets you rebase the root commit(s) on a branch. See also INCOMPATIBLE OPTIONS below, bestie.
  • --autosquash, --no-autosquash — automatically squash commits with specially formatted messages into previous commits being rebased. If a commit message starts with "squash! ", "fixup! ", or "amend! ", the rest of the title is taken as a commit specifier, which matches a previous commit if it matches the title or the hash of that commit. If no commit matches fully, matches of the specifier with the start of commit titles are considered. In the rebase todo list, the actions of squash, fixup and amend commits are changed from pick to squash, fixup or fixup -C, respectively, and they are moved right after the commit they modify. The --interactive option can be used to review and edit the todo list before proceeding. The recommended way to create commits with squash markers is by using the --squash, --fixup, --fixup=amend: or --fixup=reword: options of git-commit(1), which take the target commit as an argument and automatically fill in the title of the new commit from that. Setting configuration variable rebase.autoSquash to true enables auto-squashing by default for interactive rebase. The --no-autosquash option can be used to override that setting. See also INCOMPATIBLE OPTIONS below.
  • --autostash, --no-autostash — automatically create a temporary stash entry before the operation begins, and apply it after the operation ends. This means you can run rebase on a dirty worktree. However, use with care: the final stash application after a successful rebase might result in non-trivial conflicts. It’s giving “convenient but don’t get messy,” fr.
  • --reschedule-failed-exec, --no-reschedule-failed-exec — automatically reschedule exec commands that failed. This only makes sense in interactive mode (or when an --exec option was provided). This option applies once a rebase is started. It is preserved for the whole rebase based on, in order, the command line option provided to the initial git rebase, the rebase.rescheduleFailedExec configuration (see git-config(1) or "CONFIGURATION" below), or it defaults to false. Recording this option for the whole rebase is a convenience feature. Otherwise an explicit --no-reschedule-failed-exec at the start would be overridden by the presence of a rebase.rescheduleFailedExec=true configuration when git rebase --continue is invoked. Currently, you cannot pass --[no-]reschedule-failed-exec to git rebase --continue. The audacity, honestly.
  • --update-refs, --no-update-refs — automatically force-update any branches that point to commits that are being rebased. Any branches that are checked out in a worktree are not updated in this way. If the configuration variable rebase.updateRefs is set, then this option can be used to override and disable this setting. See also INCOMPATIBLE OPTIONS below.

💥 INCOMPATIBLE OPTIONS 💥

The following options:

  • --apply
  • --whitespace
  • -C

are incompatible with the following options:

  • --merge
  • --strategy
  • --strategy-option
  • --autosquash
  • --rebase-merges
  • --interactive
  • --exec
  • --no-keep-empty
  • --empty=
  • --[no-]reapply-cherry-picks when used without --keep-base
  • --update-refs
  • --root when used without --onto
  • --trailer

In addition, the following pairs of options are incompatible:

  • --keep-base and --onto
  • --keep-base and --root
  • --fork-point and --root

🌈 BEHAVIORAL DIFFERENCES 🌈

git rebase has two primary backends: apply and merge. (The apply backend used to be known as the am backend, but the name led to confusion as it looks like a verb instead of a noun. Also, the merge backend used to be known as the interactive backend, but it is now used for non-interactive cases as well. Both were renamed based on lower-level functionality that underpinned each.) There are some subtle differences in how these two backends behave:

Empty commits 🫧

The apply backend unfortunately drops intentionally empty commits, i.e. commits that started empty, though these are rare in practice. It also drops commits that become empty and has no option for controlling this behavior.

The merge backend keeps intentionally empty commits by default (though with -i they are marked as empty in the todo list editor, or they can be dropped automatically with --no-keep-empty).

Similar to the apply backend, by default the merge backend drops commits that become empty unless -i/--interactive is specified (in which case it stops and asks the user what to do). The merge backend also has an --empty=(drop|keep|stop) option for changing the behavior of handling commits that become empty.

