| *** mhen_ is now known as mhen | 02:24 | |
| opendevreview | Ghanshyam proposed openstack/tempest master: Remove greade jobs from integrated-gate-networking https://review.opendev.org/c/openstack/tempest/+/976203 | 02:55 |
|---|---|---|
| opendevreview | Ghanshyam proposed openstack/tempest master: Remove grenade jobs from integrated-gate-networking https://review.opendev.org/c/openstack/tempest/+/976203 | 02:57 |
| *** mhen_ is now known as mhen | 04:14 | |
| *** ykarel_ is now known as ykarel | 05:40 | |
| opendevreview | Jakub Libosvar proposed openstack/devstack master: [neutron] Configure ovn-bgp service-plugin https://review.opendev.org/c/openstack/devstack/+/963199 | 09:07 |
| opendevreview | Takashi Kajinami proposed openstack/devstack master: Drop unused PORT_INT variables https://review.opendev.org/c/openstack/devstack/+/976245 | 13:03 |
| sean-k-mooney | elodilles: fyi we are going to need to backport the removal of pkg_resouces form the requireemtn repo to unblock stable branches and that equries squashing 3 comits https://review.opendev.org/c/openstack/requirements/+/976259 | 14:17 |
| sean-k-mooney | if that passes ci ill create the 2025.1 and 2024.2 backports of the same | 14:18 |
| elodilles | sean-k-mooney: ACK, i'm trying another approach, but it would be good to progress with this somehow as we are stepping on mines everywhere ':D | 14:19 |
| sean-k-mooney | ya i also spoke to stephen about possibly capping setuptools in pbr | 14:20 |
| sean-k-mooney | and he is explorign reverting the pkg_resouce removal | 14:20 |
| sean-k-mooney | so we have like 5 diffent approchs | 14:20 |
| sean-k-mooney | we can see which one sticks | 14:21 |
| elodilles | if setuptools reverts pkg_resource removal then we might end up with the same issue in the future o:) (wasn't pkg_resources dropped a couple of months ago and then added back due to the same reason? o:)) | 14:22 |
| elodilles | but yes, it causes some pain now :S | 14:23 |
| elodilles | and probably it's not just us who encounters these issues o:) | 14:23 |
| sean-k-mooney | elodilles: no | 14:24 |
| sean-k-mooney | stephen is makign the agrument that like arg_parse | 14:24 |
| sean-k-mooney | too much depend on it to remove it ever or at least not without a pep and tighing it to a specic python version | 14:25 |
| sean-k-mooney | they undeprecated arg parse becuase of usage | 14:25 |
| sean-k-mooney | so there si precendet to reverse removal desions but at teh end of the day its not part fo the standard lib so the setuptools comunity will have the final say | 14:26 |
| sean-k-mooney | elodilles: we have mostly move away form it outside of xstatic and some of the sable branches not having backports | 14:27 |
| elodilles | ah, ok. fingers crossed then :] | 14:27 |
| sean-k-mooney | so regradelss of what happen we should complete that | 14:27 |
| elodilles | +1 | 14:27 |
| sean-k-mooney | so tempest fails on https://review.opendev.org/c/openstack/requirements/+/976259 because or pkg_resouces ... but i think that because setuptools is not capped here https://github.com/openstack/requirements/blob/master/pyproject.toml#L2 | 14:32 |
| sean-k-mooney | which is doen via setup.py on 2025.2 i think | 14:34 |
| elodilles | there is a line in log that suggests that setuptools is upgraded during devstack installation: | 14:42 |
| elodilles | inc/python:setup_devstack_virtualenv:44 : pip_install -U pip 'setuptools[core]' | 14:42 |
| elodilles | this pulls setuptools 82.0.0 | 14:42 |
| sean-k-mooney | there are multiple venvs at play | 14:46 |
| sean-k-mooney | pip uses build isolation by default | 14:46 |
