Adapting the implementation of ll_find_deltas(), create a threaded
version of the --path-walk compression step in 'git pack-objects'.
This involves adding a 'regions' member to the thread_params struct,
allowing each thread to own a section of paths. We can simplify the way
jobs are split because there is no value in extending the batch based on
name-hash the way sections of the object entry array are attempted to be
grouped. We re-use the 'list_size' and 'remaining' items for the purpose
of borrowing work in progress from other "victim" threads when a thread
has finished its batch of work more quickly.
Using the Git repository as a test repo, the p5313 performance test
shows that the resulting size of the repo is the same, but the threaded
implementation gives gains of varying degrees depending on the number of
objects being packed. (This was tested on a 16-core machine.)
Test HEAD~1 HEAD
-------------------------------------------------------------
5313.6: thin pack with --path-walk 0.01 0.01 +0.0%
5313.7: thin pack size with --path-walk 475 475 +0.0%
5313.12: big pack with --path-walk 1.99 1.87 -6.0%
5313.13: big pack size with --path-walk 14.4M 14.3M -0.4%
5313.18: repack with --path-walk 98.14 41.46 -57.8%
5313.19: repack size with --path-walk 197.2M 197.3M +0.0%
Signed-off-by: Derrick Stolee <stolee@gmail.com>
Previously, the --path-walk option to 'git pack-objects' would compute
deltas inline with the path-walk logic. This would make the progress
indicator look like it is taking a long time to enumerate objects, and
then very quickly computed deltas.
Instead of computing deltas on each region of objects organized by tree,
store a list of regions corresponding to these groups. These can later
be pulled from the list for delta compression before doing the "global"
delta search.
This presents a new progress indicator that can be used in tests to
verify that this stage is happening.
The current implementation is not integrated with threads, but could be
done in a future update.
Since we do not attempt to sort objects by size until after exploring
all trees, we can remove the previous change to t5530 due to a different
error message appearing first.
Signed-off-by: Derrick Stolee <stolee@gmail.com>
Users may want to enable the --path-walk option for 'git pack-objects' by
default, especially underneath commands like 'git push' or 'git repack'.
This should be limited to client repositories, since the --path-walk option
disables bitmap walks, so would be bad to include in Git servers when
serving fetches and clones. There is potential that it may be helpful to
consider when repacking the repository, to take advantage of improved deltas
across historical versions of the same files.
Much like how "pack.useSparse" was introduced and included in
"feature.experimental" before being enabled by default, use the repository
settings infrastructure to make the new "pack.usePathWalk" config enabled by
"feature.experimental" and "feature.manyFiles".
Signed-off-by: Derrick Stolee <stolee@gmail.com>
There are many tests that validate whether 'git pack-objects' works as
expected. Instead of duplicating these tests, add a new test environment
variable, GIT_TEST_PACK_PATH_WALK, that implies --path-walk by default
when specified.
This was useful in testing the implementation of the --path-walk
implementation, especially in conjunction with test such as:
- t0411-clone-from-partial.sh : One test fetches from a repo that does
not have the boundary objects. This causes the path-based walk to
fail. Disable the variable for this test.
- t5306-pack-nobase.sh : Similar to t0411, one test fetches from a repo
without a boundary object.
- t5310-pack-bitmaps.sh : One test compares the case when packing with
bitmaps to the case when packing without them. Since we disable the
test variable when writing bitmaps, this causes a difference in the
object list (the --path-walk option adds an extra object). Specify
--no-path-walk in both processes for the comparison. Another test
checks for a specific delta base, but when computing dynamically
without using bitmaps, the base object it too small to be considered
in the delta calculations so no base is used.
- t5316-pack-delta-depth.sh : This script cares about certain delta
choices and their chain lengths. The --path-walk option changes how
these chains are selected, and thus changes the results of this test.
- t5322-pack-objects-sparse.sh : This demonstrates the effectiveness of
the --sparse option and how it combines with --path-walk.
- t5332-multi-pack-reuse.sh : This test verifies that the preferred
pack is used for delta reuse when possible. The --path-walk option is
not currently aware of the preferred pack at all, so finds a
different delta base.
- t7406-submodule-update.sh : When using the variable, the --depth
option collides with the --path-walk feature, resulting in a warning
message. Disable the variable so this warning does not appear.
I want to call out one specific test change that is only temporary:
- t5530-upload-pack-error.sh : One test cares specifically about an
"unable to read" error message. Since the current implementation
performs delta calculations within the path-walk API callback, a
different "unable to get size" error message appears. When this
is changed in a future refactoring, this test change can be reverted.
Signed-off-by: Derrick Stolee <stolee@gmail.com>
In order to more easily compute delta bases among objects that appear at the
exact same path, add a --path-walk option to 'git pack-objects'.
This option will use the path-walk API instead of the object walk given by
the revision machinery. Since objects will be provided in batches
representing a common path, those objects can be tested for delta bases
immediately instead of waiting for a sort of the full object list by
name-hash. This has multiple benefits, including avoiding collisions by
name-hash.
