The "--recursive" flag for git-grep(1) allows users to grep for a string
across submodule boundaries. To make this work we add each submodule's
object sources to our own object database so that the objects can be
accessed directly.
The infrastructure for this depends on a global string list of submodule
paths. The caller is expected to call `add_submodule_odb_by_path()` for
each source and the object database will then eventually register all
submodule sources via `do_oid_object_info_extended()` in case it isn't
able to look up a specific object.
This reliance on global state is of course suboptimal with regards to
our libification efforts.
Refactor the logic so that the list of submodule sources is instead
tracked in the object database itself. This allows us to lose the
condition of `r == the_repository` before registering submodule sources
as we only ever add submodule sources to `the_repository` anyway. As
such, behaviour before and after this refactoring should always be the
same.
Rename the functions accordingly.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The functions `set_temporary_primary_odb()` and `restore_primary_odb()`
are responsible for managing a temporary primary source for the
database. Both of these functions implicitly rely on `the_repository`.
Refactor them to instead take an explicit object database parameter as
argument and adjust callers. Rename the functions accordingly.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There are a couple of iterator-style functions that execute a callback
for each instance of a given set, all of which currently depend on
`the_repository`. Refactor them to instead take an object database as
parameter so that we can get rid of this dependency.
Rename the functions accordingly.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The functions to manage alternates all depend on `the_repository`.
Refactor them to accept an object database as a parameter and adjust all
callers. The functions are renamed accordingly.
Note that right now the situation is still somewhat weird because we end
up using the object store path provided by the object store's repository
anyway. Consequently, we could have instead passed in a pointer to the
repository instead of passing in the pointer to the object store. This
will be addressed in subsequent commits though, where we will start to
use the path owned by the object store itself.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Get rid of our dependency on `the_repository` in `odb_mkstemp()` by
passing in the object database as a parameter and adjusting all callers.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Get rid of our dependency on `the_repository` in `assert_oid_type()` by
passing in the object database as a parameter and adjusting all callers.
Rename the function to `odb_assert_oid_type()`.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Get rid of our dependency on `the_repository` in `find_odb()` by passing
in the object database in which we want to search for the source and
adjusting all callers.
Rename the function to `odb_find_source()`.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In subsequent commits we'll get rid of our use of `the_repository` in
"odb.c" in favor of explicitly passing in a `struct object_database` or
a `struct odb_source`. In some cases though we'll need access to the
repository, for example to read a config value from it, but we don't
have a way to access the repository owning a specific object database.
Introduce parent pointers for `struct object_database` to its owning
repository as well as for `struct odb_source` to its owning object
database, which will allow us to adapt those use cases.
Note that this change requires us to pass through the object database to
`link_alt_odb_entry()` so that we can set up the parent pointers for any
source there. The callchain is adapted to pass through the object
database accordingly.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the preceding commits we have renamed the structures contained in
"object-store.h" to `struct object_database` and `struct odb_backend`.
As such, the code files "object-store.{c,h}" are confusingly named now.
Rename them to "odb.{c,h}" accordingly.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The `object_directory` structure is used as an access point for a single
object directory like ".git/objects". While the structure isn't yet
fully self-contained, the intent is for it to eventually contain all
information required to access objects in one specific location.
While the name "object directory" is a good fit for now, this will
change over time as we continue with the agenda to make pluggable object
databases a thing. Eventually, objects may not be accessed via any kind
of directory at all anymore, but they could instead be backed by any
kind of durable storage mechanism. While it seems quite far-fetched for
now, it is thinkable that eventually this might even be some form of a
database, for example.
As such, the current name of this structure will become worse over time
as we evolve into the direction of pluggable ODBs. Immediate next steps
will start to carve out proper self-contained object directories, which
requires us to pass in these object directories as parameters. Based on
our modern naming schema this means that those functions should then be
named after their subsystem, which means that we would start to bake the
current name into the codebase more and more.
Let's preempt this by renaming the structure. There have been a couple
alternatives that were discussed:
- `odb_backend` was discarded because it led to the association that
one object database has a single backend, but the model is that one
alternate has one backend. Furthermore, "backend" is more about the
actual backing implementation and less about the high-level concept.
