| *** ykarel|away is now known as ykarel | 04:44 | |
| opendevreview | Ian Wienand proposed openstack/project-config master: Remove debian-stretch disk images https://review.opendev.org/c/openstack/project-config/+/802654 | 04:49 |
|---|---|---|
| opendevreview | Ian Wienand proposed openstack/project-config master: Remove Debian stretch image builds https://review.opendev.org/c/openstack/project-config/+/802655 | 04:49 |
| *** rpittau|afk is now known as rpittau | 07:02 | |
| *** ykarel is now known as ykarel|lunch | 08:34 | |
| stephenfin | fungi: It does not, but that's the error you get with an invalid 'string.format' call, e.g. "'{y:5:2f}'.format(**{'y': 5})" | 08:57 |
| stephenfin | fungi: If it's reproducible, you can run sphinx in verbose mode to get the full traceback | 08:57 |
| *** zbr is now known as Guest2544 | 09:09 | |
| *** sshnaidm|afk is now known as sshnaidm | 09:45 | |
| *** ykarel|lunch is now known as ykarel | 10:17 | |
| *** jcapitao is now known as jcapitao_lunch | 10:42 | |
| fungi | stephenfin: oh, indeed thanks, once i swapped out my placeholders it passed fine | 11:49 |
| fungi | so that's clearly what it was | 11:49 |
| *** rlandy is now known as rlandy|ruck | 11:52 | |
| *** jcapitao_lunch is now known as jcapitao | 12:19 | |
| marios|ruck | o/ folks any idea why that one might be stuck it is holding our gate https://zuul.openstack.org/status#800848 | 13:26 |
| marios|ruck | am going to abandon and restore that, unless someone else can do anything about that | 13:26 |
| marios|ruck | problem is it has a lot of green runs below it that'll get discarded and have to rerun | 13:26 |
| marios|ruck | rlandy|ruck: ^^ | 13:26 |
| fungi | marios|ruck: don't abandon anything, probably there's a stuck node request which i can free and get that missing build run without rerunning the others | 13:29 |
| fungi | i'm checking now to see which launcher accepted that node request | 13:30 |
| marios|ruck | thank you for checking fungi ! | 13:30 |
| marios|ruck | fungi: holding the abandon for now | 13:30 |
| *** ykarel is now known as ykarel|away | 14:45 | |
| opendevreview | Radosław Piliszek proposed openstack/project-config master: Allow kolla cores to edit kolla hashtags https://review.opendev.org/c/openstack/project-config/+/802744 | 15:44 |
| zbr | pbr | 15:45 |
| zbr | stephenfin: do you happen to know if pbr got fixed on master around setuptools/dist.py UserWarnings? They are very annoying | 15:48 |
| stephenfin | which warnings are these? | 15:48 |
| zbr | the (in)famous: dist.py:697: UserWarning: Usage of dash-separated ... | 15:49 |
| zbr | usually getting rid of them requires something like https://review.opendev.org/c/openstack/pbr/+/791127/1/setup.cfg -- but that would obviously not work unless pbr knows to use the underline ones. | 15:50 |
| stephenfin | ah, it does | 15:50 |
| zbr | do we need a minimal version of pbr for that to work? | 15:50 |
| stephenfin | you can simply replace the dashes with underscores | 15:51 |
| gagehugo | Has anyone seen this error before in the docs job? Started happening recently: https://zuul.opendev.org/t/openstack/build/c3850af040d746f8969dad480e8dc923 | 15:51 |
| stephenfin | I'm not entirely sure. I'd have to check | 15:51 |
| gagehugo | I am assuming "rST localisation for language "id" not found." is the underlying cause | 15:51 |
| gagehugo | ?* | 15:52 |
| stephenfin | and I'm in a meeting for the next while. We did make changes to this some of the parser stuff recently but I don't think there was anything specific to this | 15:52 |
| stephenfin | so something like https://github.com/openstack/nova/commit/7d08a360d73dbf6b1b317867a51dfe86573b51b9 will do the trick | 15:52 |
| stephenfin | gagehugo: do you have an 'id' file or directory in doc/source/locale/? | 15:55 |
| stephenfin | that errors coming from here https://opendev.org/openstack/openstackdocstheme/src/branch/master/bin/docstheme-build-translated.sh#L153 | 15:55 |
| stephenfin | *'id' directory (not file) | 15:55 |
| stephenfin | because it's finding an 'id' directory and attempting to build the docs using the "translations" within | 15:55 |
| gagehugo | I don't believe we do | 15:56 |
| gagehugo | oh nvm we do | 15:56 |
| stephenfin | hmm, it's been there a while though | 15:57 |
| stephenfin | https://github.com/openstack/openstack-helm/commits/master/doc/source/locale/id/LC_MESSAGES | 15:57 |
| stephenfin | id is indonesia | 15:57 |
| stephenfin | https://www.science.co.il/language/Locale-codes.php | 15:57 |
| gagehugo | yeah | 15:58 |
| gagehugo | I assume our indonesia translations are missing something then? | 15:59 |
| stephenfin | okay, looks like a change in behavior introduced by a newer version of docutils which was in turn brought in by Sphinx 4.x https://lists.freebsd.org/pipermail/freebsd-ports/2021-April/120848.html | 15:59 |