Directory rename detection 🎀

Due to the lack of accurate tree information (arising from constructing fake ancestors with the limited information available in patches), directory rename detection is disabled in the apply backend. Disabled directory rename detection means that if one side of history renames a directory and the other adds new files to the old directory, then the new files will be left behind in the old directory without any warning at the time of rebasing that you may want to move these files into the new directory.

Directory rename detection works with the merge backend to provide you warnings in such cases.

Context 🦋

The apply backend works by creating a sequence of patches (by calling format-patch internally), and then applying the patches in sequence (calling am internally). Patches are composed of multiple hunks, each with line numbers, a context region, and the actual changes. The line numbers have to be taken with some offset, since the other side will likely have inserted or deleted lines earlier in the file. The context region is meant to help find how to adjust the line numbers in order to apply the changes to the right lines. However, if multiple areas of the code have the same surrounding lines of context, the wrong one can be picked. There are real-world cases where this has caused commits to be reapplied incorrectly with no conflicts reported. Setting diff.context to a larger value may prevent such types of problems, but increases the chance of spurious conflicts (since it will require more lines of matching context to apply).

The merge backend works with a full copy of each relevant file, insulating it from these types of problems.

Labelling of conflicts markers 💅

When there are content conflicts, the merge machinery tries to annotate each side’s conflict markers with the commits where the content came from. Since the apply backend drops the original information about the rebased commits and their parents (and instead generates new fake commits based off limited information in the generated patches), those commits cannot be identified; instead it has to fall back to a commit summary. Also, when merge.conflictStyle is set to diff3 or zdiff3, the apply backend will use "constructed merge base" to label the content from the merge base, and thus provide no information about the merge base commit whatsoever.

The merge backend works with the full commits on both sides of history and thus has no such limitations.

Hooks 🪝

The apply backend has not traditionally called the post-commit hook, while the merge backend has. Both have called the post-checkout hook, though the merge backend has squelched its output. Further, both backends only call the post-checkout hook with the starting point commit of the rebase, not the intermediate commits nor the final commit. In each case, the calling of these hooks was by accident of implementation rather than by design (both backends were originally implemented as shell scripts and happened to invoke other commands like git checkout or git commit that would call the hooks). Both backends should have the same behavior, though it is not entirely clear which, if any, is correct. We will likely make rebase stop calling either of these hooks in the future.

Interruptability 😭

The apply backend has safety problems with an ill-timed interrupt; if the user presses Ctrl+C at the wrong time to try to abort the rebase, the rebase can enter a state where it cannot be aborted with a subsequent git rebase --abort. The merge backend does not appear to suffer from the same shortcoming. (See https://lore.kernel.org/git/20200207132152.GC2868@szeder.dev/ for details.)

Commit Rewording 👑

When a conflict occurs while rebasing, rebase stops and asks the user to resolve. Since the user may need to make notable changes while resolving conflicts, after conflicts are resolved and the user has run git rebase --continue, the rebase should open an editor and ask the user to update the commit message. The merge backend does this, while the apply backend blindly applies the original commit message.

Miscellaneous differences ✨

There are a few more behavioral differences that most folks would probably consider inconsequential but which are mentioned for completeness:

  • Reflog: The two backends will use different wording when describing the changes made in the reflog, though both will make use of the word "rebase".
  • Progress, informational, and error messages: The two backends provide slightly different progress and informational messages. Also, the apply backend writes error messages (such as "Your files would be overwritten...") to stdout, while the merge backend writes them to stderr.
  • State directories: The two backends keep their state in different directories under .git/

💅 MERGE STRATEGIES 💅

Okay so the merge mechanism — aka the git merge and git pull commands — lets you pick backend merge strategies with the -s option. Some strategies also have their own lil options, and you pass those by giving -X<option> arguments to git merge and/or git pull. slay.

