Wednesday, 2024-09-18

cardoeJayF: you make me laugh. sushy’s style00:22
rpittaugood morning ironic! o/08:00
opendevreviewMerged openstack/tenks master: Bump minimum Python and Ansible versions  https://review.opendev.org/c/openstack/tenks/+/92162009:07
cidHey, welcome back, rpittau , \o/09:13
rpittauthanks cid :)09:57
opendevreviewWill Szumski proposed openstack/bifrost master: Fix checksum parsing for Rocky Linux  https://review.opendev.org/c/openstack/bifrost/+/92975310:09
opendevreviewMerged openstack/ironic-inspector master: Imported Translations from Zanata  https://review.opendev.org/c/openstack/ironic-inspector/+/92770110:28
opendevreviewVerification of a change to openstack/ironic stable/2023.2 failed: [CI][stable only] fix zuul config  https://review.opendev.org/c/openstack/ironic/+/92663710:59
rpittauI'm still dealing with tons of messages/mails, but I realized that we're close to final release week \o/11:22
rpittaulooking at the changes left I think  we can wait for the final release for ironic-inspector too, even if the latest release is more than 2 months old11:22
rpittauactually looking a the  list again there are a couple of changes that we may  want to add now, I'll propose a release11:26
dtantsurfolks, https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/928284 reduces the CI load on ironic-tempest-plugin quite a bit, could you check it?11:32
opendevreviewDmitry Tantsur proposed openstack/ironic-python-agent master: Trivial: fix variable in formatting  https://review.opendev.org/c/openstack/ironic-python-agent/+/92976611:35
opendevreviewDmitry Tantsur proposed openstack/bifrost master: Add support for Ubuntu 24.04  https://review.opendev.org/c/openstack/bifrost/+/92889511:40
dtantsurI think we should land ^^^ before the final release11:42
dtantsurwe're a bit overdue with 24.04 support since it's 24.09 now :)11:45
rpittaudtantsur: I think ubuntu noble support is planned for epoxy12:08
dtantsurOpenStack-wide probably, but Bifrost is on a faster cadence than OpenStack12:10
dtantsursomebody can try to install it on 24.04 already, while the current version of OpenStack won't hit any productions until next year (at best) :)12:11
rpittauright, well LGTM :)12:15
dtantsurmmm, the vmedia job fails? interesting12:26
dtantsurerror: can't find command `linuxefi'12:26
*** mnasiadka1 is now known as mnasiadka12:28
dtantsuris it time we stop using linuxefi at all?12:32
opendevreviewDmitry Tantsur proposed openstack/bifrost master: Fix missed occurrence of SETUPTOOLS_USE_DISTUTILS  https://review.opendev.org/c/openstack/bifrost/+/92976912:34
opendevreviewDmitry Tantsur proposed openstack/bifrost master: Support OVMF with 4M flash storage  https://review.opendev.org/c/openstack/bifrost/+/92977012:36
dtantsurrpittau: I've split out the relevant fixes ^^ since I don't think I can spend time on linuxefi now..12:36
rpittauack12:40
opendevreviewMerged openstack/ironic-lib stable/2023.1: Fix invalid UTF-8 characters in execute output  https://review.opendev.org/c/openstack/ironic-lib/+/92671912:49
opendevreviewMerged openstack/ironic-inspector master: Don't fail if data already migrated  https://review.opendev.org/c/openstack/ironic-inspector/+/92417212:55
opendevreviewMerged openstack/ironic stable/2023.2: [CI][stable only] fix zuul config  https://review.opendev.org/c/openstack/ironic/+/92663712:55
opendevreviewMerged openstack/sushy master: add pyproject.toml to support pip 23.1  https://review.opendev.org/c/openstack/sushy/+/92969313:09
TheJuliaGood morning even though I've already been up for like an hour since slack started exploding at 5:04 AM13:10
dtantsurouch :(13:11
dtantsurslack--13:11
TheJuliaGood conversations13:11
JayFdtantsur: in Gentoo that moved into a gnu-efi package recently13:11
TheJuliabut... yeah13:11
TheJuliaJayF: your up early too13:11
JayFI'm literally on my phone in a dark bedroom checking things13:11
TheJuliago back to bed!13:11
