Log

Author Commit Date CI Message
Edward Thomson 616628dd 2022-01-17T21:39:35 Merge branch 'main' into typos
Edward Thomson 2bfd8ddc 2022-01-17T21:05:17 Merge pull request #6175 from libgit2/ethomson/follow_redirects_initial remote: support `http.followRedirects` (`false` and `initial`) and follow initial redirects by default
Edward Thomson 6723edc7 2022-01-17T19:36:53 Merge pull request #6179 from boretrk/flags c99: change single bit flags to unsigned
Peter Pettersson a979cf3d 2021-11-17T22:19:47 c99: change single bit flags to unsigned
Edward Thomson f4fec251 2022-01-13T16:54:56 ci: minor formatting cleanups to test output
Edward Thomson adaa037d 2022-01-04T16:18:40 remote: test honoring configuration option Test that we honor `http.followRedirects` when set to initial or false.
Edward Thomson fda59a76 2022-01-04T07:05:20 remote: honor `http.followRedirects` configuration option
Edward Thomson 515daeaf 2022-01-04T06:16:30 remote: introduce `follow_redirects` connect option Give callers the ability to select how to handle redirects - either supporting redirects during the initial connection (so that, for example, `git.example.com/repo` can redirect to `github.com/example/repo`) or all/no redirects. This is for compatibility with git.
Edward Thomson 342e55ac 2021-12-18T10:13:18 url: optionally allow off-site redirects In redirect application logic, (optionally) allow off-site redirects.
Edward Thomson c104a565 2022-01-06T07:56:32 ci: stop retrying "flaky" tests We (thankfully) don't have flaky tests anymore, don't retry them.
Edward Thomson cd261af2 2021-12-18T07:42:44 ci: use poxygit v0.5.1 The new version of our test git server adds redirect support. Upgrade to it.
Edward Thomson 6fc6eeb6 2021-12-24T15:14:38 remote: introduce `git_remote_connect_options` The existing mechanism for providing options to remote fetch/push calls, and subsequently to transports, is unsatisfactory. It requires an options structure to avoid breaking the API and callback signatures. 1. Introduce `git_remote_connect_options` to satisfy those needs. 2. Add a new remote connection API, `git_remote_connect_ext` that will take this new options structure. Existing `git_remote_connect` calls will proxy to that. `git_remote_fetch` and `git_remote_push` will proxy their fetch/push options to that as well. 3. Define the interaction between `git_remote_connect` and fetch/push. Connect _may_ be called before fetch/push, but _need not_ be. The semantics of which options would be used for these operations was not specified if you specify options for both connect _and_ fetch. Now these are defined that the fetch or push options will be used _if_ they were specified. Otherwise, the connect options will be used if they were specified. Otherwise, the library's defaults will be used. 4. Update the transports to understand `git_remote_connect_options`. This is a breaking change to the systems API.
Edward Thomson f99a0d69 2022-01-04T15:17:01 remote: improved error reporting Several places in the remote code identify an error and then swallow it; return the error.
Edward Thomson f5373123 2021-12-24T15:17:52 push: don't try to connect in `finish` Push starts by connecting, don't try to connect again, just rely on the existing connection existing.
Edward Thomson 240c195f 2022-01-02T08:19:45 docs: document our API/ABI stability stance
Edward Thomson e2e3f3e4 2022-01-06T07:42:36 Merge branch 'status'
Edward Thomson 9faa9199 2022-01-06T07:42:01 status: don't use `// ...` in tests, use `/* ... */`
Edward Thomson ca9f6b98 2022-01-06T07:41:27 Merge remote-tracking branch 'pr/6159'
Edward Thomson da5385c6 2022-01-06T07:39:21 status: don't use `// ...` in tests, use `/* ... */`
Edward Thomson db3572b7 2022-01-06T07:38:07 Merge pull request #6165 from libgit2/ethomson/includeif config: handle empty conditional in includeIf
Dimitris Apostolou 90df4302 2022-01-05T12:18:05 Fix typos
Edward Thomson 0bf67234 2022-01-05T09:30:27 Merge pull request #6158 from arroz/feat/statusRenameThreshold Add `rename_threshold` to `git_status_options`.