🦋 ort 🦋

This is the default merge strategy when pulling or merging one branch. It can only resolve two heads using a 3-way merge algorithm. When there’s more than one common ancestor that can be used for 3-way merge, it creates a merged tree of the common ancestors and uses that as the reference tree for the 3-way merge. This has been reported to result in fewer merge conflicts without causing mismerges by tests done on actual merge commits taken from Linux 2.6 kernel development history. No cap, those are the receipts. Additionally this strategy can detect and handle merges involving renames. It does not make use of detected copies. The name for this algorithm is an acronym ("Ostensibly Recursive’s Twin") and came from the fact that it was written as a replacement for the previous default algorithm, recursive.

In the case where the path is a submodule, if the submodule commit used on one side of the merge is a descendant of the submodule commit used on the other side of the merge, Git attempts to fast-forward to the descendant. Otherwise, Git will treat this case as a conflict, suggesting as a resolution a submodule commit that is descendant of the conflicting ones, if one exists.

The ort strategy can take the following options:

  • ours — This option forces conflicting hunks to be auto-resolved cleanly by favoring our version. Changes from the other tree that do not conflict with our side are reflected in the merge result. For a binary file, the entire contents are taken from our side. This should not be confused with the ours merge strategy, which does not even look at what the other tree contains at all. It discards everything the other tree did, declaring our history contains all that happened in it. Do not mix these up, bestie 😭.
  • theirs — This is the opposite of ours; note that, unlike ours, there is no theirs merge strategy to confuse this merge option with. iykyk.
  • ignore-space-change, ignore-all-space, ignore-space-at-eol, ignore-cr-at-eol — Treats lines with the indicated type of whitespace change as unchanged for the sake of a three-way merge. Whitespace changes mixed with other changes to a line are not ignored. See also git-diff(1) -b, -w, --ignore-space-at-eol, and --ignore-cr-at-eol.
    • If their version only introduces whitespace changes to a line, our version is used;
    • If our version introduces whitespace changes but their version includes a substantial change, their version is used;
    • Otherwise, the merge proceeds in the usual way.
  • renormalize — This runs a virtual check-out and check-in of all three stages of any file which needs a three-way merge. This option is meant to be used when merging branches with different clean filters or end-of-line normalization rules. See "Merging branches with differing checkin/checkout attributes" in gitattributes(5) for details.
  • no-renormalize — Disables the renormalize option. This overrides the merge.renormalize configuration variable.
  • find-renames[=<n>] — Turn on rename detection, optionally setting the similarity threshold. This is the default. This overrides the merge.renames configuration variable. See also git-diff(1) --find-renames.
  • rename-threshold=<n> — Deprecated synonym for find-renames=<n>.
  • no-renames — Turn off rename detection. This overrides the merge.renames configuration variable. See also git-diff(1) --no-renames.
  • histogram — Deprecated synonym for diff-algorithm=histogram.
  • patience — Deprecated synonym for diff-algorithm=patience.
  • diff-algorithm=(histogram|minimal|myers|patience) — Use a different diff algorithm while merging, which can help avoid mismerges that occur due to unimportant matching lines (such as braces from distinct functions). See also git-diff(1) --diff-algorithm. Note that ort defaults to diff-algorithm=histogram, while regular diffs currently default to the diff.algorithm config setting.
  • subtree[=<path>] — This option is a more advanced form of subtree strategy, where the strategy makes a guess on how two trees must be shifted to match with each other when merging. Instead, the specified path is prefixed (or stripped from the beginning) to make the shape of two trees to match.

🌱 recursive 🌱

This is now a synonym for ort. It was an alternative implementation until v2.49.0, but was redirected to mean ort in v2.50.0. The previous recursive strategy was the default strategy for resolving two heads from Git v0.99.9k until v2.33.0. History lesson, but make it git.