| sean-k-mooney | so the venv used to build the wheils that installed into the devstack venv | 14:46 |
| opendevreview | Pavlo Shchelokovskyy proposed openstack/tempest master: Bump flake8-import-order to 0.19.x https://review.opendev.org/c/openstack/tempest/+/976265 | 14:46 |
| sean-k-mooney | use a diffent venv with a diffent setuptools | 14:46 |
| sean-k-mooney | stephen has 2 devstack patches up currently one to pin and a second to disbale build ioslation | 14:47 |
| sean-k-mooney | with the intent to squash them if it works | 14:47 |
| sean-k-mooney | but that only a tempory workaround obviously | 14:47 |
| sean-k-mooney | https://review.opendev.org/c/openstack/requirements/+/976259/2/setup.py | 14:48 |
| sean-k-mooney | will also work but we would have to do that in every setup.py which si not realistic but might be ok the requireemtn repo | 14:48 |
| elodilles | btw, adding 'setuptools==81.0.0' to 'deps' of tox-docs "fixes" the docs job on stable/2025, if that is enough to make the devstack patch land | 14:50 |
| sean-k-mooney | well stephen does not have a master version as far ias i can see but it think because fo greneade we might also need to merge these backward starting with 2024.2 | 14:52 |
| sean-k-mooney | https://review.opendev.org/c/openstack/devstack/+/976144?tab=change-view-tab-header-zuul-results-summary | 14:53 |
| sean-k-mooney | however that is still failing on grende so we might need to make it non voting as well to land the first part | 14:53 |
| sean-k-mooney | hum somehow https://review.opendev.org/c/openstack/devstack/+/976107 actully merged | 14:54 |
| sean-k-mooney | so i guess https://review.opendev.org/c/openstack/devstack/+/976148?tab=change-view-tab-header-zuul-results-summary is the next in line | 14:55 |
| sean-k-mooney | elodilles: so that is blocked on the tox-docs job for 2025.1 | 14:55 |
| sean-k-mooney | i think adding the pin in tox directly coudl be accpable for now | 14:56 |
| sean-k-mooney | but i dont knwo if that will actully work withoug modifying the install line | 14:56 |
| sean-k-mooney | have you teted that locally? | 14:57 |
| elodilles | it is enough to add to the deps, i think i've tested it locally and worked | 14:57 |
| elodilles | i hope i remember correctly o:) (as i was trying many things) | 14:57 |
| sean-k-mooney | stephenfin: ^ can you verify that when you have time and update/squash that patce with its parent | 14:58 |
| * stephenfin reads back | 14:59 | |
| elodilles | i think https://review.opendev.org/c/openstack/devstack/+/976106 can be merged without the patch on top of it, only by adding the 'setuptools==81.0.0' to docs job's deps | 15:00 |
| stephenfin | elodilles: sean-k-mooney: adding it to deps will do the trick, yes. However, that needs to be done for all broken projects, and I suspect there are quite a lot of them. | 15:02 |
| elodilles | yeah i see a couple | 15:02 |
| stephenfin | let's do it for devstack for now. it won't cost us much, and we can always back it out if/when the pkg_resources issues is addressed | 15:03 |
| elodilles | but that might loosen a bit the 'catch 22' if the grenade jobs start working o:) | 15:03 |
| stephenfin | *is addresssed upstream | 15:03 |
| elodilles | +1 | 15:03 |
| opendevreview | Stephen Finucane proposed openstack/devstack stable/2025.2: [Stable-Only] Pin setuptools https://review.opendev.org/c/openstack/devstack/+/976105 | 15:07 |
| opendevreview | Stephen Finucane proposed openstack/devstack stable/2025.1: [Stable-Only] Pin setuptools https://review.opendev.org/c/openstack/devstack/+/976106 | 15:07 |
| elodilles | thanks o/ | 15:09 |
| sean-k-mooney | stephenfin: your off by one by the way | 15:11 |
