| opendevreview | Jay Faulkner proposed openstack/project-config master: Ironic UM group access to bifrost UM branches https://review.opendev.org/c/openstack/project-config/+/910287 | 00:04 |
|---|---|---|
| clarkb | ianychoi: seongsoocho I don't think it would be difficult to add the governance repo to zanata. My concern is at the other end: removing projects from zanata and adding more will only make that worse | 00:11 |
| opendevreview | James E. Blair proposed zuul/zuul-jobs master: prepare-workspace-git: Add ability to define synced pojects https://review.opendev.org/c/zuul/zuul-jobs/+/887917 | 00:21 |
| opendevreview | Michal Nasiadka proposed openstack/project-config master: Add magnum-capi-helm repo to Magnum project https://review.opendev.org/c/openstack/project-config/+/910239 | 06:15 |
| mnasiadka | good morning | 06:26 |
| mnasiadka | fungi: following the discussion on openstack-discuss around Magnum and the CAPI driver discussion - we've decided to add the CAPI Helm driver (the second CAPI driver) as an out of tree driver in an opendev repo in https://review.opendev.org/q/topic:%22magnum-capi-helm-create%22 - would like a review of the project-config change if possible - thanks :-) | 06:29 |
| opendevreview | Michal Nasiadka proposed openstack/project-config master: Add magnum-capi-helm repo to Magnum project https://review.opendev.org/c/openstack/project-config/+/910239 | 06:30 |
| opendevreview | Michal Nasiadka proposed openstack/project-config master: Add magnum-capi-helm repo to Magnum project https://review.opendev.org/c/openstack/project-config/+/910239 | 06:41 |
| *** liuxie is now known as liushy | 07:49 | |
| fungi | mnasiadka: will do, i was following the change already, just waiting for the comments about importing and such to reach consensus | 13:16 |
| fungi | also it will need acknowledgement of the magnum ptl | 13:17 |
| Clark[m] | Gerrit 3.7.7 had an inappropriate/unexpected change to the jgit version. 3.8.4 does not appear to have this issue. They will make a 3.7.8 release shortly to fix this | 15:07 |
| Clark[m] | Still no 3.9.2 either | 15:08 |
| fungi | reminiscent of when they accidentally bumped the lucene version too far. was it another case of implicit change merges? | 15:17 |
| mnasiadka | fungi: thanks, will update the patch and get jakeyip’s approval | 15:22 |
| opendevreview | Merged ttygroup/gertty master: Drop prehistoric hack for urwid.GridFlow https://review.opendev.org/c/ttygroup/gertty/+/910173 | 15:31 |
| opendevreview | Merged ttygroup/gertty master: Fix urwid > 2.4.2 compatibility https://review.opendev.org/c/ttygroup/gertty/+/910265 | 15:32 |
| opendevreview | James E. Blair proposed zuul/zuul-jobs master: prepare-workspace-git: Add ability to define synced pojects https://review.opendev.org/c/zuul/zuul-jobs/+/887917 | 15:38 |
| opendevreview | Aleksei Stepanov proposed ttygroup/gertty master: Fix urwid > 2.4.2 compatibility https://review.opendev.org/c/ttygroup/gertty/+/910358 | 15:44 |
| opendevreview | Michal Nasiadka proposed openstack/project-config master: Add magnum-capi-helm repo to Magnum project https://review.opendev.org/c/openstack/project-config/+/910239 | 15:49 |
| mnasiadka | fungi: updated (https://review.opendev.org/c/openstack/project-config/+/910239) - we only want to get the 'main' branch imported to opendev (so we don't need to drop any irrelevant branches like we had to in kayobe case) - will that get renamed to 'master'? | 15:51 |
| fungi | mnasiadka: the repo will get cloned as-is and all branches/tags copied | 15:56 |
| mnasiadka | fungi: I understand, although I haven't probably seen any other projects with main branch under openstack/ namespace, so don't want to confuse anybody ;-) | 15:57 |
| fungi | so if there are branches there you don't want copied, or named something other than you want them to be once imported, it's better to correct that first on the source side. you can also do it by creating a new temporary fork and making your changes there, then importing that fork instead of the original | 15:57 |
| fungi | mnasiadka: ^ | 15:57 |
| mnasiadka | ok then, let me have a go on one of these options and come back when ready | 15:58 |
| fungi | there are a few deliverables in openstack which have a default branch named main instead of master (i think maybe sunbeam and some ofthe openstack-charms repos), but it could create some challenges for things like cross-repository testing without additional configuration to work around the branch name differences | 15:59 |
| mnasiadka | yes, I was also thinking about that | 16:00 |