Edward Thomson 715bfe10 2022-01-05T09:22:38 config: test allocation in config conditionals
Edward Thomson 83a1e614 2022-01-05T09:18:59 config: handle empty conditional in includeIf When a config file contains `[includeIf]` (with no condition), we should treat that as a falsey value. This means that we should properly parse a config value of `includeIf.path`.
Edward Thomson 82f526a9 2022-01-04T22:46:10 Merge pull request #6157 from boretrk/hmac_ctx_cleanup ntmlclient: fix linking with libressl
Miguel Arroz a3436cde 2022-01-01T15:17:08 #6154 git_status_list_new case insensitive fix
Miguel Arroz ef848891 2021-12-31T16:40:58 Add `rename_threshold` to `git_status_options`.
Peter Pettersson 6aa3603a 2021-12-31T01:50:45 ntmlclient: don't declare dummy HMAC_CTX_cleanup when building with libressl
Edward Thomson 50b4d53e 2021-12-24T08:01:57 Merge pull request #6150 from cestrand/patch-1 Remove PSGit from Language Bindings section of README
Marcin Kolenda 85442950 2021-12-24T14:24:07 Remove PSGit from Language Bindings section of README PSGit is not using libgit2 library. It is stub of a library and is not developed further.
Edward Thomson 00346ce8 2021-12-24T00:12:59 Merge pull request #6148 from libgit2/ethomson/ci ci: build with ssh on nightly
Edward Thomson c940b695 2021-12-23T22:33:32 ci: build with ssh on nightly
Edward Thomson 153f6c27 2021-12-23T22:02:11 Merge pull request #6147 from libgit2/ethomson/insteadof remote: refactor insteadof application
Edward Thomson 33d0ad9c 2021-12-23T21:18:54 remote: refactor insteadof application Using the insteadof helper would leak memory when we didn't really want the pushInsteadOf configuration. Refactor the choice into the function that allocates memory (or now, not) and use a more idiomatic `int` return code.
Edward Thomson 86e5003d 2021-12-23T14:22:34 Merge branch 'pr/6062'
Edward Thomson 942cfac1 2021-12-23T14:21:23 worktree: checkout options suggestions from code review
punkymaniac f9a98f94 2021-11-22T15:16:26 Fix test fail because of memory leak
punkymaniac 1fa8014a 2021-11-22T12:07:59 Update test for `git_str` usage instead of `git_buf` See commit f0e693b18afbe1de37d7da5b5a8967b6c87d8e53
punkymaniac 8b0ffdcb 2021-11-08T17:23:27 Add test for `git_worktree_add` no checkout
punkymaniac 58451759 2021-10-14T09:55:00 Change default checkout strategy from FORCE to SAFE Since we are able to give our own git checkout options, the default git checkout strategy will be the same as initialized in a new git_checkout_options struct.