| sean-k-mooney | its setuptools 82 that broke things not 81 right? | 15:11 |
| stephenfin | yeah, but I've already pinned < 81.0 on 2024.2 on | 15:12 |
| stephenfin | I realised that after the face | 15:12 |
| stephenfin | *fact | 15:12 |
| stephenfin | I can change | 15:12 |
| sean-k-mooney | ya that fine off by one in that direction wont break anything | 15:12 |
| stephenfin | actually, let me fix that. the commit message is wrong. We'll just a bit of confusion with the stable/2024.2 patch | 15:13 |
| sean-k-mooney | ack | 15:13 |
| opendevreview | Stephen Finucane proposed openstack/devstack stable/2025.1: [Stable-Only] Pin setuptools https://review.opendev.org/c/openstack/devstack/+/976106 | 15:14 |
| opendevreview | Stephen Finucane proposed openstack/devstack stable/2025.2: [Stable-Only] Pin setuptools https://review.opendev.org/c/openstack/devstack/+/976105 | 15:15 |
| stephenfin | done | 15:15 |
| sean-k-mooney | stephenfin: im hoping https://review.opendev.org/c/openstack/requirements/+/976259 will fix the stable branches for the requiremet repo by the way but we will see it need the devstack fix so im waiting for ti to comple now that you have updated them | 15:17 |
| sean-k-mooney | its a squash fo the 3 commits you had for fixing this on master | 15:19 |
| elodilles | sean-k-mooney: note that this also works for the requirements jobs, but it doesn't help with the cross jobs :/ --> https://review.opendev.org/c/openstack/requirements/+/976248/3/tox.ini | 15:20 |
| stephenfin | sean-k-mooney: ack. I was concerned that there might be more. I did a lot of rework of that tool recently | 15:20 |
| sean-k-mooney | elodilles: ack ya that coudl be an option too it would allow backporting the removals without squashign | 15:23 |
| elodilles | so i'd rather just pin setuptools for stable/* branches of requirements and fix the cross jobs in their repositories, that we need to do anyway :/ (so yes, it's a significant amount of fixes on stable branches /o\) | 15:23 |
| sean-k-mooney | elodilles: we are hoping we wont have to fix this in every repo | 15:24 |
| sean-k-mooney | but hope and reality may not align | 15:25 |
| elodilles | fingers crossed o:) | 15:37 |
| *** dviroel is now known as dviroel_lunch | 16:11 | |
| *** dviroel_lunch is now known as dviroel | 16:58 | |
| opendevreview | Brian Rosmaita proposed openstack/devstack-plugin-ceph master: Use debian-trixie https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/976279 | 17:01 |
| sean-k-mooney | elodilles: stephenfin: so https://review.opendev.org/c/openstack/requirements/+/976259 worked but https://review.opendev.org/c/openstack/devstack/+/976105 obviously depends on https://review.opendev.org/c/openstack/devstack/+/976106 for grenaded | 17:34 |
| sean-k-mooney | so we need to wait for the devstackpatchs to merge first | 17:34 |
| stephenfin | Sweet. We just need to merge backwards os | 17:34 |
| stephenfin | *so | 17:34 |
| sean-k-mooney | yep you can actully depend on the backport and it will actully pass | 17:35 |
| sean-k-mooney | but almost no one does that | 17:35 |
| sean-k-mooney | so ya we just need to wait | 17:36 |
| opendevreview | Merged openstack/devstack stable/2025.1: [Stable-Only] Pin setuptools https://review.opendev.org/c/openstack/devstack/+/976106 | 19:10 |
| opendevreview | Brian Rosmaita proposed openstack/devstack-plugin-ceph master: Use debian-trixie https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/976279 | 22:18 |
| opendevreview | Merged openstack/devstack stable/2025.2: [Stable-Only] Pin setuptools https://review.opendev.org/c/openstack/devstack/+/976105 | 23:09 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!