|
ac2b235e
|
2019-05-21T12:22:40
|
|
regex: use REGEX_BACKEND as the cmake option name
This avoids any misunderstanding with the REGEX keyword in cmake.
|
|
|
ce6d624a
|
2019-05-19T10:30:04
|
|
regex: optionally use PCRE2
Use PCRE2 and its POSIX compatibility layer if requested by the user.
Although PCRE2 is adequate for our needs, the PCRE2 POSIX layer as
installed on Debian and Ubuntu systems is broken, so we do not opt-in to
it by default to avoid breaking users on those platforms.
|
|
|
622166c4
|
2019-05-18T19:37:59
|
|
regex: disambiguate builtin vs system pcre
|
|
|
9ceafb57
|
2019-01-12T22:55:31
|
|
regexec: use pcre as our fallback/builtin regex
Use PCRE 8.42 as the builtin regex implementation, using its POSIX
compatibility layer. PCRE uses ASCII by default and the users locale
will not influence its behavior, so its `regcomp` implementation is
similar to `regcomp_l` with a C locale.
|
|
|
c6e48fef
|
2019-02-17T21:51:34
|
|
regex: allow regex selection in cmake
Users can now select which regex implementation they want to use: one of
the system `regcomp_l`, the system PCRE, the builtin PCRE or the
system's `regcomp`.
By default the system `regcomp_l` will be used if it exists, otherwise
the system PCRE will be used. If neither of those exist, then the
builtin PCRE implementation will be used.
The system's `regcomp` is not used by default due to problems with
locales.
|
|
|
69ecdad5
|
2019-05-19T10:09:55
|
|
regex: use system PCRE if available
Attempt to locate a system-installed version of PCRE and use its POSIX
compatibility layer, if possible.
|
|
|
fe1fb36e
|
2019-01-13T21:10:50
|
|
win32: move type definitions for improved inclusion
Move some win32 type definitions to a standalone file so that they can
be included before other header files try to use the definitions.
|
|
31f8f82a
|
2018-03-02T12:18:59
|
|
diff_driver: detect memory allocation errors when loading diff driver
When searching for a configuration key for the diff driver, we construct
the config key by modifying a buffer and then passing it to
`git_config_get_multivar_foreach`. We do not check though whether the
modification of the buffer actually succeded, so we could in theory end
up passing the OOM buffer to the config function.
Fix that by checking return codes. While at it, switch to use
`git_buf_PUTS` to avoid repetition of the appended string to calculate
its length.
|
|
|
02683b20
|
2019-01-12T23:06:39
|
|
regexec: prefix all regexec function calls with p_
Prefix all the calls to the the regexec family of functions with `p_`.
This allows us to swap out all the regular expression functions with our
own implementation. Move the declarations to `posix_regex.h` for
simpler inclusion.
|
|
|
c9f116f1
|
2019-05-12T22:06:00
|
|
Merge branch 'pr/5061'
|
|
|
ab27c835
|
2019-05-12T22:05:26
|
|
revwalk: update error message for clarity
|
|
|
1e3a639d
|
2019-05-12T21:54:39
|
|
Merge pull request #5065 from danielgindi/feature/win32_symlink_dir
Support symlinks for directories in win32
|
|
|
7f562f2c
|
2019-05-12T11:00:31
|
|
Merge pull request #5057 from eaigner/merge-rebase-onto-name
rebase: orig_head and onto accessors
|
|
6990a492
|
2019-05-06T11:39:51
|
|
revwalk: fix memory leak in error handling
This is not implemented and should fail, but it should also not leak. To
allow the memory debugger to find leaks and fix this one we test this.
|
|
336e98bb
|
2019-05-06T14:51:52
|
|
Moved dwFlags declaration to beginning of scope
|
|
|
37a7adb5
|
2019-05-05T07:49:09
|
|
Support symlinks for directories in win32
|
|
|
d55bb479
|
2019-04-26T15:59:49
|
|
git_revwalk_push_range: do not crash if range is missing
If someone passes just one ref (i.e. "master") and misses passing the
range we should be nice and return an error code instead of crashing.
|
|
|
604e2811
|
2019-05-02T12:09:23
|
|
Merge pull request #5063 from pks-t/pks/cmake-regcomp-fix
cmake: correctly detect if system provides `regcomp`
|
|
|
ee3d71fb
|
2019-04-26T08:01:56
|
|
cmake: fix include ordering issues with bundled deps
When linking against bundled libraries, we include their header
directories by using "-isystem". The reason for that is that we
want to handle our vendored library headers specially, most
importantly to ignore warnings generated by including them. By
using "-isystem", though, we screw up the order of searched
include directories by moving those bundled dependencies towards
the end of the lookup order. Like this, chances are high that any
other specified include directory contains a file that collides
with the actual desired include file.
