|
9487585d
|
2015-12-01T14:19:29
|
|
tree: mark cloned tree entries as un-pooled
When duplicating a `struct git_tree_entry` with
`git_tree_entry_dup` the resulting structure is not allocated
inside a memory pool. As we do a 1:1 copy of the original struct,
though, we also copy the `pooled` field, which is set to `true`
for pooled entries. This results in a huge memory leak as we
never free tree entries that were duplicated from a pooled
tree entry.
Fix this by marking the newly duplicated entry as un-pooled.
|
|
337b2b08
|
2015-11-30T20:53:54
|
|
Merge pull request #3508 from libgit2/cmn/tree-parse-speed
Improvements to tree parsing speed
|
|
|
bbd2fa4e
|
2015-11-30T18:05:27
|
|
object: remove unused constant OBJECT_BASE_SIZE
|
|
95ae3520
|
2015-11-30T17:32:18
|
|
tree: ensure the entry filename fits in 16 bits
Return an error in case the length is too big. Also take this
opportunity to have a single allocating function for the size and
overflow logic.
|
|
|
a27f31d8
|
2015-11-30T04:49:31
|
|
Merge pull request #3513 from ethomson/merge_recursive
Recursive Merge
|
|
|
ee42bb0e
|
2015-11-28T19:18:29
|
|
tree: make path len uint16_t and avoid holes
This reduces the size of the struct from 32 to 26 bytes, and leaves a
single padding byte at the end of the struct (which comes from the
zero-length array).
|
|
|
2580077f
|
2015-11-15T00:44:02
|
|
tree: calculate the filename length once
We already know the size due to the `memchr()` so use that information
instead of calling `strlen()` on it.
|
|
|
ed970748
|
2015-11-14T23:50:06
|
|
tree: pool the entry memory allocations
These are rather small allocations, so we end up spending a non-trivial
amount of time asking the OS for memory. Since these entries are tied to
the lifetime of their tree, we can give the tree a pool so we speed up
the allocations.
|
|
|
7132150d
|
2015-11-14T23:46:21
|
|
tree: avoid advancing over the filename multiple times
We've already looked at the filename with `memchr()` and then used
`strlen()` to allocate the entry. We already know how much we have to
advance to get to the object id, so add the filename length instead of
looking at each byte again.
|
|
|
5b9c63c3
|
2015-11-20T19:01:42
|
|
recursive merge: add a recursion limit
|
|
78859c63
|
2015-11-20T17:33:49
|
|
merge: handle conflicts in recursive base building
When building a recursive merge base, allow conflicts to occur.
Use the file (with conflict markers) as the common ancestor.
The user has already seen and dealt with this conflict by virtue
of having a criss-cross merge. If they resolved this conflict
identically in both branches, then there will be no conflict in the
result. This is the best case scenario.
If they did not resolve the conflict identically in the two branches,
then we will generate a new conflict. If the user is simply using
standard conflict output then the results will be fairly sensible.
But if the user is using a mergetool or using diff3 output, then the
common ancestor will be a conflict file (itself with diff3 output,
haha!). This is quite terrible, but it matches git's behavior.
|
|
|
76ade3a0
|
2015-11-10T21:21:26
|
|
merge: use annotated commits for recursion
Use annotated commits to act as our virtual bases, instead of regular
commits, to avoid polluting the odb with virtual base commits and
trees. Instead, build an annotated commit with an index and pointers
to the commits that it was merged from.
|
|
|
7730fe8e
|
2015-11-09T13:01:48
|
|
merge: merge annotated commits instead of regular commits
|
|
|
3f2bb387
|
2015-10-28T11:00:55
|
|
merge: octopus merge common ancestors when >2
When there are more than two common ancestors, continue merging the
virtual base with the additional common ancestors, effectively
octopus merging a new virtual base.
|
|
|
1b82f7b6
|
2015-10-27T14:24:51
|
|
merge: compute octopus merge bases
|
|
|
75dee59c
|
2015-10-26T10:37:58
|
|
merge: build virtual base of multiple merge bases
When the commits to merge have multiple common ancestors, build a
"virtual" base tree by merging the common ancestors.
|
|
|
fa78782f
|
2015-10-22T17:00:09
|
|
merge: rename `git_merge_tree_flags_t` -> `git_merge_flags_t`
|
|
|
25e84f95
|
2015-11-23T15:49:54
|
|
checkout: only consider nsecs when built that way
When examining the working directory and determining whether it's
up-to-date, only consider the nanoseconds in the index entry when
built with `GIT_USE_NSEC`. This prevents us from believing that
the working directory is always dirty when the index was originally
written with a git client that uinderstands nsecs (like git 2.x).