The objects marked as UNINTERESTING are included in these batches, so we
are guaranteeing some locality to find good delta bases.
After the individual passes are done on a per-path basis, the default
name-hash is used to find other opportunistic delta bases that did not
match exactly by the full path name.
RFC TODO: It is important to note that this option is inherently
incompatible with using a bitmap index. This walk probably also does not
work with other advanced features, such as delta islands.
Getting ahead of myself, this option compares well with --full-name-hash
when the packfile is large enough, but also performs at least as well as
the default in all cases that I've seen.
RFC TODO: this should probably be recording the batch locations to another
list so they could be processed in a second phase using threads.
RFC TODO: list some examples of how this outperforms previous pack-objects
strategies. (This is coming in later commits that include performance
test changes.)
Signed-off-by: Derrick Stolee <stolee@gmail.com>
This option is still under discussion on the Git mailing list.
We still would like to have some real-world data, and the best way to
get it is to get a Git for Windows release into users' hands so that
they can test it.
Nevertheless, without the official blessing of the Git maintainer, this
optionis experimental, and we need to be clear about that.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Add a new environment variable to opt-in to the --full-name-hash option
in 'git pack-objects'. This allows for extra testing of the feature
without repeating all of the test scenarios.
But this option isn't free. There are a few tests that change behavior
with the variable enabled.
First, there are a few tests that are very sensitive to certain delta
bases being picked. These are both involving the generation of thin
bundles and then counting their objects via 'git index-pack --fix-thin'
which pulls the delta base into the new packfile. For these tests,
disable the option as a decent long-term option.
Second, there are two tests in t5616-partial-clone.sh that I believe are
actually broken scenarios. While the client is set up to clone the
'promisor-server' repo via a treeless partial clone filter (tree:0),
that filter does not translate to the 'server' repo. Thus, fetching from
these repos causes the server to think that the client has all reachable
trees and blobs from the commits advertised as 'haves'. This leads the
server to providing a thin pack assuming those objects as delta bases.
Changing the name-hash algorithm presents new delta bases and thus
breaks the expectations of these tests. An alternative could be to set
up 'server' as a promisor server with the correct filter enabled. This
may also point out more issues with partial clone being set up as a
remote-based filtering mechanism and not a repository-wide setting. For
now, do the minimal change to make the test work by disabling the test
variable.
Signed-off-by: Derrick Stolee <stolee@gmail.com>
The pack_name_hash() method has not been materially changed since it was
introduced in ce0bd64299a (pack-objects: improve path grouping
heuristics., 2006-06-05). The intention here is to group objects by path
name, but also attempt to group similar file types together by making
the most-significant digits of the hash be focused on the final
characters.
Here's the crux of the implementation:
/*
* This effectively just creates a sortable number from the
* last sixteen non-whitespace characters. Last characters
* count "most", so things that end in ".c" sort together.
*/
while ((c = *name++) != 0) {
if (isspace(c))
continue;
hash = (hash >> 2) + (c << 24);
}
As the comment mentions, this only cares about the last sixteen
non-whitespace characters. This cause some filenames to collide more
than others. Here are some examples that I've seen while investigating
repositories that are growing more than they should be:
* "/CHANGELOG.json" is 15 characters, and is created by the beachball
[1] tool. Only the final character of the parent directory can
differntiate different versions of this file, but also only the two
most-significant digits. If that character is a letter, then this is
always a collision. Similar issues occur with the similar
"/CHANGELOG.md" path, though there is more opportunity for
differences in the parent directory.
* Localization files frequently have common filenames but differentiate
via parent directories. In C#, the name "/strings.resx.lcl" is used
for these localization files and they will all collide in name-hash.
[1] https://github.com/microsoft/beachball
I've come across many other examples where some internal tool uses a
common name across multiple directories and is causing Git to repack
poorly due to name-hash collisions.
It is clear that the existing name-hash algorithm is optimized for
repositories with short path names, but also is optimized for packing a
single snapshot of a repository, not a repository with many versions of
the same file. In my testing, this has proven out where the name-hash
algorithm does a good job of finding peer files as delta bases when
unable to use a historical version of that exact file.
However, for repositories that have many versions of most files and
directories, it is more important that the objects that appear at the
same path are grouped together.
Create a new pack_full_name_hash() method and a new --full-name-hash
option for 'git pack-objects' to call that method instead. Add a simple
pass-through for 'git repack --full-name-hash' for additional testing in
the context of a full repack, where I expect this will be most
effective.
The hash algorithm is as simple as possible to be reasonably effective:
for each character of the path string, add a multiple of that character
and a large prime number (chosen arbitrarily, but intended to be large
relative to the size of a uint32_t). Then, shift the current hash value
to the right by 5, with overlap. The addition and shift parameters are
standard mechanisms for creating hard-to-predict behaviors in the bits
of the resulting hash.
This is not meant to be cryptographic at all, but uniformly distributed
across the possible hash values. This creates a hash that appears
pseudorandom. There is no ability to consider similar file types as
being close to each other.
