Re: automatically closing stale pull requests

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 





On 10/15/2018 05:38 PM, Sage Weil wrote:
On Mon, 15 Oct 2018, Gregory Farnum wrote:
On Thu, Oct 11, 2018 at 11:44 AM Sage Weil <sweil@xxxxxxxxxx> wrote:
On Thu, 11 Oct 2018, Gregory Farnum wrote:
I don't love the idea of auto-closing because a lot of those dead PRs
are (or at least *were*) on the shoulders of reviewers and core team
members, not the submitter. Maybe we could start using a
"pending-submitter" tag or something that would be auto-closed?
Yeah... I'm thinking this is as much a forcing fuction for the maintainers
as it is for the contributors.  We can also start out with thresholds that
add the 'stale' flag and comment but don't close issues (yet) until we
have some time to do a scrub and catch up.

Of course I also have not seen a PR scrub in...years? So maybe it's
just that that needs to become a more regular part of the workflow I
and others do. :/ Is your expectation that every bug scrub includes a
review of PRs tagged with that component, and a check of untagged PRs?
Yeah, I think making PR scrubs a part of the regular process is a key
piece of this!
Well, Patrick and I started with this in our overall Ceph project
(ex-bug-only) scrub today. We used the following query and went
through the last month's worth of PRs that it matched, plus started
cleaning up some of the old ones:

https://github.com/ceph/ceph/pulls?q=is%3Apr+is%3Aopen+-label%3Acore+-label%3Acephfs+-label%3Amgr+-label%3Arbd+-label%3Argw+-label%3Aceph-volume+-label%3Abluestore+-label%3Adashboard+-label%3Acrimson+-label%3Apybind+-label%3Aperformance+-label%3Acommon+milestone%3Anone+sort%3Acreated-desc

There actually weren't too many it hit within the last month, although
I'm sure the project-specific scrubs are going to be more
time-intensive and I'm not entirely sure how often eg the pybind label
will get looked at, so there may still be some issues there.
Great!

I also opened PR with a configuration for the probot/stale plugin:

	https://github.com/ceph/ceph/pull/24598

Initially it should only warn about closing issues but not close them yet.
Feedback welcome on the comment itself!

The times/dates should be explicit when marked stale.  Instead of "...automatically marked as stale because it has not had recent activity" it should explicitly say how long it has been stale for (yes, you can figure it out if you look, but it would be better to make it explicit).  Next: "It will be closed if no further activity occurs."  It should say when it will be closed.  That can be a specific date, or it could be "in N weeks" or whatever.  It should be very clear to the submitter what criteria is being used and how much time they have to fix things if they care.


sage




[Index of Archives]     [CEPH Users]     [Ceph Large]     [Information on CEPH]     [Linux BTRFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux