| *** dasm is now known as dasm|off | 00:11 | |
| *** chandankumar is now known as chkumar|ruck | 04:47 | |
| *** chkumar|ruck is now known as chandankumar | 04:51 | |
| *** ysandeep|out is now known as ysandeep | 05:49 | |
| *** ministry is now known as __ministry | 06:48 | |
| *** jpena|off is now known as jpena | 07:01 | |
| *** ysandeep is now known as ysandeep|lunch | 07:29 | |
| *** arxcruz|out is now known as arxcruz | 08:11 | |
| *** ysandeep|lunch is now known as ysandeep | 09:50 | |
| *** icey_ is now known as icey | 09:53 | |
| *** dviroel|out is now known as dviroel | 11:31 | |
| *** ysandeep is now known as ysandeep|afk | 11:41 | |
| dpawlik | dansmith: hey, I saw your latest change https://review.opendev.org/c/openstack/devstack/+/837139 I have one question: the whole json should be parsed as a single message, right? | 12:49 |
|---|---|---|
| *** ysandeep|afk is now known as ysandeep | 12:53 | |
| *** dasm|off is now known as dasm | 13:03 | |
| *** blarnath is now known as d34dh0r53 | 13:07 | |
| dansmith | dpawlik: I'm not sure what that means.. I want to be able to graph or aggregate individual values in there | 13:27 |
| dansmith | dpawlik: I think clarkb knows better how this will all fit | 13:27 |
| stephenfin | clarkb: fungi: Just as an FYI, I tried to use pyproject.toml with pbr for sphinx-click over the long weekend. Found a couple of limitations. | 13:38 |
| dpawlik | dansmith: allright. let's see what he says | 13:38 |
| stephenfin | Firstly, I had to keep the 'setup.py' file otherwise setuptools didn't load the pbr machinery. I guess this is expected as _something_ needs to tell setuptools to use that, and there's no way to configure this via setup.cfg (that I know of) | 13:39 |
| stephenfin | I found I had to define '[project] name' in pyproject.toml or it simply wouldn't work. I also has to populate '[project] dynamic' with almost the full set of attributes otherwise setuptools issued deprecation warnings | 13:40 |
| stephenfin | I did try moving stuff from setup.cfg to pyproject.toml but it crashed and burned, likely because pbr has no idea about pyproject.toml and its "view of the world" was therefore incomplete | 13:41 |
| pdeore | fungi, something is wrong with glance team meeting etherpad https://etherpad.opendev.org/p/glance-team-meeting-agenda | 13:41 |
| fungi | stephenfin: yes, the setup.py is still necessary for now. i actually have a pyproject.toml based application using pbr since months though | 13:41 |
| pdeore | could you please help us to get it reverted? | 13:41 |
| stephenfin | other than that though, things looked pretty good | 13:41 |
| stephenfin | fungi: Oh, if you've a link I'd like to compare | 13:42 |
| fungi | stephenfin: i do still put the metadata in setup.cfg for now, though you'd need a *very* new setuptools to support metadata in pyproject.toml anyway | 13:42 |
| pdeore | fungi, ^^ | 13:43 |
| stephenfin | Yeah, I'm guessing Fedora might just have that. I'm not sure. Here's the changes I ended up with, in any case (just in case you're curious) https://github.com/click-contrib/sphinx-click/commit/4aba8f2301332e8892a7db75111bd66441553a76 | 13:44 |
| fungi | stephenfin: https://mudpy.org/gitweb?p=mudpy.git;a=tree | 13:44 |
| fungi | pdeore: i can in just a moment, need to switch computers and identify the last good revision from the timeslider | 13:45 |
| pdeore | fungi, sure np, | 13:46 |
| fungi | stephenfin: i want to test switching the metadata over to pyproject.toml, but some of my deps seem to be downgrading setuptools to too old of a version to support that yet | 13:47 |
| fungi | but maybe if i bump the minimum setuptools in pyproject.toml it'll at least give me an error | 13:49 |
| stephenfin | fungi: that's strange, you can get away without '[project] name' yet I get this issue https://paste.opendev.org/show/bK4aRjbLoLheL2fWyKPy/ | 13:52 |
| stephenfin | ah, it's the presence of '[tool.setuptools] py-modules' | 13:53 |
