We use test_commit_bulk to update the ref "main". But without qualification, that is really creating ".git/main", not an actual branch in "refs/heads/". In the end it does not really matter, because after creating "main" its only purpose is for us to point HEAD to its same commit id. And that works regardless of how we named it (the later call to update-ref just calls it "main", but that is OK; we are resolving it to an oid there, so our DWIM logic will find it in "refs/heads/"). Since it seems like going outside of "refs/" was accidental here, let's do the more normal thing and just use a branch. That makes the test less confusing and will future proof us against ref updates becoming more picky. Signed-off-by: Jeff King <peff@xxxxxxxx> --- t/t5619-clone-local-ambiguous-transport.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/t/t5619-clone-local-ambiguous-transport.sh b/t/t5619-clone-local-ambiguous-transport.sh index cce62bf78d..70db437496 100755 --- a/t/t5619-clone-local-ambiguous-transport.sh +++ b/t/t5619-clone-local-ambiguous-transport.sh @@ -21,7 +21,7 @@ test_expect_success 'setup' ' echo "secret" >sensitive/secret && git init --bare "$REPO" && - test_commit_bulk -C "$REPO" --ref=main 1 && + test_commit_bulk -C "$REPO" --ref=refs/heads/main 1 && git -C "$REPO" update-ref HEAD main && git -C "$REPO" update-server-info && -- 2.45.0.rc1.416.gbe2a76c799