punkymaniac ffead012 2021-09-09T18:02:22 Allow user checkout options on git_worktree_add Extend the `git_worktree_add_options` to include `git_checkout_options`. github issue #5949
Edward Thomson a50bbba1 2021-12-23T15:07:29 Merge pull request #6142 from libgit2/ethomson/blob_data_is_binary blob: identify binary content
Edward Thomson 05c3d972 2021-12-23T15:03:08 Merge pull request #6124 from csware/config-parsing Config parsing
Edward Thomson 3cca14b3 2021-12-23T14:13:34 Merge pull request #6125 from stforek/git_commit_summary_spaces git_commit_summary: ignore lines with spaces
Edward Thomson dca31d24 2021-12-23T14:12:23 Merge pull request #6101 from mkhl/fix/instead-of remotes: fix insteadOf/pushInsteadOf handling
Edward Thomson ab5b3f37 2021-12-23T14:09:09 Merge pull request #6095 from yoichi/better-compatiblity-for-at-time-notation Better revparse compatibility for at time notation
Edward Thomson 4b27009c 2021-12-23T14:04:43 Merge pull request #6094 from visualgitio/commit-graph-long-long Fix a long long that crept past
Edward Thomson 734468d6 2021-12-23T14:02:02 Merge pull request #6146 from Qix-/main cmake: use project-specific root variable instead of CMAKE_SOURCE_DIR
Josh Junon c5cd71b2 2021-12-23T18:23:34 cmake: use PROJECT_SOURCE_DIR of CMAKE_SOURCE_DIR Also applies to *_BINARY_DIR. This effectively reverts 84083dcc8bd41332ccac9d7b537f3e254d79011c, which broke all users of libgit2 that use it as a CMake subdirectory (via `add_subdirectory()`). This is because CMAKE_SOURCE_DIR refers to the root-most CMake directory, which in the case of `add_subdirectory()` is a parent project to libgit2 and thus the paths don't make any sense to the configuration files. Corollary, CMAKE_SOURCE_DIR only makes sense if the CMake project is always the root project - which can rarely be guaranteed. In all honesty, CMake should deprecate and eventually remove CMAKE_SOURCE_DIR and CMAKE_BINARY_DIR. It's been the source of headaches and confusion for years, they're rarely useful over CMAKE_CURRENT_(SOURCE|BINARY)_DIR or PROJECT_(SOURCE|BINARY)_DIR, and they cause a lot of confusing configuration and source code layouts to boot. Any time they are used, they break `add_subdirectory()` almost 100% of the time, cause confusing error messages, and hide subtle bugs.
Calvin Buckley 5761980d 2021-12-11T22:33:16 Simplifications to definitions to avoid UINT64_C
Edward Thomson 4591e76a 2021-12-10T15:19:59 blob: identify binary content Introduce `git_blob_data_is_binary` to examine a blob's data, instead of the blob itself. A replacement for `git_buf_is_binary`.
Sven Strickroth dff05bc3 2021-11-25T11:40:20 Multiline config values not preserved on saving (fixes issue #6088) Signed-off-by: Sven Strickroth <email@cs-ware.de>
Sven Strickroth 854164a5 2021-11-25T11:02:41 Add more tests for contents Signed-off-by: Sven Strickroth <email@cs-ware.de>
Sven Strickroth 433edb54 2021-11-25T10:22:22 Config parsing confused by continuations that start with quotes (fixes issue #6089) Signed-off-by: Sven Strickroth <email@cs-ware.de>
Edward Thomson 12b53eb0 2021-12-04T20:00:41 Merge pull request #6128 from libgit2/ethomson/object_validation Introduce `git_object_rawcontent_is_valid`
Edward Thomson 9f03ebd1 2021-11-29T13:44:42 object: introduce a raw content validation function Users may want to validate raw object content; provide them a function to do so.
Edward Thomson fc1a3f45 2021-11-29T13:36:36 object: return GIT_EINVALID on parse errors Return `GIT_EINVALID` on parse errors so that direct callers of parse functions can determine when there was a failure to parse the object. The object parser functions will swallow this error code to prevent it from propagating down the chain to end-users. (`git_merge` should not return `GIT_EINVALID` when a commit it tries to look up is not valid, this would be too vague to be useful.) The only public function that this affects is `git_signature_from_buffer`, which is now documented as returning `GIT_EINVALID` when appropriate.