🧩 resolve 🧩

This can only resolve two heads (i.e. the current branch and another branch you pulled from) using a 3-way merge algorithm. It tries to carefully detect criss-cross merge ambiguities. It does not handle renames. No rename support, just so we’re clear.

🐙 octopus 🐙

This resolves cases with more than two heads, but refuses to do a complex merge that needs manual resolution. It is primarily meant to be used for bundling topic branch heads together. This is the default merge strategy when pulling or merging more than one branch.

👑 ours 👑

This resolves any number of heads, but the resulting tree of the merge is always that of the current branch head, effectively ignoring all changes from all other branches. It is meant to be used to supersede old development history of side branches. Note that this is different from the -Xours option to the ort merge strategy. Same name energy, totally different behavior — stay woke.

🌳 subtree 🌳

This is a modified ort strategy. When merging trees A and B, if B corresponds to a subtree of A, B is first adjusted to match the tree structure of A, instead of reading the trees at the same level. This adjustment is also done to the common ancestor tree.

With the strategies that use 3-way merge (including the default, ort), if a change is made on both branches, but later reverted on one of the branches, that change will be present in the merged result; some people find this behavior confusing. It occurs because only the heads and the merge base are considered when performing a merge, not the individual commits. The merge algorithm therefore considers the reverted change as no change at all, and substitutes the changed version instead. It’s giving “math is mathing,” but make it merge logic.

🌸 NOTES 🌸

You should understand the implications of using git rebase on a repository that you share. See also RECOVERING FROM UPSTREAM REBASE below.

When the rebase is run, it will first execute a pre-rebase hook if one exists. You can use this hook to do sanity checks and reject the rebase if it isn’t appropriate. Please see the template pre-rebase hook script for an example.

Upon completion, <branch> will be the current branch. Periodt.

🎀 INTERACTIVE MODE 🎀

ok so interactive rebasing is literally where it gets fun, bestie. It means you actually get to edit the commits that are being rebased — reorder them however you want, and even yeet them entirely (that's the weeding-out-bad-or-otherwise-unwanted-patches move, and honestly it's giving spring cleaning 💅).

This mode was made for this exact workflow:

  1. have a wonderful idea ✨
  2. hack on the code
  3. prepare a series for submission
  4. submit

and step 2 is basically a bunch of these little loops stacked up:

a) regular use

  1. finish something worthy of a commit
  2. commit

b) independent fixup

  1. realize that something does not work 😭
  2. fix that
  3. commit it

Sometimes the thing you fixed in b.2 just cannot be amended into the not-quite-perfect commit it fixes, because that commit is buried deeply down in a patch series. And that, bestie, is exactly what interactive rebase is for: after you've done plenty of "a"s and "b"s, you use it to rearrange and edit commits, and squash multiple commits into one. so satisfying. 🫶

You start it with the last commit you want to keep as-is:

git rebase -i <after-this-commit>

An editor fires up (iconic) with all the commits in your current branch — merge commits get ignored, fyi — that come after the commit you gave it. You can reorder them to your heart's content and you can remove them. The list looks more or less like this:

pick deadbee The oneline of this commit
pick fa1afe1 The oneline of the next commit
...

Those oneline descriptions are purely for your pleasure — git rebase does not even look at them. It looks at the commit names (deadbee and fa1afe1 in this example), so do not delete or edit the names. no cap, that'll ruin everything 💀

Replace the command "pick" with "edit" and you're telling git rebase to stop after applying that commit, so you can edit the files and/or the commit message, amend the commit, and then continue rebasing. main character energy.

To interrupt the rebase — just like an "edit" command would do, but without cherry-picking any commit first — use the "break" command. 🦋

If you just want to edit the commit message for a commit, swap "pick" for "reword".

To drop a commit, replace "pick" with "drop", or honestly just delete the matching line. bye bestie 👋

Wanna fold two or more commits into one? Replace the command "pick" for the second and subsequent commits with "squash" or "fixup".

