Thursday, 2026-06-25

-@gerrit:opendev.org- OpenStack Proposal Bot proposed: [openstack/project-config] 994582: Normalize projects.yaml https://review.opendev.org/c/openstack/project-config/+/99458203:05
-@gerrit:opendev.org- Zuul merged on behalf of Dr. Jens Harbott: [openstack/project-config] 994689: Drop config for retired Sunbeam single charm repositories https://review.opendev.org/c/openstack/project-config/+/99468907:45
-@gerrit:opendev.org- Zuul merged: [openstack/project-config] 994582: Normalize projects.yaml https://review.opendev.org/c/openstack/project-config/+/99458208:13
-@gerrit:opendev.org- Joan Gilabert proposed: [opendev/irc-meetings] 994890: Move Watcher meeting one hour later and add chairs https://review.opendev.org/c/opendev/irc-meetings/+/99489013:06
@mordred:waterwanders.comstephenfin: merged the 3.8 patch into the fix tests pass. Assuming it goes green all combined, I think we should land that whole stack and probably cut a release13:45
@fungicide:matrix.orgyeah, i'll try to give them all another look some time today13:48
@mordred:waterwanders.comfungi: thanks! I'm about a rebase a few of them -I'll ping when it's all green14:04
-@gerrit:opendev.org- Zuul merged on behalf of Joan Gilabert: [opendev/irc-meetings] 994890: Move Watcher meeting one hour later and add chairs https://review.opendev.org/c/opendev/irc-meetings/+/99489014:07
@mordred:waterwanders.comfungi: what IRC channel should I join to chat with stephenfin about pbr things?14:08
@stephenfin:matrix.orgmordred I'm here now (I don't generally connect to Matrix)14:11
@stephenfin:matrix.orgSeems we're badly stepping on each others toes 😂 I won't push anything else today14:11
@mordred:waterwanders.comoh yay! I can also join the IRC channel, was just looking for the matrix bridge magic14:12
@mordred:waterwanders.comstephenfin: I think between the two of us we're going to get _something_ green, we seem to violently agree :)14:12
@stephenfin:matrix.orgheh14:12
@stephenfin:matrix.orgI left a review on https://review.opendev.org/c/openstack/pbr/+/989388/. I was hoping to add testing on 3.14 and 3.15-pre (non-voting) earlier in the series, then keep resolut testing to the end since it needs more investigation14:13
@stephenfin:matrix.org*resolute14:13
@stephenfin:matrix.orgDoes that make sense?14:13
@mordred:waterwanders.comahhhhhhh. nod. I grok what you're saying there. yeah, and I agree.14:13
@mordred:waterwanders.comlet me unwind the mess I just made and get it in that shape14:13
@stephenfin:matrix.orggreat. ty 🙏#14:14
@stephenfin:matrix.org* great. ty14:14
@mordred:waterwanders.comalthough - actually - since it's all in the gate (and I *think* it's going to go green this time) maybe we just shift the resolute testing into the py315 patch but otherwise dont' rebase the whoel stack?14:15
@stephenfin:matrix.orgif it works, sure, but I was under the impression more changes were needed. See fungi's comment here https://review.opendev.org/c/openstack/pbr/+/989938/7#message-9db7595df50f3eac38f173accec33558733911b414:17
@stephenfin:matrix.org...and I don't think we want to have to rebase the world while iterating on that14:18
@mordred:waterwanders.comoh - yeah- what I meant was - split the changes where they are at the end of the stack. make one "add 3.14 classifiers and add tox 3.14 testing" then make the second one "add 3.15 dev + resolute testing" - since I think the resolute and 3.15 is the one that's going to need more iteration?14:20
@mordred:waterwanders.comfungi: do we have any docs anywhere (I can't find opendev specific ones) that outline the correct process for joining openstack irc channels from matrix?14:22
@fungicide:matrix.orgmordred: "we" don't but the zuul community does, just a sec...14:25
@mordred:waterwanders.comfungi: I sorted it - do you think it would be ok to add a section to the IRC docs if I get around to it?14:26
@fungicide:matrix.orgmordred: https://zuul-ci.org/docs/zuul/latest/howtos/matrix-irc.html14:26
@fungicide:matrix.orgmordred: absolutely! that would be quite welcome i think14:26
@mordred:waterwanders.comawesome. I'll put it on to my backlog14:27
@fungicide:matrix.orgsomewhere in https://docs.opendev.org/opendev/infra-manual/latest/irc.html ideally14:27