Fix this by not treating the bundled dependencies' include
directories as system includes. This will move them to the front
of the lookup order and thus cause them to override
system-provided headers. While this may cause the compiler to
generate additional warnings when processing bundled headers,
this is a tradeoff we should make regardless to fix builds on
systems hitting this issue.
|
|
|
13cb9f7a
|
2019-02-25T11:35:16
|
|
cmake: correctly detect if system provides `regcomp`
We assume that if we are on Win32, Amiga OS, Solaris or SunOS,
that the regcomp(3P) function cannot be provided by the system.
Thus we will in these cases always include our own, bundled regex
sources to make a regcomp implementation available. This test is
obviously very fragile, and we have seen it fail on MSYS2/MinGW
systems, which do in fact provide the regcomp symbol. The effect
is that during compilation, we will use the "regex.h" header
provided by MinGW, but use symbols provided by ourselves. This
in fact may cause subtle memory layout issues, as the structure
made available via MinGW doesn't match what our bundled code
expects.
There's one more problem with our regex detection: on the listed
platforms, we will incorrectly include the bundled regex code
even in case where the system provides regcomp_l(3), but it will
never be used for anything.
Fix the issue by improving our regcomp detection code. Instead of
relying on a fragile listing of platforms, we can just use
`CHECK_FUNCTION_EXISTS` instead. This will not in fact avoid the
header-ordering problem. But we can assume that as soon as a
system-provided "regex.h" header is provided, that
`CHECK_FUNCTION_EXISTS` will now correctly find the desired
symbol and thus not include our bundled regex code.
|
|
e44110db
|
2019-03-20T12:28:45
|
|
Correctly write to missing locked global config
Opening a default config when ~/.gitconfig doesn't exist, locking it,
and attempting to write to it causes an assertion failure.
Treat non-existent global config file content as an empty string.
|
|
|
bc5b19e6
|
2019-04-29T09:01:45
|
|
Merge pull request #4561 from pks-t/pks/downcasting
[RFC] util: introduce GIT_DOWNCAST macro
|
|
e215f475
|
2019-04-21T21:36:36
|
|
rebase: orig_head and onto accessors
The rebase struct stores fields with information about the current
rebase process, which were not accessible via a public interface.
Accessors for getting the `orig_head` and `onto` branch
names and object ids have been added.
|
|
|
b3923cf7
|
2019-04-17T13:43:52
|
|
Merge pull request #5050 from libgit2/ethomson/windows_init_traversal
git_repository_init: stop traversing at windows root
|
|
|
45f24e78
|
2019-04-12T08:54:06
|
|
git_repository_init: stop traversing at windows root
Stop traversing the filesystem at the Windows directory root. We were
calculating the filesystem root for the given directory to create, and
walking up the filesystem hierarchy. We intended to stop when the
traversal path length is equal to the root path length (ie, stopping at
the root, since no path may be shorter than the root path).
However, on Windows, the root path may be specified in two different
ways, as either `Z:` or `Z:\`, where `Z:` is the current drive letter.
`git_path_dirname_r` returns the path _without_ a trailing slash, even
for the Windows root. As a result, during traversal, we need to test
that the traversal path is _less than or equal to_ the root path length
to determine if we've hit the root to ensure that we stop when our
traversal path is `Z:` and our calculated root path was `Z:\`.
|
|
cc8a9892
|
2019-04-16T18:13:31
|
|
config_file: check result of git_array_alloc
git_array_alloc can return NULL if no memory is available, causing
a segmentation fault in memset. This adds GIT_ERROR_CHECK_ALLOC
similar to how other parts of the code base deal with the return
value of git_array_alloc.
|
|
431601f2
|
2019-04-05T15:05:10
|
|
iterator: make use the `GIT_CONTAINER_OF` macro
|
|
|
b51789ac
|
2019-04-16T13:20:08
|
|
transports: make use of the `GIT_CONTAINER_OF` macro
|
|
|
2e246474
|
2019-04-16T13:19:53
|
|
refdb_fs: make use of the `GIT_CONTAINER_OF` macro
|
|
|
65203b5a
|
2019-04-16T13:21:16
|
|
config_file: make use of `GIT_CONTAINER_OF` macro
|
|
|
b5f40441
|
2019-04-16T13:21:03
|
|
util: introduce GIT_CONTAINER_OF macro
In some parts of our code, we make rather heavy use of casting
structures to their respective specialized implementation. One
example is the configuration code with the general
`git_config_backend` and the specialized `diskfile_header`
structures. At some occasions, it can get confusing though with
regards to the correct inheritance structure, which led to the
recent bug fixed in 2424e64c4 (config: harden our use of the
backend objects a bit, 2018-02-28).