- `odb_alternate` was discarded because it is a bit of a stretch to
also call the main object directory an "alternate".
Instead, pick `odb_source` as the new name. It makes it sufficiently
clear that there can be multiple sources and does not cause confusion
when mixed with the already-existing "alternate" terminology.
In the future, this change allows us to easily introduce for example a
`odb_files_source` and other format-specific implementations.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The `raw_object_store` structure is the central entry point for reading
and writing objects in a repository. The main purpose of this structure
is to manage object directories and provide an interface to access and
write objects in those object directories.
Right now, many of the functions associated with the raw object store
implicitly rely on `the_repository` to get access to its `objects`
pointer, which is the `raw_object_store`. As we want to generally get
rid of using `the_repository` across our codebase we will have to
convert this implicit dependency on this global variable into an
explicit parameter.
This conversion can be done by simply passing in an explicit pointer to
a repository and then using its `->objects` pointer. But there is a
second effort underway, which is to make the object subsystem more
selfcontained so that we can eventually have pluggable object backends.
As such, passing in a repository wouldn't make a ton of sense, and the
goal is to convert the object store interfaces such that we always pass
in a reference to the `raw_object_store` instead.
This will expose the `raw_object_store` type to a lot more callers
though, which surfaces that this type is named somewhat awkwardly. The
"raw_" prefix makes readers wonder whether there is a non-raw variant of
the object store, but there isn't. Furthermore, we nowadays want to name
functions in a way that they can be clearly attributed to a specific
subsystem, but calling them e.g. `raw_object_store_has_object()` is just
too unwieldy, even when dropping the "raw_" prefix.
Instead, rename the structure to `object_database`. This term is already
used a lot throughout our codebase, and it cannot easily be mistaken for
"object directories", either. Furthermore, its acronym ODB is already
well-known and works well as part of a function's name, like for example
`odb_has_object()`.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When a stale .midx file refers to .pack files that no longer exist,
we ended up checking for these non-existent files repeatedly, which
has been optimized by memoizing the non-existence.
* ps/midx-negative-packfile-cache:
midx: stop repeatedly looking up nonexistent packfiles
packfile: explain ordering of how we look up auxiliary pack files
"git notes --help" documentation updates.
* kh/notes-doc-fixes:
doc: notes: use stuck form throughout
doc: notes: treat --stdin equally between copy/remove
doc: notes: point out copy --stdin use with argv
doc: notes: clearly state that --stripspace is the default
doc: notes: remove stripspace discussion from other options
doc: notes: rework --[no-]stripspace
doc: notes: split out options with negated forms
doc: config: mention core.commentChar on commit.cleanup
doc: stripspace: mention where the default comes from
"git apply --index/--cached" when applying a deletion patch in
reverse failed to give the mode bits of the path "removed" by the
patch to the file it creates, which has been corrected.
* mm/apply-reverse-mode-of-deleted-path:
apply: set file mode when --reverse creates a deleted file
t4129: test that git apply warns for unexpected mode changes
Recent versions of Perl started warning against "! A =~ /pattern/"
which does not negate the result of the matching. As it turns out
that the problematic function is not even called, it was removed.
* op/cvsserver-perl-warning:
cvsserver: remove unused escapeRefName function
Avoid adding directory path to a sparse-index tree entries to the
name-hash, since they would bloat the hashtable without anybody
querying for them. This was done already for a single threaded
part of the code, but now the multi-threaded code also does the
same.
* am/sparse-index-name-hash-fix:
name-hash: don't add sparse directories in threaded lazy init
Integer overflow fix around code paths for "git multi-pack-index repack"..
* pw/midx-repack-overflow-fix:
midx docs: clarify tie breaking
midx: avoid negative array index
midx repack: avoid potential integer overflow on 64 bit systems
midx repack: avoid integer overflow on 32 bit systems
Since f93b2a0424 (reftable/basics: introduce `REFTABLE_UNUSED`
annotation, 2025-02-18), the reftable library was migrated to
use an internal version of `UNUSED`, which unconditionally sets
a GNU __attribute__ to avoid warnings function parameters that
are not being used.
