At GitHub, we had a repository that was triggering
git: merge-ort.c:3032: process_renames: Assertion `newinfo && !newinfo->merged.clean` failed.
during git replay.
This sounds similar to the somewhat recent f6ecb603ff (merge-ort: fix
directory rename on top of source of other rename/delete, 2025-08-06),
but the cause is different. Unlike that case, there are no
rename-to-self situations arising in this case, and new to this case it
can only be triggered during a replay operation on the 2nd or later
commit being replayed, never on the first merge in the sequence.
To trigger, the repository needs:
* an upstream which:
* renames a file to a different directory, e.g.
old/file -> new/file
* leaves other files remaining in the original directory (so that
e.g. "old/" still exists upstream even though file has been
removed from it and placed elsewhere)
* a topic branch being rebased where:
* a commit in the sequence:
* modifies old/file
* a subsequent commit in the sequence being replayed:
* does NOT touch *anything* under new/
* does NOT touch old/file
* DOES modify other paths under old/
* does NOT have any relevant renames that we need to detect
_anywhere_ elsewhere in the tree (meaning this interacts
interestingly with both directory renames and cached renames)
In such a case, the assertion will trigger. The fix turns out to be
surprisingly simple. I have a very vague recollection that I actually
considered whether to add such an if-check years ago when I added the
very similar one for oldinfo in 1b6b902d95 (merge-ort:
process_renames() now needs more defensiveness, 2021-01-19), but I think
I couldn't figure out a possible way to trigger it and was worried at
the time that if I didn't know how to trigger it then I wasn't so sure
that simply skipping it was correct. Waiting did give me a chance to
put more thorough tests and checks into place for the rename-to-self
cases a few months back, which I might not have found as easily
otherwise. Anyway, put the check in place now and add a test that
demonstrates the fix.
Note that this bug, as demonstrated by the conditions listed above,
runs at the intersection of relevant renames, trivial directory
resolutions, and cached renames. All three of those optimizations are
ones that unfortunately make the code (and testcases!) a bit more
complex, and threading all three makes it a bit more so. However, the
testcase isn't crazy enough that I'd expect no one to ever hit it in
practice, and was confused why we didn't see it before. After some
digging, I discovered that merge.directoryRenames=false is a workaround
to this bug, and GitHub used that setting until recently (it was a
"temporary" match-what-libgit2-does piece of code that lasted years
longer than intended). Since the conditions I gave above for triggering
this bug rule out the possibility of there being directory renames, one
might assume that it shouldn't matter whether you try to detect such
renames if there aren't any. However, due to commit a16e8efe5c
(merge-ort: fix merge.directoryRenames=false, 2025-03-13), the heavy
hammer used there means that merge.directoryRenames=false ALSO turns off
rename caching, which is critical to triggering the bug. This becomes
a bit more than an aside since...
Re-reading that old commit, a16e8efe5c (merge-ort: fix
merge.directoryRenames=false, 2025-03-13), it appears that the solution
to this latest bug might have been at least a partial alternative
solution to that old commit. And it may have been an improved
alternative (or at least help implement one), since it may be able to
avoid the heavy-handed disabling of rename cache. That might be an
interesting future thing to investigate, but is not critical for the
current fix. However, since I spent time digging it all up, at least
leave a small comment tweak breadcrumb to help some future reader
(myself or others) who wants to dig further to connect the dots a little
quicker.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
While developing commit a16e8efe5c (merge-ort: fix
merge.directoryRenames=false, 2025-03-13), I was testing things out and
had an extra condition on one of the if-blocks that I occasionally
swapped between '&& 0' and '&& 1' to see the effects of the changes. I
forgot to remove it before submitting and it wasn't caught in review.
Remove it now.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A comment at the top of t6429 mentions why the test doesn't exercise git
rebase or git cherry-pick. However, it claims that it uses `test-tool
fast-rebase`. That was true when the comment was written, but commit
f920b0289b (replay: introduce new builtin, 2023-11-24) changed it to
use git replay without updating this comment.
We could potentially just strike this second comment, since git replay
is a bona fide built-in, but perhaps the explanation about why it focuses
on git replay is still useful. Update the comment to make it accurate
again.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
CI update.
* js/ci-github-actions-update:
build(deps): bump actions/github-script from 7 to 8
build(deps): bump actions/setup-python from 5 to 6
build(deps): bump actions/checkout from 4 to 5
build(deps): bump actions/download-artifact from 4 to 5
Recent OpenSSH creates the Unix domain socket to communicate with
ssh-agent under $HOME instead of /tmp, which causes our test to
fail doe to overly long pathname in our test environment, which has
been worked around by using "ssh-agent -T".