@mordred:waterwanders.comooh - also - that's great zuul docs :)14:27
@fungicide:matrix.orgyes, corvus put a ton of work into that documentation when the zuul community moved to matrix14:28
@fungicide:matrix.orgin other news, i think gerrit is having sads again, i see my gertty repeatedly switching into offline sync mode14:28
@mordred:waterwanders.comfungi: as a stopgap I've linked to that page from the openstack IRC wiki page, as a breadcrumb. Looks like the openstack contributor guide took a similar approach. :)14:46
-@gerrit:opendev.org- Dr. Jens Harbott proposed: [openstack/project-config] 994923: Switch the requirements updates job to monthly https://review.opendev.org/c/openstack/project-config/+/99492314:53
@fungicide:matrix.orgmordred: though it backfires, because people trying to find openstack irc channels end up in the zuul matrix room asking openstack questions15:02
@fungicide:matrix.orgi'm heading out to grab lunchies, but shall return forthwith15:05
-@gerrit:opendev.org- Monty Taylor https://matrix.to/#/@mordred:inaugust.com proposed: [opendev/base-jobs] 994146: Rename opendev-buildset-registry-consumer to opendev-use-container-image https://review.opendev.org/c/opendev/base-jobs/+/99414615:42
@mordred:waterwanders.comfungi: I believe the pbr stack ending in https://review.opendev.org/c/openstack/pbr/+/989938 is worth review now. the third patch in the stack hit an unrelated tempest failure and I've rechecked it. but the first two are green and I expect the others to be green to (assuming tempest complies)15:44
@mordred:waterwanders.comfungi: the first patch is "drop python < 3.8" - it didn't start that way, it started as  https://review.opendev.org/c/openstack/pbr/+/994828 just fixing some tests, but then 3.6 was sad and it seemed to be upstream tox transtive depends. so we folded the test fixes into the drop patch which was in the pipeline already. I would have preferred to keep each one smaller, but I thought that one deserved a little commentary :) 15:51
@clarkb:matrix.orgBut of a slow start for me today. Just got back from a bike ride. There are new Gerrit releases which should fix the replication but with a tag so I'll get a change up for that shortly 15:58
@fungicide:matrix.orgmordred: awesome, i'll be getting to them shortly now that lunch is in the rear-view mirror, just need to double-check nothing else has exploded while i was away16:18
-@gerrit:opendev.org- Clark Boylan proposed: [opendev/system-config] 994938: Update Gerrit images to 3.13.7 and 3.14.1 https://review.opendev.org/c/opendev/system-config/+/99493816:38
@mordred:waterwanders.comClark: lgtm16:51
@clarkb:matrix.orgmordred: I'm way behind on the pbr stuff. THe messaging implies we dropped support for 3.8 but we really dropped support for 3.7 and older right? 3.8 is still supported?16:58
@clarkb:matrix.orgI think 3.8 is important and we will need to revert if we've dropped 3.8 support16:58
@mordred:waterwanders.comClark: that's right - dropped <= 3.7  dropped < 3.817:00
@mordred:waterwanders.comClark: I was actually about to ask you to look at https://review.opendev.org/c/openstack/pbr/+/988753 since you'd previously reviewed it to make sure you're ok with it. the rest of the stack is pretty mechanical. 17:01
@clarkb:matrix.orgok cool. I think 3.8 is still in use and dropping it will make things explode17:01
@mordred:waterwanders.comyeah. I m ean - you know me, I'm internally grumpy about dropping anything17:02
@mordred:waterwanders.combut I think overall ecosystem has made it unreasonably difficult17:02
@fungicide:matrix.orgi'm dropping mad beats, is that a problem?17:02
@clarkb:matrix.organd ya I'll do my best to rereview those changes. I'm currently head down on preparing for "how to opendev" so I'm ready for me meeting later today. https://etherpad.opendev.org/p/opendev-atlas is where I'm putting that together17:02
@mordred:waterwanders.comno, please drop more mad beats17:02
@clarkb:matrix.orgmordred: I honestly strongly disagree on that17:03
@clarkb:matrix.orgI think that we've managed to support python2 and python3 for a very long time with minimal fuss and I have yet to see anything that would prevent us from doing so going forward17:03
