|
c15ed350
|
2021-04-26T12:23:25
|
|
repo: validate repository paths
Ensure that a repository's path (at initialization or open time) is
valid. On Windows systems, this means that the longest known path
beneath the repository will fit within MAX_PATH: this is a lock file for
a loose object within the repository itself.
Other paths, like a very long loose reference, may fail to be opened
after the repository is opened. These variable length paths will be
checked when they are accessed themselves. This new functionality is
done at open to prevent needlessly checking every file in the gitdir
(eg, `MERGE_HEAD`) for its length when we could instead check once at
repository open time.
|
|
|
3589587d
|
2021-04-24T13:11:25
|
|
repo: factor the commondir detection
|
|
|
1016ad4f
|
2021-04-15T01:31:12
|
|
path: don't join paths in git_path_find_dir
Let `git_path_find_dir` simply take a `git_buf` that contains a
directory or a file, instead of trying to both join a path AND then deal
with prettifying it or its basename. This allows consumers to join
paths themselves (and apply any necessary rules - like fitting within
MAX_PATH).
|
|
|
289aaa41
|
2021-04-04T20:46:40
|
|
ignore: validate workdir paths for ignore files
|
|
|
9fb755d5
|
2021-04-04T19:59:57
|
|
attr: validate workdir paths for attribute files
We should allow attribute files - inside working directories - to have
names longer than MAX_PATH when core.longpaths is set.
`git_attr_path__init` takes a repository to validate the path with.
|
|
|
717df1a4
|
2021-04-15T01:23:07
|
|
worktree: validate worktree paths
Worktree paths need to fix within MAX_PATH always, regardless of
`core.longpaths` setting.
|
|
|
e52c2989
|
2021-04-04T19:46:08
|
|
repository: validate workdir path lengths
|
|
|
4470e48a
|
2021-04-04T14:24:35
|
|
workdir: validate working directory entry path length
|
|
|
e5851c62
|
2021-04-24T09:44:45
|
|
refs: ensure loose refs adhere to path validation
On Windows, we need to enforce MAX_PATH for loose references and their
reflogs. Ensure that any path - including the lock file - would fit
within the 260 character maximum.
We do not honor core.longpaths for loose reference files or reflogs.
core.longpaths only applies to paths in the working directory.
|
|
|
f3bcadd2
|
2021-04-05T10:57:49
|
|
submodule: validate path lengths
Validate that working directory paths honor `core.longpaths` where
appropriate. Paths to the submodule gitdirs must always honor the
operating system length restrictions; `core.longpaths` does not affect
gitdir paths.
|
|
|
91156a0f
|
2021-04-04T10:44:29
|
|
diff: use git_repository_workdir_path
The new git_repository_workdir_path function does error checking on
working directory inputs on Windows; use it to construct paths within
working directories.
|
|
|
578aeba9
|
2021-03-20T17:00:33
|
|
use git_repository_workdir_path to generate paths
Use `git_repository_workdir_path` to generate workdir paths since it
will validate the length.
|
|
|
b457fe27
|
2021-04-04T22:18:55
|
|
iterator: validate workdir paths
Supply the repository for the filesystem and workdir iterators - for
workdir iterators, this is non-null and we can lookup the core.longpaths
configuration option. (For regular filesystem iterators, this is NULL,
so core.longpaths does not apply.)
|
|
|
99ddfd5c
|
2021-03-20T16:54:09
|
|
checkout: validate path length
Ensure that we are validating working directory paths before we try to
write to them.
|
|
|
6b878db5
|
2021-03-20T16:44:08
|
|
checkout: use target path; don't assume workdir
We're not necessarily checking out into the working directory. We could
be checking out into an arbitrary location. Ensure that when we are
writing conflict data that we do it in the checkout target.
|
|
|
50e01216
|
2021-03-20T16:39:29
|
|
repo: introduce git_repository_workdir_path
Add a simple accessor for workdir paths to get an absolute on-disk path
given a repository and a relative path within it. This is useful to
avoid copy-pasta `git_buf_joinpath` and to ensure that we validate
working directory paths while honoring `core.longpaths` settings.
|
|
|
dbc03de4
|
2021-03-20T14:28:25
|
|
apply: ensure we validate paths
There was no test ensuring that we validate `.git` paths. We do, but
let's add a test to make sure that we never regress this.
|
|
|
dc1ba018
|
2021-03-20T13:01:00
|
|
path: introduce ondisk and workdir path validation
Introduce `git_path_validate_filesystem` which validates (absolute) on-disk
paths and `git_path_validate_workdir` to perform validations on (absolute)
working directory paths. These functions are useful as there may be system
limitations on on-disk paths, particularly on Windows (for example,
enforcing MAX_PATH).
