"git diff --cached" codepath did not fill the necessary stat
information for a file when fsmonitor knows it is clean and ended
up behaving as if it is not clean, which has been corrected.
* js/diff-cached-fsmonitor-fix:
diff-lib: fix check_removed when fsmonitor is on
Update "git maintainance" timers' implementation based on systemd
timers to work with WSL.
* js/systemd-timers-wsl-fix:
maintenance(systemd): support the Windows Subsystem for Linux
"git diff --no-index -R <(one) <(two)" did not work correctly,
which has been corrected.
* pw/diff-no-index-from-named-pipes:
diff --no-index: fix -R with stdin
The completion script (in contrib/) has been taught to treat the
"-t" option to "git checkout" and "git switch" just like the
"--track" option, to complete remote-tracking branches.
* js/complete-checkout-t:
completion(switch/checkout): treat --track and -t the same
"git grep -e A --no-or -e B" is accepted, even though the negation
of "or" did not mean anything, which has been tightened.
* rs/grep-no-no-or:
grep: reject --no-or
References from description of the `--patch` option in various
manual pages have been simplified and improved.
* so/diff-doc-for-patch-update:
doc/diff-options: fix link to generating patch section
Various fixes to the behaviour of "rebase -i" when the command got
interrupted by conflicting changes.
cf. <6b927687-cf6e-d73e-78fb-bd4f46736928@gmx.de>
* pw/rebase-i-after-failure:
rebase -i: fix adding failed command to the todo list
rebase --continue: refuse to commit after failed command
rebase: fix rewritten list for failed pick
sequencer: factor out part of pick_commits()
sequencer: use rebase_path_message()
rebase -i: remove patch file after conflict resolution
rebase -i: move unlink() calls
"git for-each-ref --sort='contents:size'" sorts the refs according
to size numerically, giving a ref that points at a blob twelve-byte
(12) long before showing a blob hundred-byte (100) long.
* ks/ref-filter-sort-numerically:
ref-filter: sort numerically when ":size" is used
"git diff --no-such-option" and other corner cases around the exit
status of the "diff" command has been corrected.
* jk/diff-result-code-cleanup:
diff: drop useless "status" parameter from diff_result_code()
diff: drop useless return values in git-diff helpers
diff: drop useless return from run_diff_{files,index} functions
diff: die when failing to read index in git-diff builtin
diff: show usage for unknown builtin_diff_files() options
diff-files: avoid negative exit value
diff: spell DIFF_INDEX_CACHED out when calling run_diff_index()
The use of API between two calls to require_clean_work_tree() from
the sequencer code has been cleaned up for consistency.
* ob/sequencer-empty-hint-fix:
sequencer: rectify empty hint in call of require_clean_work_tree()
transfer.unpackLimit ought to be used as a fallback, but overrode
fetch.unpackLimit and receive.unpackLimit instead.
* ts/unpacklimit-config-fix:
transfer.unpackLimit: fetch/receive.unpackLimit takes precedence
"git diff -w --exit-code" with various options did not work
correctly, which is being addressed.
* jc/diff-exit-code-with-w-fixes:
diff: the -w option breaks --exit-code for --raw and other output modes
t4040: remove test that succeeded for a wrong reason
diff: teach "--stat -w --exit-code" to notice differences
diff: mode-only change should be noticed by "--patch -w --exit-code"
diff: move the fallback "--exit-code" code down
The commit-graph verification code that detects mixture of zero and
non-zero generation numbers has been updated.
* tb/commit-graph-verify-fix:
commit-graph: avoid repeated mixed generation number warnings
t/t5318-commit-graph.sh: test generation zero transitions during fsck
commit-graph: verify swapped zero/non-zero generation cases
commit-graph: introduce `commit_graph_generation_from_graph()`
Tweak GitHub Actions CI so that pushing the same commit to multiple
branch tips at the same time will not waste building and testing
the same thing twice.
* jc/ci-skip-same-commit:
ci: avoid building from the same commit in parallel
Overly long label names used in the sequencer machinery are now
chopped to fit under filesystem limitation.
* mp/rebase-label-length-limit:
rebase: allow overriding the maximal length of the generated labels
sequencer: truncate labels to accommodate loose refs
GitHub CI workflow has learned to trigger Coverity check.
* js/ci-coverity:
coverity: detect and report when the token or project is incorrect
coverity: allow running on macOS
coverity: support building on Windows
coverity: allow overriding the Coverity project
coverity: cache the Coverity Build Tool
ci: add a GitHub workflow to submit Coverity scans
Tests with LSan from time to time seem to emit harmless message
that makes our tests unnecessarily flakey; we work it around by
filtering the uninteresting output.
* jk/test-lsan-denoise-output:
test-lib: ignore uninteresting LSan output
Flakey "git p4" tests, as well as "git svn" tests, are now skipped
in the (rather expensive) sanitizer CI job.
* js/ci-san-skip-p4-and-svn-tests:
ci(linux-asan-ubsan): let's save some time
Tests that are known to pass with LSan are now marked as such.
* tb/mark-more-tests-as-leak-free:
leak tests: mark t5583-push-branches.sh as leak-free
leak tests: mark t3321-notes-stripspace.sh as leak-free
leak tests: mark a handful of tests as leak-free
There seems to be some internal stack overflow detection in MSVC's
`malloc()` machinery that seems to be independent of the `stack reserve`
and `heap reserve` sizes specified in the executable (editable via
`EDITBIN /STACK:<n> <exe>` and `EDITBIN /HEAP:<n> <exe>`).
In the newly test cases added by `jk/tree-name-and-depth-limit`, this
stack overflow detection is unfortunately triggered before Git can print
out the error message about too-deep trees and exit gracefully. Instead,
it exits with `STATUS_STACK_OVERFLOW`. This corresponds to the numeric
value -1073741571, something the MSYS2 runtime we sadly need to use to
run Git's test suite cannot handle and which it internally maps to the
exit code 127. Git's test suite, in turn, mistakes this to mean that the
command was not found, and fails both test cases.
Here is an example stack trace from an example run:
[0x0] ntdll!RtlpAllocateHeap+0x31 0x4212603f50 0x7ff9d6d4cd49
[0x1] ntdll!RtlpAllocateHeapInternal+0x6c9 0x42126041b0 0x7ff9d6e14512
[0x2] ntdll!RtlDebugAllocateHeap+0x102 0x42126042b0 0x7ff9d6dcd8b0
[0x3] ntdll!RtlpAllocateHeap+0x7ec70 0x4212604350 0x7ff9d6d4cd49
[0x4] ntdll!RtlpAllocateHeapInternal+0x6c9 0x42126045b0 0x7ff9596ed480
[0x5] ucrtbased!heap_alloc_dbg_internal+0x210 0x42126046b0 0x7ff9596ed20d
[0x6] ucrtbased!heap_alloc_dbg+0x4d 0x4212604750 0x7ff9596f037f
[0x7] ucrtbased!_malloc_dbg+0x2f 0x42126047a0 0x7ff9596f0dee
[0x8] ucrtbased!malloc+0x1e 0x42126047d0 0x7ff730fcc1ef
[0x9] git!do_xmalloc+0x2f 0x4212604800 0x7ff730fcc2b9
[0xa] git!do_xmallocz+0x59 0x4212604840 0x7ff730fca779
[0xb] git!xmallocz_gently+0x19 0x4212604880 0x7ff7311b0883
[0xc] git!unpack_compressed_entry+0x43 0x42126048b0 0x7ff7311ac9a4
[0xd] git!unpack_entry+0x554 0x42126049a0 0x7ff7311b0628
[0xe] git!cache_or_unpack_entry+0x58 0x4212605250 0x7ff7311ad3a8
[0xf] git!packed_object_info+0x98 0x42126052a0 0x7ff7310a92da
[0x10] git!do_oid_object_info_extended+0x3fa 0x42126053b0 0x7ff7310a44e7
[0x11] git!oid_object_info_extended+0x37 0x4212605460 0x7ff7310a38ba
[0x12] git!repo_read_object_file+0x9a 0x42126054a0 0x7ff7310a6147
[0x13] git!read_object_with_reference+0x97 0x4212605560 0x7ff7310b4656
[0x14] git!fill_tree_descriptor+0x66 0x4212605620 0x7ff7310dc0a5
[0x15] git!traverse_trees_recursive+0x3f5 0x4212605680 0x7ff7310dd831