| mnasiadka | And surely we want magnum/ and magnum-capi-helm/ to be able to do cross-repository testing | 16:00 |
| fungi | right, i think if magnum's current deliverables all have master as their default branch, i'd stick with that for these for consistency | 16:01 |
| mnasiadka | actually project-config-gerrit job just failed with "ERROR: No master branch exists" ;-) | 16:02 |
| fungi | yeah, it will do that if you don't include additional configuration to tell it what default branch you're using | 16:05 |
| fungi | as a safety measure | 16:06 |
| opendevreview | Michal Nasiadka proposed openstack/project-config master: Add magnum-capi-helm repo to Magnum project https://review.opendev.org/c/openstack/project-config/+/910239 | 16:06 |
| mnasiadka | ok, forked to my private github namespace (so we don't get any additional branches and tags) and renamed main to master | 16:06 |
| mnasiadka | let me chase the PTL now | 16:06 |
| mnasiadka | fungi: thanks for all your help! | 16:06 |
| fungi | yw! | 16:06 |
| fungi | in other news, pypi now supports metadata 2.3: https://discuss.python.org/t/46137 | 16:07 |
| clarkb | fungi: yes the jgit change snuck in via a merge. However it was a merge to update jgit but rolling updates forward through the stable branches but for whatever reason the jgit version on 3.7 got modified improperly when the others didn't | 16:08 |
| fungi | for those wondering what metadata 2.2 and 2.3 bring: https://peps.python.org/pep-0643/ and https://peps.python.org/pep-0685/ | 16:11 |
| fungi | basically, adding the "dynamic" field, and becoming more strict about naming for extras | 16:14 |
| *** dasm is now known as Guest1195 | 16:51 | |
| opendevreview | James E. Blair proposed zuul/zuul-jobs master: prepare-workspace-git: Add ability to define synced pojects https://review.opendev.org/c/zuul/zuul-jobs/+/887917 | 16:56 |
| fungi | frickler: we (intentionally) don't have cyclic deps enabled in the openstack tenant | 17:27 |
| fungi | and it was only temporarily enabled in zuul's tenant in order to have a partial real-world test of the feature once that version was installed | 17:28 |
| frickler | fungi: so this is only a per-tenant flag and not per repo? then I'd indeed retract my suggestion. I would still think covering the common case happening specifically between project-config and governance would be an interesting one to cover | 17:31 |
| *** Guest1195 is now known as dasm | 17:34 | |
| clarkb | fwiw there is no circular dep in creating new openstack projects | 17:36 |
| fungi | frickler: what's specifically in need of covering there? just making sure one only merges if the other merges and vice versa? i don't think it would even work for that case if we did turn it on because the repo still won't exist in the governance test run either way | 17:37 |
| clarkb | we just allow for governance to be acked positively and proceed from there without fully merging | 17:37 |
| clarkb | its still a strict ordering | 17:37 |
| opendevreview | Michal Nasiadka proposed openstack/project-config master: Add magnum-capi-helm repo to Magnum project https://review.opendev.org/c/openstack/project-config/+/910239 | 17:52 |
| opendevreview | James E. Blair proposed zuul/zuul-jobs master: prepare-workspace-git: Add ability to define synced pojects https://review.opendev.org/c/zuul/zuul-jobs/+/887917 | 18:07 |
| clarkb | fungi: do you know if there is a way to make mailman3 archives not truncate urls when rendering them for the web? | 18:38 |
| fungi | i... do not. it'll be a hyperkitty config option if there is any | 18:38 |
| clarkb | I find it annoying that I can't copy paste out of there easily anymore. But also I've long tried to train my family to treat rendered urls that don't match destinations with much skepticism | 18:40 |
| fungi | i right-click, copy link | 18:44 |
| clarkb | ya me too now | 18:44 |
| fungi | mainly because it's hard in a lot of browsers to click-drag highlight a url without it thinking you're clicking the link to follow it | 18:45 |
| opendevreview | James E. Blair proposed zuul/zuul-jobs master: prepare-workspace-git: Add ability to define synced pojects https://review.opendev.org/c/zuul/zuul-jobs/+/887917 | 19:04 |
| opendevreview | James E. Blair proposed zuul/zuul-jobs master: prepare-workspace-git: Add ability to define synced pojects https://review.opendev.org/c/zuul/zuul-jobs/+/887917 | 19:40 |
| *** elodilles is now known as elodilles_pto | 19:46 | |
| clarkb | https://gerrit.openafs.org/#/c/15668/1 this is the openafs on arm fix | 20:49 |