dtantsuragreed ^^^13:12
TheJuliawell, more like put the phone down13:12
JayFI tried plenty lol13:12
TheJuliacuddle a cat, dog, wifey13:12
dtantsurfew things have helped my mental health as much as staring less at my phone13:12
TheJuliastuffed animal13:12
TheJuliachtulhu13:12
TheJuliaanything which is not a phone :)13:12
rpittaustuffed cthulhu :)13:12
JayFdtantsur: I apply that in an evening direction13:12
TheJuliastuffed cthulhu++13:12
JayFBasically I wake up and start work so I can be done asap13:12
* TheJulia lacks stuffed cthulhus at the moment13:13
TheJuliaI'm sort of the same way13:13
TheJuliabut I do it for overlap reasons13:13
TheJuliaThen again I got scheduled for a 6 am call tomorrow13:14
TheJuliaFall is definitely here in SoCal wine country13:17
TheJuliaone step away from hot apple cider and hot chocolate13:18
dtantsuryummy!13:18
JayFfall here is more about grey and wet lol13:20
TheJuliaWhere I am right now, it is very much about clouds rolling in off the ocean and their tops being like 150 feet up in the air13:26
TheJuliaand some fog13:26
dtantsurromantic!13:27
TheJuliaat my actual home, it is wind until the jetstream shifts13:27
TheJuliawell, wind and dust13:27
opendevreviewMerged openstack/python-ironicclient master: Fix detailed node history list  https://review.opendev.org/c/openstack/python-ironicclient/+/92849513:27
cardoerpittau: does my answer make sense on https://review.opendev.org/c/openstack/sushy/+/92970414:12
opendevreviewJulia Kreger proposed openstack/ironic-inspector master: Fix: ignore resolve error while looking up node  https://review.opendev.org/c/openstack/ironic-inspector/+/92532614:25
rpittaucardoe: my point on  pycodestyle was just about dropping it from the requirements, not dropping it entirely, do we trust the version installed by flake8 by default? we  used to constrain it for compatibility reasons14:28
cardoeflake8 has tighter constraints than sushy did14:29
rpittaualright14:37
rpittauI'm still a bit concerned about the python versions difference between local and CI, but we can give that a try14:37
cardoeI don't disagree with you. That's why it's never worked for me locally.14:37
cardoetox -e pep8 has always failed for me locally on sushy.14:37
cardoeThere's a PTG item to discuss code style and pre-commit and linting. I was just putting together some changes to highlight how the tooling code work.14:39
cardoeWe don't have to merge that change until after PTG.14:39
opendevreviewVerification of a change to openstack/sushy master failed: switch to pre-commit for style checks  https://review.opendev.org/c/openstack/sushy/+/92970414:40
rpittauyeah, I was considering that, put -1 Wfor now14:40
cardoeI've got a patch that actually tightens the code style and pins things to target a specific Python version.14:40
cardoeIt matches what other OpenStack projects have in their pipelines. clarkb shared it recently I think.14:40
cardoeBut of sushy it results in every single file being touched.14:41
opendevreviewMerged openstack/ironic-inspector master: Drop SQLALCHEMY_WARN_20  https://review.opendev.org/c/openstack/ironic-inspector/+/92953314:42
cardoeI'm also not skipping stuff like the sphinx config cause those projects don't. Plus the issues with the sphinx config are from humans hand editing the config and not how sphinx-quickstart generates it.14:42
TheJuliaAnything needing my attention after I walk Mr. Corgi before I jump into focus time for the day?14:46
clarkbcardoe: it was stephenfin and sean mooney discussing how other projects use precommit and linters. I was just calling out some of the pre commit behaviors that I find less than ideal to make sure we are ok with those concessions14:47
rpittauTheJulia: if you could check https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/928958 would be great :)14:47
cardoeclarkb: got a list of those?14:48