@clarkb:matrix.organd if it were up to me I think we should release pbr3 when we drop python217:03
@mordred:waterwanders.comyeah? I mean - the tox install for 3.6 is trying to pull virtualenv that  is trying to pull a transitive that only supports 3.817:03
@clarkb:matrix.orgbut I've only said it 100 times and no one seems to agree so meh17:03
@mordred:waterwanders.comit's funny to me that 2.7 is workign but 3.6 is not17:04
@clarkb:matrix.orgmordred: so we drop tox17:04
@fungicide:matrix.orgi do think part of why the python packaging ecosystem feels comfortable dropping support for/in old versions of toolchains is because nobody pushes back strongly to signal those things are still in use and shipped on older platforms17:04
@clarkb:matrix.orgor use old tox17:04
@mordred:waterwanders.comhrm. fair enough17:04
@fungicide:matrix.orgthat said, i'm not against dropping support for older python when we know it's not working, less so dropping support simply because we've lost the ability to easily run tests17:05
@clarkb:matrix.orguntil everything is using pyproject.toml there is no safe way to drop support for platforms in pbr17:05
@mordred:waterwanders.comI mean - I'm happy to go tilt at that windmill for a bit longer. I mostly just folded that one in because it seemed folks had already decided that was the next step17:05
@clarkb:matrix.orgso the only safe way forward imo if we want to drop old platforms is to use a new package name and then have people using pyproject.toml opt into that17:05
@clarkb:matrix.orgmordred: yes people keep trying and I keep pushing back. Keeping 3.8 was the compromise17:05
@clarkb:matrix.orgthe old version of that chagne was 3.10 and up or something17:05
@clarkb:matrix.organd I had to take a very strong stance against that move17:05
@fungicide:matrix.orgi agree that keeping support for python 3.8 right now makes sense17:06
@mordred:waterwanders.comnod, thanks, that's good context. Let me ... noodle on 3.6 solutions.17:07
@clarkb:matrix.orgI think the thing that peopel fail to grasp is that as soon as you make changes like this now all this old software is completely broken17:07
@clarkb:matrix.organd for 90% of the world thats fine. For the other 10% you've just created an unnecessary headache for typically minimal gain in the library iself17:07
@fungicide:matrix.orgthere's some stuff that becomes possible packaging-wise for projects once they do drop support for python 3.8 (in particular, dependency groups), but that doesn't mean pbr necessarily has to stop supporting 3.817:07
@mordred:waterwanders.comoh - belive me ... I'm only doing any of these patches because of all the carping pip is doing via pyproject.toml screaming about random choices pypa are shoving down my throat17:08
@clarkb:matrix.orgwhere this is going to be extra fun is python3.14 is completely broken in resolute17:08
@clarkb:matrix.orgjust wait for packages to require 3.14 or newer17:08
@mordred:waterwanders.com"you have to either provide a version field or mark that it's dynamic" <-- come on, shut the hell up17:08
@clarkb:matrix.organd your choice is broken python or broken everything17:09
@fungicide:matrix.orgsilencing setuptools deprecation warnings is not pbr's job, and in particular for some things the setuptools maintainers have gone back and forth about un-deprecating some of it, e.g. the license-tracking mess17:09
@mordred:waterwanders.comby completely broken - do you mean in a way more than just the memory garbage? it's functionally working for me?17:09
@mordred:waterwanders.comfungi: it's not - but also, if the deprecation warnings are printing for me as a pbr user, it makes my pythoning experience pretty poor, especially if, as a pbr user, I can't silence them myself17:10
@mordred:waterwanders.commost I can17:10
@fungicide:matrix.orgthe version of python inu resolute uses a garbage collection algorithm that got reverted in a later point release17:10
@clarkb:matrix.orgmordred: yes the garbage collector falls over in some very common trivial cases17:10
@mordred:waterwanders.comoh - ah - you explicitly mean the ubuntu packaged python3.14 in resolute17:10
@mordred:waterwanders.comnevermind. I don't use that for anything17:11