For working directory paths, these limitations may be per-repository, based
on the `core.longpaths` configuration setting.
|
|
|
88323cd0
|
2021-03-20T09:52:17
|
|
path: git_path_isvalid -> git_path_validate
If we want to validate more and different types of paths, the name
`git_path_validate` makes that easier and more expressive. We can add,
for example, `git_path_validate_foo` while the current name makes that
less ergonomic.
|
|
|
cb136cdd
|
2021-04-14T22:22:11
|
|
utf8: introduce git_utf8_char_length
Introduce a function to determine the number of Unicode characters in a
given UTF-8 string.
|
|
|
1d95b59b
|
2021-04-14T15:47:27
|
|
utf8: refactor utf8 functions
Move the utf8 functions into a proper namespace `git_utf8` instead of
being in the namespaceless `git__` function group. Update them to
have out-params first and use `char *` instead of `uint8_t *` to match
our API treating strings as `char *` (even if they truly contain `uchar`s
inside).
|
|
|
4f4b1139
|
2021-03-10T11:21:39
|
|
Merge pull request #5815 from libgit2/ethomson/treebuilder_write
tree: deprecate `git_treebuilder_write_with_buffer`
|
|
|
a0fca809
|
2021-03-10T11:19:14
|
|
Merge pull request #5814 from ianhattendorf/fix/winhttp-proxy-https
winhttp: skip certificate check if unable to send request
|
|
|
d7e65096
|
2021-03-04T09:44:26
|
|
Merge pull request #5764 from lhchavez/cgraph-needs-refresh
commit-graph: Introduce `git_commit_graph_needs_refresh()`
|
|
|
b99b5a81
|
2021-03-04T09:38:10
|
|
Merge pull request #5763 from lhchavez/cgraph-lookup
commit-graph: Support lookups of entries in a commit-graph
|
|
|
7eb21516
|
2021-02-28T00:20:28
|
|
tree: deprecate `git_treebuilder_write_with_buffer`
The function `git_treebuilder_write_with_buffer` is unnecessary; it
is used internally as part of treebuilder writing, but it has little
use to external callers. For callers that repeatedly write a
treebuilder, we can supply them with a buffer in the treebuilder struct
instead of recreating it. For ourselves, when we want a single buffer
in our write loop, we can use an internal function.
|
|
|
b33e018c
|
2021-03-04T09:20:34
|
|
Merge pull request #5818 from mamapanda/patch-3
merge: Check insert_head_ids error in create_virtual_base
|
|
|
80bdeca4
|
2021-03-04T09:18:30
|
|
Merge pull request #5817 from mamapanda/patch-2
Check git_signature_dup failure
|
|
dc1095a5
|
2021-03-03T14:42:12
|
|
merge: Check insert_head_ids error in create_virtual_base
insert_head_ids can fail due to allocation error
|
|
|
b9bfe812
|
2021-03-03T14:03:06
|
|
Check git_signature_dup failure
git_signature_dup can have an allocation failure
|
|
edffea15
|
2021-03-01T16:26:58
|
|
winhttp: skip certificate check if unable to send request
In some circumstances (e.g. when proxies are involved), winhttp will fail to reach the WINHTTP_CALLBACK_STATUS_SENDING_REQUEST phase. If this occurs, we'll error with ERROR_WINHTTP_INCORRECT_HANDLE_STATE when attempting to query the server certificate context (see https://docs.microsoft.com/en-us/windows/win32/api/winhttp/nf-winhttp-winhttpsendrequest#remarks).
To avoid this, verify that WinHttpSendRequest has reached the WINHTTP_CALLBACK_STATUS_SENDING_REQUEST phase before checking the certificate. Since we're using WinHTTP in synchronous mode, we know for sure that once WinHttpSendRequest returns we've either sent it successfully or not.
