pr_finish() should check if local branch is ahead of remote

[maurolepore]

Ups, I lost the PR (luckily just one line). I run pr_fetch(), committed, then run pr_finish() without running pr_push(). Silly my -- I should have pushed. I guess I didn't know what to expect and didn't read the docs (oh no!). Am I the only one learning this the hard way? Would it make sense for pr_finish() to check against the remote and ask to confirm if the local repo is behind?

[jennybc]

Yeah, it sounds like there should be more safeguards here. I think pr_pause() is what you probably wanted.

It sounds like in this case, that the local repo is ahead, but being behind is more typical (because maintainer or someone else has updated the branch). In any case, internally we do have git_branch_compare() to use in this case.

If it had been a more extensive piece of work that you "lost", you could retrieve it from the reflog. But I realize that no one likes going into the reflog 😬

[hadley]

@jennybc this was easier than expected since we already have check_branch_pushed()