Tian Yuchen [Fri, 19 Jun 2026 15:51:51 +0000 (23:51 +0800)]
environment: move ignore_case into repo_config_values
The 'core.ignorecase' configuration which is stored as the
global variable 'ignore_case' acts as a core filesystem
capability flag.
Move this global variable into 'struct repo_config_values' to tie it
to the specific repository instance it was read from. This reduces
global state and aligns with the ongoing libification effort.
To ensure code readability, the getter function
'repo_ignore_case()' is introduced.
Mentored-by: Christian Couder <christian.couder@gmail.com> Mentored-by: Ayush Chandekar <ayu.chandekar@gmail.com> Mentored-by: Olamide Caleb Bello <belkid98@gmail.com> Signed-off-by: Tian Yuchen <cat@malon.dev> Signed-off-by: Junio C Hamano <gitster@pobox.com>
'fetch.followRemoteHEAD' is added as a generic setting used by all
remotes for which 'remote.<name>.followRemoteHEAD' is undefined. If
both variables are undefined, a builtin default of "create" is in
effect, matching the previous behavior.
As mentioned in the previous patch, 'fetch.followRemoteHEAD' supports
all of the values that its 'remote' counterpart does _except_
warn-if-not-$branch, due to its tighter coupling to individual remote
repositories.
This setting interacts with the do_fetch mechanism in the same way as
the previous does, but there are opportunities for improved
user-experience discussed in [1]. See the included NEEDSWORK comment as
well.
Documentation and advice messages for both of the followRemoteHEAD
variables are reworded to better capture the relationship between the
two.
The added tests assert feature parity between the two followRemoteHEAD
variables, as well as the fact that 'remote.<name>.followRemoteHEAD'
always supersedes this new configurable default.
Matt Hunter [Fri, 19 Jun 2026 09:44:25 +0000 (05:44 -0400)]
fetch: refactor do_fetch handling of followRemoteHEAD
Update enum follow_remote_head_settings to include the value
FOLLOW_REMOTE_UNCONFIGURED as the new zero-initialized value for
followRemoteHEAD. This will allow us to distinguish between the
variable being unset vs. explicitly set to 'create', which is ultimately
the system default. The unnecessary indentation is removed.
The do_fetch function is likewise updated to perform its own decision
making to determine the effective followRemoteHEAD mode, falling back to
the system default if necessary. This will enable the next patch to
introduce a user-configurable default.
Function set_head now accepts the mode as an argument rather than only
considering the value defined by the remote.
The use of the 'warn-if-not-$branch' value is awkward in the context of
a global default, since the branches will differ between individual
remotes. For this reason, it's left out of this scheme and handling of
the no_warn_branch variable is untouched. Since a remote-specific
value for followRemoteHEAD takes priority, we can assume that if
remote->no_warn_branch is set, then the remote is also asserting
FOLLOW_REMOTE_WARN as the effective operating mode, and it will be
honored by do_fetch.
Signed-off-by: Matt Hunter <m@lfurio.us> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Matt Hunter [Fri, 19 Jun 2026 09:44:24 +0000 (05:44 -0400)]
fetch: return 0 on known git_fetch_config
The git config callback for git-fetch should only forward calls to
git_default_config when an unknown key is given. Prevent this in the
case of 'fetch.output' by returning '0', as the other known keys do.
Signed-off-by: Matt Hunter <m@lfurio.us> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Matt Hunter [Fri, 19 Jun 2026 09:44:22 +0000 (05:44 -0400)]
t5510: cleanup remote in followRemoteHEAD dangling ref test
A later patch will introduce a new test which closely mirrors this one.
Update this test to remove the 'custom-head' remote it creates.
Otherwise, the two tests will conflict with each other, as the second
one to execute will fail to create this remote (which already exists,
thanks to the first test).
Signed-off-by: Matt Hunter <m@lfurio.us> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Matt Hunter [Fri, 19 Jun 2026 09:44:21 +0000 (05:44 -0400)]
doc: explain fetchRemoteHEADWarn advice
When the user sets 'remote.<name>.followRemoteHEAD' to
'warn[-if-not-$branch]', git-fetch will report when a fetched HEAD
disagrees with the locally-configured remote's HEAD. This additional
advice instructs the user how to deal with these warnings, but was
previously undocumented in git-config.
Signed-off-by: Matt Hunter <m@lfurio.us> Signed-off-by: Junio C Hamano <gitster@pobox.com>
SubmittingPatches: note that trailer order matters
It matters where you put new trailers: they should be added in
chronological order, and each person who passes on a patch should add
their s-o-b last. You are signing off on the patch as well as the whole
message up to that point.
This also makes it clear who added what:
Acked-by: The Reviewer <r@example.org> Signed-off-by: The Contributor <c@example.org> Acked-by: The (Late) Reviewer <late@example.org> Signed-off-by: The Maintainer <m@example.org>
The first ack was added by the contributor and the second one was added
by the maintainer.
Helped-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Kristoffer Haugsbakk <code@khaugsbakk.name> Signed-off-by: Junio C Hamano <gitster@pobox.com>
SubmittingPatches: be consistent with trailer markup
The rest of this section and (most importantly) the list has decided to
use `<key>:`. So let’s use backticks (`) and a colon (:) throughout the
document.
Signed-off-by: Kristoffer Haugsbakk <code@khaugsbakk.name> Signed-off-by: Junio C Hamano <gitster@pobox.com>
This trailer comes up often enough and the use case is not fully covered
by the other trailers here. For example, it is sometimes better to use
this trailer instead of `Co-authored-by:`.
Signed-off-by: Kristoffer Haugsbakk <code@khaugsbakk.name> Signed-off-by: Junio C Hamano <gitster@pobox.com>
SubmittingPatches: discourage common Linux trailers
The Linux Kernel regularly uses trailers (or “tags”) `Fixes` and
`Link`. Sometimes people submit patches to this project with them.
They have their use in that project but it is not clear what purpose
they would serve here.
For `Fixes`: Linux has many trees, and applying patches with
cherry-picks is common. A `Fixes` trailer in commit C2 pointing to
commit C1 helps the cherry-picker figure out that she probably needs
C2 if she wants to apply C1. See linux/d5d6281a (checkpatch: check for
missing Fixes tags, 2024-06-11):[1]
Why are stable patches encouraged to have a fixes tag? Some people
mark their stable patches as "# 5.10" etc. This is useful but a
Fixes tag is still a good idea. For example, the Fixes tag helps in
review. It helps people to not cherry-pick buggy patches without
also cherry-picking the fix.
In contrast the Git project has few trees (to my knowledge), and there
is much less need to cherry-pick fixes as opposed to either using
backmerges or rebasing all of the downstream tree’s commits on top of
git.git `master` from time to time.
This project does regularly mention what commits a patch/commit fixes,
but that is done inline in the commit message proper (cf. the trailer
block of the message).
For `Link`: These are used both to link back to the patch submission as
well as with footnotes. In contrast this project has `refs/notes/amlog`
for linking back to the patch submissions, and footnotes are only used
in the commit message proper.