|
|
|
66e4e6da
|
2015-11-21T11:26:18
|
|
Merge pull request #3515 from jacquesg/unsigned-signed
Fix <0 unsigned comparison (stat.st_size should be an off_t)
|
|
e78e8fae
|
2015-11-20T20:22:38
|
|
Make stat.st_size a __int64 not a uint64_t
|
|
|
7a612b53
|
2015-11-20T17:52:00
|
|
Merge pull request #3514 from jacquesg/stat-fixes
Stat fixes
|
|
|
eb11fac6
|
2015-11-20T18:57:13
|
|
Detect stat's structure
|
|
|
2ea40fda
|
2015-11-20T13:19:23
|
|
repository: distinguish sequencer cherry-pick and revert
These are not quite like their plain counterparts and require special handling.
|
|
|
69d14948
|
2015-11-17T13:21:18
|
|
Merge pull request #3511 from ethomson/racy_fixes_2
Racy fixes for writing new indexes
|
|
|
3eac1037
|
2015-11-16T23:31:19
|
|
settings: allow users to set PROGRAMDATA
Allow users to set the `git_libgit2_opts` search path for the
`GIT_CONFIG_LEVEL_PROGRAMDATA`. Convert `GIT_CONFIG_LEVEL_PROGRAMDATA`
to `GIT_SYSDIR_PROGRAMDATA` for setting the configuration.
|
|
|
5f32c506
|
2015-11-16T18:06:52
|
|
racy: make git_index_read_index handle raciness
Ensure that `git_index_read_index` clears the uptodate bit on
files that it modifies.
Further, do not propagate the cache from an on-disk index into
another on-disk index. Although this should not be done, as
`git_index_read_index` is used to bring an in-memory index into
another index (that may or may not be on-disk), ensure that we do
not accidentally bring in these bits when misused.
|
|
|
27bc41cf
|
2015-11-13T16:31:51
|
|
index: clear uptodate bit on save
The uptodate bit should have a lifecycle of a single read->write
on the index. Once the index is written, the files within it should
be scanned for racy timestamps against the new index timestamp.
|
|
|
d1101263
|
2015-11-13T15:32:48
|
|
index: don't detect raciness in uptodate entries
Keep track of entries that we believe are up-to-date, because we
added the index entries since the index was loaded. This prevents
us from unnecessarily examining files that we wrote during the
cleanup of racy entries (when we smudge racily clean files that have
a timestamp newer than or equal to the index's timestamp when we
read it). Without keeping track of this, we would examine every
file that we just checked out for raciness, since all their timestamps
would be newer than the index's timestamp.
|
|
|
cb0ff012
|
2015-11-06T17:15:35
|
|
racy-git: do a single index->workdir diff
When examining paths that are racily clean, do a single index->workdir
diff over the entirety of the racily clean files, instead of a diff
per file.
|
|
7ff7ca62
|
2015-11-12T20:51:01
|
|
pool: Never return unaligned buffers
|
|
|
75a0ccf5
|
2015-11-12T19:53:09
|
|
Merge pull request #3170 from CmdrMoozy/nsec_fix
git_index_entry__init_from_stat: set nsec fields in entry stats
|
|
|
2c26c867
|
2015-11-12T19:22:31
|
|
Merge pull request #3499 from ethomson/ref_dir_errmsgs
Improve error messages when dirs prevent ref/reflog creation
|
|
|
ecdc0428
|
2015-11-12T19:20:36
|
|
Merge pull request #3448 from libgit2/cmn/custom-agent
Support setting custom user-agent
|
|
|
de870533
|
2015-10-02T03:43:11
|
|
settings: add a setter for a custom user-agent
|
|
|
1411cb9e
|
2015-10-02T03:57:14
|
|
winhttp: use a custom user-agent if the user has set it
We also keep the "git/1.0" prefix in order to maintain compatibility
with hosters.
|
|
|
94bac76c
|
2015-10-02T03:46:34
|
|
http: use a custom user-agent if the user has set it
We still prefix it with "git/1.0" since that's required in many
situations, but we replace the area which mentions libgit2.
|
|
|
0a700ee3
|
2015-11-03T17:34:54
|
|
reflog: error when a directory is at reflog path
When a non-empty directory exists and prevents the creation of a
reflog, provide a more informative error message.
|
|
|
b46c7ee5
|
2015-11-03T17:18:00
|
|
refs: complain when a directory exists at ref
When a (non-empty) directory exists at the reference target
location, complain with a more actionable error message.
|
|
|
ad8509ef
|
2015-11-12T11:54:06
|
|
index: overwrite the path when inserting conflicts
When we insert a conflict in a case-insensitive index, accept the
new entry's path as the correct case instead of leaving the path we
already had.