Edward Thomson 6fdb1b2f 2021-11-30T21:09:57 Merge pull request #6122 from libgit2/ethomson/cleanup Minor code cleanups
Przemyslaw Ciezkowski 1e015088 2021-11-25T15:19:17 git_commit_summary: ignore lines with spaces Fixes libgit2/libgit2#6065
Yoichi Nakayama 7bb206a7 2021-11-23T10:12:29 Merge branch 'main' into better-compatiblity-for-at-time-notation Conflicts: src/revparse.c
Edward Thomson f9c4dc10 2021-11-22T11:23:50 Merge pull request #6106 from ammgws/fixtemplateerr Fix repo init when template dir is non-existent
Edward Thomson 3461aaf7 2021-11-22T10:29:29 repo: minor formatting fix
Edward Thomson adcf638c 2021-11-21T21:34:17 filebuf: use hashes not oids The filebuf functions should use hashes directly, not indirectly using the oid functions.
Edward Thomson 91365fd8 2021-11-22T09:02:38 sha1: tests should use hashes, not oid computation The tests that examine sha1 behavior (including collision detection) should test against the hash functionality directly, not indirectly using the oid functions.
Edward Thomson fc42c28e 2021-11-19T09:36:03 util: don't include unnecessary system libraries Remove some unnecessary includes from utility code.
Edward Thomson 84083dcc 2021-11-19T08:48:08 cmake: use CMAKE_SOURCE_DIR and CMAKE_BINARY_DIR Instead of using the project-specific `libgit2_SOURCE_DIR` and `libgit2_BINARY_DIR` variables, use `CMAKE_SOURCE_DIR` and `CMAKE_BINARY_DIR`.
Edward Thomson ea62f546 2021-11-14T16:41:11 utf8: include `utf8.h` in tests
Edward Thomson 19743830 2021-11-19T08:17:28 Merge pull request #6121 from libgit2/ethomson/date util: minor cleanup and refactoring to the date class
Edward Thomson add30a83 2021-11-18T12:36:25 date: rfc2822 formatting uses a `git_buf` instead of a static string
Edward Thomson b2c40314 2021-11-18T12:19:32 date: make it a proper `git_date` utility class Instead of `git__date`, just use `git_date`.
Edward Thomson 1604be06 2021-11-17T08:54:38 Merge pull request #6120 from libgit2/ethomson/cmake_test cmake: re-enable WinHTTP
Edward Thomson 1b8d14ac 2021-11-17T08:25:56 cmake: use WINHTTP when USE_WINHTTP is specified
Edward Thomson 4a54d8e3 2021-11-17T07:27:39 cmake: enable testing at the top level Ensure that we `enable_testing()` at the top-level CMakeLists.txt or else we'll need to navigate within the build directory to the correct place in the hierarchy to run `ctest`. Now we can `ctest` at the top-level again.
Edward Thomson 7a15213e 2021-11-16T23:42:09 Merge pull request #6119 from libgit2/ethomson/tag_valid tag: set validity to 0 by default
Edward Thomson 6b2ea909 2021-11-16T23:21:29 tag: set validity to 0 by default `git_tag_name_is_valid` needs to set validity to 0 when it short-circuits.
Edward Thomson 86685f8f 2021-11-16T17:10:08 cmake: better explain why headertests exists
Peter Pettersson 7dcc29fc 2021-10-22T22:51:59 Make enum in src,tests and examples C90 compliant by removing trailing comma.
Peter Pettersson 43d9f0e3 2021-10-22T22:39:10 ntmlclient: make enum C90 compliant by removing trailing comma
Peter Pettersson 38c34498 2021-10-03T00:12:52 Make enum in includes C90 compliant by removing trailing comma.
Peter Pettersson 5221e185 2021-10-02T23:54:14 Add target for testing libgit2 headers
Edward Thomson 23c5c315 2021-11-14T08:42:12 Merge pull request #6114 from libgit2/ethomson/cmake5 cmake: further refactorings
Edward Thomson 2c154145 2021-11-11T23:09:44 cmake: move sha1 source selection into CMakeLists.txt The select hashes module selects the hash; the CMakeLists.txt selects the files to implement it.
