Re:Ask HN: How do you do code review, if at all?
by why-el on Thursday June 11, 2015 @01:14PM
(#9700973)
Only one reviewer is enough for us. Once a PR is submitted to GitHub, it should be merged by the reviewer, who usually only has to click that GitHub merge button. This means we favor a rebase-workflow, which has worked great for us.
There are two cases in which we bypass this mechanism. One is hot-fixes that can't afford to go through this process, and the other is changes too trivial to wait for a reviewer. This later case is rare, since there is always a reviewer, but we do have some remote colleagues (Australia, SF) who should use their judgment in deciding whether a PR is worth being reviewed. This is slightly subjective but so far all decisions in this regard were excellent. Once the team is bigger perhaps we'd do away with this and keep hot-fixes as the only review-by-passers.
There are two cases in which we bypass this mechanism. One is hot-fixes that can't afford to go through this process, and the other is changes too trivial to wait for a reviewer. This later case is rare, since there is always a reviewer, but we do have some remote colleagues (Australia, SF) who should use their judgment in deciding whether a PR is worth being reviewed. This is slightly subjective but so far all decisions in this regard were excellent. Once the team is bigger perhaps we'd do away with this and keep hot-fixes as the only review-by-passers.