Object-oriented programming in C is hard, but we can at least try
to have some checks when it comes to casting around stuff. Thus,
this commit introduces a `GIT_CONTAINER_OF` macro, which accepts
as parameters the pointer that is to be casted, the pointer it
should be cast to as well as the member inside of the target
structure that is the containing structure. This macro then tries
hard to detect mis-casts:
- It checks whether the source and target pointers are of the
same type. This requires support by the compiler, as it makes
use of the builtin `__builtin_types_compatible_p`.
- It checks whether the embedded member of the target structure
is the first member. In order to make this a compile-time
constant, the compiler-provided `__builtin_offsetof` is being
used for this.
- It ties these two checks together by the compiler-builtin
`__builtin_choose_expr`. Based on whether the previous two
checks evaluate to `true`, the compiler will either compile in
the correct cast, or it will output `(void)0`. The second case
results in a compiler error, resulting in a compile-time check
for wrong casts.
The only downside to this is that it relies heavily on
compiler-specific extensions. As both GCC and Clang support these
features, only define this macro like explained above in case
`__GNUC__` is set (Clang also defines `__GNUC__`). If the
compiler is not Clang or GCC, just go with a simple cast without
any additional checks.
|
|
|
ed959ca2
|
2019-04-16T12:36:24
|
|
Merge pull request #5027 from ddevault/master
patch_parse.c: Handle CRLF in parse_header_start
|
|
|
c4cd69b2
|
2019-04-07T19:10:16
|
|
Merge pull request #5039 from libgit2/ethomson/win32_hash
sha1: don't inline `git_hash_global_init` for win32
|
|
30c06b60
|
2019-03-22T23:56:10
|
|
patch_parse.c: Handle CRLF in parse_header_start
|
|
|
9d117e20
|
2019-04-05T10:22:46
|
|
ignore: treat paths with trailing "/" as directories
The function `git_ignore_path_is_ignored` is there to test the
ignore status of paths that need not necessarily exist inside of
a repository. This has the implication that for a given path, we
cannot always decide whether it references a directory or a file,
and we need to distinguish those cases because ignore rules may
treat those differently. E.g. given the following gitignore file:
*
!/**/
we'd only want to unignore directories, while keeping files
ignored. But still, calling `git_ignore_path_is_ignored("dir/")`
will say that this directory is ignored because it treats "dir/"
as a file path.
As said, the `is_ignored` function cannot always decide whether
the given path is a file or directory, and thus it may produce
wrong results in some cases. While this is unfixable in the
general case, we can do better when we are being passed a path
name with a trailing path separator (e.g. "dir/") and always
treat them as directories.
|
|
|
aeea1c46
|
2019-04-04T15:06:44
|
|
Merge pull request #4874 from tiennou/test/4615
Test that largefiles can be read through the tree API
|
|
|
6bcb7357
|
2019-04-04T14:04:59
|
|
Merge pull request #5035 from pks-t/pks/diff-with-space-in-filenames
patch_parse: fix parsing addition/deletion of file with space
|
|
|
18e836cb
|
2019-04-04T10:55:38
|
|
Merge pull request #5018 from romkatv/strings
Optimize string comparisons
|
|
|
e5aecaf6
|
2019-04-04T18:45:30
|
|
sha1: don't inline `git_hash_global_init` for win32
Users of the Win32 hash cannot be inlined, as it uses a static struct.
Don't inline it, but continue to declare the function in the header.
|
|
30a56ba6
|
2019-03-14T14:54:47
|
|
optimize string comparisons
|
|
|
9aa049d4
|
2019-03-29T13:28:59
|
|
Merge pull request #5020 from implausible/fix/gitignore-negation
Negation of subdir ignore causes other subdirs to be unignored
|
|
|
b3497344
|
2019-03-29T12:15:20
|
|
patch_parse: fix parsing addition/deletion of file with space
The diff header format is a strange beast in that it is inherently
unparseable in an unambiguous way. While parsing
a/file.txt b/file.txt
is obvious and trivially doable, parsing a diff header of
a/file b/file ab.txt b/file b/file ab.txt
is not (but in fact valid and created by git.git).
Due to that, we have relaxed our diff header parser in commit 80226b5f6
(patch_parse: allow parsing ambiguous patch headers, 2017-09-22), so
that we started to bail out when seeing diff headers with spaces in
their file names. Instead, we try to use the "---" and "+++" lines,
which are unambiguous.
In some cases, though, we neither have a useable file name from the
header nor from the "---" or "+++" lines. This is the case when we have
a deletion or addition of a file with spaces: the header is unparseable
and the other lines will simply show "/dev/null". This trips our parsing
logic when we try to extract the prefix (the "a/" part) that is being
used in the path line, where we unconditionally try to dereference a
NULL pointer in such a scenario.