[0x16] git!unpack_callback+0x441 0x4212605790 0x7ff7310b4c95
[0x17] git!traverse_trees+0x5d5 0x42126058a0 0x7ff7310dc0f2
[0x18] git!traverse_trees_recursive+0x442 0x4212605980 0x7ff7310dd831
[0x19] git!unpack_callback+0x441 0x4212605a90 0x7ff7310b4c95
[0x1a] git!traverse_trees+0x5d5 0x4212605ba0 0x7ff7310dc0f2
[0x1b] git!traverse_trees_recursive+0x442 0x4212605c80 0x7ff7310dd831
[0x1c] git!unpack_callback+0x441 0x4212605d90 0x7ff7310b4c95
[0x1d] git!traverse_trees+0x5d5 0x4212605ea0 0x7ff7310dc0f2
[0x1e] git!traverse_trees_recursive+0x442 0x4212605f80 0x7ff7310dd831
[0x1f] git!unpack_callback+0x441 0x4212606090 0x7ff7310b4c95
[0x20] git!traverse_trees+0x5d5 0x42126061a0 0x7ff7310dc0f2
[0x21] git!traverse_trees_recursive+0x442 0x4212606280 0x7ff7310dd831
[...]
[0xfad] git!cmd_main+0x2a2 0x42126ff740 0x7ff730fb6345
[0xfae] git!main+0xe5 0x42126ff7c0 0x7ff730fbff93
[0xfaf] git!wmain+0x2a3 0x42126ff830 0x7ff731318859
[0xfb0] git!invoke_main+0x39 0x42126ff8a0 0x7ff7313186fe
[0xfb1] git!__scrt_common_main_seh+0x12e 0x42126ff8f0 0x7ff7313185be
[0xfb2] git!__scrt_common_main+0xe 0x42126ff960 0x7ff7313188ee
[0xfb3] git!wmainCRTStartup+0xe 0x42126ff990 0x7ff9d5ed257d
[0xfb4] KERNEL32!BaseThreadInitThunk+0x1d 0x42126ff9c0 0x7ff9d6d6aa78
[0xfb5] ntdll!RtlUserThreadStart+0x28 0x42126ff9f0 0x0
I verified manually that `traverse_trees_cur_depth` was 562 when that
happened, which is far below the 2048 that were already accepted into
Git as a hard limit.
Despite many attempts to figure out which of the internals trigger this
`STATUS_STACK_OVERFLOW` and how to maybe increase certain sizes to avoid
running into this issue and let Git behave the same way as under Linux,
I failed to find any build-time/runtime knob we could turn to that
effect.
Note: even switching to using a different allocator (I used mimalloc
because that's what Git for Windows uses for its GCC builds) does not
help, as the zlib code used to unpack compressed pack entries _still_
uses the regular `malloc()`. And runs into the same issue.
Note also: switching to using a different allocator _also_ for zlib code
seems _also_ not to help. I tried that, and it still exited with
`STATUS_STACK_OVERFLOW` that seems to have been triggered by a
`mi_assert_internal()`, i.e. an internal assertion of mimalloc...
So the best bet to work around this for now seems to just lower the
maximum allowed tree depth _even further_ for MSVC builds.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
git merge-file knows how to merge files on the file system already. It
would be helpful, however, to allow it to also merge single blobs.
Teach it an `--object-id` option which means that its arguments are
object IDs and not files to allow it to do so.
We handle the empty blob specially since read_mmblob doesn't read it
directly and otherwise users cannot specify an empty ancestor.
Signed-off-by: brian m. carlson <bk2204@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
`git merge-file` takes three positional arguments. Each of them is
documented as `<foo-file>`. In preparation for teaching this command to
alternatively take three object IDs, make these placeholders a bit more
generic by dropping the "-file" parts. Instead, clarify early that the
three arguments are filenames. Even after the next commit, we can afford
to present this file-centric view up front and in the general
discussion, since it will remain the default one.