This puts `git_index_conflict_add()` on the same level as
`git_index_add()` in this respect.
|
|
|
16604d74
|
2015-11-11T00:36:15
|
|
index: correctly report which conflict stage has a wrong filemode
When we're at offset 'i', we're dealing with the 'i+1' stage, since
conflicts start at 1.
|
|
|
1c34b717
|
2015-11-08T05:10:18
|
|
Merge pull request #3498 from ethomson/windows_symlinks
Diff: Honor `core.symlinks=false` and fake symlinks
|
|
|
61948894
|
2015-11-08T04:54:55
|
|
Merge pull request #3500 from ethomson/submodules_with_path
Handle submodules with paths in `git_submodule_update`
|
|
|
f4b02671
|
2015-11-04T16:17:51
|
|
submodule: reload HEAD/index after reading config
Reload the HEAD and index data for a submodule after reading the
configuration. The configuration may specify a `path`, so we must
update HEAD and index data with that path in mind.
|
|
|
5aa28a8f
|
2015-11-04T14:16:24
|
|
Merge pull request #3465 from libgit2/cmn/tls-register
stream: allow registering a user-provided TLS constructor
|
|
f5f96a23
|
2015-10-09T10:41:06
|
|
Fix git_commit_summary to convert newlines to spaces even after
whitespace. Collapse spaces around newlines for the summary.
|
|
|
ec50b23a
|
2015-11-03T17:02:07
|
|
filebuf: detect directories in our way
When creating a filebuf, detect a directory that exists in our
target file location. This prevents a failure later, when we try
to move the lock file to the destination.
|
|
|
6b0fc6ab
|
2015-11-03T09:43:18
|
|
diff: on win32, treat fake "symlinks" specially
On platforms that lack `core.symlinks`, we should not go looking for
symbolic links and `p_readlink` their target. Instead, we should
examine the file's contents.
|
|
|
7fafde63
|
2015-10-13T11:25:41
|
|
stream: allow registering a user-provided TLS constructor
This allows the application to use their own TLS stream, regardless of
the capabilities of libgit2 itself.
|
|
|
d39f643a
|
2015-10-13T19:34:07
|
|
stream: accept NULL in the free function
|
|
3138ad93
|
2015-07-16T10:17:16
|
|
Add diff progress callback.
|
|
|
db1edf91
|
2015-11-02T15:09:19
|
|
Merge pull request #3491 from libgit2/cmn/config-checksum
Use checksums to detect config file changes
|
|
|
1d0bed9d
|
2015-10-30T14:02:01
|
|
merge-base: Style
|
|
|
4cacf5b5
|
2015-10-30T11:50:43
|
|
merge-base: Do not read parents from the root
|
|
|
136a71f4
|
2015-10-30T11:45:52
|
|
merge-base: Remove redundant merge bases
|
|
|
3547b122
|
2015-10-30T21:36:51
|
|
filebuf: use an internal buffer
This reduces the chances of a crash in the thread tests. This shouldn't
affect general usage too much, since the main usage of these functions
are to read into an empty buffer.
|
|
|
eb597799
|
2015-10-29T21:12:37
|
|
filebuf: use a checksum to detect file changes
Instead of relying on the size and timestamp, which can hide changes
performed in the same second, hash the file content's when we care about
detecting changes.
|
|
|
d571a54e
|
2015-10-30T19:36:16
|
|
Merge pull request #3493 from ethomson/read_index
index: read_index must update hashes
|
|
|
e0be1d60
|
2015-10-30T18:23:17
|
|
git: put NULL check at the top
|
|
|
0bf77e32
|
2015-10-30T13:07:02
|
|
index: read_index must update hashes
|
|
|
45d295e0
|
2015-10-30T17:51:50
|
|
git: accept NULL as argument to its stream free
|
|
|
fd74bd08
|
2015-10-29T20:37:48
|
|
Merge pull request #3486 from srajko/reflog-segfault-fix
Fix segfault when reading reflog with extra newlines
|
|
|
335c9e2f
|
2015-10-26T15:33:00
|
|
Prevent segfault when parsing a reflog with oid parse error
Using calloc instead of malloc because the parse error will lead to an immediate free of committer (and its properties, which can segfault on free if undefined - test_refs_reflog_reflog__reading_a_reflog_with_invalid_format_returns_error segfaulted before the fix).
