Hi Venkata and Martijn, Stale PRs can clutter the repository, and closing them helps keep it organized and ensures that only relevant and up-to-date PRs are present. An open PR should imply that work is ongoing or needs review. On the other hand, a PR that reaches a stale state rather means the proposed changes are no longer a priority or relevant. Closing them allows contributors and maintainers to focus on more critical and current tasks. Because of that, I'm inclined to close stale PRs.
Best regards, Nandor Martijn Visser <martijnvis...@apache.org> ezt írta (időpont: 2023. szept. 19., K, 23:01): > > Hi Venkata, > > Thanks for opening the discussion, I've been thinking about it quite a > bit but I'm not sure what's the right approach. > > From your proposal, the question would be "What's the added value of > closing these PRs"? I don't see an immediate value of that: it would > just close PRs where maintainers haven't been able to perform a > review, but getting a PR closed without any feedback is also > demotivating for a (potential new) contributor. I think the important > thing is that we get into a cycle where maintainers can see which PRs > are ready for review, and also a way to divide the bulk of the work. > Because doing proper reviews requires time, and these resources are > scarce. > > I do think that we can make lives a bit easier with some automation: > * There are a lot of PRs which don't follow the contribution guide (No > Jira ticket, no correct commit message etc). For the externalized > connector repositories, we've been trying Boring Cyborg to provide > information back to contributors if their PRs are as expected. If the > PR doesn't follow the contribution guide, I'm included to give such a > PR less attention review. That's primarily because there are other PRs > out there that do follow these guides. > * There are even more PRs where the CI has failed: in those cases, a > review also makes less sense, given that the PR can't be merged as is. > I do see that contributors sometimes don't know where to look for the > status of the CI, but IIRC we can't really fix that until we can > finally move to dedicated Github Action Runners instead of the current > setup with Azure, but that's primarily blocked by ASF Infra. > > I'm curious what others in the community think. > > Best regards, > > Martijn > > On Tue, Sep 19, 2023 at 10:33 PM Venkatakrishnan Sowrirajan > <vsowr...@asu.edu> wrote: > > > > Hi Flink devs, > > > > There are currently over 1,000 open pull requests > > <https://github.com/apache/flink/pulls?q=is%3Aopen+is%3Apr+sort%3Aupdated-asc> > > (PRs) in the Apache Flink repository, with only 162 having been updated in > > the last two months > > <https://github.com/apache/flink/pulls?q=is%3Aopen+is%3Apr+sort%3Aupdated-asc+updated%3A%3E2023-07-19>. > > This means that more than 85% of the PRs are stale and have not been > > touched. > > > > I suggest setting up Github actions to monitor these stale PRs, and > > automatically closing them if they have not been updated in the last 'x' > > days. Authors would still be able to reopen the closed PRs if they continue > > with their work. This would help to keep the PR queue manageable. > > > > Not sure if this has been discussed in the Apache Flink community before. > > > > Thoughts? > > > > Regards > > Venkata krishnan