In a later change, a test-tool will be added so the effectiveness of
this hash can be demonstrated directly.
For now, let's consider how effective this mechanism is when repacking a
repository with and without the --full-name-hash option. Specifically,
let's use 'git repack -adf [--full-name-hash]' as our test.
On the Git repository, we do not expect much difference. All path names
are short. This is backed by our results:
| Stage | Pack Size | Repack Time |
|-----------------------|-----------|-------------|
| After clone | 260 MB | N/A |
| Standard Repack | 127MB | 106s |
| With --full-name-hash | 126 MB | 99s |
This example demonstrates how there is some natural overhead coming from
the cloned copy because the server is hosting many forks and has not
optimized for exactly this set of reachable objects. But the full repack
has similar characteristics with and without --full-name-hash.
However, we can test this in a repository that uses one of the
problematic naming conventions above. The fluentui [2] repo uses
beachball to generate CHANGELOG.json and CHANGELOG.md files, and these
files have very poor delta characteristics when comparing against
versions across parent directories.
| Stage | Pack Size | Repack Time |
|-----------------------|-----------|-------------|
| After clone | 694 MB | N/A |
| Standard Repack | 438 MB | 728s |
| With --full-name-hash | 168 MB | 142s |
[2] https://github.com/microsoft/fluentui
In this example, we see significant gains in the compressed packfile
size as well as the time taken to compute the packfile.
Using a collection of repositories that use the beachball tool, I was
able to make similar comparisions with dramatic results. While the
fluentui repo is public, the others are private so cannot be shared for
reproduction. The results are so significant that I find it important to
share here:
| Repo | Standard Repack | With --full-name-hash |
|----------|-----------------|-----------------------|
| fluentui | 438 MB | 168 MB |
| Repo B | 6,255 MB | 829 MB |
| Repo C | 37,737 MB | 7,125 MB |
| Repo D | 130,049 MB | 6,190 MB |
Future changes could include making --full-name-hash implied by a config
value or even implied by default during a full repack.
Signed-off-by: Derrick Stolee <stolee@gmail.com>
The convention to calling into built-in command implementation has
been updated to pass the repository, if known, together with the
prefix value.
* jc/pass-repo-to-builtins:
add: pass in repo variable instead of global the_repository
builtin: remove USE_THE_REPOSITORY for those without the_repository
builtin: remove USE_THE_REPOSITORY_VARIABLE from builtin.h
builtin: add a repository parameter for builtin functions
Code clean-up.
* ps/environ-wo-the-repository: (21 commits)
environment: stop storing "core.notesRef" globally
environment: stop storing "core.warnAmbiguousRefs" globally
environment: stop storing "core.preferSymlinkRefs" globally
environment: stop storing "core.logAllRefUpdates" globally
refs: stop modifying global `log_all_ref_updates` variable
branch: stop modifying `log_all_ref_updates` variable
repo-settings: track defaults close to `struct repo_settings`
repo-settings: split out declarations into a standalone header
environment: guard state depending on a repository
environment: reorder header to split out `the_repository`-free section
environment: move `set_git_dir()` and related into setup layer
environment: make `get_git_namespace()` self-contained
environment: move object database functions into object layer
config: make dependency on repo in `read_early_config()` explicit
config: document `read_early_config()` and `read_very_early_config()`
environment: make `get_git_work_tree()` accept a repository
environment: make `get_graft_file()` accept a repository
environment: make `get_index_file()` accept a repository
environment: make `get_object_directory()` accept a repository
environment: make `get_git_common_dir()` accept a repository
...
Instead of including USE_THE_REPOSITORY_VARIABLE by default on every
builtin, remove it from builtin.h and add it to all the builtins that
include builtin.h (by definition, that means all builtins/*.c).
Also, remove the include statement for repository.h since it gets
brought in through builtin.h.
The next step will be to migrate each builtin
from having to use the_repository.
Signed-off-by: John Cai <johncai86@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In order to reduce the usage of the global the_repository, add a
parameter to builtin functions that will get passed a repository
variable.
This commit uses UNUSED on most of the builtin functions, as subsequent
commits will modify the actual builtins to pass the repository parameter
down.
Signed-off-by: John Cai <johncai86@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The `get_object_directory()` function retrieves the path to the object
directory for `the_repository`. Make it accept a `struct repository`
such that it can work on arbitrary repositories and make it part of the
repository subsystem. This reduces our reliance on `the_repository` and
clarifies scope.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The `--keep-pack` option of git-pack-objects(1) populates the arguments
into a string list. And while the list is marked as `NODUP` and thus
won't duplicate the strings, the list entries themselves still need to
be free'd. We don't though, causing a leak.
Plug it.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The function `write_reused_pack_one()` defines an header to store the
OFS_DELTA header, but uses the constant "10" instead of
"MAX_PACK_OBJECT_HEADER" (as is done elsewhere in the same patch, circa
bb514de356c (pack-objects: improve partial packfile reuse, 2019-12-18)).