NOTE: WINHTTP_CALLBACK_STATUS_SENDING_REQUEST appears to be deprecated with no direct replacement. WINHTTP_CALLBACK_STATUS_SENDREQUEST_COMPLETE is only available in async mode, and there doesn't appear to be a method of querying this flag outside of the status callback.
|
|
|
1bc8dec4
|
2021-02-28T11:02:23
|
|
Merge pull request #5797 from aaronfranke/typos
Fix some typos
|
|
|
0fb58396
|
2021-02-21T20:43:51
|
|
Merge pull request #5801 from mamapanda/patch-1
index: Check git_vector_dup error in write_entries
|
|
|
40930508
|
2021-02-18T16:36:42
|
|
index: Initialize case_sorted to GIT_VECTOR_INIT
This is for extra safety within write_entries
|
|
|
21981f28
|
2021-02-16T13:43:09
|
|
index: Check git_vector_dup error in write_entries
If allocating case_sorted.contents fails, git_vector_sort will segfault.
|
|
|
36232528
|
2021-02-16T13:08:02
|
|
refdb_fs: Check git_sortedcache wlock/rlock errors
To prevent a race condition, these wlock/rlock calls should probably be checked
for errors.
|
|
7efddeb7
|
2021-02-15T15:47:28
|
|
Fix some typos
|
|
|
02eb1711
|
2021-02-01T10:07:00
|
|
Merge pull request #5731 from KOLANICH-libs/owner_accessor
patch: add owner accessor
|
|
|
2ad15250
|
2021-02-01T09:58:25
|
|
Merge pull request #5762 from lhchavez/cgraph-parser
commit-graph: Introduce a parser for commit-graph files
|
|
|
0a710091
|
2021-02-01T09:41:42
|
|
Merge pull request #5786 from libgit2/ethomson/revparse
revspec: rename git_revparse_mode_t to git_revspec_t
|
|
|
4732e030
|
2021-01-31T00:36:54
|
|
revspec: rename git_revparse_mode_t to git_revspec_t
The information about the type of a revision spec is not information
about the parser. Name it accordingly, so that `git_revparse_mode_t`
is now `git_revspec_t`. Deprecate the old name.
|
|
2f382ab7
|
2021-01-20T04:40:40
|
|
mwindow: Fix a bug in the LRU window finding code
This change now updates the `lru_window` variable to match the current
file's MRU window. This makes it such that it doesn't always choose the
file that happened to come last in the list of window files, and instead
should now correctly choose the file with the least-recently-used one.
|
|
|
674875dc
|
2021-01-15T09:59:15
|
|
Merge pull request #5768 from lhchavez/midx-needs-refresh
midx: Fix a bug in `git_midx_needs_refresh()`
|
|
|
487f2a82
|
2021-01-10T21:32:28
|
|
Merge pull request #5775 from libgit2/ethomson/clone_branch
clone: set refs/remotes/origin/HEAD when branch is specified
|
|
|
1f32ed25
|
2021-01-04T19:33:48
|
|
commit-graph: Support lookups of entries in a commit-graph
This change introduces `git_commit_graph_entry_find()` and
`git_commit_graph_entry_parent()`. These two functions allow a much
faster lookup of commits by ID, since the ODB does not need to be
consulted, the commit object does not need to be inflated, and the
contents of the commit object do not need to be parsed.
Part of: #5757
|
|
|
3fd57a75
|
2021-01-04T18:22:43
|
|
commit-graph: Introduce a parser for commit-graph files
This change is the first in a series to add support for git's
commit-graph. This should speed up commit graph traversals by avoiding
object parsing and allowing some operations to terminate earlier.
Part of: #5757
|
|
|
1a2f9609
|
2021-01-05T06:24:26
|
|
commit-graph: Introduce `git_commit_graph_needs_refresh()`
This change introduces a function that allows the caller to know whether
the `commit-graph` file has not been modified since it was parsed.