Make the definition conditional to prevent breaking the build
with non GNU compilers.
Reported-by: "Randall S. Becker" <rsbecker@nexbridge.com>
Signed-off-by: Carlo Marcelo Arenas Belón <carenas@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* 'master' of https://github.com/j6t/git-gui:
git-gui: wire up support for the Meson build system
git-gui: stop including GIT-VERSION-FILE file
git-gui: extract script to generate macOS app
git-gui: extract script to generate macOS wrapper
git-gui: extract script to generate "tclIndex"
git-gui: extract script to generate "git-gui"
git-gui: drop no-op GITGUI_SCRIPT replacement
git-gui: make output of GIT-VERSION-GEN source'able
git-gui: prepare GIT-VERSION-GEN for out-of-tree builds
git-gui: replace GIT-GUI-VARS with GIT-GUI-BUILD-OPTIONS
* 'master' of https://github.com/j6t/gitk:
gitk: do not hard-code color of search results in commit list
gitk: place file name arguments after options in msgfmt call
gitk: Legacy widgets doesn't have combobox
* 'pks-meson-support' of github.com:pks-t/git-gui:
git-gui: wire up support for the Meson build system
git-gui: stop including GIT-VERSION-FILE file
git-gui: extract script to generate macOS app
git-gui: extract script to generate macOS wrapper
git-gui: extract script to generate "tclIndex"
git-gui: extract script to generate "git-gui"
git-gui: drop no-op GITGUI_SCRIPT replacement
git-gui: make output of GIT-VERSION-GEN source'able
git-gui: prepare GIT-VERSION-GEN for out-of-tree builds
git-gui: replace GIT-GUI-VARS with GIT-GUI-BUILD-OPTIONS
Signed-off-by: Johannes Sixt <j6t@kdbg.org>
"git receive-pack" optionally learns not to care about connectivity
check, which can be useful when the repository arranges to ensure
connectivity by some other means.
* jt/receive-pack-skip-connectivity-check:
builtin/receive-pack: add option to skip connectivity check
t5410: test receive-pack connectivity check
Remove the leftover hints to the test framework to mark tests that
do not pass the leak checker tests, as they should no longer be
needed.
* kn/passing-leak-tests:
t: remove unexpected SANITIZE_LEAK variables
The multi-pack index acts as a cache across a set of packfiles so that
we can quickly look up which of those packfiles contains a given object.
As such, the multi-pack index naturally needs to be updated every time
one of the packfiles goes away, or otherwise the multi-pack index has
grown stale.
A stale multi-pack index should be handled gracefully by Git though, and
in fact it is: if the indexed pack cannot be found we simply ignore it
and eventually we fall back to doing the object lookup by just iterating
through all packs, even if those aren't indexed.
But while this fallback works, it has one significant downside: we don't
cache the fact that a pack has vanished. This leads to us repeatedly
trying to look up the same pack only to realize that it (still) doesn't
exist.
This issue can be easily demonstrated by creating a repository with a
stale multi-pack index and a couple of objects. We do so by creating a
repository with two packfiles, both of which are indexed by the
multi-pack index, and then repack those two packfiles. Note that we have
to move the multi-pack-index before doing the final repack, as Git knows
to delete it otherwise.
$ git init repo
$ cd repo/
$ git config set maintenance.auto false
$ for i in $(seq 1000); do printf "%d-original" $i >file-$i; done
$ git add .
$ git commit -moriginal
$ git repack -dl
$ for i in $(seq 1000); do printf "%d-modified" $i >file-$i; done
$ git commit -a -mmodified
$ git repack -dl
$ git multi-pack-index write
$ mv .git/objects/pack/multi-pack-index .
$ git repack -Adl
$ mv multi-pack-index .git/objects/pack/
Commands that cause a lot of objects lookups will now repeatedly invoke
`add_packed_git()`, which leads to three failed access(3p) calls as well
as one failed stat(3p) call. The following strace for example is done
for `git log --patch` in the above repository:
% time seconds usecs/call calls errors syscall
------ ----------- ----------- --------- --------- ----------------
74.67 0.024693 1 18038 18031 access
25.33 0.008378 1 6045 6017 newfstatat
------ ----------- ----------- --------- --------- ----------------
100.00 0.033071 1 24083 24048 total
Fix the issue by introducing a negative lookup cache for indexed packs.