  • If the commits had different authors, the folded commit gets attributed to the author of the first commit.
  • The suggested commit message for the folded commit is the concatenation of the first commit's message with those identified by "squash" commands, omitting the messages of commits identified by "fixup" commands — unless fixup -c is used.
  • With fixup -c, the suggested commit message is only the message of the fixup -c commit, and an editor is opened so you can edit the message. The contents (patch) of the fixup -c commit are still incorporated into the folded commit.
  • If there's more than one fixup -c commit, the message from the final one is used.
  • You can also use fixup -C to get the same behavior as fixup -c, except without opening an editor.

git rebase will stop when "pick" has been replaced with "edit", or when a command fails due to merge errors. When you're done editing and/or resolving conflicts you can continue with git rebase --continue. slay.

For example, say you want to reorder the last 5 commits so that what was HEAD~4 becomes the new HEAD. To achieve that you'd call git rebase like this:

$ git rebase -i HEAD~5

And then move the first patch to the end of the list. ez 💖

You might also want to recreate merge commits, e.g. if your history looks like this:

               X
                \
             A---M---B
            /
    ---o---O---P---Q

Suppose you want to rebase the side branch starting at "A" to "Q". Make sure the current HEAD is "B", and call:

$ git rebase -i -r --onto Q O

Reordering and editing commits usually creates untested intermediate steps (lowkey scary ngl). You probably want to check that your history editing didn't break anything by running a test, or at least recompiling at intermediate points in history, using the "exec" command (shortcut "x"). You can do that by creating a todo list like this one:

pick deadbee Implement feature XXX
fixup f1a5c00 Fix to feature XXX
exec make
pick c0ffeee The oneline of the next commit
edit deadbab The oneline of the commit after
exec cd subdir; make test
...

The interactive rebase will stop when a command fails (i.e. exits with non-0 status) to give you a chance to fix the problem. You can continue with git rebase --continue.

The "exec" command launches the command in a shell (the default one, usually /bin/sh), so you can absolutely use shell features — things like "cd", ">", ";" ... yes bestie, all of it 🐚 The command is run from the root of the working tree.

$ git rebase -i --exec "make test"

This command lets you check that intermediate commits are compilable. The todo list becomes like that:

pick 5928aea one
exec make test
pick 04d0fda two
exec make test
pick ba46169 three
exec make test
pick f4593f9 four
exec make test

🌸 SPLITTING COMMITS 🌸

In interactive mode you can mark commits with the action "edit". But heads up — that does not necessarily mean git rebase expects the result of this edit to be exactly one commit. You can undo the commit, or you can add other commits. And you can use this to split a commit into two 👑

  • Start an interactive rebase with git rebase -i <commit>^, where <commit> is the commit you want to split. In fact, any commit range will do, as long as it contains that commit.
  • Mark the commit you want to split with the action "edit".
  • When it comes to editing that commit, execute git reset HEAD^. The effect is that the HEAD is rewound by one, and the index follows suit. However, the working tree stays the same.
  • Now add the changes to the index that you want to have in the first commit. You can use git add (possibly interactively) or git gui (or both) to do that.
  • Commit the now-current index with whatever commit message is appropriate now.
  • Repeat the last two steps until your working tree is clean.
  • Continue the rebase with git rebase --continue.
If you are not absolutely sure that the intermediate revisions are consistent (they compile, pass the testsuite, etc.) you should use git stash to stash away the not-yet-committed changes after each commit, test, and amend the commit if fixes are necessary. fr fr, don't skip this one 😭

💖 RECOVERING FROM UPSTREAM REBASE 💖

Okay so real talk, bestie: rebasing (or honestly any kind of rewriting) a branch that other people already based work on is a bad idea — anyone downstream of it gets forced to manually fix their history 😭. This section is the downstream POV on how to do that fix. But the actual real fix? Avoid rebasing the upstream in the first place, periodt. 💅

