For `review-submitted` labels, don't require that the reviewer is assigned

[tgross35]

Leaving a "request changes" review only applies the S-waiting-on-review label if the review comes from somebody who is assigned to the PR. However, this means we don't get that benefit on repos where auto-assign isn't enabled. Could this be loosened so a relabel happens when done by anybody with write access, rather than restricted to the assignee? I think removing the assignment check should suffice here since users without write access can't change leave the red/green reviews.

This could also be made configurable. IMO this probably isn't necessary since in practice leaving a "request changes" review usually always indicates action required, regardless of whether the assignee is up to date, and "comment" is used for things that aren't blocking.

Example PR: rust-lang/libc#4615

[tgross35]

I think it could make sense to do the same for review-requested, if the requested reviewer has write access.

[Urgau]

This almost feels like you want to substitute yourself from the assignee, at leats until your requested changes are handled properly.

I feel like changing the labels would not be enough for your use-case, as the assignee might review changes based on your requested changes, while you probably want to review them yourself.

[tgross35]

To clarify, are you saying that S-waiting-on-review is referring only to the assignee? I always took it more from the author's perspective, meaning that it's waiting on a review from somebody who may or may not be assigned.

I probably got that impression because it wouldn't make sense to have the S-waiting-on-author label if it's waiting on a review from somebody who doesn't happen to be assigned.

[Urgau]

To clarify, are you saying that S-waiting-on-review is referring only to the assignee?

Not really, it's as much for the reviewer/assignee as it is for the author. In other words I think of it as a shared status.

I personally won't look at a PR who is S-waiting-on-author, but will look at a PR marked as S-waiting-on-review and this last part is problematic IMO, as it's not waiting for the assignee (the assigned reviewer) but on someone else, and that isn't reflected in the status it-self.

That's why I almost feel like we want the status flip + temporary assignment. Not sure if it makes sense.