Declare the `ofs_header` field to be sized according to
`MAX_PACK_OBJECT_HEADER` (which is 10, as defined in "pack.h") instead
of the constant 10.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When reusing chunks verbatim from an existing source pack, the function
write_reused_pack() first attempts to reuse whole words (via the
function `write_reused_pack_verbatim()`), and then individual bits (via
`write_reused_pack_one()`).
In the non-MIDX case, all of this code works fine. Likewise, in the MIDX
case, processing bits individually from the first (preferred) pack works
fine. However, processing subsequent packs in the MIDX case is broken
when there are duplicate objects among the set of MIDX'd packs.
This is because we treat the individual bit positions as valid pack
positions within the source pack(s), which does not account for gaps in
the source pack, like we see when the MIDX must break ties between
duplicate objects which appear in multiple packs.
The broken code looks like:
for (; i < reuse_packfile_bitmap->word_alloc; i++) {
for (offset = 0; offset < BITS_IN_EWORD, offset++) {
/* ... */
write_reused_pack_one(reuse_packfile->p,
pos + offset - reuse_packfile->bitmap_pos,
f, pack_start, &w_curs);
}
}
, where the second argument is incorrect and does not account for gaps.
Instead, make sure that we translate bit positions in the MIDX's
pseudo-pack order to pack positions in the respective source packs by:
- Translating the bit position (pseudo-pack order) to a MIDX position
(lexical order).
- Use the MIDX position to obtain the offset at which the given object
occurs in the source pack.
- Then translate that offset back into a pack relative position within
the source pack by calling offset_to_pack_pos().
After doing this, then we can safely use the result as a pack position.
Note that when doing single-pack reuse, as well as reusing objects from
the MIDX's preferred pack, such translation is not necessary, since
either ties are broken in favor of the preferred pack, or there are no
ties to break at all (in the case of non-MIDX bitmaps).
Failing to do this can result in strange failure modes. One example that
can occur when misinterpreting bits in the above fashion is that Git
thinks it's supposed to send a delta that the caller does not want.
Under this (incorrect) assumption, we try to look up the delta's base
(so that we can patch any OFS_DELTAs if necessary). We do this using
find_reused_offset().
But if we try and call that function for an offset belonging to an
object we did not send, we'll get back garbage. This can result in us
computing a negative fixup value, which results in memory corruption
when trying to write the (patched) OFS_DELTA header.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We created a useless pseudo-merge reachability bitmap that is about
0 commits, and attempted to include commits that are not in packs,
which made no sense. These bugs have been corrected.
* tb/pseudo-merge-bitmap-fixes:
pseudo-merge.c: ensure pseudo-merge groups are closed
pseudo-merge.c: do not generate empty pseudo-merge commits
t/t5333-pseudo-merge-bitmaps.sh: demonstrate empty pseudo-merge groups
pack-bitmap-write.c: select pseudo-merges even for small bitmaps
pack-bitmap: drop redundant args from `bitmap_writer_finish()`
pack-bitmap: drop redundant args from `bitmap_writer_build()`
pack-bitmap: drop redundant args from `bitmap_writer_build_type_index()`
pack-bitmap: initialize `bitmap_writer_init()` with packing_data
In a similar fashion as the previous commit, drop a redundant argument
from the `bitmap_writer_finish()` function.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In a similar fashion as the previous commit, drop a redundant argument
from the `bitmap_writer_build()` function.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The previous commit ensures that the bitmap_writer's "to_pack" field is
initialized early on, so the "to_pack" and "index_nr" arguments to
`bitmap_writer_build_type_index()` are redundant.
Drop them and adjust the callers accordingly.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In order to determine its object order, the pack-bitmap machinery keeps
a 'struct packing_data' corresponding to the pack or pseudo-pack (when
writing a MIDX bitmap) being written.
The to_pack field is provided to the bitmap machinery by callers of
bitmap_writer_build() and assigned to the bitmap_writer struct at that
point.
But a subsequent commit will want to have access to that data earlier on
during commit selection. Prepare for that by adding a 'to_pack' argument
to 'bitmap_writer_init()', and initializing the field during that
function.
Subsequent commits will clean up other functions which take
now-redundant arguments (like nr_objects, which is equivalent to
pdata->objects_nr, or pdata itself).
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a parameter to each_ref_fn so that callers to the ref APIs
that use this function as a callback can have acess to the
unresolved value of a symbolic ref.
Signed-off-by: John Cai <johncai86@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A CPP macro USE_THE_REPOSITORY_VARIABLE is introduced to help
transition the codebase to rely less on the availability of the
singleton the_repository instance.
