| opendevreview | Stephen Finucane proposed openstack/pbr master: Fix handling of long_description https://review.opendev.org/c/openstack/pbr/+/965470 | 10:59 |
|---|---|---|
| opendevreview | Stephen Finucane proposed openstack/pbr master: Fix handling of long_description https://review.opendev.org/c/openstack/pbr/+/965470 | 11:07 |
| opendevreview | Merged openstack/ci-log-processing master: Update Opensearch and Opensearch-dasboards default to 3.1.0 https://review.opendev.org/c/openstack/ci-log-processing/+/963930 | 11:17 |
| opendevreview | Stephen Finucane proposed openstack/pbr master: Don't wrap develop on newer setuptools https://review.opendev.org/c/openstack/pbr/+/965474 | 11:28 |
| opendevreview | Stephen Finucane proposed openstack/pbr master: Don't wrap develop on newer setuptools https://review.opendev.org/c/openstack/pbr/+/965474 | 14:10 |
| opendevreview | Clark Boylan proposed openstack/pbr master: Don't wrap develop on newer setuptools https://review.opendev.org/c/openstack/pbr/+/965474 | 14:52 |
| gmaan | clarkb: Re on dropping py36 and 37, i agree on the policy for unmaintained branches testing. I am ok to move those older branch to test py3.8 onwards. | 16:05 |
| clarkb | gmaan: thanks the change is up but currently failing on monasca stuff. I think we can wait for monasca to be retired then force merge if necessary | 16:06 |
| gmaan | acl | 16:06 |
| gmaan | ack | 16:06 |
| gmaan | I think we should EOL all < antreope but i think elodilles wanted those to stay p :) | 16:07 |
| fungi | they can also drop testing | 16:08 |
| fungi | it's repeatedly insisted that's the solution when we need to remove older platforms, but then nobody actually agrees to do it | 16:08 |
| gmaan | yeah but is no testing then they qualify for the EOL | 16:08 |
| fungi | they already qualify for eol anyway | 16:09 |
| fungi | all unmaintained branches do | 16:09 |
| gmaan | yeah | 16:09 |
| fungi | reduced testing is still a middle ground for people who want to continue reviewing changes on them | 16:09 |
| gmaan | 6 unmaintained branches, it seems we are heading towards same challenge we had for EM model | 16:11 |
| fungi | which is exactly what i predicted, sadly | 16:12 |
| elodilles | i thought the main reason for moving from EM model to Unmaintained was due to people were nagging project teams that this and that EM branches did not get a bug fix, and project teams felt the pressure to maintain them, but with Unmaintained model that is not the case anymore, as it is more easy to them to point out that 'that old branch is unmaintained' | 16:15 |
| gmaan | true but I failed to do that in EM model when at the starting of EM, qa said tempest will not support EM but there was still lot of work we end up doing. | 16:16 |
| gmaan | I am keeping myself away from unmaintained but somewhere at some point things comes up | 16:17 |
| fungi | that was one reason, the other was so we could get rid of those branches sooner on projects where they're not being taken care of, where they have job configuration errors, and where they depend on platforms that we need to discontinue running | 16:17 |
| gmaan | I am not against of unmaintained model but keep growing them and no clear EOL deadline seems it will cause issues | 16:17 |
| elodilles | gmaan: yeah, that's unfortunately true, as there are things (e.g. jobs) defined on master branch of tempest, etc. :/ | 16:18 |
| fungi | in the case of projects requiring ubuntu bionic, we literally can't run those jobs any longer once zuul drops support for old enough ansible to communicate with it | 16:18 |
| gmaan | elodilles: honestly saying it make sense to keep one (or say two) old SLURP in unmaintained. | 16:19 |
| gmaan | but keeping v-z seems not what we initially planned when we moved to this model. | 16:20 |
| gmaan | I remember I mentioned to keep these before-slurp model branches for some time so that we cam transition from EM to unmaintained smoothly. But my intent was not to keep them so long | 16:21 |
| elodilles | gmaan: i understand that keeping old branches open is painful, but on the other hand there is people, like zigo, who even would want to have rocky to be open for patches o:) even without CI testing. (I'd rather see less branches open, but with a reasonable amount of testing) | 16:21 |
| gmaan | yeah, i know. that is where EM model did not success and I am saying we are heading to same situation for unmaintained model also (though project teams are not responsible but i feel it always falls on same set of contributors :) | 16:25 |