We can fix this by simply not trying to parse the prefix in cases where
we have no useable path name. That'd leave the parsed patch without
either `old_prefix` or `new_prefix` populated. But in fact such cases
are already handled by users of the patch object, which simply opt to
use the default prefixes in that case.
|
|
|
131cd9b1
|
2019-03-29T11:58:50
|
|
patch_parse: improve formatting
|
|
|
5f188c48
|
2019-03-29T11:52:39
|
|
Merge pull request #5024 from stewid/xdiff-fix-typo
xdiff: fix typo
|
|
be9a386c
|
2019-03-22T17:04:32
|
|
Each hash implementation should define `git_hash_global_init`
This means the forward declaration isn't necessary. The forward
declaration can cause compilation errors as it conflicts with the
`GIT_INLINE` declaration (the signatures are different).
|
|
1a349003
|
2019-03-20T21:20:01
|
|
xdiff: fix typo
|
|
e3d7bccb
|
2019-03-14T15:51:15
|
|
ignore: Do not match on prefix of negated patterns
Matching on the prefix of a negated pattern was triggering false
negatives on siblings of that pattern. e.g.
Given the .gitignore:
dir/*
!dir/sub1/sub2/**
The path `dir/a.text` would not be ignored.
|
|
|
7b083d3c
|
2019-03-02T18:14:36
|
|
Merge pull request #5005 from libgit2/ethomson/odb_backend_allocations
odb: provide a free function for custom backends
|
|
|
68729289
|
2019-02-25T09:25:34
|
|
Merge pull request #5000 from augfab/branch_lookup_all
Have git_branch_lookup accept GIT_BRANCH_ALL
|
|
|
459ac856
|
2019-02-23T18:42:53
|
|
odb: provide a free function for custom backends
Custom backends can allocate memory when reading objects and providing
them to libgit2. However, if an error occurs in the custom backend
after the memory has been allocated for the custom object but before
it's returned to libgit2, the custom backend has no way to free that
memory and it must be leaked.
Provide a free function that corresponds to the alloc function so that
custom backends have an opportunity to free memory before they return an
error.
|
|
|
790aae77
|
2019-02-23T18:40:43
|
|
odb: rename git_odb_backend_malloc for consistency
The `git_odb_backend_malloc` name is a system function that is provided
for custom ODB backends and allows them to allocate memory for an ODB
object in the read callback. This is important so that libgit2 can
later free the memory used by an ODB object that was read from the
custom backend.
However, the name _suggests_ that it actually allocates a
`git_odb_backend`. It does not; rename it to make it clear that it
actually allocates backend _data_.
|
|
c5d8e300
|
2019-02-21T21:46:39
|
|
branch: have git_branch_lookup accept GIT_BRANCH_ALL
|
|
|
bd132046
|
2019-02-22T20:10:52
|
|
p_fallocate: compatibility fixes for macOS
On macOS, fcntl(..., F_PREALLOCATE, ...) will only succeed when followed
by an ftruncate(), even when it reports success.
However, that syscall will fail when the file already exists. Thus, we
must ignore the error code and simply let ftruncate extend the size of
the file itself (albeit slowly).
By calling ftruncate, we also need to prevent against file shrinkage,
for compatibility with posix_ftruncate, which will only extend files,
never shrink them.
|
|
|
7ab7bf46
|
2019-02-22T11:32:01
|
|
p_fallocate: don't duplicate definitions for win32
|
|
|
32f50452
|
2019-02-22T11:22:28
|
|
p_fallocate: add Windows emulation
Emulate `p_fallocate` on Windows by seeking beyond the end of the file
and setting the size to the current seek position.
|
|
|
59001e83
|
2019-02-21T11:41:19
|
|
remote: rename git_push_transfer_progress callback
The `git_push_transfer_progress` is a callback and as such should be
suffixed with `_cb` for consistency. Rename
`git_push_transfer_progress` to `git_push_transfer_progress_cb`.
|
|
|
a1ef995d
|
2019-02-21T10:33:30
|
|
indexer: use git_indexer_progress throughout
Update internal usage of `git_transfer_progress` to
`git_indexer_progreses`.
|
|
|
4069f924
|
2019-02-22T10:56:08
|
|
Merge pull request #4901 from pks-t/pks/uniform-map-api
High-level map APIs
|
|
|
75dd7f2a
|
2019-02-22T10:13:00
|
|
Merge pull request #4984 from pks-t/pks/refdb-fs-race
refdb_fs: fix loose/packed refs lookup racing with repacks
|
|
|
c5594852
|
2019-02-22T10:06:24
|
|
Merge pull request #4998 from pks-t/pks/allocator-restructuring
Allocator restructuring
|
|
|
bbdcd450
|
2019-02-20T10:40:06
|
|
cache: fix misnaming of `git_cache_free`
Functions that free a structure's contents but not the structure
itself shall be named `dispose` in the libgit2 project, but the
function `git_cache_free` does not follow this naming pattern.