* ps/use-the-repository:
hex: guard declarations with `USE_THE_REPOSITORY_VARIABLE`
t/helper: remove dependency on `the_repository` in "proc-receive"
t/helper: fix segfault in "oid-array" command without repository
t/helper: use correct object hash in partial-clone helper
compat/fsmonitor: fix socket path in networked SHA256 repos
replace-object: use hash algorithm from passed-in repository
protocol-caps: use hash algorithm from passed-in repository
oidset: pass hash algorithm when parsing file
http-fetch: don't crash when parsing packfile without a repo
hash-ll: merge with "hash.h"
refs: avoid include cycle with "repository.h"
global: introduce `USE_THE_REPOSITORY_VARIABLE` macro
hash: require hash algorithm in `empty_tree_oid_hex()`
hash: require hash algorithm in `is_empty_{blob,tree}_oid()`
hash: make `is_null_oid()` independent of `the_repository`
hash: convert `oidcmp()` and `oideq()` to compare whole hash
global: ensure that object IDs are always padded
hash: require hash algorithm in `oidread()` and `oidclr()`
hash: require hash algorithm in `hasheq()`, `hashcmp()` and `hashclr()`
hash: drop (mostly) unused `is_empty_{blob,tree}_sha1()` functions
The pseudo-merge reachability bitmap to help more efficient storage
of the reachability bitmap in a repository with too many refs has
been added.
* tb/pseudo-merge-reachability-bitmap: (26 commits)
pack-bitmap.c: ensure pseudo-merge offset reads are bounded
Documentation/technical/bitmap-format.txt: add missing position table
t/perf: implement performance tests for pseudo-merge bitmaps
pseudo-merge: implement support for finding existing merges
ewah: `bitmap_equals_ewah()`
pack-bitmap: extra trace2 information
pack-bitmap.c: use pseudo-merges during traversal
t/test-lib-functions.sh: support `--notick` in `test_commit_bulk()`
pack-bitmap: implement test helpers for pseudo-merge
ewah: implement `ewah_bitmap_popcount()`
pseudo-merge: implement support for reading pseudo-merge commits
pack-bitmap.c: read pseudo-merge extension
pseudo-merge: scaffolding for reads
pack-bitmap: extract `read_bitmap()` function
pack-bitmap-write.c: write pseudo-merge table
pseudo-merge: implement support for selecting pseudo-merge commits
config: introduce `git_config_double()`
pack-bitmap: make `bitmap_writer_push_bitmapped_commit()` public
pack-bitmap: implement `bitmap_writer_has_bitmapped_object_id()`
pack-bitmap-write: support storing pseudo-merge commits
...
Both `oidread()` and `oidclr()` use `the_repository` to derive the hash
function that shall be used. Require callers to pass in the hash
algorithm to get rid of this implicit dependency.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Further clean-up the refs subsystem to stop relying on
the_repository, and instead use the repository associated to the
ref_store object.
* ps/refs-without-the-repository-updates:
refs/packed: remove references to `the_hash_algo`
refs/files: remove references to `the_hash_algo`
refs/files: use correct repository
refs: remove `dwim_log()`
refs: drop `git_default_branch_name()`
refs: pass repo when peeling objects
refs: move object peeling into "object.c"
refs: pass ref store when detecting dangling symrefs
refs: convert iteration over replace refs to accept ref store
refs: retrieve worktree ref stores via associated repository
refs: refactor `resolve_gitlink_ref()` to accept a repository
refs: pass repo when retrieving submodule ref store
refs: track ref stores via strmap
refs: implement releasing ref storages
refs: rename `init_db` callback to avoid confusion
refs: adjust names for `init` and `init_db` callbacks
The pack bitmap code saw some clean-up to prepare for a follow-up topic.
* tb/pack-bitmap-write-cleanups:
pack-bitmap: introduce `bitmap_writer_free()`
pack-bitmap-write.c: avoid uninitialized 'write_as' field
pack-bitmap: drop unused `max_bitmaps` parameter
pack-bitmap: avoid use of static `bitmap_writer`
pack-bitmap-write.c: move commit_positions into commit_pos fields
object.h: add flags allocated by pack-bitmap.h
The pack-bitmap-writer machinery uses a oidmap (backed by khash.h) to
map from commits selected for bitmaps (by OID) to a bitmapped_commit
structure (containing the bitmap itself, among other things like its XOR
offset, etc.)
This map was initialized at the end of `bitmap_writer_build()`. New
entries are added in `pack-bitmap-write.c::store_selected()`, which is
called by the bitmap_builder machinery (which is responsible for
traversing history and generating the actual bitmaps).
Reorganize when this field is initialized and when entries are added to
it so that we can quickly determine whether a commit is a candidate for
pseudo-merge selection, or not (since it was already selected to receive
a bitmap, and thus storing it in a pseudo-merge would be redundant).
The changes are as follows:
- Introduce a new `bitmap_writer_init()` function which initializes
the `writer.bitmaps` field (instead of waiting until the end of
`bitmap_writer_build()`).
- Add map entries in `push_bitmapped_commit()` (which is called via
`bitmap_writer_select_commits()`) with OID keys and NULL values to
track whether or not we *expect* to write a bitmap for some given
commit.
- Validate that a NULL entry is found matching the given key when we
store a selected bitmap.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Both `peel_object()` and `peel_iterated_oid()` implicitly rely on
`the_repository` to look up objects. Despite the fact that we want to
get rid of `the_repository`, it also leads to some restrictions in our
ref iterators when trying to retrieve the peeled value for a repository
other than `the_repository`.