To illustrate, suppose you are in a situation where someone develops a subsystem branch, and you are working on a topic that is dependent on this subsystem. You might end up with a history like the following:

        o---o---o---o---o---o---o---o  master
             \
              o---o---o---o---o  subsystem
                               \
                                *---*---*  topic

If subsystem is rebased against master, the following happens:

        o---o---o---o---o---o---o---o  master
             \                       \
              o---o---o---o---o       o'--o'--o'--o'--o'  subsystem
                               \
                                *---*---*  topic

If you now continue development as usual, and eventually merge topic to subsystem, the commits from subsystem will remain duplicated forever:

        o---o---o---o---o---o---o---o  master
             \                       \
              o---o---o---o---o       o'--o'--o'--o'--o'--M  subsystem
                               \                         /
                                *---*---*-..........-*--*  topic

Such duplicates are generally frowned upon because they clutter up history, making it harder to follow. To clean things up, you need to transplant the commits on topic to the new subsystem tip, i.e., rebase topic. This becomes a ripple effect: anyone downstream from topic is forced to rebase too, and so on! 🌊

There are two kinds of fixes, discussed in the following subsections:

  • Easy case: The changes are literally the same. — This happens if the subsystem rebase was a simple rebase and had no conflicts.
  • Hard case: The changes are not the same. — This happens if the subsystem rebase had conflicts, or used --interactive to omit, edit, squash, or fixup commits; or if the upstream used one of commit --amend, reset, or a full history rewriting command like filter-repo[2].

💅 The easy case 💅

Only works if the changes (patch IDs based on the diff contents) on subsystem are literally the same before and after the rebase subsystem did.

In that case, the fix is easy because git rebase knows to skip changes that are already present in the new upstream (unless --reapply-cherry-picks is given). So if you say (assuming you’re on topic)

$ git rebase subsystem

you will end up with the fixed history

        o---o---o---o---o---o---o---o  master
                                     \
                                      o'--o'--o'--o'--o'  subsystem
                                                       \
                                                        *---*---*  topic

😭 The hard case 😭

Things get more complicated if the subsystem changes do not exactly correspond to the ones before the rebase.

Note 🫶 While an "easy case recovery" sometimes appears to be successful even in the hard case, it may have unintended consequences. For example, a commit that was removed via git rebase --interactive will be resurrected!

The idea is to manually tell git rebase "where the old subsystem ended and your topic began", that is, what the old merge base between them was. You will have to find a way to name the last commit of the old subsystem, for example:

  • With the subsystem reflog: after git fetch, the old tip of subsystem is at subsystem@{1}. Subsequent fetches will increase the number. (See git-reflog(1).)
  • Relative to the tip of topic: knowing that your topic has three commits, the old tip of subsystem must be topic~3.

You can then transplant the old subsystem..topic to the new tip by saying (for the reflog case, and assuming you are on topic already):

$ git rebase --onto subsystem subsystem@{1}

The ripple effect of a "hard case" recovery is especially bad: everyone downstream from topic will now have to perform a "hard case" recovery too! 💀

🦋 REBASING MERGES 🦋

The interactive rebase command was originally designed to handle individual patch series. As such, it makes sense to exclude merge commits from the todo list, as the developer may have merged the then-current master while working on the branch, only to rebase all the commits onto master eventually (skipping the merge commits).

However, there are legitimate reasons why a developer may want to recreate merge commits: to keep the branch structure (or "commit topology") when working on multiple, inter-related branches. 👑

In the following example, the developer works on a topic branch that refactors the way buttons are defined, and on another topic branch that uses that refactoring to implement a "Report a bug" button. The output of git log --graph --format=%s -5 may look like this:

*   Merge branch 'report-a-bug'
|\
| * Add the feedback button
* | Merge branch 'refactor-button'
|\ \
| |/
| * Use the Button class for all buttons
| * Extract a generic Button class from the DownloadButton one

The developer might want to rebase those commits to a newer master while keeping the branch topology, for example when the first topic branch is expected to be integrated into master much earlier than the second one, say, to resolve merge conflicts with changes to the DownloadButton class that made it into master.