† 1: Commit linux/d5d6281a has “linux” in front of it since this commit
is from the Linux Kernel, not Git. Example of a Linux tree—as well
as an example of `Link`—is [2].
SubmittingPatches: encourage trailer use for substantial help
Trailers beyond the mandatory s-o-b are regularly used based on my
last two years of reading the mailing list. Moreover, reviewers might
encourage it.[1]
This is also in line with the project crediting both commit authors and
people mentioned in trailers each release; “Nobody is THE one making
contribution”.[2]
Adding trailers is already encouraged, but in the section `send-patches`.
Let’s replace “If you like” with outright encouragement in this section
so that all trailer discussion (except s-o-b; see `sign-off` section) is
contained in this section; a link to from `send-patches` makes this
information equally visible.
Now we need to make a heading for `commit-trailers` in order for the
HTML output to make sense.
At the same time, it is important to temper this recommendation to a
significant enough contribution; in my experience beginners can be eager
to add a trailer for everyone who replies with an action point that is
followed up on.
Let’s also spell out that these trailers should follow the Git author/
committer format. One might naturally just write the name, but in that
case it will not be picked up by:
git shortlog --group=trailer:<key>
and normalization via `.mailmap` will not work.
Also introduce the list of common trailers as such. Granted, this is
already implied by the later paragraph about “create your own trailer”,
so this just frontloads this information.
On platforms where `unsigned long` and `size_t` differ in bit size, we
want to clamp the buffers we pass to zlib to the former's size, as per d05d666977 (git-zlib: handle data streams larger than 4GB, 2026-05-08).
The logic introduced in that commit performs a clamping to the bits,
though, which fails to do what is needed here: If too many bytes are
available in the buffers, we need to clamp to the maximum value of an
`unsigned long`. Otherwise, we ask zlib to use too small buffers, in the
worst case using 0 as the size (think: a value whose 32 lowest bits are
all zero).
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
gitlab-ci: migrate Windows builds away from Chocolatey
The Windows builds in GitLab CI use Chocolatey to install dependencies.
Unfortunately, Chocolatey seems to be very unreliable, which causes the
jobs to fail very regularly. This is a limitation that seems to be
somewhat known [1]:
As an organization, you want 100% reliability (or at least that
potential), and you may want full trust and control as well. This is
something you can get with internally hosted packages, and you are
unlikely to achieve from use of the Community Package Repository.
So using the Community Package Repository is kind of discouraged in case
one wants reliability. We _do_ want reliability though, and we cannot
easily switch to an enterprise license to fix this issue.
Introduce a new script that downloads and installs dependencies
directly. This has a couple of benefits:
- We can drop our dependency on Chocolatey completely, thus improving
reliability.
- We can easily cache the installers.
- We get direct control over the exact versions we install.
- Installing dependencies is sped up from roundabout 3 minutes to 1
minute.
Junio C Hamano [Wed, 17 Jun 2026 18:10:12 +0000 (11:10 -0700)]
Merge branch 'en/commit-graph-timestamp-fix'
compute_reachable_generation_numbers() in commit-graph used a 32-bit
integer to accumulate parent generations, which is OK for generation
number v1 (topological levels), but with generation number v2
(adjusted committer timestamps), it truncated timestamps beyond
2106. Fixed by widening the accumulator to timestamp_t.
* en/commit-graph-timestamp-fix:
commit-graph: use timestamp_t for max parent generation accumulator
Junio C Hamano [Wed, 17 Jun 2026 18:10:12 +0000 (11:10 -0700)]
Merge branch 'dl/posix-unused-warning-clang'
The UNUSED macro in 'compat/posix.h' has been updated to use a
newly introduced GIT_CLANG_PREREQ macro for compiler version
checks, and the existing GIT_GNUC_PREREQ macro has been modernized
to use explicit major/minor comparisons rather than bit-shifting.
* dl/posix-unused-warning-clang:
compat/posix.h: simplify GIT_GNUC_PREREQ() comparison
compat/posix.h: clean up GIT_GNUC_PREREQ() and UNUSED
compat/posix.h: enable UNUSED warning messages for Clang
Junio C Hamano [Wed, 17 Jun 2026 18:10:11 +0000 (11:10 -0700)]
Merge branch 'td/ls-files-pathspec-prefilter'
`git ls-files --modified` and `git ls-files --deleted` have been
optimized to filter with pathspec before calling lstat() when there is
only a single pathspec item, avoiding unnecessary filesystem access
for entries that will not be shown.
* td/ls-files-pathspec-prefilter:
ls-files: filter pathspec before lstat
Junio C Hamano [Wed, 17 Jun 2026 18:10:11 +0000 (11:10 -0700)]
Merge branch 'td/describe-tag-iteration'
'git describe' has been taught to pass the 'refs/tags/' prefix down to
the ref iterator when '--all' is not requested, avoiding unnecessary
iteration over non-tag refs.
* td/describe-tag-iteration:
describe: limit default ref iteration to tags
Junio C Hamano [Wed, 17 Jun 2026 18:10:11 +0000 (11:10 -0700)]
Merge branch 'ps/transport-helper-tsan-fix'
The TSAN race in transfer_debug() within transport-helper.c has been
resolved by initializing the debug flag early in
bidirectional_transfer_loop() before spawning worker threads, allowing
the removal of a TSAN suppression.
* ps/transport-helper-tsan-fix:
transport-helper: fix TSAN race in transfer_debug()
Junio C Hamano [Wed, 17 Jun 2026 12:39:13 +0000 (05:39 -0700)]
Merge branch 'ab/index-pack-retain-child-bases'
"git index-pack" has been optimized by retaining child bases in the
delta cache instead of immediately freeing them, letting the existing
cache limit policy decide eviction.
* ab/index-pack-retain-child-bases:
index-pack: retain child bases in delta cache
odb/source-packed: drop pointer to "files" parent source
Over the last commits we have turned the packfile store into a proper
object database source that can be used as a standalone backend. As
such, it is no longer necessary to have it coupled to the "files" parent
source.
Remove the pointer to the owning "files" source so that the "packed"
source can be used as a standalone entity.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
midx: refactor interfaces to work on "packed" source
Our interfaces used to interact with MIDXs all work on top of the
generic `struct odb_source`. This doesn't make much sense though: a MIDX
is strictly tied to the "packed" source, so passing in a generic source
gives the false sense that it may also work with a different type of
source.
Fix this conceptual weirdness and instead require the caller to pass in
a "packed" source explicitly. This also makes the next commit easier to
implement, where we drop the pointer to the "files" source in the
"packed" source.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
odb/source-packed: wire up `freshen_object()` callback
Move `packfile_store_freshen_object()` and from "packfile.c" into
"odb/source-packed.c" and wire it up as the `freshen_object()` callback
of the "packed" source.