Signed-off-by: Martin Ågren <martin.agren@gmail.com>
Signed-off-by: brian m. carlson <bk2204@github.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Back in 32f3c541e3 (multi-pack-index: write pack names in chunk,
2018-07-12) the MIDX's "Packfile Names" (or "PNAM", for short) chunk was
described as containing an array of string entries. e0d1bcf825 notes
that this is the only chunk in the MIDX format's specification that is
not guaranteed to be 4-byte aligned, and so should be placed last.
This isn't quite accurate: the entries within the PNAM chunk are not
guaranteed to be 4-byte aligned since they are arbitrary strings, but
the chunk itself is 4-byte aligned since the ending is padded with NUL
bytes.
That padding has always been there since 32f3c541e3 via
midx.c::write_midx_pack_names(), which ended with:
i = MIDX_CHUNK_ALIGNMENT - (written % MIDX_CHUNK_ALIGNMENT)
if (i < MIDX_CHUNK_ALIGNMENT) {
unsigned char padding[MIDX_CHUNK_ALIGNMENT];
memset(padding, 0, sizeof(padding))
hashwrite(f, padding, i);
written += i;
}
In fact, 32f3c541e3's log message itself describes the chunk in its
first paragraph with:
Since filenames are not well structured, add padding to keep good
alignment in later chunks.
So these have always been externally aligned. Correct the corresponding
part of our documentation to reflect that.
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
e0d1bcf825 (multi-pack-index: add format details, 2018-07-12) describes
the MIDX's "PNAM" chunk as having entries which are "null-terminated
strings".
This is a typo, as strings are terminated with a NUL character, which is
a distinct concept from "NULL" or "null", which we typically reserve for
the void pointer to address 0.
Correct the documentation accordingly.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Convert tests that use `test_path_is_file` and `test_path_is_missing` to
instead use a set of helpers `test_ref_exists` and `test_ref_missing`.
These helpers are implemented via the newly introduced `git show-ref
--exists` command. Thus, we can avoid intimate knowledge of how the ref
backend stores references on disk.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
While we have multiple ways to show the value of a given reference, we
do not have any way to check whether a reference exists at all. While
commands like git-rev-parse(1) or git-show-ref(1) can be used to check
for reference existence in case the reference resolves to something
sane, neither of them can be used to check for existence in some other
scenarios where the reference does not resolve cleanly:
- References which have an invalid name cannot be resolved.
- References to nonexistent objects cannot be resolved.
- Dangling symrefs can be resolved via git-symbolic-ref(1), but this
requires the caller to special case existence checks depending on
whether or not a reference is symbolic or direct.
Furthermore, git-rev-list(1) and other commands do not let the caller
distinguish easily between an actually missing reference and a generic
error.
Taken together, this seems like sufficient motivation to introduce a
separate plumbing command to explicitly check for the existence of a
reference without trying to resolve its contents.
This new command comes in the form of `git show-ref --exists`. This
new mode will exit successfully when the reference exists, with a
specific exit code of 2 when it does not exist, or with 1 when there
has been a generic error.
Note that the only way to properly implement this command is by using
the internal `refs_read_raw_ref()` function. While the public function
`refs_resolve_ref_unsafe()` can be made to behave in the same way by
passing various flags, it does not provide any way to obtain the errno
with which the reference backend failed when reading the reference. As
such, it becomes impossible for us to distinguish generic errors from
the explicit case where the reference wasn't found.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The synopsis treats the `--verify` and the implicit mode the same. They
are slightly different though:
- They accept different sets of flags.
- The implicit mode accepts patterns while the `--verify` mode
accepts references.
Split up the synopsis for these two modes such that we can disambiguate
those differences.
While at it, drop "--quiet" from the pattern mode's synopsis. It does
not make a lot of sense to list patterns, but squelch the listing output
itself. The description for "--quiet" is adapted accordingly.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The git-show-ref(1) command has three different modes, of which one is
implicit and the other two can be chosen explicitly by passing a flag.
But while these modes are standalone and cause us to execute completely
separate code paths, we gladly accept the case where a user asks for
both `--exclude-existing` and `--verify` at the same time even though it
is not obvious what will happen. Spoiler: we ignore `--verify` and
execute the `--exclude-existing` mode.