Fix this by renaming it to `git_cache_dispose` and adjusting all
callers to make use of the new name.
|
|
|
765ff6e0
|
2019-02-21T12:35:48
|
|
allocators: make crtdbg allocator reuse its own realloc
In commit 6e0dfc6ff (Make stdalloc__reallocarray call
stdalloc__realloc, 2019-02-16), we have changed the stdalloc
allocator to reuse `stdalloc__realloc` to implement
`stdalloc__reallocarray`. This commit is making the same change
for the Windows-specific crtdbg allocator to avoid code
duplication.
|
|
|
48727e5d
|
2019-02-21T12:27:42
|
|
allocators: extract crtdbg allocator into its own file
The Windows-specific crtdbg allocator is currently mixed into the
crtdbg stacktracing compilation unit, making it harder to find
than necessary. Extract it and move it into the new "allocators/"
subdirectory to improve discoverability.
This change means that the crtdbg compilation unit is now
compiled unconditionally, whereas it has previously only been
compiled on Windows platforms. Thus we now have additional guards
around the code so that it will only be compiled if
GIT_MSVC_CRTDBG is defined. This also allows us to move over the
fallback-implementation of `git_win32_crtdbg_init_allocator` into
the same compilation unit.
|
|
|
b63396b7
|
2019-02-21T12:13:59
|
|
allocators: move standard allocator into subdirectory
Right now, our two allocator implementations are scattered around
the tree in "stdalloc.h" and "win32/w32_crtdbg_stacktrace.h".
Start grouping them together in a single directory "allocators/",
similar to how e.g. our streams are organized.
|
|
|
9eb098d8
|
2019-02-21T11:37:04
|
|
Merge pull request #4991 from libgit2/ethomson/inttypes
Remove public 'inttypes.h' header
|
|
|
247e6d90
|
2019-02-18T07:22:20
|
|
Remove public 'inttypes.h' header
Remove an `inttypes.h` header that is too large in scope, and far too
public.
For Visual Studio 2012 and earlier (ie, `_MSC_VER < 1800`), we do need
to include `stdint.h` in our public headers, for types like `uint32_t`.
Internally, we also need to define `PRId64` as a printf formatting
string when it is not available.
|
|
|
554b3b9a
|
2019-02-21T10:31:21
|
|
Merge pull request #4996 from eaigner/master
Prevent reading out of bounds memory
|
|
|
014d4955
|
2019-02-20T15:30:11
|
|
apply: prevent OOB read when parsing source buffer
When parsing the patch image from a string, we split the string
by newlines to get a line-based view of it. To split, we use
`memchr` on the buffer and limit the buffer length by the
original length provided by the caller. This works just fine for
the first line, but for every subsequent line we need to actually
subtract the amount of bytes that we have already read.
The above issue can be easily triggered by having a source buffer
with at least two lines, where the second line does _not_ end in
a newline. Given a string "foo\nb", we have an original length of
five bytes. After having extracted the first line, we will point
to 'b' and again try to `memchr(p, '\n', 5)`, resulting in an
out-of-bounds read of four bytes.
Fix the issue by correctly subtracting the amount of bytes
already read.
|
|
6b3730d4
|
2019-02-16T19:55:30
|
|
Fix a memory leak in odb_otype_fast()
This change frees a copy of a cached object in odb_otype_fast().
|
|
|
12c6e1fa
|
2019-02-20T10:54:00
|
|
Merge pull request #4986 from lhchavez/realloc
Make stdalloc__reallocarray call stdalloc__realloc
|
|
|
9f388e9f
|
2019-02-20T10:51:33
|
|
Merge pull request #4990 from libgit2/remove_time_monotonic
Remove `git_time_monotonic`
|
|
|
e6c6d3bb
|
2019-02-17T22:31:37
|
|
Remove `git_time_monotonic`
`git_time_monotonic` was added so that non-native bindings like rugged
could get high-resolution timing for benchmarking. However, this is
outside the scope of libgit2 *and* rugged decided not to use this
function in the first place.
Google suggests that absolutely _nobody_ is using this function and we
don't want to be in the benchmarking business. Remove the function.
|
|
|
dd45539d
|
2019-02-16T22:06:58
|
|
Fix a _very_ improbable memory leak in git_odb_new()
This change fixes a mostly theoretical memory leak in got_odb_new()
that can only manifest if git_cache_init() fails due to running out of
memory or not being able to acquire its lock.
|
|
|
6e0dfc6f
|
2019-02-16T20:26:17
|
|
Make stdalloc__reallocarray call stdalloc__realloc
This change avoids calling realloc(3) in more than one place.
|
|
|
df42f368
|
2018-12-01T10:54:57
|
|
idxmap: remove legacy low-level interface
Remove the low-level interface that was exposing implementation details of
`git_idxmap` to callers. From now on, only the high-level functions shall be
used to retrieve or modify values of a map. Adjust remaining existing callers.