Note that this removes the last external caller of `find_pack_entry()`
from "packfile.c", which means that we can now make this function
static.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
odb/source-packed: wire up `find_abbrev_len()` callback
Move `packfile_store_find_abbrev_len()` and its associated helpers from
"packfile.c" into "odb/source-packed.c" and wire it up as the
`find_abbrev_len()` callback of the "packed" source.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
odb/source-packed: wire up `count_objects()` callback
Move `packfile_store_count_objects()` from "packfile.c" into
"odb/source-packed.c" and wire it up as the `count_objects()` callback
of the "packed" source.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
odb/source-packed: wire up `for_each_object()` callback
Move `packfile_store_for_each_object()` and its associated helpers from
"packfile.c" into "odb/source-packed.c" and wire it up as the
`for_each_object()` callback of the "packed" source.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
odb/source-packed: wire up `read_object_stream()` callback
Wire up the `read_object_stream()` callback for the packed source and
call it in the "files" source via the `odb_source_read_object_stream()`
interface.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
odb/source-packed: wire up `read_object_info()` callback
Move the logic to read object info from a "packed" source into
"odb/source-packed.c" and wire it up as the `read_object_info()`
callback.
Note that we also move around the supporting `find_pack_entry()`, but we
still have to expose it to other callers that exist in "packfile.c".
This will be fixed in subsequent commits though, where all callers in
"packfile.c" will have been moved into "odb/source-packed.c", and at
that point we'll be able to make `find_pack_entry()` file-local again.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
packfile: use higher-level interface to implement `has_object_pack()`
In `has_object_pack()` we're checking whether a specific object exists
as part of a packfile. This is done by calling the low-level function
`find_pack_entry()`, but this function will eventually be moved into
"odb/source-packed.c" and made file-local.
Refactor the code to use `packfile_store_read_object_info()` instead.
This refactoring is functionally equivalent as that function will call
`find_pack_entry()` itself and then return immediately when it ain't got
no object info pointer as parameter.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Move the logic to prepare and reprepare the "packed" source into
"odb/source-packed.c" and wire it up as the `reprepare()` callback.
Note that "preparing" a source is not yet generic. Eventually, it would
probably make sense to turn the existing `reprepare()` callback into a
`prepare()` callback with an optional flag to force re-preparing. But
this step will be handled in a separate patch series.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
odb/source-packed: start converting to a proper `struct odb_source`
Start converting `struct odb_source_packed` into a proper pluggable
`struct odb_source` by embedding the base struct and assigning it the
new `ODB_SOURCE_PACKED` type. Furthermore, wire up lifecycle management
of this source by implementing the `free` callback and taking ownership
of the chdir notifications.
Note that the packed source is not yet functional as a standalone `struct
odb_source`, as it's missing all of the callback implementations. These
will be wired up in subsequent commits.
Further note that we're also registering a `chdir_notify` callback to
reparent our path. This wasn't previously necessary (and still isn't at
this point in time) because all paths are taken from the owning "files"
source, and that source already handles the reparenting for us. But a
subsequent commit will change that so that we're using the path of the
"packed" source, and once that happens we'll need it to be updated when
changing the working directory.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
odb/source-packed: store pointer to "files" instead of generic source
The `struct odb_source_packed` holds a pointer to its owning parent
source. The way that Git is currently structured, this parent is always
the "files" source. In subsequent commits we're going to detangle that
so that the "packed" source doesn't have any owning parent source at
all, which makes it usable as a completely standalone source.
Detangling this mess is somewhat intricate though, and is made even more
intricate because it's not always clear which kind of source one is
holding at a specific point in time -- either the parent "files" source,
or the child "packed" source.
Make this relationship more explicit by storing a pointer to the "files"
source instead of storing a pointer to a generic `struct odb_source`.
This will help make subsequent steps a bit clearer by making it more
obvious whether we're using the generic "base" source or the owning
"files" source.
Note that this is a temporary step, only. At the end of this series
we will have dropped the "files" pointer completely.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
packfile: move packed source into "odb/" subsystem
In subsequent patches we'll be turning `struct odb_source_packed` into a
proper `struct odb_source`. As a first step towards this goal, move its
struct out of "packfile.{c,h}" and into "odb/source-packed.{c,h}".
This detaches the implementation of the packfile object source from the
generic packfile code, following the same convention already used by the
"files" and "in-memory" sources.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the next commit we're about to introduce the "packed" object database
source. This source will embed a packfile list, and consequently we'll
have to include "packfile.h" to make the struct definition available.
This will unfortunately lead to a cyclic dependency that we cannot
resolve with a forward declaration.
Split out the code that relates to the packfile list into a separate
compilation unit so that both "packfile.h" and "odb/source-packed.h" can
include it.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
packfile: rename `struct packfile_store` to `odb_source_packed`
Not too long ago, we have introduced the packfile store in b7983adb51
(packfile: introduce a new `struct packfile_store`, 2025-09-23). This
struct is responsible for managing all of our access to packfiles and is
used as one of the two sources of objects for the "files" source.
Back when I introduced this structure I didn't have the clear vision yet
that it will eventually also turn into a proper object database source,
and how exactly that infrastructure will look like. Now though it's
becoming increasingly clear that it does make sense to treat it just the
same as any of our other ODB sources.
The consequence is that the naming is now a bit out-of-date: it's just
another source and will be turned into a proper `struct odb_source` over
the next couple of commits, but it's not named accordingly.
Rename the structure to `odb_source_packed` to align it with this goal
and to bring it in line with the other sources we already have.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Without NO_RUST defined, the varint encoder/decoder lives in the
RUST_LIB, which needs to be linked. Symptom:
cc [... -o contrib/credential/osxkeychain/git-credential-osxkeychain [...]
Undefined symbols for architecture x86_64:
"_decode_varint", referenced from:
_read_untracked_extension in libgit.a[x86_64][63](dir.o)
_read_untracked_extension in libgit.a[x86_64][63](dir.o)
_read_one_dir in libgit.a[x86_64][63](dir.o)
_read_one_dir in libgit.a[x86_64][63](dir.o)
_load_cache_entry_block in libgit.a[x86_64][174](read-cache.o)
"_encode_varint", referenced from:
_write_untracked_extension in libgit.a[x86_64][63](dir.o)
_write_untracked_extension in libgit.a[x86_64][63](dir.o)
_write_untracked_extension in libgit.a[x86_64][63](dir.o)
_write_one_dir in libgit.a[x86_64][63](dir.o)
_write_one_dir in libgit.a[x86_64][63](dir.o)
_do_write_index in libgit.a[x86_64][174](read-cache.o)
ld: symbol(s) not found for architecture x86_64
While it is curious why these functions are needed at all (osxkeychain
does not read or write the index), the compile error is a real problem.
Instead of trying to play games to add `GITLIBS` while filtering out
`common-main.o`, replace the `$(LIB_FILE) $(EXTLIBS)` construct with the
much shorter `$(LIBS)` construct that _already_ filters out
`common-main.o` and adds the Rust library when needed.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Philip Oakley [Tue, 16 Jun 2026 14:49:57 +0000 (14:49 +0000)]
hash-object: add a >4GB/LLP64 test case using filtered input
To verify that the `clean` side of the `clean`/`smudge` filter code is
correct with regards to LLP64 (read: to ensure that `size_t` is used
instead of `unsigned long`), here is a test case using a trivial filter,
specifically _not_ writing anything to the object store to limit the
scope of the test case.
