git-replay(1)

NAME#

git-replay - EXPERIMENTAL: Replay commits on a new base, works with bare repos too

SYNOPSIS#

(EXPERIMENTAL!) \ngit replay\n ([--contained] --onto <newbase> | --advance <branch>) <revision-range>...

DESCRIPTION#

Takes ranges of commits and replays them onto a new location. Leaves the working tree and the index untouched, and updates no references. The output of this command is meant to be used as input to git update-ref --stdin, which would update the relevant branches (see the OUTPUT section below).

THIS COMMAND IS EXPERIMENTAL. THE BEHAVIOR MAY CHANGE.

OPTIONS#

--onto <newbase>

When --onto is specified, the update-ref command(s) in the output will update the branch(es) in the revision range to point at the new commits, similar to the way how git rebase --update-refs updates multiple branches in the affected range.

--advance <branch>

When --advance is specified, the update-ref command(s) in the output will update the branch passed as an argument to --advance to point at the new commits (in other words, this mimics a cherry-pick operation).

<revision-range>

Commit Limiting#

Besides specifying a range of commits that should be listed using the special notations explained in the description, additional commit limiting may be applied.

Using more options generally further limits the output (e.g. --since=<date1> limits to commits newer than <date1>, and using it with --grep=<pattern> further limits to commits whose log message has a line that matches <pattern>), unless otherwise noted.

Note that these are applied before commit ordering and formatting options, such as --reverse.

-<number>, -n <number>, --max-count=<number>

--skip=<number>

--since=<date>, --after=<date>

--since-as-filter=<date>

--until=<date>, --before=<date>

--author=<pattern>, --committer=<pattern>

--grep-reflog=<pattern>

--grep=<pattern>

When --notes is in effect, the message from the notes is matched as if it were part of the log message.

--all-match

--invert-grep

-i, --regexp-ignore-case

--basic-regexp

-E, --extended-regexp

-F, --fixed-strings

-P, --perl-regexp

Support for these types of regular expressions is an optional compile-time dependency. If Git wasn’t compiled with support for them providing this option will cause it to die.

--remove-empty

--merges

--no-merges

--min-parents=<number>, --max-parents=<number>, --no-min-parents, --no-max-parents

--no-min-parents and --no-max-parents reset these limits (to no limit) again. Equivalent forms are --min-parents=0 (any commit has 0 or more parents) and --max-parents=-1 (negative numbers denote no upper limit).

--first-parent

--exclude-first-parent-only

--not

--all

--branches[=<pattern>]

--tags[=<pattern>]

--remotes[=<pattern>]

--glob=<glob-pattern>

--exclude=<glob-pattern>

The patterns given should not begin with refs/heads, refs/tags, or refs/remotes when applied to --branches, --tags, or --remotes, respectively, and they must begin with refs/ when applied to --glob or --all. If a trailing /* is intended, it must be given explicitly.

--exclude-hidden=[fetch|receive|uploadpack]

--reflog

--alternate-refs

--single-worktree

--ignore-missing

--bisect

--stdin

--cherry-mark

--cherry-pick

For example, if you have two branches, A and B, a usual way to list all commits on only one side of them is with --left-right (see the example below in the description of the --left-right option). However, it shows the commits that were cherry-picked from the other branch (for example, “3rd on b” may be cherry-picked from branch A). With this option, such pairs of commits are excluded from the output.

--left-only, --right-only

For example, --cherry-pick --right-only A...B omits those commits from B which are in A or are patch-equivalent to a commit in A. In other words, this lists the + commits from git cherry A B. More precisely, --cherry-pick --right-only --no-merges gives the exact list.

--cherry

-g, --walk-reflogs

With --pretty format other than oneline and reference (for obvious reasons), this causes the output to have two extra lines of information taken from the reflog. The reflog designator in the output may be shown as ref@{<Nth>} (where <Nth> is the reverse-chronological index in the reflog) or as ref@{<timestamp>} (with the <timestamp> for that entry), depending on a few rules:

Under --pretty=oneline, the commit message is prefixed with this information on the same line. This option cannot be combined with --reverse. See also git-reflog(1).