Part of: #5757
|
|
|
923c0f7b
|
2021-01-07T17:34:17
|
|
clone: set refs/remotes/origin/HEAD when branch is specified
When a branch is specified to check out in clone, update the remote
tracking `HEAD` to point to it. This mimics git's behavior, when
`git clone -b <name>` is used.
|
|
|
c6cf7f0e
|
2021-01-07T17:00:57
|
|
Merge pull request #5769 from lhchavez/pwrite-pread
Use `p_pwrite`/`p_pread` consistently throughout the codebase
|
|
|
ff6f6754
|
2021-01-07T05:44:16
|
|
Use `p_pwrite`/`p_pread` consistently throughout the codebase
This change stops using the seek+read/write combo to perform I/O with an
offset, since this is faster by one system call (and also more atomic
and therefore safer).
|
|
|
d50d3db6
|
2021-01-07T05:43:30
|
|
midx: Fix a bug in `git_midx_needs_refresh()`
The very last check in the freshness check for the midx was wrong ><
This was also because this function was not tested.
|
|
|
3f4bc213
|
2021-01-07T14:05:02
|
|
repo: ignore empty init.defaultbranch
The init.defaultbranch option may be set, but empty. In this case, we
should ignore it instead of trying to set our default branch to
`refs/heads/`.
|
|
|
32350e89
|
2021-01-07T14:01:30
|
|
remote: don't update invalid refs
If a symbolic reference points to something invalid, then do not try to
update it.
|
|
|
c31032a3
|
2021-01-07T10:15:25
|
|
Merge pull request #5760 from libgit2/ethomson/tttoo_many_ttts
blob: fix name of `GIT_BLOB_FILTER_ATTRIBUTES_FROM_HEAD`
|
|
|
27e34f9b
|
2021-01-06T11:49:00
|
|
Merge pull request #5759 from JoshuaS3/master
Add documentation for git_blob_filter_options.version
|
|
|
855f2998
|
2021-01-05T14:45:14
|
|
blob: fix name of `GIT_BLOB_FILTER_ATTRIBUTES_FROM_HEAD`
`GIT_BLOB_FILTER_ATTTRIBUTES_FROM_HEAD` is misspelled, it should be
`GIT_BLOB_FILTER_ATTRIBUTES_FROM_HEAD`, and it would be if it were not
for the MacBook Pro keyboard and my inattentiveness.
|
|
|
d9c15387
|
2021-01-05T14:29:58
|
|
blob: add git_blob_filter_options_init
The `git_blob_filter_options_init` function should be included, to allow
callers in FFI environments to let us initialize an options structure
for them.
|
|
|
f0d7922c
|
2021-01-05T12:12:19
|
|
Merge pull request #5583 from 0xdky/dhruva/build-with-nommap
Build with NO_MMAP
|
|
|
754fa526
|
2021-01-04T06:10:10
|
|
Use an option instead of a flag for USE_BUNDLED_ZLIB
Now `USE_BUNDLED_ZLIB` can be set to the string `Chromium` to enable the
Chromium implementation of zlib.
|
|
4ce8e01a
|
2020-06-17T14:31:11
|
|
Support build with NO_MMAP to disable use of system mmap
* Use pread/pwrite to avoid updating position in file descriptor
* Emulate missing pread/pwrite on win32 using overlapped file IO
|
|
|
a9f57a89
|
2020-11-23T10:14:00
|
|
Remove broken support for write in emulated mmap
* Emulated mmap based write without pagefault handling is not
possible since IO happens outside of call to mmap and data is
written to mapped memory
* Potential emulation using userfaultfd() might be possible
|
|
|
83265b3e
|
2020-12-18T06:50:22
|
|
zlib: Add support for building with Chromium's zlib implementation
This change builds libgit2 using Chromium's zlib implementation by
invoking cmake with `-DUSE_BUNDLED_ZLIB=ON -DUSE_CHROMIUM_ZLIB=ON`,
which is ~10% faster than the bundled zlib for the core::zstream suite.
This version of zlib has some optimizations:
a) Decompression (Intel+ARM): inflate_fast, adler32, crc32, etc.
b) Compression (Intel): fill_window, longest_match, hash function, etc.
Due to the introduction of SIMD optimizations, and to get the maximum
performance out of this fork of zlib, this requires an x86_64 processor
with SSE4.2 and CLMUL (anything Westmere or later, ~2010). The Chromium
zlib implementation also supports ARM with NEON, but it has not been
enabled in this patch.
Performance
===========
TL;DR: Running just `./libgit2_clar -score::zstream` 100 times in a loop
took 0:56.30 before and 0:50.67 after (~10% reduction!).
The bundled and system zlib implementations on an Ubuntu Focal system
perform relatively similar (the bundled one is marginally better due to
the compiler being able to inline some functions), so only the bundled
and Chromium zlibs were compared.
