[pull] master from git:master#74
Merged
pull[bot] merged 20 commits intoturkdevops:masterfrom Jul 8, 2025
Merged
Conversation
The branch structure has both branch->merge_name and branch->merge for tracking the merge information. The former is allocated by add_merge() and stores the names read from the configuration file. The latter is allocated by set_merge() which is called by branch_get() when an external caller requests a branch. This leads to the confusing situation where branch->merge_nr tracks both the size of branch->merge (once its allocated) and branch->merge_name. The branch_release() function incorrectly assumes that branch->merge is always set when branch->merge_nr is non-zero, and can potentially crash if read_config() is called without branch_get() being called on every branch. In addition, branch_release() fails to free some of the memory associated with the structure including: * Failure to free the refspec_item containers in branch->merge[i] * Failure to free the strings in branch->merge_name[i] * Failure to free the branch->merge_name parent array. The set_merge() function sets branch->merge_nr to 0 when there is no valid remote_name, to avoid external callers seeing a non-zero merge_nr but a NULL merge array. This results in failure to release most of the merge data as well. These issues could be fixed directly, and indeed I initially proposed such a change at [1] in the past. While this works, there was some confusion during review because of the inconsistencies. Instead, its time to clean up the situation properly. Remove branch->merge_name entirely. Instead, allocate branch->merge earlier within add_merge() instead of within set_merge(). Instead of having set_merge() copy from merge_name[i] to merge[i]->src, just have add_merge() directly initialize merge[i]->src. Modify the add_merge() to call xstrdup() itself, instead of having the caller of add_merge() do so. This makes it more obvious which code owns the memory. Update all callers which use branch->merge_name[i] to use branch->merge[i]->src instead. Add a merge_clear() function which properly releases all of the merge-related memory, and which sets branch->merge_nr to zero. Use this both in branch_release() and in set_merge(), fixing the leak when set_merge() finds no valid remote_name. Add a set_merge variable to the branch structure, which indicates whether set_merge() has been called. This replaces the previous use of a NULL check against the branch->merge array. With these changes, the merge array is always allocated when merge_nr is non-zero. This use of refspec_item to store the names should be safe. External callers should be using branch_get() to obtain a pointer to the branch, which will call set_merge(), and the callers internal to remote.c already handle the partially initialized refpsec_item structure safely. This end result is cleaner, and avoids duplicating the merge names twice. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Link: [1] https://lore.kernel.org/git/20250617-jk-submodule-helper-use-url-v2-1-04cbb003177d@gmail.com/ Signed-off-by: Junio C Hamano <gitster@pobox.com>
The remote_clear() function failed to free the remote->push and remote->fetch refspec fields. This should be caught by the leak sanitizer. However, for callers which use ``the_repository``, the values never go out of scope and the sanitizer doesn't complain. A future change is going to add a caller of read_config() for a submodule repository structure, which would result in the leak sanitizer complaining. Fix remote_clear(), updating it to properly call refspec_clear() for both the push and fetch members. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Both submodule--helper.c and submodule-config.c have an implementation of starts_with_dot_slash and starts_with_dot_dot_slash. The dir.h header has starts_with_dot(_dot)_slash_native, which sets PATH_MATCH_NATIVE. Move the helpers to dir.h as static inlines. I thought about renaming them to postfix with _platform but that felt too long and ugly. On the other hand it might be slightly confusing with _native. This simplifies a submodule refactor which wants to use the helpers earlier in the submodule--helper.c file. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The remotes_remote_get_1 (and its caller, remotes_remote_get, have an implicit dependency on the_repository due to calling read_branches_file() and read_remotes_file(), both of which use the_repository. The branch_get() function calls set_merge() which has an implicit dependency on the_repository as well. Because of this use of the_repository, the helper functions cannot be used in code paths which operate on other repositories. A future refactor of the submodule--helper will want to make use of some of these functions. Refactor to break the dependency by passing struct repository *repo instead of struct remote_state *remote_state in a few places. The public callers and many other helper functions still depend on the_repository. A repo-aware function will be exposed in a following change for git submodule--helper. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The repo_get_default_remote() function in submodule--helper currently tries to figure out the proper remote name to use for a submodule based on a few factors. First, it tries to find the remote for the currently checked out branch. This works if the submodule is configured to checkout to a branch instead of a detached HEAD state. In the detached HEAD state, the code calls back to using "origin", on the assumption that this is the default remote name. Some users may change this, such as by setting clone.defaultRemoteName, or by changing the remote name manually within the submodule repository. As a first step to improving this situation, refactor to reuse the logic from remotes_remote_for_branch(). This function uses the remote from the branch if it has one. If it doesn't then it checks to see if there is exactly one remote. It uses this remote first before attempting to fall back to "origin". To allow using this helper function, introduce a repo_default_remote() helper to remote.c which takes a repository structure. This helper will load the remote configuration and get the "HEAD" branch. Then it will call remotes_remote_for_branch to find the default remote. Replace calls of repo_get_default_remote() with the calls to this new function. To maintain consistency with the existing callers, continue copying the returned string with xstrdup. This isn't a perfect solution for users who change remote names, but it should help in cases where the remote name is changed but users haven't added any additional remotes. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