Under --pretty=reference, this information will not be shown at all.

--merge

--boundary

History Simplification#

Sometimes you are only interested in parts of the history, for example the commits modifying a particular <path>. But there are two parts of History Simplification, one part is selecting the commits and the other is how to do it, as there are various strategies to simplify the history.

The following options select the commits to be shown:

<paths>

--simplify-by-decoration

Note that extra commits can be shown to give a meaningful history.

The following options affect the way the simplification is performed:

Default mode

--show-pulls

--full-history

--dense

--sparse

--simplify-merges

--ancestry-path[=<commit>]

A more detailed explanation follows.

Suppose you specified foo as the <paths>. We shall call commits that modify foo !TREESAME, and the rest TREESAME. (In a diff filtered for foo, they look different and equal, respectively.)

In the following, we will always refer to the same example history to illustrate the differences between simplification settings. We assume that you are filtering for a file foo in this commit graph:

\n
          .-A---M---N---O---P---Q
\n
         /     /   /   /   /   /
\n
        I     B   C   D   E   Y
\n
         \   /   /   /   /   /
\n
          `-------------'   X

The horizontal line of history A---Q is taken to be the first parent of each merge. The commits are:

rev-list walks backwards through history, including or excluding commits based on whether --full-history and/or parent rewriting (via --parents or --children) are used. The following settings are available.

Default mode

This results in:

\n
          .-A---N---O
\n
         /     /   /
\n
        I---------D

Note how the rule to only follow the TREESAME parent, if one is available, removed B from consideration entirely. C was considered via N, but is TREESAME. Root commits are compared to an empty tree, so I is !TREESAME.

Parent/child relations are only visible with --parents, but that does not affect the commits selected in default mode, so we have shown the parent lines.

--full-history without parent rewriting

\n
        I  A  B  N  D  O  P  Q

M was excluded because it is TREESAME to both parents. E, C and B were all walked, but only B was !TREESAME, so the others do not appear.

Note that without parent rewriting, it is not really possible to talk about the parent/child relationships between the commits, so we show them disconnected.

--full-history with parent rewriting

Merges are always included. However, their parent list is rewritten: Along each parent, prune away commits that are not included themselves. This results in

\n
          .-A---M---N---O---P---Q
\n
         /     /   /   /   /
\n
        I     B   /   D   /
\n
         \   /   /   /   /
\n
          `-------------'

Compare to --full-history without rewriting above. Note that E was pruned away because it is TREESAME, but the parent list of P was rewritten to contain E's parent I. The same happened for C and N, and X, Y and Q.

In addition to the above settings, you can change whether TREESAME affects inclusion:

--dense

--sparse

Note that without --full-history, this still simplifies merges: if one of the parents is TREESAME, we follow only that one, so the other sides of the merge are never walked.

--simplify-merges

Then simplify each commit C to its replacement C' in the final history according to the following rules:

The effect of this is best shown by way of comparing to --full-history with parent rewriting. The example turns into:

\n
          .-A---M---N---O
\n
         /     /       /
\n
        I     B       D
\n
         \   /       /
\n
          `---------'

Note the major differences in N, P, and Q over --full-history:

There is another simplification mode available:

--ancestry-path[=<commit>]

As an example use case, consider the following commit history:

\n
            D---E-------F
\n
           /     \       \
\n
          B---C---G---H---I---J
\n
         /                     \
\n
        A-------K---------------L--M

A regular D..M computes the set of commits that are ancestors of M, but excludes the ones that are ancestors of D. This is useful to see what happened to the history leading to M since D, in the sense that “what does M have that did not exist in D”. The result in this example would be all the commits, except A and B (and D itself, of course).

When we want to find out what commits in M are contaminated with the bug introduced by D and need fixing, however, we might want to view only the subset of D..M that are actually descendants of D, i.e. excluding C and K. This is exactly what the --ancestry-path option does. Applied to the D..M range, it results in:

\n
                E-------F
\n
                 \       \
\n
                  G---H---I---J
\n
                               \
\n
                                L--M

We can also use --ancestry-path=D instead of --ancestry-path which means the same thing when applied to the D..M range but is just more explicit.