As in previous commits, the `big` file from previous test cases is
reused if available, to save setup time, otherwise re-generated.
Signed-off-by: Philip Oakley <philipoakley@iee.email> Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Philip Oakley [Tue, 16 Jun 2026 14:49:56 +0000 (14:49 +0000)]
hash-object: add another >4GB/LLP64 test case
To complement the `--stdin` and `--literally` test cases that verify
that we can hash files larger than 4GB on 64-bit platforms using the
LLP64 data model, here is a test case that exercises `hash-object`
_without_ any options.
Just as before, we use the `big` file from the previous test case if it
exists to save on setup time, otherwise generate it.
Signed-off-by: Philip Oakley <philipoakley@iee.email> Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Philip Oakley [Tue, 16 Jun 2026 14:49:55 +0000 (14:49 +0000)]
hash-object --stdin: verify that it works with >4GB/LLP64
Just like the `hash-object --literally` code path, the `--stdin` code
path also needs to use `size_t` instead of `unsigned long` to represent
memory sizes, otherwise it would cause problems on platforms using the
LLP64 data model (such as Windows).
To limit the scope of the test case, the object is explicitly not
written to the object store, nor are any filters applied.
The `big` file from the previous test case is reused to save setup time;
To avoid relying on that side effect, it is generated if it does not
exist (e.g. when running via `sh t1007-*.sh --long --run=1,41`).
Signed-off-by: Philip Oakley <philipoakley@iee.email> Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Philip Oakley [Tue, 16 Jun 2026 14:49:54 +0000 (14:49 +0000)]
hash algorithms: use size_t for section lengths
Continue walking the code path for the >4GB `hash-object --literally`
test to the hash algorithm step for LLP64 systems.
This patch lets the SHA1DC code use `size_t`, making it compatible with
LLP64 data models (as used e.g. by Windows).
The interested reader of this patch will note that we adjust the
signature of the `git_SHA1DCUpdate()` function without updating _any_
call site. This certainly puzzled at least one reviewer already, so here
is an explanation:
This function is never called directly, but always via the macro
`platform_SHA1_Update`, which is usually called via the macro
`git_SHA1_Update`. However, we never call `git_SHA1_Update()` directly
in `struct git_hash_algo`. Instead, we call `git_hash_sha1_update()`,
which is defined thusly:
i.e. it contains an implicit downcast from `size_t` to `unsigned long`
(before this here patch). With this patch, there is no downcast anymore.
With this patch, finally, the t1007-hash-object.sh "files over 4GB hash
literally" test case is fixed.
Signed-off-by: Philip Oakley <philipoakley@iee.email> Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Philip Oakley [Tue, 16 Jun 2026 14:49:53 +0000 (14:49 +0000)]
object-file.c: use size_t for header lengths
Continue walking the code path for the >4GB `hash-object --literally`
test. The `hash_object_file_literally()` function internally uses both
`hash_object_file()` and `write_object_file_prepare()`. Both function
signatures use `unsigned long` rather than `size_t` for the mem buffer
sizes. Use `size_t` instead, for LLP64 compatibility.
While at it, convert those function's object's header buffer length to
`size_t` for consistency. The value is already upcast to `uintmax_t` for
print format compatibility.
Note: The hash-object test still does not pass. A subsequent commit
continues to walk the call tree's lower level hash functions to identify
further fixes.
Signed-off-by: Philip Oakley <philipoakley@iee.email> Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Philip Oakley [Tue, 16 Jun 2026 14:49:52 +0000 (14:49 +0000)]
hash-object: demonstrate a >4GB/LLP64 problem
On LLP64 systems, such as Windows, the size of `long`, `int`, etc. is
only 32 bits (for backward compatibility). Git's use of `unsigned long`
for file memory sizes in many places, rather than size_t, limits the
handling of large files on LLP64 systems (commonly given as `>4GB`).
Provide a minimum test for handling a >4GB file. The `hash-object`
command, with the `--literally` and without `-w` option avoids
writing the object, either loose or packed. This avoids the code paths
hitting the `bigFileThreshold` config test code, the zlib code, and the
pack code.
Subsequent patches will walk the test's call chain, converting types to
`size_t` (which is larger in LLP64 data models) where appropriate.
Signed-off-by: Philip Oakley <philipoakley@iee.email> Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano [Tue, 16 Jun 2026 16:01:03 +0000 (09:01 -0700)]
Merge branch 'mm/subprocess-handshake-fix'
The subprocess handshake during startup has been made gentler by using
packet_read_line_gently() instead of packet_read_line() to prevent the
parent Git process from dying abruptly when a configured subprocess
(e.g., a clean/smudge filter) fails to start.
* mm/subprocess-handshake-fix:
sub-process: use gentle handshake to avoid die() on startup failure
Junio C Hamano [Tue, 16 Jun 2026 16:01:02 +0000 (09:01 -0700)]
Merge branch 'ps/t7527-fix-tap-output'
A recent regression in t7527 that broke TAP output has been fixed,
some other test noise that also broke TAP output has been silenced,
and 'prove' is now configured to fail on invalid TAP output to
prevent future regressions.
* ps/t7527-fix-tap-output:
t: let prove fail when parsing invalid TAP output
t/lib-git-p4: silence output when killing p4d and its watchdog
t/test-lib: silence EBUSY errors on Windows during test cleanup
t7810: turn MB_REGEX check into a lazy prereq
t7527: fix broken TAP output
ci: unify Linux images across GitLab and GitHub
gitlab-ci: add missing Linux jobs
gitlab-ci: rearrange Linux jobs to match GitHub's order
Junio C Hamano [Tue, 16 Jun 2026 16:01:02 +0000 (09:01 -0700)]
Merge branch 'jk/describe-contains-all-match-fix'
The 'git describe --contains --all' command has been fixed to
properly honor the '--match' and '--exclude' options by passing
them down to 'git name-rev' with the appropriate reference
prefixes.
* jk/describe-contains-all-match-fix:
describe: fix --exclude, --match with --contains and --all
Junio C Hamano [Tue, 16 Jun 2026 16:01:02 +0000 (09:01 -0700)]
Merge branch 'mf/revision-max-count-oldest'
"git rev-list" (and "git log" family of commands) learned a new "--max-count-oldest"
that picks oldest N commits in the range instead of the usual newest.
Junio C Hamano [Tue, 16 Jun 2026 16:01:02 +0000 (09:01 -0700)]
Merge branch 'js/win-kill-child-more-gently'
Advanced emulation of kill() used on Windows in GfW has been
upstreamed to improve the symptoms like left-behind .lock files and
that fails to let the child clean-up itself when it gets killed.
* js/win-kill-child-more-gently:
mingw: really handle SIGINT
mingw: kill child processes in a gentler way
Jeff King [Tue, 16 Jun 2026 12:35:16 +0000 (08:35 -0400)]
read_gitfile(): simplify NOT_A_REPO error message
If a .git file is well-formed but points to a directory that is not
itself a valid repository, then we say:
fatal: not a git repository: <pointed-to-repo>
without mentioning the .git file that pointed us there in the first
place. Doing so could better help the user understand the source of the
problem.