@clarkb:matrix.orglike using dequeue you will oom17:11
@mordred:waterwanders.comI either use python via pyenv or via the python-base container image17:11
@mordred:waterwanders.combut good to know there's a good reason for my life choices :)17:11
@clarkb:matrix.orgya I think we may need to rebuild our base container images so that we get updated 3.14 with the old GC17:12
@fungicide:matrix.orgthe way i view pbr's position, oneof its main fatures is in providing backward compatibility to distutils(2) features that setuptools decided to drop (or never implemented)17:12
@clarkb:matrix.orgswitching things to 3.14 is one of my idaes for new contributor effort so I'll make a note about that prestep17:12
@mordred:waterwanders.comwe base those on the more upstream python ones right? so the python in there is from python not from distro iirc (still might need a bump for the  newer point release)17:12
@fungicide:matrix.orgso some of those things that setuptools is emitting deprecation warnings about may be an opportunity for pbr to shim the old configs into something newer setuptools will accept, i guess17:13
@clarkb:matrix.orgmordred: yes the container images are upstream python builds within the image17:14
@clarkb:matrix.orgthey dno't use the distro packages that the python is running within17:14
@mordred:waterwanders.comyeah. I love those images17:14
-@gerrit:opendev.org- Monty Taylor https://matrix.to/#/@mordred:inaugust.com proposed: [zuul/zuul-jobs] 994946: Add a flag to ensure-tox to allow upgrading pip https://review.opendev.org/c/zuul/zuul-jobs/+/99494617:57
-@gerrit:opendev.org- Monty Taylor https://matrix.to/#/@mordred:inaugust.com proposed: [zuul/zuul-jobs] 994946: Add a flag to ensure-tox to allow upgrading pip https://review.opendev.org/c/zuul/zuul-jobs/+/99494618:44
-@gerrit:opendev.org- Scott Little proposed:19:02
- [openstack/project-config] 994821: add repo build-config for StarlingX https://review.opendev.org/c/openstack/project-config/+/994821
- [openstack/project-config] 994823: Remove branch deletion powers from three new repo https://review.opendev.org/c/openstack/project-config/+/994823
@mordred:waterwanders.comfungi: fwiw - the zuul-jobs patch above is a followup to our discussion above. https://review.opendev.org/c/openstack/pbr/+/994828 now works (it's still doing tempest tests, but the py36 job that was failing earlier is green) - so once that's fully reported green I think we can reparent the various fixes on top of that and shift the drop-support discussion to the side to be able to have fully separately and not tie it in to "fix the gate"19:05
@fungicide:matrix.org994946? i'll check that one out too, thanks!19:13
-@gerrit:opendev.org- Zuul merged on behalf of Scott Little: [openstack/project-config] 994821: add repo build-config for StarlingX https://review.opendev.org/c/openstack/project-config/+/99482119:17
-@gerrit:opendev.org- Zuul merged on behalf of Scott Little: [openstack/project-config] 994823: Remove branch deletion powers from three new repo https://review.opendev.org/c/openstack/project-config/+/99482319:22
@scott.little:matrix.orghow long for group starlingx-build-config-core tp become visible ?19:35
@clarkb:matrix.orgit has been variable recently with the crawlers making gitea updates slow at times19:37
@clarkb:matrix.orgprobably aroudn an hour or two if things are slow right now. As little as 10 minutes if not19:37
@clarkb:matrix.orgyou can follow it on the zuul status page https://zuul.opendev.org/t/openstack/status?change=994821%2C319:38
@clarkb:matrix.orggitea13 just did x/daisycloud which I think is at the tail end of the list so hopefully not too much longer19:41
@fungicide:matrix.orgscott.little: yeah, once zuul reports back on the change that the deploy jobs have completed, then i can add you19:46
@fungicide:matrix.orgpossibly sooner19:48
-@gerrit:opendev.org- Zuul merged on behalf of Monty Taylor https://matrix.to/#/@mordred:inaugust.com: [zuul/zuul-jobs] 994946: Add a flag to ensure-tox to allow upgrading pip https://review.opendev.org/c/zuul/zuul-jobs/+/99494620:29
@clarkb:matrix.orgfungi: did the group membership get modified?20:46
@clarkb:matrix.org(I can do it if it helps which is hwy i ask)20:46