For a more balanced comparison (to ensure that nothing regressed
overall), `libgit2_clar` under `perf` was also run, and the zlib-related
functions were compared.
Bundled
-------
```shell
cmake \
-DUSE_BUNDLED_ZLIB=ON \
-DUSE_CHROMIUM_ZLIB=OFF \
-DCMAKE_BUILD_TYPE="RelWithDebInfo" \
-DCMAKE_C_FLAGS="-fPIC -fno-omit-frame-pointer" \
-GNinja \
..
ninja
perf record --call-graph=dwarf ./libgit2_clar
perf report --children
```
```
Samples: 87K of event 'cycles', Event count (approx.): 75923450603
Children Self Command Shared Objec Symbol
+ 4.14% 0.01% libgit2_clar libgit2_clar [.] git_zstream_get_output_chunk
+ 2.91% 0.00% libgit2_clar libgit2_clar [.] git_zstream_get_output
+ 0.69% 0.00% libgit2_clar libgit2_clar [.] git_zstream_get_output (inlined)
0.17% 0.00% libgit2_clar libgit2_clar [.] git_zstream_init
0.02% 0.00% libgit2_clar libgit2_clar [.] git_zstream_reset
0.00% 0.00% libgit2_clar libgit2_clar [.] git_zstream_eos
0.00% 0.00% libgit2_clar libgit2_clar [.] git_zstream_done
0.00% 0.00% libgit2_clar libgit2_clar [.] git_zstream_free (inlined)
Samples: 87K of event 'cycles', Event count (approx.): 75923450603
Children Self Command Shared Objec Symbol
+ 3.12% 0.01% libgit2_clar libgit2_clar [.] deflate
+ 2.65% 1.48% libgit2_clar libgit2_clar [.] deflate_slow
+ 1.60% 0.55% libgit2_clar libgit2_clar [.] inflate
+ 0.53% 0.00% libgit2_clar libgit2_clar [.] write_deflate
0.49% 0.36% libgit2_clar libgit2_clar [.] inflate_fast
0.46% 0.02% libgit2_clar libgit2_clar [.] deflate_fast
0.19% 0.19% libgit2_clar libgit2_clar [.] inflate_table
0.16% 0.01% libgit2_clar libgit2_clar [.] inflateInit_
0.15% 0.00% libgit2_clar libgit2_clar [.] inflateInit2_ (inlined)
0.10% 0.00% libgit2_clar libgit2_clar [.] deflateInit_
0.10% 0.00% libgit2_clar libgit2_clar [.] deflateInit2_
0.03% 0.00% libgit2_clar libgit2_clar [.] deflateReset (inlined)
0.02% 0.00% libgit2_clar libgit2_clar [.] deflateReset
0.02% 0.00% libgit2_clar libgit2_clar [.] inflateEnd
0.02% 0.00% libgit2_clar libgit2_clar [.] deflateEnd
0.01% 0.00% libgit2_clar libgit2_clar [.] deflateResetKeep
0.01% 0.01% libgit2_clar libgit2_clar [.] inflateReset2
0.01% 0.00% libgit2_clar libgit2_clar [.] deflateReset (inlined)
0.00% 0.00% libgit2_clar libgit2_clar [.] inflateStateCheck (inlined)
0.00% 0.00% libgit2_clar libgit2_clar [.] inflateReset (inlined)
0.00% 0.00% libgit2_clar libgit2_clar [.] inflateStateCheck (inlined)
0.00% 0.00% libgit2_clar libgit2_clar [.] deflateStateCheck (inlined)
0.00% 0.00% libgit2_clar libgit2_clar [.] inflateResetKeep (inlined)
```
Chromium
--------
```shell
cmake \
-DUSE_BUNDLED_ZLIB=ON \
-DUSE_CHROMIUM_ZLIB=ON \
-DCMAKE_BUILD_TYPE="RelWithDebInfo" \
-DCMAKE_C_FLAGS="-fPIC -fno-omit-frame-pointer" \
-GNinja \
..
