[Meta] PR sign-off policy

[kailuowang]

Looks like we are blocked from making progress towards RC1 due to the fact that we are down to 2 active maintainers together with the 2 maintainer sign-offs policy - PRs submitted by one maintainer often have to wait for a long time for a 3rd maintainer to review and approve.

This has been an on-going issue for a while and luckily we had some awesome new maintainers and what a difference they've made. But now we are down to 2 active maintainers again.

I see three ways going forward (there could be other ones that didn't occur to me):

1. find a third active maintainer (or "reactivate" one of our existing ones). By "active", I mean, dedicate at least 2 hours+/week for cats PR review. That is about 25 minutes/day during weekdays.
2. The "at least 2 maintainers sign-offs" policy was installed when cats was undergoing vigorous development with more active maintainers. Now cats is more stable and mature we could change the sign-off policy so that for PRs submitted by maintainers, only one more maintainer sign-off is required for merging. We still require two maintainer sign-offs for PR merging, it's just that a maintainer authoring the PR counts as one sign-off. TBC, this loosening does NOT apply to the technical decision process. For significant technical decisions (or any significant decisions regarding the project), we still require meticulous discussions and strong consensus among the team and the community. That has been the tradition in the past and will remain so in the future. The loosening in this proposal ONLY applies to the more mechanic code review process on the implementation of the technical decisions.
3. It is what it is, everyone chill.

Thoughts? Opinions? Suggestions?

[tpolecat]

I'm fine with (2) but will also try to find time to look at PRs more often. A problem with having a bunch of committers is that it's really easy to assume someone else will do X, forAll X; especially since we're all so busy.

[iravid]

@alexandru has expressed interest in helping with maintainership of cats a while back (I cannot recall where exactly). Perhaps he could help?

(apologies @alexandru if I am misrepresenting you... :-))

[johnynek]

As a note, I review most PRs that have green CI and I think should be merged most mornings. I do spend probably 15 minutes per day on cats (sometimes more sometimes less).

I would prefer option 1 or 3. Reducing the number of sign offs I fear will reduce the stability of the library. I think it should be pretty hard to get things merg d and make changes at this stage.

[LukaJCB]

I think 2 is a good idea, because it'll allow us to move more quickly. Alternatively we could be less aggressive with invalidating existing reviews on every commit?

To address the point made by @tpolecat, maybe we should use the request-review-feature more extensively.

I'm going to be very busy in November with work + uni and 3 conferences so I really hope we can get RC1 out ASAP :)

[kailuowang]

@johnynek a side topic, since you are here, it seems all your feedbacks on #1927 and #1837 have been addressed, do you want to give a quick sign-off and/or merge on those two or you need more time w/ them? That would unblock RC1 progress.
Sorry for distracting the conversation, I tried to reach to you through gitter but seems you don't go there often.

[rossabaker]

Model (2) has treated http4s well.

I am willing to reactivate. I focused on the ecosystem while it needed more help than cats, but now the ecosystem needs a stable cats. But I am worried about missing tribal wisdom that may have emerged during my dormancy: is the guidelines document reasonably up-to-date?

I would weigh in on any ticket that requested my review, but I'm unsure who would request it: a contributor wouldn't know who to pick from a large list, and I don't know how best to express capacity to a Review Scheduler.

[kailuowang]

Thank you @johnynek for quickly reviewed 3 PRs. Now we are possibly 1 PR away from RC1.

[kailuowang]

@rossabaker and @tpolecat thank you for offering more reviews. It will make a big difference.
The guidelines seems up-to-date but just as incomplete as they used to be. I think it terms of conventions in the code bases, not much changed though. If you guys want, I can send PR review requests your way from time to time. No pressure on keeping up with them though, if you don't have the time simply reassign it back to me (if I am still assignable) or just remove yourself.

[rossabaker]

👍 I will happily accept review assignments, and generally try to watch this repo and cats-dev a bit closer. Thank you for your very hard work getting us this close!

[alexandru]

I can accept review assignments as well.

I'm not sure if I can commit to 25 minutes per day, due to my other commitments, but if it's needed for getting to Cats 1.0, then I can help. Please make me a member of the typelevel org.

And thanks for your hard work @kailuowang and for keeping us on track, you're awesome.

[kailuowang]

Thanks @alexandru for your offering and kind words. With @rossabaker and @tpolecat reviewing PRs, I think we should be good for 1.0. It would be awesome if you'd still like to help maintain Cats going forward. Let me know - I'll be more than happy to start the process (nothing required on your end though).

[ceedubs]

First of all, I want to really thank @kailuowang @LukaJCB and @johnynek for keeping Cats moving. You've put in hard work, and it's greatly appreciated.

I tend to agree with @johnynek's comments. I know that the slow pace can be frustrating, but I think that it's better than not getting several sets of eyes on changes. It also sounds like we have some people who are willing to step up and spend some more time on Cats.

[kailuowang]

@ceedubs yeah, I agree we no longer need to change this policy. I think the issue has been mostly addressed thanks to more maintainers engaging in reviewing. We have been able to move fast since then.

[ceedubs]

@kailuowang do you think that we should close this out?