You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The CLA [not] signed label handling has been removed because the new CLA bot doesn't use those labels, so it now won't close any stale PRs: python/cpython#91429.
What I wrote in #93 (comment) isn't accurate on two counts:
only-pr-labels doesn't restrict to processing PRs only, it overrides only-labels for PRs, and for issues it will use the only-labels default instead. Anyway, it doesn't matter because only-pr-labels was deleted in the PR.
Aside: One option is to also mark issues as stale, but that's another conversation. We should also figure out a way to reinstate closing old PRs that don't have the CLA signed (i.e. redo #93), that's also another conversation.
If set to a negative number like -1, no issues or pull requests will be marked as stale automatically.
In that case, you can still add the stale label manually to mark as stale.
As I mentioned on Discord, during the migration I kept both the stale and the pending label. My intention was to use them as follow:
pending
the issue/PR is invalid in its current state (e.g. not enough info, CLA not signed, etc.) and will automatically be closed after a fixed period unless the problem is addressed and the label removed
it can be set by both humans (e.g. on a new issue with not enough info) or bots/actions (e.g. CLA still not signed after a period of time) and it can be removed by both (bots for things like a now-signed CLA and humans for the other cases)
stale
the issue/PR has seen no activity for a certain period of time
it should be set automatically by bots/actions and removed automatically on action
Would this work with the current workflow?
My reasoning is that:
stale doesn't apply to new issues with not enough info, but we might want to close them automatically too
I think we shouldn't close issues/PRs automatically just because they are stale, but we can do it if a human set them to pending
In theory we might get rid of stale altogether if pending is applied by bots/action automatically when the CLA is not signed after a certain period, and search queries can be used to search for issues that haven't been updated in a while (providing moreflexiblity at the price of simplicity since you can't filter for label:stale anymore, but have to use updated:<2022-03-13)
Activity
erlend-aasland commentedon Apr 12, 2022
Yeah, I had a look at the docs myself, and that seems to be the correct change for this particular problem.
Regarding the "close-stale-prs-action": Perhaps it's worth it opening a new issue instead of re-opening #93 (that issue is pretty noisy)?
ezio-melotti commentedon Apr 12, 2022
As I mentioned on Discord, during the migration I kept both the
staleand thependinglabel. My intention was to use them as follow:pendingstaleWould this work with the current workflow?
My reasoning is that:
staledoesn't apply to new issues with not enough info, but we might want to close them automatically toopendingIn theory we might get rid of
stalealtogether ifpendingis applied by bots/action automatically when the CLA is not signed after a certain period, and search queries can be used to search for issues that haven't been updated in a while (providing more flexiblity at the price of simplicity since you can't filter forlabel:staleanymore, but have to useupdated:<2022-03-13)hugovk commentedon Apr 13, 2022
This issue is for the pressing need to prevent the current action processing issues, I've created PR python/cpython#91501 for that.
Yes, let's open a new issue for closing stale PRs.
Should we discuss this as part of a new issue or closing stale PRs?
hugovk commentedon Apr 13, 2022
-> #441