Refactor these functions such that both take a repository as argument
and remove the now-unnecessary restrictions.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Now that there is clearer memory ownership around the bitmap_writer
structure, introduce a bitmap_writer_free() function that callers may
use to free any memory associated with their instance of the
bitmap_writer structure.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The `max_bitmaps` parameter in `bitmap_writer_select_commits()` was
introduced back in 7cc8f97108 (pack-objects: implement bitmap writing,
2013-12-21), making it original to the bitmap implementation in Git
itself.
When that patch was merged via 0f9e62e084 (Merge branch
'jk/pack-bitmap', 2014-02-27), its sole caller in builtin/pack-objects.c
passed a value of "-1" for `max_bitmaps`, indicating no limit.
Since then, the only other caller (in midx.c, added via c528e17966
(pack-bitmap: write multi-pack bitmaps, 2021-08-31)) also uses a value
of "-1" for `max_bitmaps`.
Since no callers have needed a finite limit for the `max_bitmaps`
parameter in the nearly decade that has passed since 0f9e62e084, let's
remove the parameter and any dead pieces of code connected to it.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The pack-bitmap machinery uses a structure called 'bitmap_writer' to
collect the data necessary to write out .bitmap files. Since its
introduction in 7cc8f971085 (pack-objects: implement bitmap writing,
2013-12-21), there has been a single static bitmap_writer structure,
which is responsible for all bitmap writing-related operations.
In practice, this is OK, since we are only ever writing a single .bitmap
file in a single process (e.g., `git multi-pack-index write --bitmap`,
`git pack-objects --write-bitmap-index`, `git repack -b`, etc.).
However, having a single static variable makes issues like data
ownership unclear, when to free variables, what has/hasn't been
initialized unclear.
Refactor this code to be written in terms of a given bitmap_writer
structure instead of relying on a static global.
Note that this exposes the structure definition of the bitmap_writer at
the pack-bitmap.h level. We could work around this by, e.g., forcing
callers to declare their writers as:
struct bitmap_writer *writer;
bitmap_writer_init(&bitmap_writer);
and then declaring `bitmap_writer_init()` as taking in a double-pointer
like so:
void bitmap_writer_init(struct bitmap_writer **writer);
which would avoid us having to expose the definition of the structure
itself. This patch takes a different approach, since future patches
(like for the ongoing pseudo-merge bitmaps work) will want to modify the
innards of this structure (in the previous example, via pseudo-merge.c).
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Apply the rules that rewrite callers of "refs" interfaces to explicitly
pass `struct ref_store`. The resulting patch has been applied with the
`--whitespace=fix` option.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Work to support a repository that work with both SHA-1 and SHA-256
hash algorithms has started.
* eb/hash-transition: (30 commits)
t1016-compatObjectFormat: add tests to verify the conversion between objects
t1006: test oid compatibility with cat-file
t1006: rename sha1 to oid
test-lib: compute the compatibility hash so tests may use it
builtin/ls-tree: let the oid determine the output algorithm
object-file: handle compat objects in check_object_signature
tree-walk: init_tree_desc take an oid to get the hash algorithm
builtin/cat-file: let the oid determine the output algorithm
rev-parse: add an --output-object-format parameter
repository: implement extensions.compatObjectFormat
object-file: update object_info_extended to reencode objects
object-file-convert: convert commits that embed signed tags
object-file-convert: convert commit objects when writing
object-file-convert: don't leak when converting tag objects
object-file-convert: convert tag objects when writing
object-file-convert: add a function to convert trees between algorithms
object: factor out parse_mode out of fast-import and tree-walk into in object.h
cache: add a function to read an OID of a specific algorithm
tag: sign both hashes
commit: export add_header_signature to support handling signatures on tags
...
Now that multi-pack reuse is supported, enable it via the
feature.experimental configuration in addition to the classic
`pack.allowPackReuse`.
This will allow more users to experiment with the new behavior who might
not otherwise be aware of the existing `pack.allowPackReuse`
configuration option.
The enum with values NO_PACK_REUSE, SINGLE_PACK_REUSE, and
MULTI_PACK_REUSE is defined statically in builtin/pack-objects.c's
compilation unit. We could hoist that enum into a scope visible from the
repository_settings struct, and then use that enum value in
pack-objects. Instead, define a single int that indicates what
pack-objects's default value should be to avoid additional unnecessary
code movement.
Though `feature.experimental` implies `pack.allowPackReuse=multi`, this
can still be overridden by explicitly setting the latter configuration
to either "single" or "false". Tests covering all of these cases are
showin t5332.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Streaming spans of packfile data used to be done only from a
single, primary, pack in a repository with multiple packfiles. It
has been extended to allow reuse from other packfiles, too.