#3458
|
|
|
1b4449b4
|
2015-10-28T10:53:03
|
|
pool: fix documentation
|
|
|
d845abe6
|
2015-10-28T14:49:28
|
|
merge: Do not mallocz unecessary entries
|
|
|
340b15b7
|
2015-10-28T14:31:09
|
|
pool: update comment
|
|
|
d3416dfe
|
2015-10-28T10:50:25
|
|
pool: Dot not assume mallocs are zeroed out
|
|
|
410efda8
|
2015-10-28T10:28:43
|
|
pool: Take into account malloc overhead & pool page size
|
|
|
4f971852
|
2015-10-28T10:15:24
|
|
repository: plug memory leak
cc @carlosmn
|
|
|
1e5e02b4
|
2015-10-27T17:26:04
|
|
pool: Simplify implementation
|
|
|
7a02e93e
|
2015-10-27T22:42:40
|
|
merge: Plug memory leak
|
|
|
a1f5d691
|
2015-10-27T22:42:15
|
|
merge: Implement `GIT_MERGE_TREE_SKIP_REUC`
|
|
|
d307a013
|
2015-10-27T22:17:32
|
|
reuc: Be smarter when inserting new REUC entries
Inserting new REUC entries can quickly become pathological given that
each insert unsorts the REUC vector, and both subsequent lookups *and*
insertions will require sorting it again before being successful.
To avoid this, we're switching to `git_vector_insert_sorted`: this keeps
the REUC vector constantly sorted and lets us use the `on_dup` callback
to skip an extra binary search on each insertion.
|
|
|
e324005e
|
2015-10-23T15:13:38
|
|
Merge pull request #3481 from ethomson/xdiff_include
xdiff: reference util.h in parent directory
|
|
|
8683d31f
|
2015-10-22T14:39:20
|
|
merge: add GIT_MERGE_TREE_FAIL_ON_CONFLICT
Provide a new merge option, GIT_MERGE_TREE_FAIL_ON_CONFLICT, which
will stop on the first conflict and fail the merge operation with
GIT_EMERGECONFLICT.
|
|
|
dc2cf3eb
|
2015-10-22T18:35:43
|
|
Merge pull request #3480 from ethomson/nsecs
Nanoseconds in the index: ignore for diffing
|
|
|
c7b336b0
|
2015-10-22T10:29:51
|
|
xdiff: reference util.h in parent directory
Although CMake will correctly configure include directories for us,
some people may use their own build system, and we should reference
`util.h` based on where it actually lives.
|
|
|
7499eae9
|
2015-10-22T09:30:41
|
|
diff: ignore nsecs when diffing
Although our index contains the literal time present in the index,
we do not read nanoseconds from disk, and thus we should not use
them in any comparisons, lest we always think our working directory
is dirty.
Guard this behind a `GIT_USE_NSECS` for future improvement.
|
|
|
44b1e3e3
|
2015-10-21T13:43:22
|
|
Merge pull request #3475 from libgit2/cmn/programdata-config
config: add a ProgramData level
|
|
|
8c7c5fa5
|
2015-10-20T17:42:42
|
|
config: add a ProgramData level
This is where portable git stores the global configuration which we can
use to adhere to it even though git isn't quite installed on the system.
|
|
|
128e94bb
|
2015-10-21T12:04:53
|
|
index: Remove unneeded consts
|
|
|
307c4a2b
|
2015-10-21T11:58:44
|
|
signature: Strip crud just like Git does
|
|
|
4280fabb
|
2015-10-15T07:10:48
|
|
Merge pull request #3466 from libgit2/cmn/quick-parse-64
revwalk: make commit list use 64 bits for time
|
|
|
8321596a
|
2015-10-15T12:22:10
|
|
Merge pull request #3444 from ethomson/add_preserves_conflict_mode
Preserve modes from a conflict in `git_index_insert`
|
|
|
a0a1b19a
|
2015-10-14T19:31:54
|
|
odb: Prioritize alternate backends
For most real use cases, repositories with alternates use them as main
object storage. Checking the alternate for objects before the main
repository should result in measurable speedups.