Let's explicitly detect this invalid usage and die in case both modes
were requested.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The patterns subcommand is the last command that still uses global
variables to track its options. Convert it to use a structure instead
with the same motivation as preceding commits.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The `show_one()` function implicitly receives a bunch of options which
are tracked via global variables. This makes it hard to see which
subcommands of git-show-ref(1) actually make use of these options.
Introduce a `show_one_options` structure that gets passed down to this
function. This allows us to get rid of more global state and makes it
more explicit which subcommands use those options.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When passing patterns to git-show-ref(1) we're checking whether any
reference matches -- if none do, we indicate this condition via an
unsuccessful exit code.
We're using a global variable to count these matches, which is required
because the counter is getting incremented in a callback function. But
now that we have the `struct show_ref_data` in place, we can get rid of
the global variable and put the counter in there instead.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It's not immediately obvious options which options are applicable to
what subcommand in git-show-ref(1) because all options exist as global
state. This can easily cause confusion for the reader.
Refactor options for the `--exclude-existing` subcommand to be contained
in a separate structure. This structure is stored on the stack and
passed down as required. Consequently, it clearly delimits the scope of
those options and requires the reader to worry less about global state.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When passing patterns to `git show-ref` we have some code that will
cause us to die if `verify && !quiet` is true. But because `verify`
indicates a different subcommand of git-show-ref(1) that causes us to
execute `cmd_show_ref__verify()` and not `cmd_show_ref__patterns()`, the
condition cannot ever be true.
Let's remove this dead code.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix a leaking string buffer in `git show-ref --exclude-existing`. While
the buffer is technically not leaking because its variable is declared
as static, there is no inherent reason why it should be.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
While not immediately obvious, git-show-ref(1) actually implements three
different subcommands:
- `git show-ref <patterns>` can be used to list references that
match a specific pattern.
- `git show-ref --verify <refs>` can be used to list references.
These are _not_ patterns.
- `git show-ref --exclude-existing` can be used as a filter that
reads references from standard input, performing some conversions
on each of them.
Let's make this more explicit in the code by splitting up the three
subcommands into separate functions. This also allows us to address the
confusingly named `patterns` variable, which may hold either patterns or
reference names.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The `pattern` variable is a global variable that tracks either the
reference names (not patterns!) for the `--verify` mode or the patterns
for the non-verify mode. This is a bit confusing due to the slightly
different meanings.
Convert the variable to be local. While this does not yet fix the double
meaning of the variable, this change allows us to address it in a
subsequent patch more easily by explicitly splitting up the different
subcommands of git-show-ref(1).
Note that we introduce a `struct show_ref_data` to pass the patterns to
`show_ref()`. While this is overengineered now, we will extend this
structure in a subsequent patch.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The `--missing` object option in rev-list currently works only with
missing blobs/trees. For missing commits the revision walker fails with
a fatal error.
Let's extend the functionality of `--missing` option to also support
commit objects. This is done by adding a `missing_objects` field to
`rev_info`. This field is an `oidset` to which we'll add the missing
commits as we encounter them. The revision walker will now continue the
traversal and call `show_commit()` even for missing commits. In rev-list
we can then check if the commit is a missing commit and call the
existing code for parsing `--missing` objects.
A scenario where this option would be used is to find the boundary
objects between different object directories. Consider a repository with
a main object directory (GIT_OBJECT_DIRECTORY) and one or more alternate
object directories (GIT_ALTERNATE_OBJECT_DIRECTORIES). In such a
repository, using the `--missing=print` option while disabling the
alternate object directory allows us to find the boundary objects
between the main and alternate object directory.
Helped-by: Patrick Steinhardt <ps@pks.im>
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Karthik Nayak <karthik.188@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The `show_commit()` function already depends on `finish_commit()`, and
in the upcoming commit, we'll also add a dependency on
`finish_object__ma()`. Since in C symbols must be declared before
they're used, let's move `show_commit()` below both `finish_commit()`
and `finish_object__ma()`, so the code is cleaner as a whole without the
need for declarations.
