meta: handling abandoned PRs

[refack]

This came up WRT to #23615 (comment), and specifically WRT to code & learn PRs.

What do we do with abandoned PRs, that is PR that are either incomplete, have outstanding changes requested, or questions left unanswered. Where the author disengaged, or is simply unresponsive for a long time?
And how much time needs to pass before we consider a PR as such.

[cjihrig]

I think if a collaborator wants to shepherd an abandoned PR, then more power to them.

[jasnell]

A PR that was opened 5 days ago by a code-and-learn participant does not really qualify as "abandoned". People get busy and have other things going on in their lives. It's worth being patient.

Generally the way we've handled stalled PRs in the past is to either: (1) ping the author after about a month or more to see if they are still interested in progressing the PR forward, (2) if the PR is in relatively good shape (only style nits remaining) and passes CI, then landing it is fine (we've done this in the past), (3) if the PR still needs work, we label it "stalled" and wait to see if the author or anyone else wishes to pick it up, (4) after several months time, if it becomes obvious that the PR is going no where after it's been stalled for a bit, we can go ahead and close it.

[refack]

My opinion:
Either close them or some Collaborator adopts them (into a new PR).
Especially for Code and Learn PR, IMHO we are doing a disservice to the participants if we over-reach our role as stewards and actually do the work for them. That missed the "and Learn" part of the process. It's my opinion that interacting with reviewers, and learning GitHub etiquette is a major part of the learning experience, and we should not "rob" the participants of it.

[jasnell]

It's only been five days. Give it some time. Part of GitHub etiquette is being patient.

[refack]

A PR that was opened 5 days ago by a code-and-learn participant does not really qualify as "abandoned". People get busy and have other things going on in their lives. It's worth being patient.

I think that it that specific C&L case the steward jump the gun and didn't give the original author a chance to engage and continue the learning process. The change on it's own is not urgent, and IMHO the main benefit we can get from it, is the engagement with the original author.
If time passes and it seems the the author abandoned this PR, I say we close it.

[refack]

It's only been five days. Give it some time. Part of GitHub etiquette is being patient.

I'm totally behind being patient, IMHO we should give that author the time to follow up on their own.

[refack]

P.S. AFAIK We also don't have a policy for dismissing other collaborator's reviews without giving them a chance to respond.

[Trott]

This conversation seems to have...not quite run its course, but stalled out. I'm going to close it with a stalled label. Feel free to re-open if there's still something to address here (like policy changes/additions, although feel free to just PR your ideas in).