| clarkb | once that merges I figure we can ask debian for a backport? | 20:49 |
| clarkb | I assume "doesn't work at all" is a good reason to patch the software on debian stable | 20:49 |
| fungi | likely | 20:56 |
| fungi | or we can resurrect our ppa, worst case | 20:56 |
| clarkb | fungi: left some comments on the git-review change. I don't feel they deserve a -1 but wanted to make sure you saw them before I +2'd | 21:10 |
| fungi | saw, thanks! | 21:15 |
| fungi | any opinion on the (much simpler) parent change? | 21:16 |
| opendevreview | James E. Blair proposed zuul/zuul-jobs master: prepare-workspace-git: Add ability to define synced pojects https://review.opendev.org/c/zuul/zuul-jobs/+/887917 | 21:16 |
| clarkb | oh I missedthere were two /me looks | 21:17 |
| fungi | that one corrects a problem i ran into when implementing the embedded hook change | 21:17 |
| clarkb | is S_IMODE not documented in python's docs? help() returns a short blurb about it atleast | 21:30 |
| fungi | is it missing from the module documentation? | 21:30 |
| fungi | keep in mind that's a function in os.path.stat not os.stat | 21:31 |
| clarkb | ya I can't find it in either but maybe I'm confusing myself | 21:32 |
| clarkb | ya I don't see it in either | 21:32 |
| clarkb | fungi: ok posted a small thing that can probably be improved | 21:33 |
| clarkb | actually one sec. Just realized another thing | 21:33 |
| fungi | mmm, actually the module index doesn't mention os.path.stat at all | 21:34 |
| clarkb | fungi: I wonder if this is old code that had docs removed but not the implementation? | 21:34 |
| clarkb | fungi: if we are worried about that we can just &0o777 the st_mode instead | 21:37 |
| clarkb | I posted about the other thing now too | 21:37 |
| fungi | i think the concern with not pumping the current mask through os.path.stat.S_IMODE() is that you may end up with additional bits os.chmod() doesn't support and chokes on for some platforms | 21:38 |
| clarkb | side note it would be cool if python preserved int representation when set with a particular represetation like 0o for octal | 21:38 |
| clarkb | fungi: right. And the function is there on my pythons. I'm just concerend that python may have meant to delete it. I think we should have something like that method in use | 21:39 |
| fungi | you can wrap in oct() to get back a str representation in octal notation | 21:39 |
| clarkb | we can keep S_IMODE and then deal with it if it ever bcomes a problem | 21:39 |
| corvus | looks like os.path.stat is also just the stat module -- https://docs.python.org/3/library/stat.html | 21:40 |
| fungi | huh | 21:40 |
| clarkb | oh and it is documented there | 21:40 |
| clarkb | so the problem was in confusion over what the canonical name for the function is | 21:40 |
| fungi | indeed, looks like maybe it was os.path.stat in ancient 2.x and they left it in as an alias | 21:40 |
| fungi | i can shorten it | 21:40 |
| clarkb | ++ | 21:40 |
| fungi | thanks corvus! | 21:41 |
| corvus | np | 21:41 |
| clarkb | heh when you run help(stat.S_IMODE) or help(os.path.stat.S_IMODE) they both say `Help on built-in function S_IMODE in module _stat:` which is still confusing | 21:42 |
| fungi | the _ leader is an indication that cpython has a compiled c implementation of the module its using rather than a pure python fallback it may also ship | 21:53 |
| fungi | or at least that's what i've understood from discussions i've seen in the past | 21:53 |
| clarkb | today I learned | 21:54 |
| fungi | i'm not sure how consistently that convention is followed/enforced though | 21:55 |
| fungi | but yeah, there are parallel pure python and c implementations of a lot of the stdlib modules, and specific installations may use one or the other depending on if the platform could compile the c version successfully | 21:56 |
| fungi | aids in portability, but does also mean potential for behavior differences in some cases | 21:56 |
| opendevreview | James E. Blair proposed zuul/zuul-jobs master: prepare-workspace-git: Add ability to define synced pojects https://review.opendev.org/c/zuul/zuul-jobs/+/887917 | 22:14 |
| clarkb | I'm running out of daylight for monitoring gitea upgrades if approved now. But feel free to review and I can approve first thing tomorrow | 23:49 |
| tonyb | clarkb: Sounds good. | 23:51 |
| tonyb | clarkb: requirements chnages, git-review, gitea ... did I commit to reviewing anything else in the meeting that you recall? | 23:51 |
| clarkb | that was it :) thanks! | 23:51 |
| tonyb | \o/ | 23:52 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!