If we instead are interested in a given topic within this range, and all commits affected by that topic, we may only want to view the subset of D..M which contain that topic in their ancestry path. So, using --ancestry-path=H D..M for example would result in:

\n
                E
\n
                 \
\n
                  G---H---I---J
\n
                               \
\n
                                L--M

Whereas --ancestry-path=K D..M would result in

\n
                K---------------L--M

Before discussing another option, --show-pulls, we need to create a new example history.

A common problem users face when looking at simplified history is that a commit they know changed a file somehow does not appear in the file’s simplified history. Let’s demonstrate a new example and show how options such as --full-history and --simplify-merges works in that case:

\n
          .-A---M-----C--N---O---P
\n
         /     / \  \  \/   /   /
\n
        I     B   \  R-'`-Z'   /
\n
         \   /     \/         /
\n
          \ /      /\        /
\n
           `---X--'  `---Y--'

For this example, suppose I created file.txt which was modified by A, B, and X in different ways. The single-parent commits C, Z, and Y do not change file.txt. The merge commit M was created by resolving the merge conflict to include both changes from A and B and hence is not TREESAME to either. The merge commit R, however, was created by ignoring the contents of file.txt at M and taking only the contents of file.txt at X. Hence, R is TREESAME to X but not M. Finally, the natural merge resolution to create N is to take the contents of file.txt at R, so N is TREESAME to R but not C. The merge commits O and P are TREESAME to their first parents, but not to their second parents, Z and Y respectively.

When using the default mode, N and R both have a TREESAME parent, so those edges are walked and the others are ignored. The resulting history graph is:

\n
        I---X

When using --full-history, Git walks every edge. This will discover the commits A and B and the merge M, but also will reveal the merge commits O and P. With parent rewriting, the resulting graph is:

\n
          .-A---M--------N---O---P
\n
         /     / \  \  \/   /   /
\n
        I     B   \  R-'`--'   /
\n
         \   /     \/         /
\n
          \ /      /\        /
\n
           `---X--'  `------'

Here, the merge commits O and P contribute extra noise, as they did not actually contribute a change to file.txt. They only merged a topic that was based on an older version of file.txt. This is a common issue in repositories using a workflow where many contributors work in parallel and merge their topic branches along a single trunk: many unrelated merges appear in the --full-history results.

When using the --simplify-merges option, the commits O and P disappear from the results. This is because the rewritten second parents of O and P are reachable from their first parents. Those edges are removed and then the commits look like single-parent commits that are TREESAME to their parent. This also happens to the commit N, resulting in a history view as follows:

\n
          .-A---M--.
\n
         /     /    \
\n
        I     B      R
\n
         \   /      /
\n
          \ /      /
\n
           `---X--'

In this view, we see all of the important single-parent changes from A, B, and X. We also see the carefully-resolved merge M and the not-so-carefully-resolved merge R. This is usually enough information to determine why the commits A and B "disappeared" from history in the default view. However, there are a few issues with this approach.

The first issue is performance. Unlike any previous option, the --simplify-merges option requires walking the entire commit history before returning a single result. This can make the option difficult to use for very large repositories.

The second issue is one of auditing. When many contributors are working on the same repository, it is important which merge commits introduced a change into an important branch. The problematic merge R above is not likely to be the merge commit that was used to merge into an important branch. Instead, the merge N was used to merge R and X into the important branch. This commit may have information about why the change X came to override the changes from A and B in its commit message.

--show-pulls

When a merge commit is included by --show-pulls, the merge is treated as if it "pulled" the change from another branch. When using --show-pulls on this example (and no other options) the resulting graph is:

\n
        I---X---R---N

Here, the merge commits R and N are included because they pulled the commits X and R into the base branch, respectively. These merges are the reason the commits A and B do not appear in the default history.

