Overhaul branch protection rules

[tsalo]

Summary

I think we should drop the number of required approvals from 2 to 1 and we should stop dismissing reviews as stale on new commits. I think we can trust contributors to request new reviews if their changes invalidate old reviews.

[handwerkerd]

I agree. I could see requiring only 1 review, but adding to the contributors guide that 2 reviews are expected for PRs that include substantive functionality or code changes. We can trust contributors to self-identify which ones those are.

[tsalo]

From today's call, everyone's good with this. I'll handle the documentation changes and under-the-hood settings.