* tb/multi-pack-verbatim-reuse: (26 commits)
t/perf: add performance tests for multi-pack reuse
pack-bitmap: enable reuse from all bitmapped packs
pack-objects: allow setting `pack.allowPackReuse` to "single"
t/test-lib-functions.sh: implement `test_trace2_data` helper
pack-objects: add tracing for various packfile metrics
pack-bitmap: prepare to mark objects from multiple packs for reuse
pack-revindex: implement `midx_pair_to_pack_pos()`
pack-revindex: factor out `midx_key_to_pack_pos()` helper
midx: implement `midx_preferred_pack()`
git-compat-util.h: implement checked size_t to uint32_t conversion
pack-objects: include number of packs reused in output
pack-objects: prepare `write_reused_pack_verbatim()` for multi-pack reuse
pack-objects: prepare `write_reused_pack()` for multi-pack reuse
pack-objects: pass `bitmapped_pack`'s to pack-reuse functions
pack-objects: keep track of `pack_start` for each reuse pack
pack-objects: parameterize pack-reuse routines over a single pack
pack-bitmap: return multiple packs via `reuse_partial_packfile_from_bitmap()`
pack-bitmap: simplify `reuse_partial_packfile_from_bitmap()` signature
ewah: implement `bitmap_is_empty()`
pack-bitmap: pass `bitmapped_pack` struct to pack-reuse functions
...
Remove unused header "#include".
* en/header-cleanup:
treewide: remove unnecessary includes in source files
treewide: add direct includes currently only pulled in transitively
trace2/tr2_tls.h: remove unnecessary include
submodule-config.h: remove unnecessary include
pkt-line.h: remove unnecessary include
line-log.h: remove unnecessary include
http.h: remove unnecessary include
fsmonitor--daemon.h: remove unnecessary includes
blame.h: remove unnecessary includes
archive.h: remove unnecessary include
treewide: remove unnecessary includes in source files
treewide: remove unnecessary includes from header files
Each of these were checked with
gcc -E -I. ${SOURCE_FILE} | grep ${HEADER_FILE}
to ensure that removing the direct inclusion of the header actually
resulted in that header no longer being included at all (i.e. that
no other header pulled it in transitively).
...except for a few cases where we verified that although the header
was brought in transitively, nothing from it was directly used in
that source file. These cases were:
* builtin/credential-cache.c
* builtin/pull.c
* builtin/send-pack.c
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Now that both the pack-bitmap and pack-objects code are prepared to
handle marking and using objects from multiple bitmapped packs for
verbatim reuse, allow marking objects from all bitmapped packs as
eligible for reuse.
Within the `reuse_partial_packfile_from_bitmap()` function, we no longer
only mark the pack whose first object is at bit position zero for reuse,
and instead mark any pack contained in the MIDX as a reuse candidate.
Provide a handful of test cases in a new script (t5332) exercising
interesting behavior for multi-pack reuse to ensure that we performed
all of the previous steps correctly.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In e704fc7978 (pack-objects: introduce pack.allowPackReuse, 2019-12-18),
the `pack.allowPackReuse` configuration option was introduced, allowing
users to disable the pack reuse mechanism.
To prepare for debugging multi-pack reuse, allow setting configuration
to "single" in addition to the usual bool-or-int values.
"single" implies the same behavior as "true", "1", "yes", and so on. But
it will complement a new "multi" value (to be introduced in a future
commit). When set to "single", we will only perform pack reuse on a
single pack, regardless of whether or not there are multiple MIDX'd
packs.
This requires no code changes (yet), since we only support single pack
reuse.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As part of the multi-pack reuse effort, we will want to add some tests
that assert that we reused a certain number of objects from a certain
number of packs.
We could do this by grepping through the stderr output of
`pack-objects`, but doing so would be brittle in case the output format
changed.
Instead, let's use the trace2 mechanism to log various pieces of
information about the generated packfile, which we can then use to
compare against desired values.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In addition to including the number of objects reused verbatim from a
reuse-pack, include the number of packs from which objects were reused.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The function `write_reused_pack_verbatim()` within
`builtin/pack-objects.c` is responsible for writing out a continuous
set of objects beginning at the start of the reuse packfile.
In the existing implementation, we did something like:
while (pos < reuse_packfile_bitmap->word_alloc &&
reuse_packfile_bitmap->words[pos] == (eword_t)~0)
pos++;
if (pos)
/* write first `pos * BITS_IN_WORD` objects from pack */
as an optimization to record a single chunk for the longest continuous
prefix of objects wanted out of the reuse pack, instead of having a
chunk for each individual object. For more details, see bb514de356
(pack-objects: improve partial packfile reuse, 2019-12-18).
In order to retain this optimization in a multi-pack reuse world, we can
no longer assume that the first object in a pack is on a word boundary
in the bitmap storing the set of reusable objects.
Assuming that all objects from the beginning of the reuse packfile up to
the object corresponding to the first bit on a word boundary are part of
the result, consume whole words at a time until the last whole word
belonging to the reuse packfile. Copy those objects to the resulting
packfile, and track that we reused them by recording a single chunk.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The function `write_reused_pack()` within `builtin/pack-objects.c` is
responsible for performing pack-reuse on a single pack, and has two main
functions:
- it dispatches a call to `write_reused_pack_verbatim()` to see if we
can reuse portions of the packfile in whole-word chunks
- for any remaining objects (that is, any objects that appear after
the first "gap" in the bitmap), call write_reused_pack_one() on that
object to record it for reuse.