JayFcardoe: rpittau: if we want more consistent behavior in CI runs on things like the pep8 or docs target, I think we'll have to set a basepython. I often have to patch my tox.ini to avoid it using an unsupported newer version of python.14:49
cardoeI do think in the end if OpenStack as a whole wants to adopt linters and such that exist upstream and OpenStack specific style isn't in them. Either gonna work to add that style upstream or just adopt the available style.14:49
stephenfincardoe: https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/Q5M7RGCSSFHFIK3RA7MTSY4K6KKOECNE/#GGF3YWEKYMC7GVTGIRCIFM53YUU4TSAK14:49
rpittauJayF: just for the pep8 job sounds good actually14:49
cardoeJayF: 100% agreed. We need a base that's constant.14:50
cardoeJayF: but if we use ruff, we don't need that.14:50
clarkbthe need for a git repo meta package to install things frp pypi is just silly imo and introduces extra potential failure points14:50
clarkbbut as long as I'm not asked to debug "github failed" (which it does fairly often) I don't care too much14:50
rpittauJayF: if you have a moment can you please havea look at https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/928284 ?14:50
stephenfincardoe: iiuc clarkb's comments were mainly around how pre-commit sources/installs things (from Git rather than PyPI) and the implications of same, as opposed to the linters themselves14:51
stephenfin(and yes, clarkb I do need to repond to your reply. Just been a busy week so far 0:) )14:51
TheJuliarpittau: done14:52
rpittauTheJulia: thanks :)14:52
clarkbstephenfin: yup exactly, when it comes to the linters themselves I think the bset thing is the thing that introduces the least friction for reviewers during review and when writing things whatever that may be for the group of people14:52
stephenfin++14:52
stephenfinJayF: Do you find basepython affects linters? Personally I've always seen the issue with docs, which mainly stems from us holding back various deps in u-c14:53
cardoeclarkb: ++ agreed. asottile won't change anything btw. It's his way of the highway.14:53
cardoestephenfin: when using pycodestyle it 100% does.14:53
stephenfin(Asking out of curiosity. Running Fedora 40 with py3.12 means I nearly always have to set basepython to build docs, but never anything else)14:54
cardoeThe biggest change is when Python introduced its new AST parser.14:55
stephenfincardoe: Can honestly say I've never run into that. Very interesting. Any idea why?14:55
stephenfinI know mypy can be version dependent also but you can (and should) configure the version it's testing again to side-step that14:56
cardoeYou can see it on sushy today. Run tox -e pep8 with a Python 3.8 vs 3.1214:56
stephenfine.g. https://opendev.org/openstack/openstacksdk/commit/7db657a8c4e1f2f291c6aa9fb185e33ce483508514:56
* stephenfin looks15:01
stephenfincardoe: Ah, that's fair. You're running into an area of the code I fixed a bug in recently, namely, Python 3.12 added new tokens for f-strings. Bumping to flake8 7.x (maybe earlier) will definitely fix things, but point taken :)15:04
stephenfinThe bug fwiw https://github.com/PyCQA/flake8/pull/1949 From experience on that, agreed on things being way or the highway 😅15:05
opendevreviewTakashi Natsume proposed openstack/ironic-inspector master: Fix a wrong assertion method  https://review.opendev.org/c/openstack/ironic-inspector/+/92981515:08
cardoeI'm just throwing some stuff out there and seeing what folks are happy with. I ultimately want this to be stupid easy for maintainers and even easier for contributors to avoid a friction to contribution and friction to maintainers reviewing/merging.15:09
cardoeI was a maintainer on the Xen Project for a while. They had their own code style that they stuck to. Things were debated for years about someone needs to contribute their format to clang-format so that their style could be in there. No one wanted to work on it. They got a Summer of Code student that did it. Got it mostly there into a release. By the time it finally got over the finish line the style had slightly changed.15:14