When --show-pulls is paired with --simplify-merges, the graph includes all of the necessary information:

\n
          .-A---M--.   N
\n
         /     /    \ /
\n
        I     B      R
\n
         \   /      /
\n
          \ /      /
\n
           `---X--'

Notice that since M is reachable from R, the edge from N to M was simplified away. However, N still appears in the history as an important commit because it "pulled" the change R into the main branch.

The --simplify-by-decoration option allows you to view only the big picture of the topology of the history, by omitting commits that are not referenced by tags. Commits are marked as !TREESAME (in other words, kept after history simplification rules described above) if (1) they are referenced by tags, or (2) they change the contents of the paths given on the command line. All other commits are marked as TREESAME (subject to be simplified away).

Commit Ordering#

By default, the commits are shown in reverse chronological order.

--date-order

--author-date-order

--topo-order

For example, in a commit history like this:

\n
    ---1----2----4----7
\n
        \              \
\n
         3----5----6----8---

where the numbers denote the order of commit timestamps, git rev-list and friends with --date-order show the commits in the timestamp order: 8 7 6 5 4 3 2 1.

With --topo-order, they would show 8 6 5 3 7 4 2 1 (or 8 7 4 2 6 5 3 1); some older commits are shown before newer ones in order to avoid showing the commits from two parallel development track mixed together.

--reverse

Object Traversal#

These options are mostly targeted for packing of Git repositories.

--no-walk[=(sorted|unsorted)]

--do-walk

Commit Formatting#

--pretty[=<format>], --format=<format>

See the "PRETTY FORMATS" section for some additional details for each format. When =<format> part is omitted, it defaults to medium.

Note: you can specify the default pretty format in the repository configuration (see git-config(1)).

--abbrev-commit

This should make "--pretty=oneline" a whole lot more readable for people using 80-column terminals.

--no-abbrev-commit

--oneline

--encoding=<encoding>

--expand-tabs=<n>, --expand-tabs, --no-expand-tabs

By default, tabs are expanded in pretty formats that indent the log message by 4 spaces (i.e. medium, which is the default, full, and fuller).

--notes[=<ref>]

By default, the notes shown are from the notes refs listed in the core.notesRef and notes.displayRef variables (or corresponding environment overrides). See git-config(1) for more details.

With an optional <ref> argument, use the ref to find the notes to display. The ref can specify the full refname when it begins with refs/notes/; when it begins with notes/, refs/ and otherwise refs/notes/ is prefixed to form the full name of the ref.

Multiple --notes options can be combined to control which notes are being displayed. Examples: "--notes=foo" will show only notes from "refs/notes/foo"; "--notes=foo --notes" will show both notes from "refs/notes/foo" and from the default notes ref(s).

--no-notes

--show-notes-by-default

--show-notes[=<ref>], --[no-]standard-notes

--show-signature

--relative-date

--date=<format>

--date=relative shows dates relative to the current time, e.g. “2 hours ago”. The -local option has no effect for --date=relative.

--date=local is an alias for --date=default-local.

--date=iso (or --date=iso8601) shows timestamps in a ISO 8601-like format. The differences to the strict ISO 8601 format are:

--date=iso-strict (or --date=iso8601-strict) shows timestamps in strict ISO 8601 format.

--date=rfc (or --date=rfc2822) shows timestamps in RFC 2822 format, often found in email messages.

--date=short shows only the date, but not the time, in YYYY-MM-DD format.

--date=raw shows the date as seconds since the epoch (1970-01-01 00:00:00 UTC), followed by a space, and then the timezone as an offset from UTC (a + or - with four digits; the first two are hours, and the second two are minutes). I.e., as if the timestamp were formatted with strftime("%s %z")). Note that the -local option does not affect the seconds-since-epoch value (which is always measured in UTC), but does switch the accompanying timezone value.

--date=human shows the timezone if the timezone does not match the current time-zone, and doesn’t print the whole date if that matches (ie skip printing year for dates that are "this year", but also skip the whole date itself if it’s in the last few days and we can just say what weekday it was). For older dates the hour and minute is also omitted.

