I like the flagging, but not closing approach. I also like Robert's proposal to ping, but not close - it means the people with the power to do something about the stale issues face the consequence of letting them languish :)
Mike On Fri, Feb 6, 2026 at 9:42 AM Yufei Gu <[email protected]> wrote: > I think we don't even need to flag issues as stale, but I'm fine to do > that. Approved the PR. Thanks a lot for acting on it, Alex! > > Yufei > > > On Fri, Feb 6, 2026 at 6:10 AM Alexandre Dutra <[email protected]> wrote: > > > Hi all, > > > > Alright, here is a PR: https://github.com/apache/polaris/pull/3683 > > > > It still marks issues stale but doesn't close them. It also fixes some > > misconfigurations. > > > > If that's not what you folks want, let's continue this discussion in the > > PR. > > > > Thanks, > > Alex > > > > On Fri, Feb 6, 2026 at 12:04 PM Jean-Baptiste Onofré <[email protected]> > > wrote: > > > > > > Hi > > > > > > My take on this one is that we have consensus to close PRs but not > > issues. > > > > > > I would suggest to start like this and revisit later if needed. > > > > > > Regards > > > JB > > > > > > Le ven. 6 févr. 2026 à 11:58, Alexandre Dutra <[email protected]> a > > écrit : > > > > > > > Hi all, > > > > > > > > What are the next steps here? It seems we have more people in favor > of > > > > NOT closing than closing. Is my understanding correct? > > > > > > > > Thanks, > > > > Alex > > > > > > > > On Thu, Feb 5, 2026 at 12:37 AM Yufei Gu <[email protected]> > wrote: > > > > > > > > > > Thanks Alex for raising this. +1 to not closing the issue > > automatically. > > > > > > > > > > Yufei > > > > > > > > > > > > > > > On Wed, Feb 4, 2026 at 4:44 AM Francois Papon <[email protected]> > > wrote: > > > > > > > > > > > Hi Alexandre, > > > > > > > > > > > > Sounds good to me, we can try this approach and see if it fit in > > the > > > > > > future. > > > > > > > > > > > > regards, > > > > > > > > > > > > François > > > > > > > > > > > > Le 04/02/2026 à 11:44, Alexandre Dutra a écrit : > > > > > > > Hi all, > > > > > > > > > > > > > > I'm open to either approach, but I wanted to explain why having > > this > > > > > > > job running might not be such a big deal: > > > > > > > > > > > > > > Since many community contributors subscribe to all Polaris > > > > > > > notifications, any stale issue notification from the CI job > will > > be > > > > > > > received by many people. This provides us with an opportunity > to > > > > > > > evaluate whether an issue should be reopened or not. > > > > > > > > > > > > > > In fact, we just saw this in action: the job flagged this issue > > [1] > > > > as > > > > > > > stale 8 hours ago, and Robert immediately unflagged it :-) > > > > > > > > > > > > > > While frequently unflagging issues could become a burden, for > the > > > > time > > > > > > > being, the effort required seems minimal. > > > > > > > > > > > > > > Just my 2 cents. > > > > > > > > > > > > > > Thanks, > > > > > > > Alex > > > > > > > > > > > > > > [1]: > > > > > > > > https://github.com/apache/polaris/issues/3086#issuecomment-3844852237 > > > > > > > > > > > > > > > > > > > > > On Wed, Feb 4, 2026 at 9:42 AM Francois Papon > > > > > > > <[email protected]> wrote: > > > > > > >> Hi, > > > > > > >> > > > > > > >> I am agree with JB, closing PR automatically after 14 days can > > be a > > > > > > >> little agressive and whitout review, the users will not > > understand > > > > why. > > > > > > >> > > > > > > >> Closing a PR without answer activity after a delay from the > user > > > > make > > > > > > >> more sense to me. > > > > > > >> > > > > > > >> regards, > > > > > > >> > > > > > > >> François > > > > > > >> [email protected] > > > > > > >> [email protected] > > > > > > >> > > > > > > >> Le 04/02/2026 à 09:33, Jean-Baptiste Onofré a écrit : > > > > > > >>> Hi Alex, > > > > > > >>> > > > > > > >>> Thanks for starting this discussion! > > > > > > >>> > > > > > > >>> While I am comfortable with automatically closing PRs (as the > > > > author > > > > > > can > > > > > > >>> always comment to keep them open), I don't believe we should > > > > > > automatically > > > > > > >>> close issues. Issues are typically opened for a good reason > and > > > > should > > > > > > be > > > > > > >>> reviewed, reproduced, and investigated. I prefer having > > reviewers > > > > > > manually > > > > > > >>> close issues when appropriate. > > > > > > >>> > > > > > > >>> Regards, > > > > > > >>> JB > > > > > > >>> > > > > > > >>> On Tue, Feb 3, 2026 at 10:52 PM Alexandre Dutra < > > [email protected] > > > > > > > > > > > wrote: > > > > > > >>> > > > > > > >>>> Hi team, > > > > > > >>>> > > > > > > >>>> A recent PR [1] fixed a misconfiguration that had prevented > > the > > > > > > >>>> automatic closing of stale issues from working in Polaris. > > While > > > > this > > > > > > >>>> feature seems to have been intended from the start, its > > > > re-enablement > > > > > > >>>> raises a fundamental question: should we be closing stale > > issues > > > > > > >>>> automatically? > > > > > > >>>> > > > > > > >>>> Arguments for closing include: > > > > > > >>>> > > > > > > >>>> - It helps to clear out issues that are no longer being > > actively > > > > > > >>>> worked on or are irrelevant. > > > > > > >>>> > > > > > > >>>> - The system provides a 14-day grace period before an issue > is > > > > closed, > > > > > > >>>> giving anyone the chance to comment and keep the issue open. > > > > > > >>>> > > > > > > >>>> Arguments against closing include: > > > > > > >>>> > > > > > > >>>> - An issue might still be valid even if the original > reporter > > has > > > > > > >>>> become inactive. Closing it could lead to losing track of > > > > important, > > > > > > >>>> unresolved problems. > > > > > > >>>> > > > > > > >>>> What are your thoughts on this? > > > > > > >>>> > > > > > > >>>> Thanks, > > > > > > >>>> Alex > > > > > > >>>> > > > > > > >>>> [1]: https://github.com/apache/polaris/pull/3636 > > > > > > >>>> > > > > > > > > > > > > >
