I do this, but not in a way then end up on master. It's a driving force behind my preference for squash-and-rebase merge patterns. A good bugfix PR is often tw

I do this, but not in a way then end up on master. It's a driving force behind my preference for squash-and-rebase merge patterns. A good bugfix PR is often two commits then: one with a test to catch the breakage, another to fix it so the tests pass. Reviewers can see the failing-then-passing CI job logs, so if they agree your test catches the bug, they have additional CI-automated validation your fix worked. Then as long as you squash when completing the merge, you get the best of both worlds.