In theory the most helpful thing we could do is mention both paths,
like:
gitfile '<gitfile>' points to invalid repository: <pointed-to-repo>
But there's another catch: when we generate the error, we don't always
know the pointed-to repository! This leads to a potential segfault.
The message comes from read_gitfile_error_die(). Originally we only
called that function from inside read_gitfile_gently(), passing in both
the gitfile path and the pointed-to path. But that changed in 1dd27bfbfd
(setup: improve error diagnosis for invalid .git files, 2026-03-04).
Since then, the caller in setup_git_directory_gently(), even if it wants
to die on error, always passes in the "return_error_code" flag, asking
the function to instead return a numeric error code. And then it calls
read_gitfile_error_die() itself, passing NULL for the pointed-to path.
If we get the READ_GITFILE_ERR_NOT_A_REPO code, we form a message using
that NULL pointer, and either segfault or get garbage like "not a git
repository: (null)", depending on the platform.
We could fix this by having the function pass out both the numeric error
code and the pointed-to path. But that creates a new headache: we have
to allocate that string on the heap and pass ownership back to the
caller. So now every caller has to be aware of it (and either free the
result, or signal that they are not interested by using an extra
parameter).
Instead, let's just drop the pointed-to path from the error message
entirely, and mention only the gitfile. This fixes the NULL dereference
without introducing any more complexity. The user-facing error message
is not as detailed as it could be, but is better than the original.
Since it mentions the gitfile, a user investigating the situation can
look there to find the pointed-to path (whereas you could not go the
other way from the original message).
There's an existing test in t0002 which triggers this case, but we
didn't notice the problem because it checks only that we said "not a
repository", and not the full string. So if we print "(null)" it is
happy. It will probably crash on some non-glibc platforms, but nobody
seems to have reported it yet (the breakage is recent-ish as of v2.54).
I'm also somewhat surprised that building with ASan/UBSan doesn't catch
this, but it doesn't seem to (and I found an open issue with somebody
asking for NULL printf checks to be implemented in the sanitizers).
We'll tweak the test to match the new error, but there's no need to beef
it up further, since we're not showing the pointed-to path at all.
We also racily trigger this in t7450. During parallel cloning we might
see one of several errors, including this one. And so we must update
that message, too (you can otherwise find the failure pretty quickly by
running t7450 with --stress).
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the preceding commit we have extended our documentation to recommend
b4 for sending patch series to the mailing list. Introduce configuration
so that it knows to honor preferences of the Git project by default. For
now, this configuration does two things:
- It configures "send-same-thread = shallow", which tells b4 to always
send subsequent versions of the same patch series as a reply to the
cover letter of the first version.
- It configures "prep-cover-template", which tells b4 to use a custom
template for the cover letter. The most important change compared to
the default template is that our custom template also includes a
range-diff.
There's potentially more things that we may want to configure going
forward, like for example auto-configuration of folks to Cc on certain
patches. But these two tweaks feel like a good place to start.
Note that these values only serve as defaults, and users may want to
tweak those defaults based on their own preference. Luckily, users can
do that without having to touch `.b4-config` at all, as b4 allows them
to override values via Git configuration:
```
$ git config set b4.prep-cover-template /does/not/exist
$ b4 send --dry-run
ERROR: prep-cover-template says to use x, but it does not exist
```
So this gives users an easy way to override our defaults without having
to touch ".b4-config", which would dirty the tree.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The b4 tool originates from the Linux kernel community and is intended
to help mailing-list based workflows. It automates a lot of the annoying
bookkeeping tasks that contributors typically need to do: tracking the
list of recipients, Message-IDs, range-diffs and the like. In addition
to that, b4 also has many other subcommands that help the maintainer and
reviewers.
The Git project uses the same infrastructure as the kernel, so this tool
is also a very good fit for us. Adapt "MyFirstContribution" to
explicitly recommend its use.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
MyFirstContribution: recommend shallow threading of cover letters
The "MyFirstContribution" document recommends the use of deep threading
of cover letters: every cover letter of subsequent iterations shall be
linked to the cover letter of the preceding version. The result of this
is that eventually, threads with many versions are getting nested so
deep that it becomes hard to follow.
Adapt the recommendation to instead propose shallow threading of cover
letters: instead of linking the cover letter to the previous cover
letter, the user is supposed to always link it to the first cover
letter. This still makes it easy to follow the iterations, but has the
benefit of nesting to a much shallower level.
Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
René Scharfe [Sun, 14 Jun 2026 16:28:34 +0000 (18:28 +0200)]
cat-file: speed up default format
eb54a3391b (cat-file: skip expanding default format, 2022-03-15) added
special handling for the default batch format. In the meantime it has
fallen behind the code path for handling arbitrary formats. Bring it up
to speed by using the new and more efficient strbuf_add_oid_hex() and
strbuf_add_uint() instead of strbuf_addf():
Benchmark 1: ./git_main cat-file --batch-all-objects --batch-check='%(objectname) %(objecttype) %(objectsize)'
Time (mean ± σ): 1.051 s ± 0.003 s [User: 1.027 s, System: 0.023 s]
Range (min … max): 1.049 s … 1.058 s 10 runs
Benchmark 2: ./git_main cat-file --batch-all-objects --batch-check='%(objectname)-%(objecttype)-%(objectsize)'
Time (mean ± σ): 1.012 s ± 0.002 s [User: 0.988 s, System: 0.023 s]
Range (min … max): 1.010 s … 1.018 s 10 runs
Benchmark 3: ./git cat-file --batch-all-objects --batch-check='%(objectname) %(objecttype) %(objectsize)'
Time (mean ± σ): 979.0 ms ± 1.1 ms [User: 954.1 ms, System: 23.2 ms]
Range (min … max): 977.7 ms … 980.8 ms 10 runs
Summary
./git cat-file --batch-all-objects --batch-check='%(objectname) %(objecttype) %(objectsize)' ran
1.03 ± 0.00 times faster than ./git_main cat-file --batch-all-objects --batch-check='%(objectname)-%(objecttype)-%(objectsize)'
1.07 ± 0.00 times faster than ./git_main cat-file --batch-all-objects --batch-check='%(objectname) %(objecttype) %(objectsize)'
Signed-off-by: René Scharfe <l.s.r@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
odb: use size_t for object_info.sizep and the size APIs
When `js/objects-larger-than-4gb-on-windows` widened the streaming,
index-pack and unpack-objects code paths, in the interest of keeping the
patches somewhat reasonably-sized, it left the public ODB API still
typed in `unsigned long`. In particular `struct object_info::sizep` and
the four wrappers built on top of it (`odb_read_object`,
`odb_read_object_peeled`, `odb_read_object_info`, `odb_pretend_object`)
still return the unpacked size through `unsigned long *`, so on Windows
`cat-file -s` and the `git add` / `git status` paths for a >4 GiB blob
silently cap at 4 GiB.