This cache works by simply storing an invalid pointer for a missing pack
when `prepare_midx_pack()` fails to look up the pack. Most users of the
`packs` array don't need to be adjusted, either, as they all know to
call `prepare_midx_pack()` before accessing the array.
With this change in place we can now see a significantly reduced number
of syscalls:
% time seconds usecs/call calls errors syscall
------ ----------- ----------- --------- --------- ----------------
73.58 0.000323 5 60 28 newfstatat
26.42 0.000116 5 23 16 access
------ ----------- ----------- --------- --------- ----------------
100.00 0.000439 5 83 44 total
Furthermore, this change also results in a speedup:
Benchmark 1: git log --patch (revision = HEAD~)
Time (mean ± σ): 50.4 ms ± 2.5 ms [User: 22.0 ms, System: 24.4 ms]
Range (min … max): 45.4 ms … 54.9 ms 53 runs
Benchmark 2: git log --patch (revision = HEAD)
Time (mean ± σ): 12.7 ms ± 0.4 ms [User: 11.1 ms, System: 1.6 ms]
Range (min … max): 12.4 ms … 15.0 ms 191 runs
Summary
git log --patch (revision = HEAD) ran
3.96 ± 0.22 times faster than git log --patch (revision = HEAD~)
In the end, it should in theory never be necessary to have this negative
lookup cache given that we know to update the multi-pack index together
with repacks. But as the change is quite contained and as the speedup
can be significant as demonstrated above, it does feel sensible to have
the negative lookup cache regardless.
Based-on-patch-by: Jeff King <peff@peff.net>
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When adding a packfile to an object database we perform four syscalls:
- Three calls to access(3p) are done to check for auxiliary data
structures.
- One call to stat(3p) is done to check for the ".pack" itself.
One curious bit is that we perform the access(3p) calls before checking
for the packfile itself, but if the packfile doesn't exist we discard
all results. The access(3p) calls are thus essentially wasted, so one
may be triggered to reorder those calls so that we can short-circuit the
other syscalls in case the packfile does not exist.
The order in which we look up files is quite important though to help
avoid races:
- When installing a packfile we move auxiliary data structures into
place before we install the ".idx" file.
- When deleting a packfile we first delete the ".idx" and ".pack"
files before deleting auxiliary data structures.
As such, to avoid any races with concurrently created or deleted packs
we need to make sure that we _first_ read auxiliary data structures
before we read the corresponding ".idx" or ".pack" file. Otherwise it
may easily happen that we return a populated but misclassified pack.
Add a comment to `add_packed_git()` to make future readers aware of this
ordering requirement.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
gitcli(7) recommends the *stuck form*. `--ref` is the only one which
does not use it.
Signed-off-by: Kristoffer Haugsbakk <code@khaugsbakk.name>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
46538012d9 (notes remove: --stdin reads from the standard input,
2011-05-18) added `--stdin` for the `remove` subcommand, documenting it
in the “Options” section. But `copy --stdin` was added before that, in
160baa0d9c (notes: implement 'git notes copy --stdin', 2010-03-12).
Treat this option equally between the two subcommands:
• remove: mention `--stdin` on the subcommand as well, like for `copy`
• copy: mention it as well under the option documentation
Signed-off-by: Kristoffer Haugsbakk <code@khaugsbakk.name>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Unlike `remove --stdin`, this option cannot be combined with object
names given via the command line.
Signed-off-by: Kristoffer Haugsbakk <code@khaugsbakk.name>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Cleaning up whitespace in metadata is typical porcelain behavior and
this default does not need to be pointed out.[1] Only speak up when
the default `--stripspace` is not used.
Also remove all misleading mentions of comment lines in the process;
see the previous commit.
Also remove the period that trails the parenthetical here.
† 1: See `-F` in git-commit(1) which has nothing to say about whitespace
cleanup. The cleanup discussion is on `--cleanup`.