Because of this, we're changing the sorting algorithm to prioritize
alternates *in cases where two backends have the same priority*. This
means that the pack backend for the alternate will be checked before the
pack backend for the main repository *but* both of them will be checked
before any loose backends.
|
|
|
43820f20
|
2015-10-14T19:24:07
|
|
odb: Be smarter when refreshing backends
In the current implementation of ODB backends, each backend is tasked
with refreshing itself after a failed lookup. This is standard Git
behavior: we want to e.g. reload the packfiles on disk in case they have
changed and that's the reason we can't find the object we're looking
for.
This behavior, however, becomes pathological in repositories where
multiple alternates have been loaded. Given that each alternate counts
as a separate backend, a miss in the main repository (which can
potentially be very frequent in cases where object storage comes from
the alternate) will result in refreshing all its packfiles before we
move on to the alternate backend where the object will most likely be
found.
To fix this, the code in `odb.c` has been refactored as to perform the
refresh of all the backends externally, once we've verified that the
object is nowhere to be found.
If the refresh is successful, we then perform the lookup sequentially
through all the backends, skipping the ones that we know for sure
weren't refreshed (because they have no refresh API).
The on-disk pack backend has been adjusted accordingly: it no longer
performs refreshes internally.
|
|
|
5ffdea6f
|
2015-10-14T16:49:01
|
|
revwalk: make commit list use 64 bits for time
We moved the "main" parsing to use 64 bits for the timestamp, but the
quick parsing for the revwalk did not. This means that for large
timestamps we fail to parse the time and thus the walk.
Move this parser to use 64 bits as well.
|
|
|
ae195a71
|
2015-09-29T12:46:41
|
|
blame: guard xdiff calls for large files
|
|
|
6c014bcc
|
2015-09-29T12:18:17
|
|
diff: don't feed large files to xdiff
|
|
|
e4352066
|
2015-09-28T18:25:24
|
|
merge_file: treat large files as binary
xdiff craps the bed on large files. Treat very large files as binary,
so that it doesn't even have to try.
Refactor our merge binary handling to better match git.git, which
looks for a NUL in the first 8000 bytes.
|
|
|
46c0e6e3
|
2015-09-28T16:34:29
|
|
xdiff: convert size variables to size_t
|
|
|
4bc9b74c
|
2015-09-28T16:24:50
|
|
GITERR_CHECK_ALLOC_ADDn: multi-arg adders
|
|
28659e50
|
2015-10-01T18:36:10
|
|
diff: refactor complex timestamp check into its own function
|
|
|
973a09a4
|
2015-10-01T18:35:21
|
|
index: don't populate nsec values if GIT_USE_NSEC is off
|
|
|
c7b17fb5
|
2015-10-01T18:01:32
|
|
Merge branch 'master' into nsec_fix_next
|
|
d3b29fb9
|
2015-10-01T00:50:37
|
|
refdb and odb backends must provide `free` function
As refdb and odb backends can be allocated by client code, libgit2
can’t know whether an alternative memory allocator was used, and thus
should not try to call `git__free` on those objects.
Instead, odb and refdb backend implementations must always provide
their own `free` functions to ensure memory gets freed correctly.
|
|
|
ba1a5553
|
2015-09-30T17:44:10
|
|
Merge pull request #3446 from ethomson/portability
portability: use `CHECK_FUNCTION_EXISTS` for checking whether functions exist...
|
|
|
21515f22
|
2015-09-29T15:49:16
|
|
index: also try conflict mode when inserting
When we do not trust the on-disk mode, we use the mode of an existing
index entry. This allows us to preserve executable bits on platforms
that do not honor them on the filesystem.
If there is no stage 0 index entry, also look at conflicts to attempt
to answer this question: prefer the data from the 'ours' side, then
the 'theirs' side before falling back to the common ancestor.
|
|
|
e683d152
|
2015-09-30T05:49:04
|
|
qsort_r/qsort_s: detect their support
|
|
|
8649dfd8
|
2015-09-29T13:36:37
|
|
p_futimes: support using futimens when available
|
|
|
146a96de
|
2015-09-30T09:41:25
|
|
openssl: don't try to teardown an unconnected SSL context
SSL_shutdown() does not like it when we pass an unitialized ssl context
to it. This means that when we fail to connect to a host, we hide the
error message saying so with OpenSSL's indecipherable error message.
|