stephenfinMakes sense. fwiw, in SDK deliverables we have had been using the pre-commit model described in that mail and bumping the versions often (to the point that I'll likely get a bot to do it at some point)15:14
TheJuliastupid easy is double++ good15:22
TheJuliaanything which creates friction or pain == bad15:22
* TheJulia gets away with making such broad statements for $reasons ;)15:24
masgharDo we have a recommended python version for ironic master branch?15:36
rpittaumasghar: in theory what's supported based on setup.cfg, but probably better what's actually tested in CI, so Python 3.9 (from bifrost CS9 jobs) and Python 3.10 (from ubuntu jobs)15:40
masgharrpittau: thanks! I have python 3.9 locally, so it should be okay15:42
masgharI ran tox -e py3 on master branch and got 2 failing tests, and my first thought was python version15:42
rpittaummm there should be no difference in test results between py3.9 to py3.1115:45
rpittaugood night! o/15:48
masgharalright, o/15:49
TheJuliaI *highly* recommend 3.11 since it gives much obvious "you have a typo" errors15:57
clarkbwe've found it to be much quicker in some circumstances too15:58
TheJuliaYes, the GIL changes I beleive16:09
TheJuliaor that might have been 3.1216:09
clarkb3.13 makes the GIL optional but is still enabled by default. but ya in general since like 3.9 they have been making small performance improvements with each release. 3.11 seemed to have the biggest impact for zuul16:10
TheJuliaI thought the gil stuff already shipped in 3.11 or 3.12. Maybe I'm thinking the dynamic recompile stuffs16:13
opendevreviewMerged openstack/ironic-tempest-plugin master: Add tempest tests for runbooks  https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/92895816:57
opendevreviewMerged openstack/ironic master: Add microversion headers to root ('/') endpoint  https://review.opendev.org/c/openstack/ironic/+/92849116:59
opendevreviewMerged openstack/ironic master: Drop SQLALCHEMY_WARN_20  https://review.opendev.org/c/openstack/ironic/+/92939616:59
JayFwe need to not forget release mappings this release, we often do 17:51
opendevreviewMerged openstack/ironic-inspector master: Fix: ignore resolve error while looking up node  https://review.opendev.org/c/openstack/ironic-inspector/+/92532617:55
TheJuliaheh18:14
TheJuliayeah18:14
TheJuliahttps://github.com/openstack/ironic/blob/master/ironic/common/release_mappings.py#L685 https://github.com/openstack/releases/blob/master/deliverables/dalmatian/ironic.yaml#L9-L2218:16
TheJuliaseriously guys....18:16
TheJuliaseriously18:16
JayFI was worried about that18:35
* JayF had a chat about how this worked with c i d and adam today so it came to mind18:35
TheJuliaalmost have a patch18:35
TheJuliaoh, looks like because we did 25.0 and then 26.0, we broke a test18:37
opendevreviewJulia Kreger proposed openstack/ironic master: Update release mappings  https://review.opendev.org/c/openstack/ironic/+/92983918:45
TheJuliano sense to keep the test, I guess18:45
TheJuliaJayF: ^^18:48
TheJuliaadded rpittau as a reviewer18:52
JayFyeah I pointed adamcarthur5 at that config + ironic local dev, we're trying to find a way to emulate a really, really old ironic API microversion19:17
JayFbasically trying to ensure https://github.com/openstack-exporter/openstack-exporter/pull/378 doesn't break people who have older clouds19:17
JayF(and we'll be adding more metrics that make sense, and I'm sure those will get newer and newer as we go)19:18
JayFso I'm glad we thought about it and found this out before cutting an integrated release :D 19:18
opendevreviewMerged openstack/ironic-tempest-plugin master: CI: scale back the number of inspector and scope enforced jobs  https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/92828420:15

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!