Widen the field and the four wrappers. The previous commits already
widened the `unpack_entry()` cascade and pack-objects' in-core size
accessors, so most of the cascade arrives here with no further work: the
temporary shims in `packed_object_info_with_index_pos()` and in
`unpack_entry()`'s delta-base recovery path go away, the two
`SET_SIZE(entry, cast_size_t_to_ulong(canonical_size))` calls in
`check_object()` and the matching one in `drop_reused_delta()` collapse
to plain `SET_SIZE`, and `oe_get_size_slow()`'s tail
`cast_size_t_to_ulong()` is gone too.
What remains narrow are the boundaries this series does not
intend to touch: the diff, blame, textconv and fast-import machinery.
Even so, this patch is unfortunately quite large.
Assisted-by: Opus 4.7 Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
packfile,delta: drop the `cast_size_t_to_ulong()` wrappers
When I started the transition from `unsigned long` to `size_t`, in the
interest of keeping the patches reviewable, I introduced these calls to
prevent data type narrowing from silently failing to handle large object
sizes. I also introduced `*_sz()` variants that would allow most of the
callers to keep using that `unsigned long` that the 90s kindly asked to
be returned.
After the preceding commits, the only places that called the narrow
wrappers either no longer exist or already use the `_sz` form
internally, so the wrappers just narrow values back through
`cast_size_t_to_ulong()` for no reason.
Drop them and rename the `_sz` variants back to the natural names.
Assisted-by: Opus 4.7 Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
`pack-objects` stores per-entry object sizes in either the 31-bit
`size_` member of the `struct object_entry` or, when the value does not
fit, the `pack->delta_size[]` spill array. The accessors (`oe_size`,
`oe_delta_size`, `oe_get_size_slow`, `oe_size_*_than`) and the setters
(`oe_set_size`, `oe_set_delta_size`) used `unsigned long` for the spill
type, which on Windows means the spill silently caps at 4 GiB per entry.
That is what made `upload-pack` die with "object too large to read on
this platform" when serving the >4 GiB blob in `t5608` tests 5 and 6
when run with `GIT_TEST_CLONE_2GB`.
Widen them all to `size_t` (including `pack->delta_size`) and drop the
three `cast_size_t_to_ulong()` calls in `check_object()` that guarded
`in_pack_size`. The two `SET_SIZE(entry, canonical_size)` calls in the
same function stay cast-free as before, since `canonical_size` is still
`unsigned long` until a later commit widens `object_info::sizep`.
Assisted-by: Opus 4.7 Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
packfile: widen unpack_entry()'s size out-parameter to size_t
The topic `js/objects-larger-than-4gb-on-windows` widened the streaming,
index-pack and unpack-objects paths to `size_t` but deliberately stopped
at the in-memory `unpack_entry()` cascade, which still hands back the
unpacked size through `unsigned long *`. On Windows that boundary
truncates above 4 GiB because that data type is only 32 bits wide on
that platform.
Widen the code path. Except `packed_object_info_with_index_pos()`: It
cannot yet pass `oi->sizep` directly because the field is still
`unsigned long *`; bridge it with a `size_t` temporary that narrows
back, and let a later commit drop the bridge once the field is wide
too. `gfi_unpack_entry()` keeps its narrow signature because fast-import
tracks sizes through `unsigned long` everywhere it crosses subsystem
boundaries, keeping its signature allows the scope of this commit to be
somewhat reasonable, still.
Assisted-by: Opus 4.7 Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
pack-objects(check_pack_inflate()): use size_t instead of unsigned long
`write_reuse_object()` learned to track its packed-object size as
`size_t` in 606c192380 (odb, packfile: use size_t for streaming
object sizes, 2026-05-08), but the comparison sink it feeds,
`check_pack_inflate()`, still takes the expected decompressed size
as `unsigned long`. The call site bridges the mismatch with
`cast_size_t_to_ulong()`, which on Windows turns a >4 GiB object
into an immediate die().
That function only uses `expect` once: as the right-hand side of a
`stream.total_out == expect` equality test against zlib's counter.
zlib's own `total_out` counter is `uLong` and is therefore still
32-bit-bound on Windows. Widening `expect` to `size_t` cannot fix that,
but it is a strict improvement nonetheless: instead of dying outright,
an oversized object now simply makes the equality fail and lets
`write_reuse_object()` fall back to `write_no_reuse_object()`, which
decompresses and re-deflates the content (and which the larger
pack-objects widening series targets separately).
Drop the `cast_size_t_to_ulong()` shim at the call site now that
the receiving parameter speaks the same type as `entry_size`.
Assisted-by: Opus 4.7 Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
`patch_delta()` takes the source and delta sizes by value and writes
back the reconstructed target size through an `unsigned long *`. That
datatype cannot represent a value that exceeds 4 GiB on systems where
`unsigned long` is 32-bit (notably 64-bit Windows builds), though, even
though the delta encoding itself, the on-disk layout, and the in-memory
buffers happily carry such sizes. A `size_t` companion to
`get_delta_hdr_size()`, `get_delta_hdr_size_sz()`, was introduced in 17fa077596 (delta, packfile: use size_t for delta header sizes,
2026-05-08) precisely so that `patch_delta()` could be widened without
changing the on-the-wire decoding helper's signature.
Widen `patch_delta()`'s three size parameters to `size_t` and switch
its internal use of `get_delta_hdr_size()` to the `_sz` variant.
Then propagate the wider type through the callers.
Assisted-by: Opus 4.7 Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
On Windows, `unsigned long` and `long` are 32 bits even on 64-bit
builds. The MSVC compatibility header has shimmed `ftruncate()` with
#define ftruncate _chsize
ever since `compat/msvc-posix.h` was introduced. `_chsize()` takes a
32-bit `long` for the new length, which silently truncates files (and
the requested size) to 2 GiB. That is enough to make t7508 test 126
"git add fails gracefully with 4 GiB and 8 GiB files" fail under
MSVC: `test-tool truncate` creates a sparse 4 GiB or 8 GiB file via
the shimmed `ftruncate()`, and the test never gets off the ground.
`_chsize_s()` is the modern replacement, accepts a 64-bit `__int64`
length, and is the only sensible target on Windows. The catch is that
it does not follow the POSIX `-1` + `errno` convention: it returns
`0` on success and an errno value (a small positive integer) on
failure. A plain `#define ftruncate _chsize_s` would therefore
silently break callers that test the return value as `< 0` or against
`-1`, of which there are several: `http.c`, `parallel-checkout.c`,
and `t/helper/test-truncate.c` among them.
Introduce a `static inline` wrapper that calls `_chsize_s()`, copies
its errno return into `errno`, and translates the result to the
familiar `-1` / `0` convention, then point `ftruncate` at the
wrapper. Place the wrapper after `#include "mingw-posix.h"` so the
`off_t` parameter resolves to the already-widened `off64_t` rather
than the 32-bit `_off_t` from `compat/vcbuild/include/unistd.h`.
MinGW is unaffected: its `ftruncate()` already takes `off_t` and
routes through `ftruncate64()` when `_FILE_OFFSET_BITS=64`, which is
the default in our build.
Assisted-by: Opus 4.7 Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano [Mon, 15 Jun 2026 14:42:00 +0000 (07:42 -0700)]
Merge branch 'ob/more-repo-config-values'
Many core configuration variables have been migrated from global
variables into 'repo_config_values' to tie them to a specific
repository instance, avoiding cross-repository state leakage.
