Hacker News new | ask | show | jobs
by OJFord 705 days ago
I agree, but I usually explain (and do) this from the side of fixing a bug, but where the test suite is currently passing: first commit adds the failing test (shows that it would have caught the error), second commit makes it pass.

Also agree with GP that each commit on master should be passing/deployable/etc., but I don't see why they can't be merge commits of a branch that wasn't like that.
1 comments
epage 705 days ago
That still interferes with `git bisect`. Make the test pass in history but then make it fail in your working directory and work to get it to pass before committing.
link
OJFord 704 days ago
No it doesn't? Only on your unmerged branch anyway, which seems either no big deal or desirable to me.
link