Would a ‚time based closure’ not be close to no 2, weakened to something like likely not to be merged?
If a PR was really excellent, would it not be looked at by somebody in good time? Oscar made an additional point about GSoC I am convinced that the flood of low level (Oscar’s judgement. I do not have the skills to judge them) PRs is due to the fact that the submitters want to participate in GSoC. I cannot judge the pros of sympy being “in” GSoC vs. the drawback of the flood of PRs, but surely the experts must have opinions based on their experience on this question. Peter From: [email protected] <[email protected]> On Behalf Of Jason Moore Sent: Sunday, February 1, 2026 7:53 AM To: [email protected] Subject: Re: [sympy] AI generated pull requests In the past we've used the "closed" designation on a PR to mean: 1) this is merged into master and 2) this will definitely not be merged into master. If we close PRs based on inactivity time, then we have PRs labeled "closed" which are neither 1 or 2, they still have the state "could be or might be merged to master or might be rejected" but now we've labeled them with "closed" which would seemingly imply 1 or 2. So it seems to me if you close based on inactivity time, then the meaning of "open" or "closed" PR no longer has distinct meanings. Jason moorepants.info <http://moorepants.info> +01 530-601-9791 On Sun, Feb 1, 2026 at 3:53 AM '[email protected] <mailto:[email protected]> ' via sympy <[email protected] <mailto:[email protected]> > wrote: I like Oscar's idea of a pre-warning that the PR will be closed in a month with suggestions about things to check if the author wishes to pursue getting the PR merged. I can understand from the maintainers' view not wanting the time frame for automatic closure of to be too long. I only have short bursts of time I can spend contributing to the various open-source projects I am involved in, so am likely to have PRs that would not be completed in a two month time frame. However, from my perspective that could easily be worked around by opening a new PR referring back to the old one, when I am able to cycle back around to work on it again. Thus, I think that having PRs that I am continuing to work on closed after 2 months of inactivity would be fine. It is unlikely to deter me from continuing on projects I am interested in over the long term. I do think the shorter time frame will clean up the repo and may be effective at discouraging people who are not truly serious about working on sympy. I strongly suggest that automatic closure of PRs after a period of inactivity be implemented. From my perspective as a sometime contributor, I am not sure that the time frame matters as much as the fact that it should happen. The time frame should be chosen to work best for the development cycle the core maintainers can manage. Jonathan On Saturday, January 31, 2026 at 5:10:44 PM UTC-6 Oscar wrote: On Sat, 31 Jan 2026 at 22:37, '[email protected]' via sympy <[email protected] <mailto:[email protected]> > wrote: > > As a minor contributor, I strongly agree with the idea that PRs without > activity for some period of time should be automatically closed. I am not > sure the core maintainers and reviewers need to be notified. The originator > of the PR should be notified with a message explaining that it was closed > because of inactivity over the last XX period of time. They should be > encouraged to review the PR carefully and decide if they have the time and > interest in adjusting the code so that it meets all requirements for merging > and address any concerns raised in the PR before it was closed. If so, they > should open a new pull request with code updated to pass all tests when built > with the current development branch and refer to the old PR in case reviewers > want to look at the history. > > I would be in favor of the period of inactivity being in the range of 6 > months to a year. This would potentially close both pull requests I currently > have open (https://github.com/sympy/sympy/pull/28258 and > https://github.com/sympy/sympy/pull/24574). This seems reasonable, because > although I would be interested in pursuing both of them, the reality is that > my primary job is as a Chemistry Professor/Computational Quantum Chemist and > I am unlikely to have much time for work on either of these until at least > the end of the current semester. I do not object to opening a new PR or > reopening the old one when I get back to being able to consider the code. I'm not sure how this would work in practice but for example if someone else closes your PR then I think it isn't possible for you to reopen it unless you are a member (some who can merge PRs). You could of course comment that you would like to reopen and then a member could do it and we could make sure that the bot that closes the PR would leave a message explaining that. I'm not sure about 6 months. I think that basically after 1 month of inactivity the PR is usually forgotten but the amount of time passed is still small enough that the PR could easily be revived if people were reminded and actually wanted to continue with it. If all maintainers simply happened to overlook a PR then 1 month is probably a good amount of time for there to be some kind of notification so perhaps a bot could comment then that it has been inactive for 1 month and then if no one does anything then at the 2 month mark it can be closed. It would probably be helpful for some people if the 1 month bot message explains some common reasons to the author like "if CI checks have failed and there are red crosses everywhere then that might explain why no one has reviewed your PR". This is something that is more common right now in the AI age that someone has opened a PR with broken code and then all of the checks have failed but it almost seems like the author has not seen that all the tests have failed. I would not generally bother commenting to say something like "as you can see all the tests have failed" but perhaps after 1 month it might be good to point that out to the author. If we wait longer than 2 months then what are we actually waiting for? No one is actually going to go back more than 2 months looking for PRs to review. If anyone wants to revive it further in the future then it can just be reopened but if no one expresses any positive interest in doing that then what benefit do we get from keeping a 2 month old PR in the "open" state rather than the "closed" state? The closed PRs are still there with all their code and message for everyone to see. -- Oscar -- You received this message because you are subscribed to the Google Groups "sympy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected] <mailto:[email protected]> . To view this discussion visit https://groups.google.com/d/msgid/sympy/e5ae466e-1cf1-40db-bfdc-5a8ba525ae5bn%40googlegroups.com <https://groups.google.com/d/msgid/sympy/e5ae466e-1cf1-40db-bfdc-5a8ba525ae5bn%40googlegroups.com?utm_medium=email&utm_source=footer> . -- You received this message because you are subscribed to the Google Groups "sympy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected] <mailto:[email protected]> . To view this discussion visit https://groups.google.com/d/msgid/sympy/CAP7f1AhcGTF_wTD96CyUFk3uPRpE3d_kxWm3-sW_5dLe20AXVg%40mail.gmail.com <https://groups.google.com/d/msgid/sympy/CAP7f1AhcGTF_wTD96CyUFk3uPRpE3d_kxWm3-sW_5dLe20AXVg%40mail.gmail.com?utm_medium=email&utm_source=footer> . -- You received this message because you are subscribed to the Google Groups "sympy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/d/msgid/sympy/050001dc934c%24551688d0%24ff439a70%24%40gmail.com.