|
|
|
bd66925a
|
2018-12-01T10:29:32
|
|
oidmap: remove legacy low-level interface
Remove the low-level interface that was exposing implementation details of
`git_oidmap` to callers. From now on, only the high-level functions shall be
used to retrieve or modify values of a map. Adjust remaining existing callers.
|
|
|
4713e7c8
|
2018-12-01T09:58:30
|
|
offmap: remove legacy low-level interface
Remove the low-level interface that was exposing implementation details of
`git_offmap` to callers. From now on, only the high-level functions shall be
used to retrieve or modify values of a map. Adjust remaining existing callers.
|
|
|
fdfabdc4
|
2018-12-01T09:49:10
|
|
strmap: remove legacy low-level interface
Remove the low-level interface that was exposing implementation details of
`git_strmap` to callers. From now on, only the high-level functions shall be
used to retrieve or modify values of a map. Adjust remaining existing callers.
|
|
|
6a9117f5
|
2018-12-01T10:18:42
|
|
cache: use iteration interface for cache eviction
To relieve us from memory pressure, we may regularly call `cache_evict_entries`
to remove some entries from it. Unfortunately, our cache does not support a
least-recently-used mode or something similar, which is why we evict entries
completeley at random right now. Thing is, this is only possible due to the map
interfaces exposing the entry indices, and we intend to completely remove those
to decouple map users from map implementations. As soon as that is done, we are
unable to do this random eviction anymore.
Convert this to make use of an iterator for now. Obviously, there is no random
eviction possible like that anymore, but we'll always start by evicting from the
beginning of the map. Due to hashing, one may hope that the selected buckets
will be evicted at least in some way unpredictably. But more likely than not,
this will not be the case. But let's see what happens and if any users complain
about degraded performance. If so, we might come up with a different scheme than
random removal, e.g. by using an LRU cache.
|
|
|
c976b4f9
|
2018-12-01T10:18:26
|
|
indexer: use map iterator to delete expected OIDs
To compute whether there are objects missing in a packfile, the indexer keeps
around a map of OIDs that it still expects to see. This map does not store any
values at all, but in fact the keys are owned by the map itself. Right now, we
free these keys by iterating over the map and freeing the key itself, which is
kind of awkward as keys are expected to be constant.
We can make this a bit prettier by inserting the OID as value, too. As we
already store the `NULL` pointer either way, this does not increase memory
usage, but makes the code a tad more clear. Furthermore, we convert the
previously existing map iteration via indices to make use of an iterator,
instead.
|
|
|
18cf5698
|
2018-12-01T09:37:40
|
|
maps: provide high-level iteration interface
Currently, our headers need to leak some implementation details of maps due to
their direct use of indices in the implementation of their foreach macros. This
makes it impossible to completely hide the map structures away, and also makes
it impossible to include the khash implementation header in the C files of the
respective map only.
This is now being fixed by providing a high-level iteration interface
`map_iterate`, which takes as inputs the map that shall be iterated over, an
iterator as well as the locations where keys and values shall be put into. For
simplicity's sake, the iterator is a simple `size_t` that shall initialized to
`0` on the first call. All existing foreach macros are then adjusted to make use
of this new function.
|
|
|
c50a8ac2
|
2018-12-01T08:59:24
|
|
maps: use high-level function to check existence of keys
Some callers were still using the tightly-coupled pattern of `lookup_index` and
`valid_index` to verify that an entry exists in a map. Instead, use the more
high-level `exists` functions to decouple map users from its implementation.
|
|
|
84a089da
|
2018-12-01T08:50:36
|
|
maps: provide return value when deleting entries
Currently, the delete functions of maps do not provide a return value. Like
this, it is impossible to tell whether the entry has really been deleted or not.
Change the implementation to provide either a return value of zero if the entry
has been successfully deleted or `GIT_ENOTFOUND` if the key could not be found.
Convert callers to the `delete_at` functions to instead use this higher-level
interface.
|
|
|
8da93944
|
2018-12-01T10:52:44
|
|
idxmap: have `resize` functions return proper error code
The currently existing function `git_idxmap_resize` and
`git_idxmap_icase_resize` do not return any error codes at all due to their
previous implementation making use of a macro. Due to that, it is impossible to
see whether the resize operation might have failed due to an out-of-memory
situation.
Fix this by providing a proper error code. Adjust callers to make use of it.
|
|
|
661fc57b
|
2018-12-01T01:16:25
|
|
idxmap: introduce high-level setter for key/value pairs
Currently, one would use the function `git_idxmap_insert` to insert key/value
pairs into a map. This function has historically been a macro, which is why its
syntax is kind of weird: instead of returning an error code directly, it instead
has to be passed a pointer to where the return value shall be stored. This does
not match libgit2's common idiom of directly returning error codes.