* ob/more-repo-config-values:
environment: move "warn_on_object_refname_ambiguity" into `struct repo_config_values`
environment: move "sparse_expect_files_outside_of_patterns" into `struct repo_config_values`
environment: move "core_sparse_checkout_cone" into `struct repo_config_values`
environment: move "precomposed_unicode" into `struct repo_config_values`
environment: move "pack_compression_level" into `struct repo_config_values`
environment: move `zlib_compression_level` into `struct repo_config_values`
environment: move "check_stat" into `struct repo_config_values`
environment: move "trust_ctime" into `struct repo_config_values`
Junio C Hamano [Mon, 15 Jun 2026 14:41:59 +0000 (07:41 -0700)]
Merge branch 'ps/setup-centralize-odb-creation'
The setup logic to discover and configure repositories has been
refactored, and the initialization of the object database has been
centralized.
* ps/setup-centralize-odb-creation:
setup: construct object database in `apply_repository_format()`
repository: stop reading loose object map twice on repo init
setup: stop initializing object database without repository
setup: stop creating the object database in `setup_git_env()`
repository: stop initializing the object database in `repo_set_gitdir()`
setup: deduplicate logic to apply repository format
setup: drop `setup_git_env()`
t0001: plug test gaps for git-init(1) with GIT_OBJECT_DIRECTORY
Junio C Hamano [Mon, 15 Jun 2026 14:41:59 +0000 (07:41 -0700)]
Merge branch 'hn/config-typo-advice'
"git config foo.bar=baz" is not likely to be a request to read the
value of such a variable with '=' in its name; rather it is plausible
that the user meant "git config set foo.bar baz". Give advice when
giving an error message.
* hn/config-typo-advice:
config: improve diagnostic for "set" with missing value
config: add git_config_key_is_valid() for quiet validation
Junio C Hamano [Mon, 15 Jun 2026 14:41:59 +0000 (07:41 -0700)]
Merge branch 'ls/doc-raw-timestamp-prefix'
Documentation and tests have been added to clarify that Git's internal
raw timestamp format requires a `@` prefix for values less than
100,000,000 to prevent ambiguity with other formats like YYYYMMDD.
* ls/doc-raw-timestamp-prefix:
doc: document and test `@` prefix for raw timestamps
Junio C Hamano [Mon, 15 Jun 2026 14:41:59 +0000 (07:41 -0700)]
Merge branch 'jc/submitting-patches-cover-letter'
Guidelines on how to write a cover letter for a multi-patch series
have been added to SubmittingPatches, which also got a new marker
to separate the section for typofixes.
* jc/submitting-patches-cover-letter:
SubmittingPatches: describe cover letter
SubmittingPatches: separate typofixes section
Elijah Newren [Sun, 14 Jun 2026 06:57:50 +0000 (06:57 +0000)]
commit-graph: use timestamp_t for max parent generation accumulator
compute_reachable_generation_numbers() computes each commit's
generation as
max(c->date, max(parent.generation)) + 1
by walking its parents and accumulating their generations into a
local
uint32_t max_gen = 0;
while info->get_generation() returns timestamp_t and
compute_generation_from_max() already takes its max_gen parameter
as timestamp_t. For v1 (topological levels) the narrowing is
harmless because GENERATION_NUMBER_V1_MAX is less than 2^30, but
for v2 (corrected committer dates) it silently truncates any
parent generation that does not fit in 32 bits, i.e. any parent
whose committer timestamp is at or beyond 2106-02-07 UTC
(>= 2^32).
The truncated max then causes child commits to end up with a
corrected committer date that matches the parent's instead of being
at least 1 higher. The bad value gets written into the commit-graph
and causes problems later, and can be noticed by running `git
commit-graph verify`.
Widen the accumulator to timestamp_t.
This is solely an in-memory arithmetic fix with no on-disk format
change: the on-disk format already encodes timestamp_t values and
existing readers handle them unchanged. This merely allows the code to
compute the correct value to write to disk.
The narrowing was introduced in 80c928d947c2 (commit-graph:
simplify compute_generation_numbers(), 2023-03-20), which rewired
v2 to use the shared compute_reachable_generation_numbers()
helper; the helper's local accumulator had been declared uint32_t
in the immediately preceding 368d19b0b7fa (commit-graph: refactor
compute_topological_levels(), 2023-03-20) when only v1 was using
it, where it was harmless.
Add a new test with a future-dated parent and a present-day child;
without the above fix, `git commit-graph verify` reports the
descendant's stored generation as below parent + 1.
Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Elijah Newren [Sun, 14 Jun 2026 06:37:26 +0000 (06:37 +0000)]
cache-tree: fix verify_cache() to catch non-adjacent D/F conflicts
verify_cache() checks that the index does not contain both "path" and
"path/file" before writing a tree. It does this by comparing only
adjacent entries, relying on the assumption that "path/file" would
immediately follow "path" in sorted order. Unfortunately, this
assumption does not always hold. For example:
docs <-- submodule entry
docs-internal/README.md <-- intervening entry
docs/requirements.txt <-- D/F conflict, NOT adjacent to "docs"
When this happens, verify_cache() silently misses the D/F conflict and
write-tree produces a corrupt tree object containing duplicate entries
(one for the submodule "docs" and one for the tree "docs").
I could not find any caller in current git that both allows the index to
get into this state and then tries to write it out without doing other
checks beyond the verify_cache() call in cache_tree_update(), but
verify_cache() is documented as a safety net for preventing corrupt
trees and should actually provide that guarantee. A downstream consumer
that relied solely on cache_tree_update()'s internal checking via
verify_cache() to prevent duplicate tree entries was bitten by the gap.
Add a test that constructs a corrupt index directly (bypassing the D/F
checks in add_index_entry) and verifies that write-tree now rejects it.
Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Elijah Newren [Sun, 14 Jun 2026 06:37:25 +0000 (06:37 +0000)]
merge-ort: abort merge when trees have duplicate entries
Trees with duplicate entries are malformed; fsck reports "contains
duplicate file entries" for them. merge-ort has from the beginning
assumed that we would never hit such trees. It was written with the
assumption that traverse_trees() calls collect_merge_info_callback() at
most once per path. The "sanity checks" in that callback (added in d2bc1994f363 (merge-ort: implement a very basic collect_merge_info(),
2020-12-13)) verify properties of each individual call but not that
invariant. The strmap_put() in setup_path_info() silently overwrites
the entry from any prior call for the same path, because it assumed
there would be no other path. Unfortunately, supplemental data
structures for various optimizations could still be tweaked before the
extra paths were overwritten, and those data structures not matching
expected state could trip various assertions.
Change the return type of setup_path_info() from void to int to allow us
to detect this case, and abort the merge with a clear error message when
it occurs.
Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Elijah Newren [Sun, 14 Jun 2026 06:37:24 +0000 (06:37 +0000)]
merge-ort: free diff pairs queue in clear_or_reinit_internal_opts()
clear_or_reinit_internal_opts() is responsible for cleaning up the
various data structures in merge_options_internal. It already handles
many renames-related structures (dirs_removed, dir_renames,
relevant_sources, cached_pairs, deferred, etc.) but does not free
renames->pairs[].queue.
