[PATCH v5 0/3] fetch: skip unnecessary tasks when using --negotiate-only
- From
Glen Choo <chooglen@google.com>
- Date
- Jan 18, 2022, 18:54 UTC
- Message-ID
- <20220118185453.33710-1-chooglen@google.com>
- In-Reply-To
- <20220113004501.78822-1-chooglen@google.com>
`git fetch --negotiate-only` is used internally by push negotation and it behaves very differently from other uses of `git fetch`, e.g. it does not update refs or fetch objects. But because of how cmd_fetch() is written, `git fetch --negotiate-only` performs tasks that it shouldn't. This is results in behavior that is unnecessary at best, and incorrect at worst:
* Submodules are updated if enabled by recurse.submodules=true, but negotiation fetch doesn't actually update the repo, so this doesn't make sense. * Commit graphs will be written if enabled by fetch.writeCommitGraph=true, but this is unnecessary because no objects are fetched [1]. * gc is run, but according to the commit message in [2], we only do this because we expect `git fetch` to introduce objects.
Make `git fetch --negotiate-only` handle these tasks more rigorously by doing the following:
* Make cmd_fetch() skip irrelevant tasks if we know for certain that objects will not be fetched * Disable submodule recursion and die() if a user explicitly asks for it
[1] This is also confirmed by Documentation/config/fetch.txt, which states that Git should only write commit graphs if a pack-file is downloaded. [2] 131b8fcbfb (fetch: run gc --auto after fetching, 2013-01-26)
Changes since v4: * drop an unnecessary block (thanks Junio!)
Changes since v3: * change commit message subject: builtin/fetch -> fetch --negotiate-only * move the 'goto cleanup' to _after_ the submodule updating task because we may want to update submodules even if objects were not fetched (as pointed out by Junio, thanks!) * disable submodule recursion in the patch that checks for --negotiate-only + --recurse-submodules, so we never silently ignore --recurse-submodules. * incorporate some of Jonathan's suggestions (thanks!)
Changes since v2: * added a prepatory patch that introduces a "goto cleanup" * drop an unnecessary line move (as suggested by Junio) * check for user-given --recurse-submodules and die() (as suggested by Jonathan and Junio) * update --negotiate-only's documentation
Changes since v1: * added more context to commit message * added a NEEDSWORK comment
Glen Choo (3): fetch: use goto cleanup in cmd_fetch() fetch: skip tasks related to fetching objects fetch --negotiate-only: do not update submodules
Documentation/fetch-options.txt | 1 + builtin/fetch.c | 40 ++++++++++++++++++++++++++++++--- t/t5516-fetch-push.sh | 12 ++++++++++ t/t5702-protocol-v2.sh | 12 ++++++++++ 4 files changed, 62 insertions(+), 3 deletions(-)
Range-diff against v4:
1: ffa1a24109 = 1: ffa1a24109 fetch: use goto cleanup in cmd_fetch()
2: b0c73e8135 = 2: b0c73e8135 fetch: skip tasks related to fetching objects
3: 914d30866f ! 3: f929297961 fetch --negotiate-only: do not update submodules
@@ builtin/fetch.c: int cmd_fetch(int argc, const char **argv, const char *prefix)
+ if (recurse_submodules_cli != RECURSE_SUBMODULES_DEFAULT)
+ recurse_submodules = recurse_submodules_cli;
+
-+ if (negotiate_only) {
++ if (negotiate_only)
+ switch (recurse_submodules_cli) {
+ case RECURSE_SUBMODULES_OFF:
+ case RECURSE_SUBMODULES_DEFAULT: {
@@ builtin/fetch.c: int cmd_fetch(int argc, const char **argv, const char *prefix)
+ default:
+ die(_("--negotiate-only and --recurse-submodules cannot be used together"));
+ }
-+ }
++
+
if (recurse_submodules != RECURSE_SUBMODULES_OFF) {
int *sfjc = submodule_fetch_jobs_config == -1base-commit: 90d242d36e248acfae0033274b524bfa55a947fd
-- 2.33.GIT