Just for reference, I ran a script I wrote using Stefan’s as inspiration, and there are 413 PRs without any activity for the past 6 months.
Bernardo > On Apr 14, 2025, at 6:39 AM, Štefan Miklošovič <smikloso...@apache.org> wrote: > > > > On Mon, Apr 14, 2025 at 3:22 PM Josh McKenzie <jmcken...@apache.org > <mailto:jmcken...@apache.org>> wrote: >>> Funny that people don't forget to create a PR when trying to make a change >>> but as soon as it is delivered the respective PR is "memory holed". >> We use the PR mechanisms for review but don't use the PR mechanism for >> merge. Makes sense that we open them since they're part of our workflow and >> forget to close them. >> >> I'd much prefer a workflow where we just used the industry standard tools >> for both opening and closing (i.e. had per-branch patches we merged using gh >> after review passed and linked CI passed). But I suspect that's another >> [DISCUSS] thread and >> I should appropriately don metaphorical flame retardant protective gear >> before wading back into that particular dumpster fire. >> > > genuinely chuckled :D Look ... just try to put a JIRA number into a title and > bonus points for closing it afterwards. If a PR is merged, and a ticket is > going to be resolved, people still need to put there a commit url from GitHub > etc ... so maybe internalizing it a little bit more that a PR might be closed > too would be great. > > Unless we start to merge the PRs by "pushing buttons" I don't think this is > going to be resolved. > > What is interesting is that there is automatic creation of a link into a JIRA > ticket when a PR is created (I guess that works by scanning a title of a PR > and linking it to a ticket? Or does it look into the name of a branch of that > PR?). Anyway, I would expect that the same is done when a JIRA is closed - > that it would go over the links of PRs and close them. When it can work one > way, why cannot it work the other way around as well? > >> :D >> >> On Mon, Apr 14, 2025, at 8:27 AM, Mick Semb Wever wrote: >>> >>> >>> On Mon, 14 Apr 2025 at 10:23, Štefan Miklošovič <smikloso...@apache.org >>> <mailto:smikloso...@apache.org>> wrote: >>> BTW If you still do not want to take care of closing it, that is also fine, >>> because we have a script at least. >>> >>> >>> >>> Relying on the PR name seems a bit brittle. Maybe it wouldn't take much to >>> improve it. >>> e.g. would it be possible to also auto-detect which PRs, still open, have >>> no changes to merge ? This is an easy indicator that the PR has otherwise >>> been merged. >>> Stale PRs with file conflicts is another lhf category that can get closed >>> out. >>> >>> Not putting the work on you Stefan, just brainstorming… >>