In the normal code path, resolve_and_process_renames() frees
pairs[s].queue and reinitializes it with diff_queue_init() before
clear_or_reinit_internal_opts() runs, so the omission is harmless.
However, if collect_merge_info() encounters an error and returns early
(before resolve_and_process_renames() is ever called), any diff pairs
already queued by collect_rename_info()/add_pair() will have their
backing array leaked.
Fix this by freeing renames->pairs[].queue in the cleanup function.
In the normal path the pointer is already NULL (from the earlier
diff_queue_init() in resolve_and_process_renames()), so free(NULL) is
a safe no-op.
Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Elijah Newren [Sun, 14 Jun 2026 06:37:23 +0000 (06:37 +0000)]
merge-ort: drop unnecessary show_all_errors from collect_merge_info()
collect_merge_info() has set info.show_all_errors = 1 since d2bc1994f363 (merge-ort: implement a very basic collect_merge_info(),
2020-12-13). This setting was copied from unpack-trees.c where it
controls batching of error messages for porcelain display, but
merge-ort has no such error-batching logic and never needed it.
With show_all_errors set, traverse_trees() captures a negative callback
return but continues processing remaining entries rather than stopping
immediately. Removing the setting restores the default behavior where
a negative return from collect_merge_info_callback() breaks out of the
traversal loop right away, allowing a future commit to exit early when
a corrupt tree is detected.
Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Elijah Newren [Sun, 14 Jun 2026 06:37:22 +0000 (06:37 +0000)]
merge-ort: propagate callback errors from traverse_trees_wrapper()
traverse_trees_wrapper() saves entries from a first pass through
traverse_trees() and then replays them through the real callback
(collect_merge_info_callback). However, the replay loop silently
discards the callback return value. This is not a deferred error;
it is an ignored error.
Today the only originator of a negative return in this entire call
graph is traverse_trees()'s "exceeded maximum allowed tree depth"
check; everything else (collect_merge_info_callback,
traverse_trees_wrapper, the inner traverse_trees recursion) only
relays that. So in current Git, the visible effect of dropping the
replay callback's return value is narrow but bad: a tree nested past
core.maxTreeDepth has its -1 swallowed, the subtree below the limit
is silently pruned, and the merge completes as if that were the
correct result.
A later patch in this series will teach collect_merge_info_callback()
to return -1 on an additional path -- detecting duplicate
entries in malformed trees -- which is similarly handled today by
just ignoring the problem (resulting in mostly a "last one wins" rule,
though the non-last entry can mutate various state flags).
Capture the return value, stop the loop on negative returns, and
propagate the error to the caller. The callback returns a positive mask
value on success, so normalize non-negative returns to
0 for the caller.
Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Dominik Loidolt [Sat, 13 Jun 2026 12:27:09 +0000 (14:27 +0200)]
compat/posix.h: enable UNUSED warning messages for Clang
Use a dedicated Clang version check for the UNUSED macro.
Commit 7c07f36ad2 (git-compat-util.h: GCC deprecated message arg only in
GCC 4.5+, 2022-10-05) restricted use of the deprecated attribute's
message argument in the UNUSED macro to GCC 4.5 or newer.
Clang identifies itself as GNUC 4.2.1 for compatibility, so
GIT_GNUC_PREREQ(4, 5) does not detect whether Clang supports the
deprecated("...") form. Add GIT_CLANG_PREREQ() macro and use it to
enable the UNUSED warning message for Clang 2.9 and newer.
Signed-off-by: Dominik Loidolt <dominik.loidolt@univie.ac.at> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Tamir Duberstein [Fri, 12 Jun 2026 21:27:44 +0000 (17:27 -0400)]
ref-filter: restore prefix-scoped iteration
dabecb9db2 (for-each-ref: introduce a '--start-after' option,
2025-07-15) changed branch, remote-tracking branch, and tag enumeration
from constructing an iterator with the namespace prefix to constructing
an unscoped iterator and seeking to the prefix.
Review of --start-after noted that the construction prefix and seek
position represent different state and are easy to conflate [1]. It also
noted that future branch or tag support would need to retain the
namespace prefix while moving the cursor [2].
The files backend constructs its loose-ref iterator with cache priming
enabled. cache_ref_iterator_begin() immediately applies the construction
prefix through cache_ref_iterator_set_prefix(), reading loose refs
beneath it before packed refs are opened. An empty prefix therefore
reads every loose ref, and a later seek cannot undo that I/O.
For the current single-kind filters, construct the iterator with the
namespace prefix when start_after is not set. Leave the existing
start_after path unchanged; no current command combines it with these
filters, and future support must carry the prefix separately from the
cursor.
With 10,000 unrelated loose refs in the files backend, the p6300 tests
improve as follows:
before after
branch 2.74 s 0.11 s
branch --remotes 2.81 s 0.12 s
tag 3.01 s 0.11 s
Tamir Duberstein [Fri, 12 Jun 2026 04:31:51 +0000 (21:31 -0700)]
ls-files: filter pathspec before lstat
In --deleted and --modified modes, show_files() calls lstat() for each
index entry before show_ce() applies the pathspec. prune_index() avoids
most of these calls for pathspecs with a common directory prefix, but
not for a top-level name or leading wildcard.
Match before lstat() to avoid accessing the worktree for entries that
cannot be shown. Treat this as a prefilter: do not update ps_matched,
and retain the match in show_ce() so --error-unmatch is satisfied only
by entries that the selected modes actually show.
Prefilter only a single pathspec item, bounding the added work for each
index entry. Applying match_pathspec() to multiple arguments can cost
more than the lstat() calls it avoids. In a synthetic repository with
10,000 clean files, passing every path to ls-files --modified increased
runtime from 112.5 ms to 494.1 ms when the prefilter was unconditional.
With $parent and $this exported as paths to binaries built from the
parent and this commit, on a repository with 881,290 index entries:
Tian Yuchen [Wed, 10 Jun 2026 12:43:52 +0000 (20:43 +0800)]
environment: move 'protect_hfs' and 'protect_ntfs' into 'repo_config_values'
Move the global 'protect_hfs' and 'protect_ntfs' configurations
into the repository-specific 'repo_config_values' struct.
This will help with the elimination of 'the_repository'
To ensure code readability, the getter functions
'repo_protect_hfs()' and 'repo_protect_ntfs()'
have been introduced.
For now, associated functions access this configuration by
explicitly falling back to 'the_repository', which needs to
be addressed in the future.
Note: In 't/helper/test-path-utils.c', there is a function
'protect_ntfs_hfs_benchmark()' where these two global
variables are used as loop iterators. New local variables
have been created to replace them.
Mentored-by: Christian Couder <christian.couder@gmail.com> Mentored-by: Ayush Chandekar <ayu.chandekar@gmail.com> Mentored-by: Olamide Caleb Bello <belkid98@gmail.com> Signed-off-by: Tian Yuchen <cat@malon.dev> Signed-off-by: Junio C Hamano <gitster@pobox.com>
projects / thirdparty / git.git / log