ninja
perf record --call-graph=dwarf ./libgit2_clar
perf report --children
```
```
Samples: 97K of event 'cycles', Event count (approx.): 80862210917
Children Self Command Shared Objec Symbol
+ 3.31% 0.00% libgit2_clar libgit2_clar [.] git_zstream_get_output_chunk
+ 2.27% 0.01% libgit2_clar libgit2_clar [.] git_zstream_get_output
+ 0.55% 0.00% libgit2_clar libgit2_clar [.] git_zstream_get_output (inlined)
0.18% 0.00% libgit2_clar libgit2_clar [.] git_zstream_init
0.02% 0.00% libgit2_clar libgit2_clar [.] git_zstream_reset
0.00% 0.00% libgit2_clar libgit2_clar [.] git_zstream_free (inlined)
0.00% 0.00% libgit2_clar libgit2_clar [.] git_zstream_done
0.00% 0.00% libgit2_clar libgit2_clar [.] git_zstream_free
Samples: 97K of event 'cycles', Event count (approx.): 80862210917
Children Self Command Shared Objec Symbol
+ 2.55% 0.01% libgit2_clar libgit2_clar [.] deflate
+ 2.25% 1.41% libgit2_clar libgit2_clar [.] deflate_slow
+ 1.10% 0.52% libgit2_clar libgit2_clar [.] inflate
0.36% 0.00% libgit2_clar libgit2_clar [.] write_deflate
0.30% 0.03% libgit2_clar libgit2_clar [.] deflate_fast
0.28% 0.15% libgit2_clar libgit2_clar [.] inflate_fast_chunk_
0.19% 0.19% libgit2_clar libgit2_clar [.] inflate_table
0.17% 0.01% libgit2_clar libgit2_clar [.] inflateInit_
0.16% 0.00% libgit2_clar libgit2_clar [.] inflateInit2_ (inlined)
0.15% 0.00% libgit2_clar libgit2_clar [.] deflateInit_
0.15% 0.00% libgit2_clar libgit2_clar [.] deflateInit2_
0.11% 0.01% libgit2_clar libgit2_clar [.] adler32_z
0.09% 0.09% libgit2_clar libgit2_clar [.] adler32_simd_
0.05% 0.00% libgit2_clar libgit2_clar [.] deflateReset (inlined)
0.05% 0.00% libgit2_clar libgit2_clar [.] deflate_read_buf
0.03% 0.00% libgit2_clar libgit2_clar [.] inflateEnd
0.02% 0.00% libgit2_clar libgit2_clar [.] deflateReset
0.01% 0.00% libgit2_clar libgit2_clar [.] deflateEnd
0.01% 0.01% libgit2_clar libgit2_clar [.] inflateReset2
0.01% 0.00% libgit2_clar libgit2_clar [.] inflateReset (inlined)
0.00% 0.00% libgit2_clar libgit2_clar [.] adler32
0.00% 0.00% libgit2_clar libgit2_clar [.] inflateResetKeep (inlined)
0.00% 0.00% libgit2_clar libgit2_clar [.] deflateResetKeep
0.00% 0.00% libgit2_clar libgit2_clar [.] inflateStateCheck (inlined)
0.00% 0.00% libgit2_clar libgit2_clar [.] inflateStateCheck (inlined)
0.00% 0.00% libgit2_clar libgit2_clar [.] inflateStateCheck (inlined)
0.00% 0.00% libgit2_clar libgit2_clar [.] deflateStateCheck (inlined)
```
|
|
|
fe41e582
|
2020-12-23T12:32:58
|
|
Merge pull request #5741 from libgit2/ethomson/ipv6
Handle ipv6 addresses
|
|
|
27301cd0
|
2020-12-17T20:53:11
|
|
net: function to identify ipv6 addresses in URLs
|
|
|
953dee20
|
2020-12-17T20:51:37
|
|
net: is_default_port is a bool
|
|
|
b7ffc63b
|
2020-12-17T18:38:01
|
|
winhttp: handle ipv6 addresses
|
|
|
2807de5c
|
2020-12-17T15:18:31
|
|
http: handle ipv6 addresses
|
|
3433acd9
|
2020-12-21T21:27:58
|
|
Wrap newer hostkeys in #ifdefs
This allows the library to be built using a pre-1.9.0 version of libssh2.
|
|
|
ed7b20e7
|
2020-12-21T17:26:34
|
|
Add support for additional hostkey types.
