git remote prune does not handle negative refspecs properly. If a negative refspec got added after a fetch, a user may expect that the resulting refs that were previously fetched would now be considered stale, and thus removed by git remote prune. Show that this isn't the case with a new test case. Signed-off-by: Jacob Keller <jacob.keller@xxxxxxxxx> --- t/t5505-remote.sh | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/t/t5505-remote.sh b/t/t5505-remote.sh index 8cad753023ef..0810e7d8d0b7 100755 --- a/t/t5505-remote.sh +++ b/t/t5505-remote.sh @@ -399,6 +399,19 @@ test_expect_success 'prune --dry-run' ' test_cmp expect output ' +cat >expect <<EOF +Pruning origin +URL: $(pwd)/one + * [would prune] origin/side2 +EOF + +test_expect_failure 'prune --dry-run negative refspec' ' + test_config -C test --add remote.origin.fetch ^refs/heads/side2 && + git -C test remote prune --dry-run origin >output && + git -C test rev-parse refs/remotes/origin/side2 && + test_cmp expect output +' + test_expect_success 'add --mirror && prune' ' mkdir mirror && git -C mirror init --bare && -- 2.36.1
