Hacker News new | ask | show | jobs
by facorreia 985 days ago
It’s very weird that PR approvals remain “approved” after changes to the PR that was approved.
3 comments

This can go either way, and it's ultimately a social problem, not a technical one.

I find that I fairly often notice some minor issues, like a typo in a comment. I do want the submitter to fix those, but making them go through a full review cycles is an unfortunate waste of time as I usually trust them to be adult enough to just make the fix and submit.

So I find myself wanting to say "Approve modulo these minor issues". If a project wants to allow this, it needs to either not require approvals, or allow approvals to remain even after changes to a PR.

There's a setting to automatically revoke approvals as "stale" after new changes get pushed
Typical PR flow on my team for PRs to dev branch is reviewer approves with comments, then submitter addresses comments with new commits and merges (unless they judge the new commits are significant changes). This makes it lighter weight to offer comments since it doesn't mean another delay for a rereview. The reviewer gets emailed about any new commits that come in, so all changes do get seen.

Github makes it easy to review just the changes since your last approval, a feature which I think obviates the need for changesets as in the OP.