Specifically: ECDSA_256, ECDSA_384, ECDSA_521 and ED25519.
|
|
|
f4b473fa
|
2020-12-20T21:25:22
|
|
Merge pull request #5746 from libgit2/ethomson/configmapcache
repository: use intptr_t's in the config map cache
|
|
|
1b708683
|
2020-12-20T20:42:40
|
|
config_cache: cleanup config map lookup
|
|
|
f52f593e
|
2020-12-20T16:33:42
|
|
repository: use intptr_t's in the config map cache
Since we're using atomic primitives to read and write into the config
map cache, we need to read/write something pointer-sized. Use an
`intptr_t` for the config map cache.
|
|
|
5ab0736b
|
2020-12-19T09:30:26
|
|
Add tests for `git__multiply_int64_overflow`
As it turns out, the implementation of `git__multiply_int64_overflow` is
full of edge cases and tricky arithmetic. That means that it should have
unit tests.
As a result, a bug in `git__strntol64` was found (and fixed!) in
clang+32-bit.
|
|
|
e9b98cd5
|
2020-12-18T19:43:13
|
|
Third attempt to fix the 32-bit version of `git__multiply_int64_overflow`
This change should now fix the issue for realsies.
`./libgit2_clar -score::strtol` passes on a 32-bit Docker.
|
|
|
7f8ae019
|
2020-12-18T07:03:21
|
|
Avoid using `__builtin_mul_overflow` with the clang+32-bit combo
This causes clang to produce an undefined reference to `__mulodi4`. This
could be fixed by statically linking some compiler-rt libraries to
provide this symbol, but let's first stop the bleeding since doing the
correct long-term fix requires some non-trivial CMake knowledge which I
lack.
|
|
|
1865806e
|
2020-12-16T00:16:10
|
|
Merge pull request #5740 from libgit2/ethomson/zlib
pack: continue zlib while we can make progress
|
|
|
93f61c5a
|
2020-12-15T23:03:03
|
|
pack: continue zlib while we can make progress
Continue the zlib stream as long as we can make progress; stop when we
stop getting output _or_ when zlib stops taking input from us.
|
|
|
d185ab24
|
2020-12-13T15:18:12
|
|
Merge pull request #5727 from lhchavez/make-git-diff-fast
Cache the parsed submodule config when diffing
|
|
|
851e778e
|
2020-12-13T13:46:16
|
|
Merge pull request #5735 from lhchavez/faster-strntol64
Make git__strntol64() ~70%* faster
|
|
|
86a1cdd3
|
2020-12-13T13:44:56
|
|
Merge pull request #5384 from ianhattendorf/fix/winhttp-client-cert
winhttp: support optional client cert
|
|
|
e99e833f
|
2020-12-12T16:51:15
|
|
Make git__strntol64() ~70%* faster
This change uses compiler intrinsics to detect overflows instead of
using divisions to detect potential overflow. This makes the code faster
and makes it easier to read as a bonus side-effect!
Some of the things this quickens:
* Config parsing.
* Tree parsing.
* Smart protocol negotiation.
\* Measured by running `libgit2_clar` with `-fno-optimize-sibling-calls
-fno-omit-frame-pointer` under `perf(1)`:
```shell
$ perf diff --symbols=git__strntol64 --compute=ratio \
--percentage=absolute baseline.data perf.data
\# Event 'cycles'
\#
\# Baseline Ratio Shared Object
\# ........ .............. .............
\#
0.25% 0.321836 libgit2_clar
```
|
|
|
1ab2f577
|
2020-12-11T05:13:16
|
|
Fix an oops
|
|
58d757b1
|
2016-11-14T16:52:33
|
|
patch: add owner accessor
|
|
|
6cd0c853
|
2020-12-11T05:08:45
|
|
Small refactor to make thing tidier
Also repurposed an unused function and deleted another one.
|
|
|
41da4e16
|
2020-12-10T19:52:01
|
|
Cache the parsed submodule config when diffing
This change makes that anything that calls `git_diff__from_iterators`
(any of the `git_diff_xxx` functions) only need to parse the
`.gitmodules` file once. This can be avoided by calling
`git_repository_submodule_cache_all(...)`, but we can do that safely for
the user with no change in semantics.
Fixes: #5725
|
|
|
08f28ff5
|
2020-07-10T08:57:05
|
|
alloc: set up an allocator that fails before library init
We require the library to be initialized with git_libgit2_init before it
is functional. However, if a user tries to uses the library without
doing so - as they might when getting started with the library for the
first time - we will likely crash.