A future refactor got get_default_remote_submodule() is going to depend on resolve_relative_url(). That function depends on get_default_remote(). Move get_default_remote_submodule() after resolve_relative_url() first to make the additional functionality easier to review. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The get_default_remote_submodule() function performs a lookup to find the appropriate remote to use within a submodule. The function first checks to see if it can find the remote for the current branch. If this fails, it then checks to see if there is exactly one remote. It will use this, before finally falling back to "origin" as the default. If a user happens to rename their default remote from origin, either manually or by setting something like clone.defaultRemoteName, this fallback will not work. In such cases, the submodule logic will try to use a non-existent remote. This usually manifests as a failure to trigger the submodule update. The parent project already knows and stores the submodule URL in either .gitmodules or its .git/config. Add a new repo_remote_from_url() helper which will iterate over all the remotes in a repository and return the first remote which has a matching URL. Refactor get_default_remote_submodule to find the submodule and get its URL. If a valid URL exists, first try to obtain a remote using the new repo_remote_from_url(). Fall back to the repo_default_remote() otherwise. The fallback logic is kept in case for some reason the user has manually changed the URL within the submodule. Additionally, we still try to use a remote rather than directly passing the URL in the fetch_in_submodule() logic. This ensures that an update will properly update the remote refs within the submodule as expected, rather than just fetching into FETCH_HEAD. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since df076bd ([PATCH] GIT: Listen on IPv6 as well, if available., 2005-07-23), any file descriptor assigned to a listening socket was validated to be within the range to be used in an FDSET later. 6573faf (NO_IPV6 support for git daemon, 2005-09-28), moves to use poll() instead of select(), that doesn't have that restriction, so remove the original check. Signed-off-by: Carlo Marcelo Arenas Belón <carenas@gmail.com> Acked-by: Phillip Wood <phillip.wood123@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit c800963 ("fetch-pack, send-pack: clean up shallow oid array", 2024-09-25) cleaned up the shallow oid array in cmd_send_pack, but didn't clean up extra_have, which is still leaked at program exit. I suspect the particular tests in t5539 don't trigger any additions to the extra_have array, which explains why the tests can pass leak free despite this gap. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since df076bd ([PATCH] GIT: Listen on IPv6 as well, if available., 2005-07-23), the original error checking was included in an inner loop unchanged, where its effect was different. Instead of retrying, after a EINTR during accept() in the listening socket, it will advance to the next one and try with that instead, leaving the client waiting for another round. Make sure to retry with the same listener socket that failed originally. To avoid an unlikely busy loop, fallback to the old behaviour after a couple of attempts. Signed-off-by: Carlo Marcelo Arenas Belón <carenas@gmail.com> Acked-by: Phillip Wood <phillip.wood@dunelm.org.uk> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 'branch' section of the git-config documentation was missing inline code formatting and emphasis for the <name> placeholder. Both changes improve readability, especially when viewed online. Signed-off-by: Jakub Ječmínek <kuba@kubajecminek.cz> Signed-off-by: Junio C Hamano <gitster@pobox.com>
FreeBSD 13.4 is no longer supported, and 13.5 will be the last release from that series, so jump instead to 14.3 which should be supported for another 10 months and will be at that point the oldest supported release with the interim release of 15. While at it, move some variables to the environment and make sure to skip a git grep test that assumes glibc regex. Signed-off-by: Carlo Marcelo Arenas Belón <carenas@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Previous commit has plugged one leak in the normal code path, but there is an early exit that leaves without releasing any resources acquired in the function. Signed-off-by: Junio C Hamano <gitster@pobox.com>
Updating submodules from the upstream did not work well when submodule's HEAD is detached, which has been improved. * jk/submodule-remote-lookup-cleanup: submodule: look up remotes by URL first submodule: move get_default_remote_submodule() submodule--helper: improve logic for fallback remote name remote: remove the_repository from some functions dir: move starts_with_dot(_dot)_slash to dir.h remote: fix tear down of struct remote remote: remove branch->merge_name and fix branch_release()
Remove unnecessary check from "git daemon" code. * cb/daemon-fd-check-fix: daemon: remove unnecesary restriction for listener fd
Leakfix. * jk/fix-leak-send-pack: send-pack: clean-up even when taking an early exit send-pack: clean up extra_have oid array
When "git daemon" sees a signal while attempting to accept() a new client, instead of retrying, it skipped it by mistake, which has been corrected. * cb/daemon-retry-interrupted-accept: daemon: correctly handle soft accept() errors in service_loop
Doc markup fix. * jj/doc-branch-markup-fix: doc: improve formatting in branch section
CI updates. * cb/ci-freebsd-update-to-14.3: ci: update FreeBSD image to 14.3
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
See Commits and Changes for more details.
Created by
pull[bot] (v2.0.0-alpha.2)
Can you help keep this open source service alive? 💖 Please sponsor : )