Signed-off-by: Karthik Nayak <karthik.188@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The bit `do_not_die_on_missing_tree` is used in revision.h to ensure the
revision walker does not die when encountering a missing tree. This is
currently exclusively set within `builtin/rev-list.c` to ensure the
`--missing` option works with missing trees.
In the upcoming commits, we will extend `--missing` to also support
missing commits. So let's rename the bit to
`do_not_die_on_missing_objects`, which is object type agnostic and can
be used for both trees/commits.
Signed-off-by: Karthik Nayak <karthik.188@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* ps/do-not-trust-commit-graph-blindly-for-existence:
commit: detect commits that exist in commit-graph but not in the ODB
commit-graph: introduce envvar to disable commit existence checks
Commit graphs can become stale and contain references to commits that do
not exist in the object database anymore. Theoretically, this can lead
to a scenario where we are able to successfully look up any such commit
via the commit graph even though such a lookup would fail if done via
the object database directly.
As the commit graph is mostly intended as a sort of cache to speed up
parsing of commits we do not want to have diverging behaviour in a
repository with and a repository without commit graphs, no matter
whether they are stale or not. As commits are otherwise immutable, the
only thing that we really need to care about is thus the presence or
absence of a commit.
To address potentially stale commit data that may exist in the graph,
our `lookup_commit_in_graph()` function will check for the commit's
existence in both the commit graph, but also in the object database. So
even if we were able to look up the commit's data in the graph, we would
still pretend as if the commit didn't exist if it is missing in the
object database.
We don't have the same safety net in `parse_commit_in_graph_one()`
though. This function is mostly used internally in "commit-graph.c"
itself to validate the commit graph, and this usage is fine. We do
expose its functionality via `parse_commit_in_graph()` though, which
gets called by `repo_parse_commit_internal()`, and that function is in
turn used in many places in our codebase.
For all I can see this function is never used to directly turn an object
ID into a commit object without additional safety checks before or after
this lookup. What it is being used for though is to walk history via the
parent chain of commits. So when commits in the parent chain of a graph
walk are missing it is possible that we wouldn't notice if that missing
commit was part of the commit graph. Thus, a query like `git rev-parse
HEAD~2` can succeed even if the intermittent commit is missing.
It's unclear whether there are additional ways in which such stale
commit graphs can lead to problems. In any case, it feels like this is a
bigger bug waiting to happen when we gain additional direct or indirect
callers of `repo_parse_commit_internal()`. So let's fix the inconsistent
behaviour by checking for object existence via the object database, as
well.
This check of course comes with a performance penalty. The following
benchmarks have been executed in a clone of linux.git with stable tags
added:
Benchmark 1: git -c core.commitGraph=true rev-list --topo-order --all (git = master)
Time (mean ± σ): 2.913 s ± 0.018 s [User: 2.363 s, System: 0.548 s]
Range (min … max): 2.894 s … 2.950 s 10 runs
Benchmark 2: git -c core.commitGraph=true rev-list --topo-order --all (git = pks-commit-graph-inconsistency)
Time (mean ± σ): 3.834 s ± 0.052 s [User: 3.276 s, System: 0.556 s]
Range (min … max): 3.780 s … 3.961 s 10 runs
Benchmark 3: git -c core.commitGraph=false rev-list --topo-order --all (git = master)
Time (mean ± σ): 13.841 s ± 0.084 s [User: 13.152 s, System: 0.687 s]
Range (min … max): 13.714 s … 13.995 s 10 runs
Benchmark 4: git -c core.commitGraph=false rev-list --topo-order --all (git = pks-commit-graph-inconsistency)
Time (mean ± σ): 13.762 s ± 0.116 s [User: 13.094 s, System: 0.667 s]
Range (min … max): 13.645 s … 14.038 s 10 runs
Summary
git -c core.commitGraph=true rev-list --topo-order --all (git = master) ran
1.32 ± 0.02 times faster than git -c core.commitGraph=true rev-list --topo-order --all (git = pks-commit-graph-inconsistency)
4.72 ± 0.05 times faster than git -c core.commitGraph=false rev-list --topo-order --all (git = pks-commit-graph-inconsistency)
4.75 ± 0.04 times faster than git -c core.commitGraph=false rev-list --topo-order --all (git = master)
We look at a ~30% regression in general, but in general we're still a
whole lot faster than without the commit graph. To counteract this, the
new check can be turned off with the `GIT_COMMIT_GRAPH_PARANOIA` envvar.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Our `lookup_commit_in_graph()` helper tries to look up commits from the
commit graph and, if it doesn't exist there, falls back to parsing it
from the object database instead. This is intended to speed up the
lookup of any such commit that exists in the database. There is an edge
case though where the commit exists in the graph, but not in the object
database. To avoid returning such stale commits the helper function thus
double checks that any such commit parsed from the graph also exists in
the object database. This makes the function safe to use even when
commit graphs aren't updated regularly.