This rebase can be performed using the --rebase-merges option. It will generate a todo list looking like this:

label onto

# Branch: refactor-button
reset onto
pick 123456 Extract a generic Button class from the DownloadButton one
pick 654321 Use the Button class for all buttons
label refactor-button

# Branch: report-a-bug
reset refactor-button # Use the Button class for all buttons
pick abcdef Add the feedback button
label report-a-bug

reset onto
merge -C a1b2c3 refactor-button # Merge 'refactor-button'
merge -C 6f5e4d report-a-bug # Merge 'report-a-bug'

In contrast to a regular interactive rebase, there are label, reset and merge commands in addition to pick ones. ✨

The label command associates a label with the current HEAD when that command is executed. These labels are created as worktree-local refs (refs/rewritten/<label>) that will be deleted when the rebase finishes. That way, rebase operations in multiple worktrees linked to the same repository do not interfere with one another. If the label command fails, it is rescheduled immediately, with a helpful message how to proceed.

The reset command resets the HEAD, index and worktree to the specified revision. It is similar to an exec git reset --hard <label>, but refuses to overwrite untracked files. If the reset command fails, it is rescheduled immediately, with a helpful message how to edit the todo list (this typically happens when a reset command was inserted into the todo list manually and contains a typo).

The merge command will merge the specified revision(s) into whatever is HEAD at that time. With -C <original-commit>, the commit message of the specified merge commit will be used. When the -C is changed to a lower-case -c, the message will be opened in an editor after a successful merge so that the user can edit the message.

If a merge command fails for any reason other than merge conflicts (i.e. when the merge operation did not even start), it is rescheduled immediately.

By default, the merge command will use the ort merge strategy for regular merges, and octopus for octopus merges. One can specify a default strategy for all merges using the --strategy argument when invoking rebase, or can override specific merges in the interactive list of commands by using an exec command to call git merge explicitly with a --strategy argument. Note that when calling git merge explicitly like this, you can make use of the fact that the labels are worktree-local refs (the ref refs/rewritten/onto would correspond to the label onto, for example) in order to refer to the branches you want to merge.

Note: the first command (label onto) labels the revision onto which the commits are rebased; The name onto is just a convention, as a nod to the --onto option.

It is also possible to introduce completely new merge commits from scratch by adding a command of the form merge <merge-head>. This form will generate a tentative commit message and always open an editor to let the user edit it. This can be useful e.g. when a topic branch turns out to address more than a single concern and wants to be split into two or even more topic branches. Consider this todo list:

pick 192837 Switch from GNU Makefiles to CMake
pick 5a6c7e Document the switch to CMake
pick 918273 Fix detection of OpenSSL in CMake
pick afbecd http: add support for TLS v1.3
pick fdbaec Fix detection of cURL in CMake on Windows

The one commit in this list that is not related to CMake may very well have been motivated by working on fixing all those bugs introduced by switching to CMake, but it addresses a different concern. To split this branch into two topic branches, the todo list could be edited like this:

label onto

pick afbecd http: add support for TLS v1.3
label tlsv1.3

reset onto
pick 192837 Switch from GNU Makefiles to CMake
pick 918273 Fix detection of OpenSSL in CMake
pick fdbaec Fix detection of cURL in CMake on Windows
pick 5a6c7e Document the switch to CMake
label cmake

reset onto
merge tlsv1.3
merge cmake

🎀 CONFIGURATION 🎀

okay so literally everything under this line? it's cherry-picked from the git-config(1) docs — same content, same facts, just vibing here so you don't have to tab-hop. bestie, it's all right here:

