| jakeyip | hi, magnum-ui has a stuck reno change, I spent some time on it but can't figure out. hopefully someone can help? I've jotted down some comments on https://review.opendev.org/c/openstack/magnum-ui/+/909519 | 10:25 |
|---|---|---|
| fungi | jakeyip: sphinx is complaining because the generated rst ends up with two entries for bug/1638863 | 10:56 |
| fungi | i'm not real familiar with reno, but a quick git grep on master for bug/1638863 turns up two different release notes using that same name to refer to different urls | 10:57 |
| fungi | https://bugs.launchpad.net/magnum/+bug/1638863 vs https://bugs.launchpad.net/magnum-ui/+bug/1638863 | 10:57 |
| fungi | maybe that's the reason? | 10:57 |
| frickler | fungi: jakeyip: I think that is yet another bug in reno, it lists all older tags for unmaintained/yoga, so the head-of-branch detection is somehow broken there | 11:20 |
| frickler | I thought we had fixed this when unmaintained/yoga was first made, but must be some special case here | 11:20 |
| frickler | same happens for unmaintained/zed on 917397 | 11:21 |
| frickler | elodilles: tonyb: ^^ I'd really appreciate if the unmaintainer team could look into this, otherwise likely the next best solution might be to simply eol those branches? | 11:23 |
| frickler | not sure if the not-cleaned-up stable/newton branch might also play a role in this game | 11:24 |
| frickler | jakeyip: actually there is a newton-eol tag, so let me just delete the stable/newton branch in gerrit and see if that helps | 11:37 |
| frickler | hmm, didn't help, but good to have that cleaned up anyway, I guess | 11:50 |
| opendevreview | Merged openstack/project-config master: Clean up diskimages from inmotion cloud https://review.opendev.org/c/openstack/project-config/+/921073 | 15:25 |
| haleyb | fungi: so i have a follow-on question regarding regex - is there an inverse to the irrelevant-files section? i.e. if this file/dir is touched i always want to trigger this job, even if it matches some pattern in irrelevant-files. | 15:57 |
| clarkb | haleyb: no because irrelevant files are evaluated after files so they win | 15:58 |
| haleyb | clarkb: ack, just trying to find a way to write these things on one line, or some other option | 15:59 |
| clarkb | There are several appraoches that can be taken. One is to rethink the use of irrelevant files entirely (use positive matchins in files exclusively). Another is to accept that running some jobs more often is ok (I think ironic took this appraoch with bandit since it is super quick and low cost). One more drastic appraoch would be to structure the code so that it can be matched | 16:00 |
| clarkb | using simpler rules (though I don't think I would advocate for this). And finally there is probably improvements that could be made to zuul to make the expressions around this stuff richer. Though I'm not sure anyone in the zuul community is treating that as a priority right now (so not a quick fix) | 16:00 |
| haleyb | clarkb: ack, thanks, at least something to discuss in our CI meeting as irrelevant-files is huge in some cases | 16:03 |
| clarkb | haleyb: looking at https://review.opendev.org/c/openstack/neutron/+/921309/5/zuul.d/base.yaml and naively comparing to the neutron tree you might be able to have files: ^.*.py$ and zuul.d/project.yaml then irrelevant files simply filters out the .py containing dirs that you don't want to trigger on (stuff like unittests in the functional job) | 16:08 |
| clarkb | then you've got like 5 rules total rather than 30 or whatever the current set is | 16:09 |
| clarkb | reading between the lines there it seems like you're trying to say this job should only run if code changes and not release notes/docs/ansible/etc/etc/etc. So express that directly and then trim out anything that overmatches with irrelevant files | 16:10 |
| haleyb | clarkb: yes, someone just mentioned doing something similar in neutron channel. i think we have options to fix this up better | 16:10 |
| haleyb | and hope we don't break things... | 16:10 |
| opendevreview | Merged openstack/project-config master: Remove the inmotion cloud entirely https://review.opendev.org/c/openstack/project-config/+/921074 | 17:01 |
| tonyb | frickler: I can look at it tomorrow. | 18:36 |
| *** dasm is now known as Guest8716 | 20:46 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!