This commit introduces some guard rails - now instead of having _no_
allocator by default, we'll have an allocator that always fails, and
never tries to set an error message (since the thread-local state is
set up by git_libgit2_init). We've modified the error retrieval
function to (try to) ensure that the library has been initialized before
getting the thread-local error message.
(Unfortunately, we cannot determine if the thread local storage has
actually been configured, this does require initialization by
git_libgit2_init. But a naive attempt should be good enough for most
cases.)
|
|
|
6c51014d
|
2020-07-11T14:24:17
|
|
libgit2: provide init_count of the library
A function to provide the initialization count of the library; this is
subject to race conditions but is useful for a naive determination as to
whether the library has been initialized or not.
|
|
20ce17f0
|
2020-08-06T16:24:27
|
|
Replace global storage TLS with new interface
|
|
|
fd2d4759
|
2020-12-05T22:22:38
|
|
threads: git_tls_data to git_tlsdata
Use a no-allocation approach to the TLS data abstraction.
|
|
42d5f110
|
2020-12-08T14:33:31
|
|
win32: provide macros for calling conventions
Provide more clarity for Win32 calling conventions - now that we always
compile to __cdecl on Win32, we'll define that as the the libgit2
calling convention. Also offer NTAPI (__stdcall) calling conventions
for things that need callbacks from NTAPI code (eg fiber-local storage).
|
|
|
971b7187
|
2020-08-06T16:24:18
|
|
Implement generic TLS interface
This adds a generic TLS interface for anyone to store TLS data. It is designed to work regardless of whether threading support is built into the library or not.
Nobody in the library should directly interface with the data on the TLS struct, so it's been built to be opaque even in the library.
Requires the allocator to be initialized before use.
|
|
|
404dd024
|
2020-12-05T15:57:48
|
|
threads: rename thread files to thread.[ch]
|
|
|
ab772974
|
2020-12-05T15:49:30
|
|
threads: give atomic functions the git_atomic prefix
|
|
|
37763d38
|
2020-12-05T15:26:59
|
|
threads: rename git_atomic to git_atomic32
Clarify the `git_atomic` type and functions now that we have a 64 bit
version as well (`git_atomic64`).
|
|
|
9800728a
|
2020-12-05T16:08:34
|
|
util: move git_online_cpus into util
The number of CPUs is useful information for creating a thread pool or a
number of workers, but it's not really about threading directly. Evict
it from the thread file
|
|
|
fe07ef8a
|
2020-12-05T21:38:29
|
|
packbuilder: use git__noop on non-threaded systems
Our git_packbuilder__cache_lock function returns a value; use git__noop.
|
|
|
8413c0f9
|
2020-12-05T21:32:48
|
|
util: move git__noop into the util header
The git__noop function is more largely useful; move it into the
util header. (And reduce the number of underscores.)
|
|
|
322c15ee
|
2020-08-01T18:24:41
|
|
Make the pack and mwindow implementations data-race-free
This change fixes a packfile heap corruption that can happen when
interacting with multiple packfiles concurrently across multiple
threads. This is exacerbated by setting a lower mwindow open file limit.
This change:
* Renames most of the internal methods in pack.c to clearly indicate
that they expect to be called with a certain lock held, making
reasoning about the state of locks a bit easier.
* Splits the `git_pack_file` lock in two: the one in `git_pack_file`
only protects the `index_map`. The protection to `git_mwindow_file` is
now in that struct.
* Explicitly checks for freshness of the `git_pack_file` in
`git_packfile_unpack_header`: this allows the mwindow implementation
to close files whenever there is enough cache pressure, and
`git_packfile_unpack_header` will reopen the packfile if needed.
* After a call to `p_munmap()`, the `data` and `len` fields are poisoned
with `NULL` to make use-after-frees more evident and crash rather than
being open to the possibility of heap corruption.
* Adds a test case to prevent this from regressing in the future.
Fixes: #5591
|
|
|
4ae41f9c
|
2020-08-02T16:26:25
|
|
Make the odb race-free
This change adds all the necessary locking to the odb to avoid races in
the backends.
Part of: #5592
|
|
|
ef6de8d5
|
2020-11-27T14:46:38
|
|
Merge pull request #5704 from lhchavez/ssh-raw-certificate
Also add the raw hostkey to `git_cert_hostkey`
|