diff mbox series

fetch: fix duplicate remote parallel fetch bug

Message ID 20230119220538.1522464-1-calvinwan@google.com (mailing list archive)
State Accepted
Commit 06a668cb90a6e8628f295adb6177855bb0a85a4a
Headers show
Series fetch: fix duplicate remote parallel fetch bug | expand

Commit Message

Calvin Wan Jan. 19, 2023, 10:05 p.m. UTC
Fetching in parallel from a remote group with a duplicated remote results
in the following:

error: cannot lock ref '<ref>': is at <oid> but expected <oid>

This doesn't happen in serial since fetching from the same remote that
has already been fetched from is a noop. Therefore, remove any duplicated
remotes after remote groups are parsed.

Signed-off-by: Calvin Wan <calvinwan@google.com>
---
 builtin/fetch.c          | 1 +
 t/t5506-remote-groups.sh | 9 +++++++++
 2 files changed, 10 insertions(+)

Comments

Junio C Hamano Jan. 19, 2023, 10:40 p.m. UTC | #1
Calvin Wan <calvinwan@google.com> writes:

> Fetching in parallel from a remote group with a duplicated remote results
> in the following:
>
> error: cannot lock ref '<ref>': is at <oid> but expected <oid>
>
> This doesn't happen in serial since fetching from the same remote that
> has already been fetched from is a noop. Therefore, remove any duplicated
> remotes after remote groups are parsed.
>
> Signed-off-by: Calvin Wan <calvinwan@google.com>
> ---
>  builtin/fetch.c          | 1 +
>  t/t5506-remote-groups.sh | 9 +++++++++
>  2 files changed, 10 insertions(+)
>
> diff --git a/builtin/fetch.c b/builtin/fetch.c
> index b06e454cbd..508ab2670c 100644
> --- a/builtin/fetch.c
> +++ b/builtin/fetch.c
> @@ -2225,6 +2225,7 @@ int cmd_fetch(int argc, const char **argv, const char *prefix)
>  			argv++;
>  		}
>  	}
> +	string_list_remove_duplicates(&list, 0);

As it always is possible to edit .git/config manually, it is
necessary to perform deduplication like this patch does on the
consumer side of the list, but do you know if our tool create
duplication, or is it entirely something the end-user does manually?

If it is the former, I am wondering if we should also fix such a
code path that does so in the first place.

Will queue.  Thanks.
Calvin Wan Jan. 19, 2023, 11:44 p.m. UTC | #2
> As it always is possible to edit .git/config manually, it is
> necessary to perform deduplication like this patch does on the
> consumer side of the list, but do you know if our tool create
> duplication, or is it entirely something the end-user does manually?

I checked git-remote and there is protection against duplication
there, but I'm unsure if there are other places where remotes are
being added/renamed. I discovered the bug initially by using
git-config.
diff mbox series

Patch

diff --git a/builtin/fetch.c b/builtin/fetch.c
index b06e454cbd..508ab2670c 100644
--- a/builtin/fetch.c
+++ b/builtin/fetch.c
@@ -2225,6 +2225,7 @@  int cmd_fetch(int argc, const char **argv, const char *prefix)
 			argv++;
 		}
 	}
+	string_list_remove_duplicates(&list, 0);
 
 	if (negotiate_only) {
 		struct oidset acked_commits = OIDSET_INIT;
diff --git a/t/t5506-remote-groups.sh b/t/t5506-remote-groups.sh
index 5bac03ede8..0e176175a3 100755
--- a/t/t5506-remote-groups.sh
+++ b/t/t5506-remote-groups.sh
@@ -99,4 +99,13 @@  test_expect_success 'updating remote name updates that remote' '
 	! repo_fetched two
 '
 
+test_expect_success 'updating group in parallel with a duplicate remote does not fail (fetch)' '
+	mark fetch-group-duplicate &&
+	update_repo one &&
+	git config --add remotes.duplicate one &&
+	git config --add remotes.duplicate one &&
+	git -c fetch.parallel=2 remote update duplicate &&
+	repo_fetched one
+'
+
 test_done