| stephenfin | that points to this docutils change https://sourceforge.net/p/docutils/code/8465/ | 15:59 |
| clarkb | zbr: stephenfin: I'm still not entirely convinced that pbr/setuptools are avoiding conflict in that space because setuptools should be ignoring setup.cfg when pbr is used. However, it seems to work in all the cases we are using it so a technicality may not matter here | 15:59 |
| clarkb | basically setup.cfg as handled by pbr is incompatible with setup.cfg as handled by setuptools for a number of reasons (still not super happy they did that since pbr absed it on their original docs) and that warning implies setuptools is reading the file as if it can do so safely | 16:00 |
| stephenfin | clarkb: possibly. I touch pbr infrequently enough to have forgotten most of the context by the next time I work on it :-( (hence all my docs changes in the past, mostly as notes to myself) | 16:00 |
| clarkb | ya me too | 16:01 |
| fungi | clarkb: i think it's distutils which ends up complaining | 16:02 |
| gagehugo | stephenfin: thanks for the help! | 16:08 |
| stephenfin | gagehugo: Okay, so docutils actually provides a list of supported languages and Indonesian isn't among them https://repo.or.cz/docutils.git/tree/HEAD:/docutils/docutils/languages | 16:13 |
| stephenfin | so I guess the solution is 'rm -rf doc/source/locale/id' /o\ | 16:13 |
| gagehugo | ha | 16:14 |
| gagehugo | ok | 16:14 |
| gagehugo | I assume opendev does support it though? | 16:14 |
| stephenfin | you mean docs.openstack.org? | 16:14 |
| stephenfin | if not, what? | 16:14 |
| gagehugo | I mean the proposal bot was the thing that added it | 16:15 |
| gagehugo | I think anyway | 16:15 |
| stephenfin | Ah, yes. Those are generated by a bot based on translations located in Transifex or whatever service we use for translations now | 16:15 |
| stephenfin | I don't really know who, if anyone, is still doing those translations | 16:16 |
| gagehugo | oh ok | 16:16 |
| stephenfin | Is was ianychoi and company but I think they've all moved on | 16:16 |
| gagehugo | wasn't sure if removing it would cause issues | 16:16 |
| stephenfin | the bot might re-add them | 16:16 |
| stephenfin | guess wait and see :) | 16:16 |
| stephenfin | if they do, we can figure out how to go delete them at source | 16:16 |
| fungi | zbr: for a minimum pbr version, if you're trying to tackle deprecation warnings you probably want at least 5.5.1 as it's the first version to include https://review.opendev.org/752794 | 16:17 |
| gagehugo | ok | 16:17 |
| clarkb | stephenfin: gagehugo: if you can convince docutils to just ignore that dir that might be better? then the data is still there for people who want it rather tha ndiscarded | 16:17 |
| clarkb | certainly in a far less useable form though | 16:17 |
| zbr | fungi: thanks for the hint. my first attempt to locate it in release notes failed. | 16:18 |
| stephenfin | yes, we could modify https://opendev.org/openstack/openstackdocstheme/src/branch/master/bin/docstheme-build-translated.sh to skip unsupported locales | 16:18 |
| *** rpittau is now known as rpittau|afk | 16:21 | |
| zbr | fungi: so the appropiate way to do it looks more like https://review.opendev.org/c/openstack/tripleo-repos/+/802750 ? | 16:22 |
| stephenfin | I fricking hate bash arrays :-( | 16:22 |
| * zbr wonders who loves bash arrays... | 16:23 | |
| stephenfin | :) | 16:23 |
| fungi | zbr: i've got a personal pbr-using project where i export PYTHONWARNINGS=error globally in tox for all testenvs, and current pbr is no longer a problem for me. my current challenges are around the pip versions included in various interpreter builds (i build all my python interpreters from upstream source) vendoring old versions of things which touch distutils or urllib3, pyyaml using | 16:24 |
| fungi | outdated wheel metadata and missing cython directives | 16:24 |
| fungi | my PYTHONWARNINGS exclusion list is still very ugly as a result: | 16:25 |
| fungi | ignore::FutureWarning:Cython.Compiler.Main, ignore::DeprecationWarning:setuptools.config, ignore::DeprecationWarning:distutils.command.install, ignore::ImportWarning:importlib._bootstrap, ignore::DeprecationWarning:pip._vendor.urllib3.connection, ignore::DeprecationWarning:pip._vendor.urllib3.util.ssl_, ignore:The "license_file" option is deprecated. Use "license_files" | 16:25 |
| fungi | instead.:DeprecationWarning:wheel.bdist_wheel, ignore:The distutils package is deprecated and slated for removal in Python 3.12. Use setuptools or check PEP 632 for potential alternatives:DeprecationWarning, ignore:Creating a LegacyVersion has been deprecated and will be removed in the next major release:DeprecationWarning | 16:25 |
| zbr | for not openstack projects, i no longer use pbr, went full on setuptools-scm for 2+ years now. Not painless, but as you know no support for release notes. | 16:25 |