Introduce a new function `git_idxmap_set`, which takes as parameters the map,
key and value and directly returns an error code. Convert all callers of
`git_idxmap_insert` to make use of it.
|
|
|
d00c24a9
|
2019-01-23T10:49:25
|
|
idxmap: introduce high-level getter for values
The current way of looking up an entry from a map is tightly coupled with the
map implementation, as one first has to look up the index of the key and then
retrieve the associated value by using the index. As a caller, you usually do
not care about any indices at all, though, so this is more complicated than
really necessary. Furthermore, it invites for errors to happen if the correct
error checking sequence is not being followed.
Introduce new high-level functions `git_idxmap_get` and `git_idxmap_icase_get`
that take a map and a key and return a pointer to the associated value if such a
key exists. Otherwise, a `NULL` pointer is returned. Adjust all callers that can
trivially be converted.
|
|
|
b9d0b664
|
2018-12-17T09:10:53
|
|
offmap: introduce high-level setter for key/value pairs
Currently, there is only one caller that adds entries into an offset map, and
this caller first uses `git_offmap_put` to add a key and then set the value at
the returned index by using `git_offmap_set_value_at`. This is just too tighlty
coupled with implementation details of the map as it exposes the index of
inserted entries, which we really do not care about at all.
Introduce a new function `git_offmap_set`, which takes as parameters the map,
key and value and directly returns an error code. Convert the caller to make use
of it instead.
|
|
|
aa245623
|
2018-11-30T18:28:05
|
|
offmap: introduce high-level getter for values
The current way of looking up an entry from a map is tightly coupled with the
map implementation, as one first has to look up the index of the key and then
retrieve the associated value by using the index. As a caller, you usually do
not care about any indices at all, though, so this is more complicated than
really necessary. Furthermore, it invites for errors to happen if the correct
error checking sequence is not being followed.
Introduce a new high-level function `git_offmap_get` that takes a map and a key
and returns a pointer to the associated value if such a key exists. Otherwise,
a `NULL` pointer is returned. Adjust all callers that can trivially be
converted.
|
|
|
2e0a3048
|
2019-01-23T10:48:55
|
|
oidmap: introduce high-level setter for key/value pairs
Currently, one would use either `git_oidmap_insert` to insert key/value pairs
into a map or `git_oidmap_put` to insert a key only. These function have
historically been macros, which is why their syntax is kind of weird: instead of
returning an error code directly, they instead have to be passed a pointer to
where the return value shall be stored. This does not match libgit2's common
idiom of directly returning error codes.Furthermore, `git_oidmap_put` is tightly
coupled with implementation details of the map as it exposes the index of
inserted entries.
Introduce a new function `git_oidmap_set`, which takes as parameters the map,
key and value and directly returns an error code. Convert all trivial callers of
`git_oidmap_insert` and `git_oidmap_put` to make use of it.
|
|
|
9694ef20
|
2018-12-17T09:01:53
|
|
oidmap: introduce high-level getter for values
The current way of looking up an entry from a map is tightly coupled with the
map implementation, as one first has to look up the index of the key and then
retrieve the associated value by using the index. As a caller, you usually do
not care about any indices at all, though, so this is more complicated than
really necessary. Furthermore, it invites for errors to happen if the correct
error checking sequence is not being followed.
Introduce a new high-level function `git_oidmap_get` that takes a map and a key
and returns a pointer to the associated value if such a key exists. Otherwise,
a `NULL` pointer is returned. Adjust all callers that can trivially be
converted.
|
|
|
03555830
|
2019-01-23T10:44:33
|
|
strmap: introduce high-level setter for key/value pairs
Currently, one would use the function `git_strmap_insert` to insert key/value
pairs into a map. This function has historically been a macro, which is why its
syntax is kind of weird: instead of returning an error code directly, it instead
has to be passed a pointer to where the return value shall be stored. This does
not match libgit2's common idiom of directly returning error codes.
Introduce a new function `git_strmap_set`, which takes as parameters the map,
key and value and directly returns an error code. Convert all callers of
`git_strmap_insert` to make use of it.
|
|
|
ef507bc7
|
2019-01-23T10:44:02
|
|
strmap: introduce `git_strmap_get` and use it throughout the tree
The current way of looking up an entry from a map is tightly coupled with the
map implementation, as one first has to look up the index of the key and then
retrieve the associated value by using the index. As a caller, you usually do
not care about any indices at all, though, so this is more complicated than
really necessary. Furthermore, it invites for errors to happen if the correct
error checking sequence is not being followed.