| fungi | oh, yes that's probably going to require a host of other stuff | 13:54 |
| stephenfin | that appears to be where the '[project] dynamic' stuff is coming from also | 13:56 |
| fungi | pdeore: looks like the usual case we see of some user with an autotranslation browser extension has accidentally translated a chunk of the document into standard chinese. i think you'll want it rolled back to revision 103282 from 2022-04-14 since the translation happened immediately after that | 13:57 |
| fungi | #status log Restored the glance-team-meeting-agenda Etherpad to revision 103282 at pdeore's request | 14:00 |
| opendevstatus | fungi: finished logging | 14:00 |
| fungi | pdeore: it's probably no coincidence that i just did the same for glance's ptg pad on friday | 14:00 |
| fungi | i suspect it was the same user who visited both pads, so maybe it's someone active in glance? it would be awesome if we could figure out what browser extension is doing that | 14:01 |
| fungi | stephenfin: if you skim the recent entries for the setuptools changelog (particularly for 61.0.0) you'll see that it treats presence of [tool.setuptools] in pyproject.toml as a switch to turn on a bunch of validation and new behaviors | 14:17 |
| fungi | https://setuptools.pypa.io/en/latest/history.html#v61-0-0 | 14:17 |
| * stephenfin looks | 14:17 | |
| stephenfin | My one big criticism of this pyproject.toml thing is that its mere presence seems to completely change the build behavior (to isolated builds) | 14:18 |
| stephenfin | sounds like setuptools is following a similar path :-( | 14:19 |
| fungi | yep | 14:19 |
| fungi | though for the stuff i've been testing with it so far, build isolation hasn't caused a problem | 14:20 |
| *** dviroel is now known as dviroel|afk | 14:24 | |
| *** dviroel|afk is now known as dviroel|mtg | 14:39 | |
| *** ysandeep- is now known as ysandeep|out | 14:45 | |
| clarkb | dansmith: dpawlik: yes I think you index it not like line by line logs but as a single json document. I would use a separate set of indexes too. Then you can query the info out of that index and get per document data | 14:50 |
| clarkb | dansmith: dpawlik: basically using elasticsaerch more properly as a document store | 14:50 |
| dpawlik | clarkb: ok, thanks | 14:51 |
| dpawlik | dansmith: I will prepare a patch for logsender that will parse the json and send it today/tomorrow | 14:51 |
| dansmith | thanks! | 14:52 |
| *** dviroel|mtg is now known as dviroel | 14:54 | |
| *** dviroel is now known as dviroel|lunch | 15:06 | |
| bcafarel | gmann: hi is https://review.opendev.org/q/I941ef5ea90970a0901236afe81c551aaf24ac1d8 is the proper fix for git issue on ussuri and older branches? | 16:05 |
| bcafarel | (hoping to clear some of the pending train backports we have) | 16:06 |
| *** jpena is now known as jpena|off | 16:08 | |
| fungi | it's less of a fix and more of a workaround, but properly fixing is going to need a hefty overhaul of how devstack installs and runs services | 16:09 |
| *** dviroel|lunch is now known as dviroel | 16:14 | |
| clarkb | bcafarel: fungi: but ya that second iteration should work on bionic which is what ussuri and older run on | 16:16 |
| clarkb | the first iteration did not, it only worked on focal | 16:16 |
| bcafarel | sounds good | 16:24 |
| *** dasm is now known as dasm|off | 17:06 | |
| *** dviroel is now known as dviroel|out | 20:41 | |
| gmann | fungi: can you +W in this https://review.opendev.org/c/openstack/project-config/+/836703 | 21:35 |
| gmann | bcafarel: this one basically https://review.opendev.org/c/openstack/devstack/+/837749 | 21:35 |
| bcafarel | gmann: thanks for the squashed backports! | 22:07 |
| opendevreview | Merged openstack/project-config master: Remove tempest-lib from infra https://review.opendev.org/c/openstack/project-config/+/836703 | 22:31 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!