@fungicide:matrix.orgnot yet that i can see20:51
@fungicide:matrix.orgoh, infra-prod-manage-projects timed out20:52
@fungicide:matrix.orgi was expecting https://review.opendev.org/admin/groups/starlingx-build-config-core to eventually not 404 and then i had the ssh api call to add the initial group member already queued up20:54
@clarkb:matrix.orgfungi: yes that is expected but I think it ends up runnnig manage-projects20:55
@clarkb:matrix.orgbut I could be wrong about this. This is what I was trying to point out yesterday20:55
@clarkb:matrix.orgsince you had approved another acl update yesterday after I noticed that things were slow and may not be applying as expected20:55
@fungicide:matrix.orgyesterday i think we were saved by two back-to-back jobs and the second one did succeed20:55
@fungicide:matrix.orgtoday there were two back-to-back as well and the second one hasn't reported from deploy yet20:55
@clarkb:matrix.orgI do half wonder if maybe gitea13 is slow for some infrastructure reason20:57
@clarkb:matrix.orgsince all of the others seem to be quick. But historically slowness like that has been luck of the draw with haproxy load balancing20:57
@mordred:waterwanders.comfungi, Clark when next you're bored, the pbr "fix tests" patch (994828) went green so I rebased the rest of the stack on top of it.  so I think that stack is ready to review. by the end of the stack the pyproject warnings should be gone, py3.14 and resolute are added, and we didn't have to drop support for anything. Also, thanks for landing that zuul-jobs patch21:09
@mordred:waterwanders.comblast. I spoke too soon. I've uncovered a bug in patch #2 in that stack. The fix tests patch is fine.21:13
@fungicide:matrix.orgback on the manage-projects failure, the second change did pass and the new project group was created so i set the initial member on it21:41
@jim:acmegating.comdoes anyone else want to review this change to the 16gb labels (discussed in the meeting)? https://review.opendev.org/99481822:04
@mordred:waterwanders.comlgtm22:07
@fungicide:matrix.orglgtm2, approved22:19
-@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [opendev/zuul-providers] 994818: Use v3-starter-8 for vexxhost 16GB and remove rax 16GB https://review.opendev.org/c/opendev/zuul-providers/+/99481822:20
@mordred:waterwanders.comcorvus: mostly mentioning in case there's a real problem - but I pushed up a new version of 989910 (patchset 9) at 15:25 - so almost 20 minutes ago - and zuul status is still showing jobs in progress for patchset 822:44
@mordred:waterwanders.comobviously I can, you know, just wait, that's not important - but 20 minutes feels weird22:45
@fungicide:matrix.orgmordred: was the plan to rebase the rest of the longer pbr stack starting with 994826 onto the shorter one since you squashed 994828 into 988753 which is currently a separate series?22:57
@fungicide:matrix.orgor did you just miss clearing the pyproject.toml diff out of an update to 994828?22:59
@fungicide:matrix.organyway, most of both stacks lgtm23:16
@fungicide:matrix.orgmaybe we can get them finalized and merged tomorrow, then release when folks feel comfortable (we could try a prerelease again, though testing that way last time we did it was complicated if memory serves)23:18
@fungicide:matrix.orgin the meantime, i'm going to evening. later, all23:19
@clarkb:matrix.orgwe had to push thinsg pretty hard to actually use a prerelease version right?23:19
@clarkb:matrix.orggnite23:19
@clarkb:matrix.orgI think pyproject.toml may make the forcing of a prerelease version easier at least23:19
@clarkb:matrix.orgbut htat only check the software using new python not old. For old we probably have to preinstall pbr so it doesn't try to install it itself23:20
@mordred:waterwanders.comfungi: ah - no the pyproject.toml stuff in 994828 is still needed to pass one of the validation tests (the harness has updated and is stricter now) I left 998753 to the side since I think "fix current testing" and "drop support for old versions" can (and should) be separate concerns. At least now that we have a path to the repo working _without_ 99875323:33
@mordred:waterwanders.comI've marked 988753 as WIP for now23:34

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!