Introduce a new high-level function `git_strmap_get` that takes a map and a key
and returns a pointer to the associated value if such a key exists. Otherwise,
a `NULL` pointer is returned. Adjust all callers that can trivially be
converted.
|
|
|
7e926ef3
|
2018-11-30T12:14:43
|
|
maps: provide a uniform entry count interface
There currently exist two different function names for getting the entry count
of maps, where offmaps offset and string maps use `num_entries` and OID maps use
`size`. In most programming languages with built-in map types, this is simply
called `size`, which is also shorter to type. Thus, this commit renames the
other two functions `num_entries` to match the common way and adjusts all
callers.
|
|
|
351eeff3
|
2019-01-23T10:42:46
|
|
maps: use uniform lifecycle management functions
Currently, the lifecycle functions for maps (allocation, deallocation, resize)
are not named in a uniform way and do not have a uniform function signature.
Rename the functions to fix that, and stick to libgit2's naming scheme of saying
`git_foo_new`. This results in the following new interface for allocation:
- `int git_<t>map_new(git_<t>map **out)` to allocate a new map, returning an
error code if we ran out of memory
- `void git_<t>map_free(git_<t>map *map)` to free a map
- `void git_<t>map_clear(git<t>map *map)` to remove all entries from a map
This commit also fixes all existing callers.
|
|
|
94743daf
|
2019-02-15T11:16:46
|
|
refdb_fs: fix race when migrating loose to packed refs in iteration
Right now, we first load the packed refs cache and only afterwards load the
loose references. This is susceptible to a race when the loose ref is being
migrated to a packed cache by e.g. git-pack-refs(1):
libgit2 git-pack-refs
1. We load the packed ref, which
does not yet have the migrated
reference.
2. git-pack-refs updates the packed ref
file to have the migrated ref.
3. git-pack-refs deletes the old loose
ref.
4. We look up the loose ref.
So we now do not find the reference at all and will never iterate over it.
Fix the issue by reversing the order: instead of first loading the packed refs,
we will now look up the loose reference first. If it has already been deleted,
then it must already be present in the packed-refs by definition, as git.git
will only delete the reference after updating the packed refs file.
|
|
|
3ff0e3b5
|
2019-02-15T11:16:05
|
|
refdb_fs: remove ordering dependency on loose/packed refs loading
Right now, loading loose refs has the side-effect of setting the
`PACKREF_SHADOWED` flag for references that exist both in the loose and the
packed refs. Because of this, we are force do first look up packed refs and only
afterwards loading the packed refs. This is susceptible to a race, though, when
refs are being repacked: when first loading the packed cache, then it may not
yet have the migrated loose ref. But when now trying to look up the loose
reference afterwards, then it may already have been migrated. Thus, we would
fail to find this reference in this scenario.
Remove this ordering dependency to allow fixing the above race. Instead of
setting the flag when loading loose refs, we will now instead set it lazily when
iterating over the loose refs. This even has the added benefit of not requiring
us to lock the packed refs cache, as we already have an owned copy of it.
|
|
|
83333814
|
2019-02-15T10:56:50
|
|
refdb_fs: do not lazily copy packed ref cache
When creating a new iterator, we eagerly load loose refs but only lazily create
a copy of packed refs. The lazy load only happens as soon as we have iterated
over all loose refs, opening up a potentially wide window for races. This
may lead to an inconsistent view e.g. when the caller decides to reload packed
references somewhen between iterating the loose refs, which is unexpected.
Fix the issue by eagerly copying the sorted cache. Note that right now, we are
heavily dependent on ordering here: we first need to reload packed refs, then we
have to load loose refs and only as a last step are we allowed to copy the
cache. This is because loading loose refs has the side-effect of setting the
`PACKED_SHADOWED` flag in the packed refs cache, which we require to avoid
outputting packed refs that already exist as loose refs.
|
|
|
32063d82
|
2019-02-15T10:41:30
|
|
refdb_fs: refactor error handling in iterator creation
Refactor the error handling in `refdb_fs_backend__iterator` to always return the
correct error code returned by the failing function.
|
|
|
8c773438
|
2019-02-15T10:15:39
|
|
refdb_fs: fix potential race with ref repacking in `exists` callback
When repacking references, git.git will first update the packed refs and only
afterwards delete any existing loose references that have now been moved to the
new packed refs file. Due to this, there is a potential for racing if one first
reads the packfile (which has not been updated yet) and only then trying to read
the loose reference (which has just been deleted). In this case, one will
incorrectly fail to lookup the reference and it will be reported as missing.
Naturally, this is exactly what we've been doing in `refdb_fs_backend__exists`.
Fix the race by reversing the lookup: we will now first check if the loose
reference exists and only afterwards refresh the packed file.
|
|
|
bda08397
|
2019-02-14T16:57:47
|
|
Merge pull request #4982 from pks-t/pks/worktree-add-bare-head
Enable creation of worktree from bare repo's default branch
|