We're about to introduce the same pattern into other parts of our code
base though, namely `repo_parse_commit_internal()`. Here the extra
sanity check is a bit of a tougher sell: `lookup_commit_in_graph()` was
a newly introduced helper, and as such there was no performance hit by
adding this sanity check. If we added `repo_parse_commit_internal()`
with that sanity check right from the beginning as well, this would
probably never have been an issue to begin with. But by retrofitting it
with this sanity check now we do add a performance regression to
preexisting code, and thus there is a desire to avoid this or at least
give an escape hatch.
In practice, there is no inherent reason why either of those functions
should have the sanity check whereas the other one does not: either both
of them are able to detect this issue or none of them should be. This
also means that the default of whether we do the check should likely be
the same for both. To err on the side of caution, we thus rather want to
make `repo_parse_commit_internal()` stricter than to loosen the checks
that we already have in `lookup_commit_in_graph()`.
The escape hatch is added in the form of a new GIT_COMMIT_GRAPH_PARANOIA
environment variable that mirrors GIT_REF_PARANOIA. If enabled, which is
the default, we will double check that commits looked up in the commit
graph via `lookup_commit_in_graph()` also exist in the object database.
This same check will also be added in `repo_parse_commit_internal()`.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As an attempt to come up with a useful mechanism to ensure that
certain messages are left untranslated [*], we earlier wrote
GIT_TEST_GETTEXT_POISON off as a failed experiment.
But the output from the test helper was easier to use while
debugging failed tests, compared to the same test writtein with the
plain-vanilla "grep". Especially when a test that expects a certain
string to appear in the output (e.g. "this test must fail with this
message") fails, "grep message output" would just silently fail and
in a &&-chained sequence of commands, it is hard to tell which step
failed. test_i18ngrep explicitly said "we wanted to see a line that
match this pattern but did not see a hit in this file".
What we have as test_i18ngrep in our tree still retains this verbose
output (even though we got rid of the "poison" support). Let's
rename it to test_grep (because it is no longer about i18n at all)
and then make test_i18ngrep a thin wrapper around it. Existing
callers of test_i18ngrep can be mechanically rewritten to instead
use test_grep over time, but it does not have to be done in this
commit.
[Footnote]
* The idea was that human-facing messages are often translated, but
there are messages that should never be translated. We use
"grep" only for the latter kind of messages, and then run tests
in "poison" mode that spew garbage for translatable messages. If
such a test run fails, it means these messages tested with "grep"
were marked for translation by mistake. test_i18ngrep was to be
used for other messages that are to be translated, and was to
always "succeed" when runing under the "poison" mode.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Information on how users are accessing hosted repositories can be
helpful to server operators. For example, being able to broadly
differentiate between fetches and initial clones; the use of shallow
repository features; or partial clone filters.
a29263c (fetch-pack: add tracing for negotiation rounds, 2022-08-02)
added some information on have counts to fetch-pack itself to help
diagnose negotiation; but from a git-upload-pack (server) perspective,
there's no means of accessing such information without using
GIT_TRACE_PACKET to examine the protocol packets.
Improve this by emitting a Trace2 JSON event from upload-pack with
summary information on the contents of a fetch request.
* haves, wants, and want-ref counts can help determine (broadly) between
fetches and clones, and the use of single-branch, etc.
* shallow clone depth, tip counts, and deepening options.
* any partial clone filter type.
Signed-off-by: Robert Coup <robert@coup.net.nz>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