| fungi | does setuptools-scm create pep 440 compliant dev version numbers, and does it provide you with a way to encode the git commit? | 16:26 |
| fungi | i remember those being reasons i didn't use it in the past | 16:27 |
| zbr | those warnings are there for a reason, we need to be very careful which ones we hide. | 16:28 |
| fungi | yep, normally i would exclude by full message plus module, but a limitation of the PYTHONWARNINGS envvar is that it uses comma as a separator so can't represent messages containing commas (because it also lacks a quoting/escaping mechanism) | 16:29 |
| zbr | fungi: yes but not both at the same time, if you include the commit id, is no longer pepe 440 compliant. you can easily extend the versioning if you really want. | 16:30 |
| fungi | also some of those warnings are unlikely to ever be "fixed" because they're specific to outdated versions of modules vendored into outdated tools vendored into old python release branches | 16:30 |
| zbr | my preference was to avoid commit it and produce simple versions like `5.1.2.dev1` | 16:30 |
| fungi | zbr: yeah, one of the things i like about pbr is that it writes additional metadata you can reference so the commit id is accessible in the package info but not as part of the version string | 16:31 |
| zbr | i think is possible but i did not check, maybe already does this and i not know. indeed useful when implementing `--version`. | 16:31 |
| fungi | right, that's how i end up leveraging it | 16:32 |
| zbr | the maintainance improved recently so it should be quite easy to add features like this if needed. | 16:32 |
| fungi | and yeah, maybe it's gained that capability recently, i haven't tried it for a while | 16:33 |
| fungi | i also rely on pbr's authors and changelog file generation | 16:33 |
| zbr | one thing that i did not check yet is if we can remove the runtime dependency on setuptools-scm and only use it as a build-dependency, once somene built a package wheel or sdist, you should not need it anymore. | 16:34 |
| zbr | one bug that is not yet addressed is versionign of github tar archives, as they lack tags. i guess pbr has the same issue, unless github finds a way to add extra info to these nobody can guess these. | 16:35 |
| clarkb | that is beacuse github tar archives are not sdists | 16:36 |
| clarkb | if you want sdists in your github releases you have to manually add them yourself | 16:36 |
| fungi | yes, the problem with "github tarballs" is that they're missing egg-info metadata files | 16:36 |
| fungi | an sdist is not merely a tarball, it's a tarball with some very specific files located inside it | 16:37 |
| fungi | you *could* commit those metadata files into your git repository and they would probably work, but then you run into the commit/versioning race conditions pbr (and setuptools-scm) exists to solve | 16:37 |
| *** marios|ruck is now known as marios|out | 16:38 | |
| fungi | zbr: as for being only a build-time dependency and not a runtime dependency, you can do that with pbr just fine (it's my personal preferred way to use pbr). just use native python stdlib mechanisms to interface with package/module metadata | 16:39 |
| fungi | i only include pbr in my setup_requires, never in my install_requires | 16:39 |
| fungi | and none of my code imports pbr | 16:40 |
| stephenfin | gagehugo: clarkb: https://review.opendev.org/c/openstack/openstackdocstheme/+/802758 | 16:50 |
| stephenfin | just FYI | 16:50 |
| clarkb | stephenfin: fwiw I think you can do a for X in $FOO ; do stuff ;done and avoid the array stuff entirely as long as the entries are whitespace separated | 16:53 |
| stephenfin | o rly? | 16:53 |
| * stephenfin tries | 16:53 | |
| clarkb | ya so LANGS="x y z" ; for X in $LANGS; do echo $X ; done | 16:53 |
| gagehugo | stephenfin: We're gonna just remove the id dir for now to unblock people, but that works too, can always revert the change on our end | 16:53 |
| stephenfin | gagehugo: yes, I think that makes sense | 16:54 |
| clarkb | stephenfin: also is your last condition inverted from what it should be? | 16:54 |
| clarkb | I think you may want if ! $supported ; then | 16:55 |
| stephenfin | whoops, so it is /o\ | 16:55 |
| opendevreview | Jeremy Stanley proposed openstack/project-config master: Make Sunny an operator on Kata IRC channels https://review.opendev.org/c/openstack/project-config/+/802765 | 17:59 |
| *** mtreinish_ is now known as mtreinish | 18:52 | |
| *** prometheanfire is now known as Guest2621 | 19:34 | |
| opendevreview | Merged openstack/project-config master: Make Sunny an operator on Kata IRC channels https://review.opendev.org/c/openstack/project-config/+/802765 | 19:35 |
| *** timburke_ is now known as timburke | 20:55 | |
| opendevreview | Merged openstack/project-config master: Revert "nodepool: set inmotion cloud to zero" https://review.opendev.org/c/openstack/project-config/+/801010 | 22:06 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!