  • rebase.backend — the default backend for rebasing. your choices are apply or merge. and like… in the future, if the merge backend picks up all the remaining capabilities of the apply backend, this setting might just become unused. so watch this space, i guess 💅
  • rebase.stat — whether to show a diffstat of what changed upstream since the last rebase. defaults to false (so, off, if you're not about that diffstat life).
  • rebase.autoSquash — set it to true and you get the --autosquash option of git-rebase(1) turned on by default for interactive mode ✨ and yes, you can override it with --no-autosquash.
  • rebase.autoStash — set to true and git automatically creates a temporary stash entry before the operation starts, then applies it after the operation ends. translation: you can run rebase on a dirty worktree 🫶 but use with care, okay?? the final stash application after a successful rebase might result in non-trivial conflicts 😭. this one can be overridden by the --no-autostash and --autostash options of git-rebase(1). defaults to false.
  • rebase.updateRefs — set to true and the --update-refs option is enabled by default. easy.
  • rebase.missingCommitsCheck — the vibe-check on missing commits:
    • set it to "warn" and git rebase -i prints a warning if some commits are removed (like, say, a line got deleted) — but the rebase still proceeds, it's not gonna stop you 🙈
    • set it to "error" and it prints that same warning and then stops the rebase, and you can fix it with git rebase --edit-todo
    • set it to "ignore" and no checking happens at all
    • and if you want to drop a commit with no warning or error? use the drop command in the todo list, iconically clean
    • defaults to "ignore"
  • rebase.instructionFormat — a format string, exactly as specified in git-log(1), used for the todo list during an interactive rebase. heads up: the format automatically has the commit hash prepended to it. free hash, no cap 🎁
  • rebase.abbreviateCommands — set to true and git rebase uses abbreviated command names in the todo list, giving you something like:
p deadbee The oneline of the commit
p fa1afe1 The oneline of the next commit
...

instead of:

pick deadbee The oneline of the commit
pick fa1afe1 The oneline of the next commit
...

defaults to false. (yes it's just pick → p. she's efficient. she's minimal. 💅)

  • rebase.rescheduleFailedExec — automatically reschedule exec commands that failed. this only makes sense in interactive mode (or when an --exec option was provided). literally the same as specifying the --reschedule-failed-exec option 🔁
  • rebase.forkPoint — if set to false, sets the --no-fork-point option by default.
  • rebase.rebaseMerges — whether and how to set the --rebase-merges option by default. the values are rebase-cousins, no-rebase-cousins, or a boolean:
    • true or no-rebase-cousins → same as --rebase-merges=no-rebase-cousins
    • rebase-cousins → same as --rebase-merges=rebase-cousins
    • false → same as --no-rebase-merges
    • and passing --rebase-merges on the command line, with or without an argument, overrides any rebase.rebaseMerges configuration. command line stays main character 👑
  • rebase.maxLabelLength — when generating label names from commit subjects, truncate the names to this length. by default the names are truncated to a little less than NAME_MAX (so that e.g. .lock files can still be written for the corresponding loose refs — thoughtful, honestly 🌸)
  • sequence.editor — the text editor used by git rebase -i for editing the rebase instruction file. the value is meant to be interpreted by the shell when it's used 🐚 it can be overridden by the GIT_SEQUENCE_EDITOR environment variable. when it's not configured, the default commit message editor is used instead.

👑 GIT 👑

part of the git(1) suite. that's it, that's the whole section. she's short, she's iconic, she belongs to the family fr fr 💖

🦋 NOTES 🦋

  1. revert-a-faulty-merge How-To file:///usr/share/doc/git-doc/howto/revert-a-faulty-merge.html
  2. filter-repo https://github.com/newren/git-filter-repo

and that's a wrap on the whole page, bestie!! 🎀✨ you made it to the very end, which is genuinely a flex. this lil beauty is Git 2.56.0, dated 2026-09-27, and it's the GIT-REBASE(1) page — go forth and rebase responsibly, or at least rebase with confidence 💜🫶🌸

📦 where this page comes from