Signed-off-by: Kristoffer Haugsbakk <code@khaugsbakk.name>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Document this option by copying the bullet list from git-stripspace(1).
A bullet list is cleaner when there are this many points to consider.
We also get a more standardized description of the multiple-blank-lines
behavior. Compare the repeating (git-notes(1)):
empty lines other than a single line between paragraphs
With (git-stripspace(1)):
multiple consecutive empty lines
And:
leading [...] whitespace
With:
empty lines from the beginning
Leading whitespace in the form of spaces (indentation) are not removed.
However, empty lines at the start of the message are removed.
Note that we drop the mentions of comment line handling because they are
wrong; this option does not control how lines which can be recognized as
comment lines are handled. Only interactivity controls that:
• Comment lines are stripped after editing interactively
• Lines which could be recognized as comment lines are left alone when
the message is given non-interactively
So it is misleading to document the comment line behavior on
this option.
Further, the text is wrong:
Lines starting with `#` will be stripped out in non-editor cases
like `-m`, [...]
Comment lines are still indirectly discussed on other options. We will
deal with them in the next commit.
Signed-off-by: Kristoffer Haugsbakk <code@khaugsbakk.name>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Mention it in parentheses since we are in a configuration context.
Refer to the default as such, not as “the” character.
Also don’t mention `#` again; just say “comment character”.
Signed-off-by: Kristoffer Haugsbakk <code@khaugsbakk.name>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Also quote `#` in line with the modern formatting convention.
Signed-off-by: Kristoffer Haugsbakk <code@khaugsbakk.name>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Prefix '#' to the commit title in the "rebase -i" todo file, just
like a merge commit being replayed.
* en/sequencer-comment-messages:
sequencer: make it clearer that commit descriptions are just comments
Assorted fixes for issues found with CodeQL.
* js/misc-fixes:
sequencer: stop pretending that an assignment is a condition
bundle-uri: avoid using undefined output of `sscanf()`
commit-graph: avoid using stale stack addresses
trace2: avoid "futile conditional"
Avoid redundant conditions
fetch: avoid unnecessary work when there is no current branch
has_dir_name(): make code more obvious
upload-pack: rename `enum` to reflect the operation
commit-graph: avoid malloc'ing a local variable
fetch: carefully clear local variable's address after use
commit: simplify code
The code path to access the "packed-refs" file while "fsck" is
taught to mmap the file, instead of reading the whole file in the
memory.
* sj/use-mmap-to-check-packed-refs:
packed-backend: mmap large "packed-refs" file during fsck
packed-backend: extract snapshot allocation in `load_contents`
packed-backend: fsck should warn when "packed-refs" file is empty
"git apply" and "git add -i/-p" code paths no longer unnecessarily
expand sparse-index while working.
* ds/sparse-apply-add-p:
p2000: add performance test for patch-mode commands
reset: integrate sparse index with --patch
git add: make -p/-i aware of sparse index
apply: integrate with the sparse index
Updates to meson-based build procedure.
* rj/build-tweaks-part2:
configure.ac: upgrade to a compilation check for sysinfo
meson.build: correct setting of GIT_EXEC_PATH
meson: correct path to system config/attribute files
meson: correct install location of YAML.pm
meson.build: quote the GITWEBDIR build configuration
"git merge-tree" learned an option to see if it resolves cleanly
without actually creating a result.
* en/merge-tree-check:
merge-tree: add a new --quiet flag
merge-ort: add a new mergeability_only option
Support to create a loose object file with unknown object type has
been dropped.
* jk/no-funny-object-types:
object-file: drop support for writing objects with unknown types
hash-object: handle --literally with OPT_NEGBIT
hash-object: merge HASH_* and INDEX_* flags
hash-object: stop allowing unknown types
t: add lib-loose.sh
t/helper: add zlib test-tool
oid_object_info(): drop type_name strbuf
fsck: stop using object_info->type_name strbuf
oid_object_info_convert(): stop using string for object type
cat-file: use type enum instead of buffer for -t option
object-file: drop OBJECT_INFO_ALLOW_UNKNOWN_TYPE flag
cat-file: make --allow-unknown-type a noop
object-file.h: fix typo in variable declaration