Prepare this function for multi-pack reuse by removing the assumption
that the bit position corresponding to the first object being reused
from a given pack must be at bit position zero.
The changes in this function are mostly straightforward. Initialize `i`
to the position of the first word to contain bits corresponding to that
reuse pack. In most situations, we throw the initialized value away,
since we end up replacing it with the return value from
write_reused_pack_verbatim(), moving us past the section of whole words
that we reused.
Likewise, modify the per-object loop to ignore any bits at the beginning
of the first word that do not belong to the pack currently being reused,
as well as skip to the "done" section once we have processed the last
bit corresponding to this pack.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Further prepare pack-objects to perform verbatim pack-reuse over
multiple packfiles by converting functions that take in a pointer to a
`struct packed_git` to instead take in a pointer to a `struct
bitmapped_pack`.
The additional information found in the bitmapped_pack struct (such as
the bit position corresponding to the beginning of the pack) will be
necessary in order to perform verbatim pack-reuse.
Note that we don't use any of the extra pieces of information contained
in the bitmapped_pack struct, so this step is merely preparatory and
does not introduce any functional changes.
Note further that we do not change the argument type to
write_reused_pack_one(). That function is responsible for copying
sections of the packfile directly and optionally patching any OFS_DELTAs
to account for not reusing sections of the packfile in between a delta
and its base.
As such, that function is (and should remain) oblivious to multi-pack
reuse, and does not require any of the extra pieces of information
stored in the bitmapped_pack struct.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When reusing objects from a pack, we keep track of a set of one or more
`reused_chunk`s, corresponding to sections of one or more object(s) from
a source pack that we are reusing. Each chunk contains two pieces of
information:
- the offset of the first object in the source pack (relative to the
beginning of the source pack)
- the difference between that offset, and the corresponding offset in
the pack we're generating
The purpose of keeping track of these is so that we can patch an
OFS_DELTAs that cross over a section of the reuse pack that we didn't
take.
For instance, consider a hypothetical pack as shown below:
(chunk #2)
__________...
/
/
+--------+---------+-------------------+---------+
... | <base> | <other> | (unused) | <delta> | ...
+--------+---------+-------------------+---------+
\ /
\______________/
(chunk #1)
Suppose that we are sending objects "base", "other", and "delta", and
that the "delta" object is stored as an OFS_DELTA, and that its base is
"base". If we don't send any objects in the "(unused)" range, we can't
copy the delta'd object directly, since its delta offset includes a
range of the pack that we didn't copy, so we have to account for that
difference when patching and reassembling the delta.
In order to compute this value correctly, we need to know not only where
we are in the packfile we're assembling (with `hashfile_total(f)`) but
also the position of the first byte of the packfile that we are
currently reusing. Currently, this works just fine, since when reusing
only a single pack those two values are always identical (because
verbatim reuse is the first thing pack-objects does when enabled after
writing the pack header).
But when reusing multiple packs which have one or more gaps, we'll need
to account for these two values diverging.
Together, these two allow us to compute the reused chunk's offset
difference relative to the start of the reused pack, as desired.
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The routines pack-objects uses to perform verbatim pack-reuse are:
- write_reused_pack_one()
- write_reused_pack_verbatim()
- write_reused_pack()
, all of which assume that there is exactly one packfile being reused:
the global constant `reuse_packfile`.
Prepare for reusing objects from multiple packs by making reuse packfile
a parameter of each of the above functions in preparation for calling
these functions in a loop with multiple packfiles.
Note that we still have the global "reuse_packfile", but pass it through
each of the above function's parameter lists, eliminating all but one
direct access (the top-level caller in `write_pack_file()`). Even after
this series, we will still have a global, but it will hold the array of
reusable packfiles, and we'll pass them one at a time to these functions
in a loop.
Note also that we will eventually need to pass a `bitmapped_pack`
instead of a `packed_git` in order to hold onto additional information
required for reuse (such as the bit position of the first object
belonging to that pack). But that change will be made in a future commit
so as to minimize the noise below as much as possible.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Further prepare for enabling verbatim pack-reuse over multiple packfiles
by changing the signature of reuse_partial_packfile_from_bitmap() to
populate an array of `struct bitmapped_pack *`'s instead of a pointer to
a single packfile.
Since the array we're filling out is sized dynamically[^1], add an
additional `size_t *` parameter which will hold the number of reusable
packs (equal to the number of elements in the array).
Note that since we still have not implemented true multi-pack reuse,
these changes aren't propagated out to the rest of the caller in
builtin/pack-objects.c.
In the interim state, we expect that the array has a single element, and
we use that element to fill out the static `reuse_packfile` variable
(which is a bog-standard `struct packed_git *`). Future commits will
continue to push this change further out through the pack-objects code.
[^1]: That is, even though we know the number of packs which are
candidates for pack-reuse, we do not know how many of those
candidates we can actually reuse.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