* ps/t7528-ssh-agent-uds-workaround:
t7528: work around ETOOMANY in OpenSSH 10.1 and newer
The "--short" option of "git status" that meant output for humans
and "-z" option to show NUL delimited output format did not mix
well, and colored some but not all things. The command has been
updated to color all elements consistently in such a case.
* jk/status-z-short-fix:
status: make coloring of "-z --short" consistent
An earlier addition to "git diff --no-index A B" to limit the
output with pathspec after the two directories misbehaved when
these directories were given with a trailing slash, which has been
corrected.
* jk/diff-no-index-with-pathspec-fix:
diff --no-index: fix logic for paths ending in '/'
Windows "real-time monitoring" interferes with the execution of
tests and affects negatively in both correctness and performance,
which has been disabled in Gitlab CI.
* ps/gitlab-ci-disable-windows-monitoring:
gitlab-ci: disable realtime monitoring to unbreak Windows jobs
The code to squelch output from "git diff -w --name-status"
etc. for paths that "git diff -w -p" would have stayed silent
leaked output from dry-run patch generation, which has been
corrected.
* jc/diff-from-contents-fix:
diff: make sure the other caller of diff_flush_patch_quietly() is silent
Recently we attempted to improve "git diff -w" and friends to
handle cases where patch output would be suppressed, but it
introduced a bug that emits unnecessary output, which has been
corrected.
* jk/diff-from-contents-fix:
diff: restore redirection to /dev/null for diff_from_contents
In t7528 we spawn an SSH agent to verify that we can sign a commit via
it. This test has started to fail on some machines:
+++ ssh-agent
unix_listener_tmp: path "/home/pks/Development/git/build/test-output/trash directory.t7528-signed-commit-ssh/.ssh/agent/s.UTulegefEg.agent.UrPHumMXPq" too long for Unix domain socket
main: Couldn't prepare agent socket
As it turns out this is caused by a change in OpenSSH 10.1 [1]:
* ssh-agent(1), sshd(8): move agent listener sockets from /tmp to
under ~/.ssh/agent for both ssh-agent(1) and forwarded sockets
in sshd(8).
Instead of creating the socket in "/tmp", OpenSSH now creates the socket
in our home directory. And as the home directory gets modified to be
located in our test output directory we end up with paths that are
somewhat long. But Linux has a rather short limit of 108 characters for
socket paths, and other systems have even lower limits, so it is very
easy now to exceed the limit and run into the above error.
Work around the issue by using `ssh-agent -T`, which instructs it to
use the old behaviour and create the socket in "/tmp" again. This switch
has only been introduced with 10.1 though, so for older versions we have
to fall back to not using it. That's fine though, as older versions know
to put the socket into "/tmp" already.
An alternative approach would be to abbreviate the socket name itself so
that we create it as e.g. "sshsock" in the trash directory. But taking
the above example we'd still end up with a path that is 91 characters
long. So we wouldn't really have a lot of headroom, and it is quite
likely that some developers would see the issue on their machines.
[1]: https://www.openssh.com/txt/release-10.1
Reported-by: Xi Ruoyao <xry111@xry111.site>
Suggested-by: brian m. carlson <sandals@crustytoothpaste.net>
Helped-by: Jeff King <peff@peff.net>
Helped-by: Lauri Tirkkonen <lauri@hacktheplanet.fi>
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Earlier, we added is a protection for the loop that computes "git
diff --quiet -w" to ensure calls to the diff_flush_patch_quietly()
helper stays quiet. Do the same for another loop that deals with
options like "--name-status" to make calls to the same helper.
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When running "git status -z --short", the marker on modified index
entries (e.g., "M") is colorized, but the "??" marker for untracked
entries is not. Let's fix the "??" entries to show color here.
At first glance you might think that neither should be colorized, as
usually one would use "-z" to get machine-readable output. But this is a
tricky and unusual case. We have two output formats, "--short" and
"--porcelain" which are substantially similar, but differ in that
"--short" is for humans who want something short and "--porcelain" is
for machines. And "-z" by itself, without any other output option, does
default to "--porcelain", so "git status -z" will not colorize anything.
But if you explicitly ask for "-z" and "--short" together, then that is
asking for the human-readable output, but separated by NULs. This is
unlikely to be useful directly, but could for example be used if the
output will be shown to a human outside of the terminal. At any rate,
the current behavior is clearly wrong (since we colorize some things but
not others), and I think colorizing everything is the least-surprising
thing we can do here.