| elodilles | gmaan: tbh neither i expect that pre-slurp branches will be open for long, though i'd like to help keeping open a couple more, e.g. zed and yoga for some time. if we don't give the possibility for these branches to be open for bug fixes, then people will less likely turn up backporting fixes. (i know, this isn't really happening, but i'm hopelessly optimistic o:)) | 16:28 |
| fungi | honestly, i'm fine with the opendev-side concerns being ignored, it simply gives us license to cut off resources without bothering to coordinate it. if branches have job config errors we can just delete them, or remove those projects entirely from zuul. when we're ready to drop platforms we'll just drop them | 16:30 |
| gmaan | elodilles: ++ for yoga/zed but my main concern is unmaintained->EOL is not happening. or maybe I need to wait when it happen for first time | 16:30 |
| gmaan | I am just afraid if that never happen :) | 16:31 |
| fungi | if openstack wants its projects to be able to merge changes, then they'll prioritize cleanup once opendev stops running jobs for them on any branch | 16:32 |
| fungi | it would be unfortunate if the desire to keep unmaintained branches open while not devoting time to keeping them in shape resulted in stable and master branches no longer being usable either | 16:32 |
| elodilles | gmaan: true, ussuri, train, stein, rocky, queens, pike and ocata were EM branches, not Unmaintained. but those went to EOL. furthermore, quite many projects' zed-victoria branches were EOL'd, though not entire $SERIES, i know | 16:34 |
| elodilles | fungi: i don't want to ignore opendev-side concerns, and yes, please go ahead of cutting off resources if those cannot be kept. i understand that these things have to be done. and i was trying to clean up newly turning up zuul config errors because of missing job definitons nodesets etc and i'll try to do that in the future as well | 16:39 |
| fungi | okay, well just to be clear the ubuntu-bionic nodes are going away very soon. clarkb has a change up to move python 2.7 jobs to a newer platform | 16:41 |
| elodilles | i'm not insisting on keeping open branches forever, i'm just trying to keep them as long as it is possible (with some value). | 16:41 |
| fungi | remember that zuul has to read configuration from every branch of every project, so the longer these branches stay open and the more of them there are, the greater the complexity of zuul's config state becomes | 16:42 |
| clarkb | which would be fine if people were more proactively caring for it | 16:44 |
| clarkb | I think the issue is that for 90% of the community anything older than 6 months stops mattering | 16:44 |
| clarkb | this is why I'ev been trying to get us to anticipate future cleanups and address them early when people still care | 16:44 |
| clarkb | we won't predict everything but a lot of stuff is probably doable | 16:45 |
| clarkb | in this case the python27 jobs could've been updated to focal then jammy over time as an example | 16:45 |
| opendevreview | Merged openstack/pbr master: Fix handling of long_description https://review.opendev.org/c/openstack/pbr/+/965470 | 16:47 |
| opendevreview | Merged openstack/pbr master: Misc fixes https://review.opendev.org/c/openstack/pbr/+/965313 | 16:47 |
| elodilles | yes, unfortunately there are not many maintainers, and even stable-maint cores mostly look at bug fix backports and active CI errors, not really thinking ahead of fixing possible changes of CI. and even less actively the zuul config errors. | 16:56 |
| tonyb | elodilles: You can add me back into stable-main-core if you'd like. | 17:13 |
| tonyb | elodilles: tangent to zuul config errors | 17:14 |
| fungi | the pending pbr change is just waiting for tempest to finish now. once it merges i'll push up the release request | 17:31 |
| elodilles | tonyb: you were not removed from stable-maint-core, you are still there ;) | 17:44 |
| tonyb | elodilles: Oh! Interesting. I was looking the other day and didn't have +2 .... must be a different issue. | 17:44 |
| elodilles | tonyb: *some* of the python-*client project's have such ACLs that stable-maint-core group is not added, maybe you looked at one such project | 17:47 |
| tonyb | Yeah. I jumped to the wrong conclusion | 17:47 |
| fungi | looks like tempest is wrapping up, i'll get the pbr release request prepped so i can paste in the commit id once it's final | 18:24 |
| clarkb | thanks! | 18:25 |
| opendevreview | Merged openstack/pbr master: Don't wrap develop on newer setuptools https://review.opendev.org/c/openstack/pbr/+/965474 | 18:29 |
| clarkb | fungi: ^ finally | 18:29 |
| fungi | https://review.opendev.org/c/openstack/releases/+/965848 pbr 7.0.3 [NEW] | 18:30 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!