Edward Thomson c7f6ecb3 2021-11-11T22:54:14 cmake: use a string not a boolean for LIBGIT2_FILENAME
Edward Thomson 395b3dc4 2021-11-11T22:10:51 cmake: refactor global variables Update the global variables `LIBGIT2_OBJECTS` to `LIBGIT2_DEPENDENCY_OBJECTS` for clarity and consistency.
Edward Thomson c3fec456 2021-11-11T22:37:36 cmake: reformat modules Apply the standard project cmake formatting to the modules.
Edward Thomson b608af6c 2021-11-14T07:23:01 Merge pull request #6116 from lhchavez/drop-volatile-qualifier-in-atomic-exchange
Edward Thomson 03132b36 2021-11-14T07:22:24 Merge pull request #6115 from lhchavez/fix-indentation-threadstate
lhchavez 5675312e 2021-11-14T02:27:10 Fix a gcc 11 warning in src/thread.h When building under gcc 11, there is a warning about an incompatible pointer type, since [`__atomic_exchange`](https://gcc.gnu.org/onlinedocs/gcc/_005f_005fatomic-Builtins.html) does not take `volatile` pointers: ``` In file included from ../src/common.h:81, from ../src/transports/winhttp.c:8: ../src/thread-utils.h: In function ‘git___swap’: ../src/thread-utils.h:168:9: warning: argument 3 of ‘__atomic_exchange’ discards ‘volatile’ qualifier [-Wincompatible-pointer-types] 168 | __atomic_exchange(ptr, &newval, &foundval, __ATOMIC_SEQ_CST); | ^~~~~~~~~~~~~~~~~ ``` This change drops the `volatile` qualifier so that the pointer type matches what `__atomic_exchange` expects.
lhchavez 0c4d24da 2021-11-14T02:24:39 Fix a gcc 11 warning in src/threadstate.c When building under gcc 11, there is a warning about a misaligned guard clause because there were mixed spaces and tabs: ``` [128/634] Building C object src/CMakeFiles/git2internal.dir/threadstate.c.o ../src/threadstate.c: In function ‘threadstate_dispose’: ../src/threadstate.c:39:5: warning: this ‘if’ clause does not guard... [-Wmisleading-indentation] 39 | if (threadstate->error_t.message != git_str__initstr) | ^~ ../src/threadstate.c:41:9: note: ...this statement, but the latter is misleadingly indented as if it were guarded by the ‘if’ 41 | threadstate->error_t.message = NULL; | ^~~~~~~~~~~ ../src/threadstate.c: At top level: ``` This change indents the code with tabs for consistency with the rest of the code, which makes the warning go away.
Jason Nader bc0d1ad2 2021-11-05T03:08:43 tests: try to init with non-existent template path
Jason Nader 709b1b62 2021-11-05T02:28:37 repository: do not copy templates if dir nonexistent This mimics the behaviour of git which just prints a warning and continues with the repo initialisation.
Edward Thomson 043f3123 2021-11-11T18:21:35 Merge pull request #6113 from libgit2/ethomson/cmake3 Add missing-declarations warning globally
Edward Thomson 4a6ef5a4 2021-11-11T17:04:24 cmake: move missing-declarations warning to top-level We should enforce declarations throughout the code-base, including examples, fuzzers and tests, not just in the `src` tree.
Edward Thomson 8d2b3110 2021-11-11T19:58:49 tests: include function declarations
Edward Thomson bfdc0971 2021-11-11T13:32:00 test: correct test names
Edward Thomson ca14942e 2021-11-11T13:28:08 tests: declare functions statically where appropriate
Edward Thomson 3f024b6d 2021-11-11T12:57:44 clar: declare summary functions as static
Edward Thomson eabbee04 2021-11-11T12:48:19 example: declare print_usage function
Edward Thomson 8be22614 2021-11-11T12:47:02 examples: remove unused function
Edward Thomson 2374ba8d 2021-11-11T17:15:47 fuzzers: declare static functions as static