Reported-by: Langbart <Langbart@protonmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In --quiet mode, since we produce only an exit code for "something was
changed" and no actual output, we can often get by with just a
tree-level diff. However, certain options require us to actually look at
the file contents (e.g., if we are ignoring whitespace changes). We have
a flag "diff_from_contents" for that, and if it is set we call
diff_flush() on each path.
To avoid producing any output (since we were asked to be --quiet), we
traditionally just redirected the output to /dev/null. That changed in
b55e6d36eb (diff: ensure consistent diff behavior with ignore options,
2025-08-08), which replaced that with a "dry_run" flag. In theory, with
dry_run set, we should produce no output. But it carries a risk of
regression: if we forget to respect dry_run in any of the output paths,
we'll accidentally produce output.
And indeed, there is at least one such regression in that commit, as it
covered only the case where we actually call into xdiff, and not
creation or deletion diffs, where we manually generate the headers. We
even test this case in t4035, but only with diff-tree, which does not
show the bug by default because it does not require diff_from_contents.
But git-diff does, because it allows external diff programs by default
(so we must dig into each diff filepair to decide if it requires running
an external diff that may declare two distinct blobs to actually be the
same).
We should fix all of those code paths to respect dry_run correctly, but
in the meantime we can protect ourselves more fully by restoring the
redirection to /dev/null. This gives us an extra layer of protection
against regressions dues to other code paths we've missed.
Though the original issue was reported with "git diff" (and due to its
default of --ext-diff), I've used "diff-tree -w" in the new test. It
triggers the same issue, but I think the fact that "-w" implies
diff_from_contents is a bit more obvious, and fits in with the rest of
t4035.
Reported-by: Jake Zimmerman <jake@zimmerman.io>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Update old-style shell path checks to use the modern test
helpers 'test_path_is_file' and 'test_path_is_dir' for improved
runtime diagnosis.
Signed-off-by: Solly <solobarine@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Documentation mark-up fix.
* ja/doc-markup-attached-paragraph-fix:
doc: fix indentation of refStorage item in git-config(1)
doc: change the markup of paragraphs following a nested list item
Clarify the "--merge-base" command line option in "git merge-tree".
* en/doc-merge-tree-describe-merge-base:
Documentation/git-merge-tree.adoc: clarify the --merge-base option
Deal more gracefully with directory / file conflicts when the files
backend is used for ref storage, by failing only the ones that are
involved in the conflict while allowing others.
* kn/refs-files-case-insensitive:
refs/files: handle D/F conflicts during locking
refs/files: handle F/D conflicts in case-insensitive FS
refs/files: use correct error type when lock exists
refs/files: catch conflicts on case-insensitive file-systems
"git rebase -i" failed to clean-up the commit log message when the
command commits the final one in a chain of "fixup" commands, which
has been corrected.
* pw/rebase-i-cleanup-fix:
sequencer: remove VERBATIM_MSG flag
rebase -i: respect commit.cleanup when picking fixups
Some among "git add -p" and friends ignored color.diff and/or
color.ui configuration variables, which is an old regression, which
has been corrected.
* jk/add-i-color:
contrib/diff-highlight: mention interactive.diffFilter
add-interactive: manually fall back color config to color.ui
add-interactive: respect color.diff for diff coloring
stash: pass --no-color to diff plumbing child processes
A corner case bug in "git log -L..." has been corrected.
* sg/line-log-boundary-fixes:
line-log: show all line ranges touched by the same diff range
line-log: fix assertion error
A broken or malicious "git fetch" can say that it has the same
object for many many times, and the upload-pack serving it can
exhaust memory storing them redundantly, which has been corrected.
* ps/upload-pack-oom-protection:
upload-pack: don't ACK non-commits repeatedly in protocol v2
t5530: modernize tests
Fixes multiple crashes around midx write-out codepaths.
* ds/midx-write-fixes:
midx-write: simplify error cases
midx-write: reenable signed comparison errors
midx-write: use uint32_t for preferred_pack_idx
midx-write: use cleanup when incremental midx fails
midx-write: put failing response value back
midx-write: only load initialized packs
"git repack --path-walk" lost objects in some corner cases, which
has been corrected.
cf. <CABPp-BHFxxGrqKc0m==TjQNjDGdO=H5Rf6EFsf2nfE1=TuraOQ@mail.gmail.com>
* ds/path-walk-repack-fix:
path-walk: create initializer for path lists
path-walk: fix setup of pending objects
Under a race against another process that is repacking the
repository, especially a partially cloned one, "git fetch" may
mistakenly think some objects we do have are missing, which has
been corrected.
* jk/fetch-check-graph-objects-fix:
fetch-pack: re-scan when double-checking graph objects
Elijah Newren