--date=unix shows the date as a Unix epoch timestamp (seconds since 1970). As with --raw, this is always in UTC and therefore -local has no effect.

--date=format:... feeds the format ... to your system strftime, except for %s, %z, and %Z, which are handled internally. Use --date=format:%c to show the date in your system locale’s preferred format. See the strftime manual for a complete list of format placeholders. When using -local, the correct syntax is --date=format-local:....

--date=default is the default format, and is based on ctime(3) output. It shows a single line with three-letter day of the week, three-letter month, day-of-month, hour-minute-seconds in "HH:MM:SS" format, followed by 4-digit year, plus timezone information, unless the local time zone is used, e.g. Thu Jan 1 00:00:00 1970 +0000.

--parents

--children

--left-right

For example, if you have this topology:

\n
             y---b---b  branch B
\n
            / \ /
\n
           /   .
\n
          /   / \
\n
         o---x---a---a  branch A

you would get an output like this:

\n
        $ git rev-list --left-right --boundary --pretty=oneline A...B
\n
        >bbbbbbb... 3rd on b
\n
        >bbbbbbb... 2nd on b
\n
        <aaaaaaa... 3rd on a
\n
        <aaaaaaa... 2nd on a
\n
        -yyyyyyy... 1st on b
\n
        -xxxxxxx... 1st on a

--graph

This enables parent rewriting, see History Simplification above.

This implies the --topo-order option by default, but the --date-order option may also be specified.

--show-linear-break[=<barrier>]

OUTPUT#

When there are no conflicts, the output of this command is usable as input to git update-ref --stdin. It is of the form:

update refs/heads/branch1 ${NEW_branch1_HASH} ${OLD_branch1_HASH}
update refs/heads/branch2 ${NEW_branch2_HASH} ${OLD_branch2_HASH}
update refs/heads/branch3 ${NEW_branch3_HASH} ${OLD_branch3_HASH}

where the number of refs updated depends on the arguments passed and the shape of the history being replayed. When using --advance, the number of refs updated is always one, but for --onto, it can be one or more (rebasing multiple branches simultaneously is supported).

EXIT STATUS#

For a successful, non-conflicted replay, the exit status is 0. When the replay has conflicts, the exit status is 1. If the replay is not able to complete (or start) due to some kind of error, the exit status is something other than 0 or 1.

EXAMPLES#

To simply rebase mybranch onto target:

$ git replay --onto target origin/main..mybranch
update refs/heads/mybranch ${NEW_mybranch_HASH} ${OLD_mybranch_HASH}

To cherry-pick the commits from mybranch onto target:

$ git replay --advance target origin/main..mybranch
update refs/heads/target ${NEW_target_HASH} ${OLD_target_HASH}

Note that the first two examples replay the exact same commits and on top of the exact same new base, they only differ in that the first provides instructions to make mybranch point at the new commits and the second provides instructions to make target point at them.

What if you have a stack of branches, one depending upon another, and you’d really like to rebase the whole set?

$ git replay --contained --onto origin/main origin/main..tipbranch
update refs/heads/branch1 ${NEW_branch1_HASH} ${OLD_branch1_HASH}
update refs/heads/branch2 ${NEW_branch2_HASH} ${OLD_branch2_HASH}
update refs/heads/tipbranch ${NEW_tipbranch_HASH} ${OLD_tipbranch_HASH}

When calling git replay, one does not need to specify a range of commits to replay using the syntax A..B; any range expression will do:

$ git replay --onto origin/main ^base branch1 branch2 branch3
update refs/heads/branch1 ${NEW_branch1_HASH} ${OLD_branch1_HASH}
update refs/heads/branch2 ${NEW_branch2_HASH} ${OLD_branch2_HASH}
update refs/heads/branch3 ${NEW_branch3_HASH} ${OLD_branch3_HASH}

This will simultaneously rebase branch1, branch2, and branch3, all commits they have since base, playing them on top of origin/main. These three branches may have commits on top of base that they have in common, but that does not need to be the case.

GIT#

Part of the git(1) suite