| *** e-tienne has quit IRC | 00:41 | |
| *** weshay_pto is now known as weshay | 13:07 | |
| *** herald85 has joined #ara | 13:20 | |
| *** weshay is now known as weshay1-1 | 13:34 | |
| *** weshay1-1 is now known as weshay | 14:05 | |
| *** etienne has joined #ara | 15:23 | |
| openstackgerrit | David Moreau Simard proposed openstack/ara-plugins master: Remove leftover .gitkeep file https://review.openstack.org/625302 | 16:36 |
|---|---|---|
| openstackgerrit | David Moreau Simard proposed openstack/ara-plugins master: Add helper commands to determine plugin location https://review.openstack.org/625303 | 16:36 |
| openstackgerrit | David Moreau Simard proposed openstack/ara-plugins master: Improve README to include docs for enabling ara-plugins https://review.openstack.org/625304 | 16:37 |
| *** herald85 has quit IRC | 16:37 | |
| *** sshnaidm|off has quit IRC | 17:10 | |
| *** sshnaidm|off has joined #ara | 17:25 | |
| openstackgerrit | David Moreau Simard proposed openstack/ara-clients master: Use 'ara-clients' instead of __name__ when determining client version https://review.openstack.org/625316 | 17:29 |
| openstackgerrit | David Moreau Simard proposed openstack/ara-clients master: Use 'ara-clients' instead of __name__ when determining client version https://review.openstack.org/625316 | 17:30 |
| *** etienne has quit IRC | 18:34 | |
| apollo13 | dmsimard: gerrit no longer shows +1/+2 for me when I try to reply to a merge request, just a comment box | 19:08 |
| apollo13 | dmsimard: oh, you already merged it :D | 19:08 |
| dmsimard | apollo13: tried to wait but needed to make progress :( | 19:09 |
| apollo13 | or rather I can't read my emails | 19:09 |
| apollo13 | dmsimard: no problem | 19:09 |
| dmsimard | apollo13: btw I'm going to send a patch to add a lot of docs on top of https://review.openstack.org/#/c/624443/ soon | 19:09 |
| apollo13 | I think I looked at them anyways and gave my feedback; will look through the open ones today | 19:09 |
| dmsimard | it looks like dynaconf actually works pretty well despite a little workaround | 19:10 |
| dmsimard | Would very much appreciate your input once I have the docs patch up | 19:11 |
| apollo13 | I'll look at it right now, I should understand it without the docs too | 19:11 |
| *** gvincent has quit IRC | 19:12 | |
| apollo13 | dmsimard: the only thing I dislike about dynaconf is how it integrates into django | 19:13 |
| dmsimard | apollo13: sure -- I do have to fix the linters failure and address some things I realized when writing docs | 19:13 |
| dmsimard | apollo13: agreed that the django implementation could be better but it's not /that/ bad, if anything we're empowered to improve it | 19:15 |
| dmsimard | you might even have an opinion on https://github.com/rochacbruno/dynaconf/issues/89 | 19:15 |
| apollo13 | not yet no (just read the ticket) | 19:16 |
| openstackgerrit | David Moreau Simard proposed openstack/ara-server master: WIP: Basic documentation https://review.openstack.org/623350 | 19:20 |
| dmsimard | ^ it's still WIP but is better than nothing to give a bit of context into how the config might work | 19:20 |
| apollo13 | does dynaconf patch os.environ? | 19:21 |
| dmsimard | I don't think ? Why ? | 19:22 |
| apollo13 | https://review.openstack.org/#/c/624443/4/ara/server/settings.py | 19:22 |
| apollo13 | because you use os.environ there | 19:22 |
| apollo13 | shouldn't that be dynaconf.somethingsomething so it could also read from a conf file? | 19:22 |
| dmsimard | apollo13: the os.environ usage to get the database credentials is the compromise I settled with to work around the bug I linked | 19:24 |
| dmsimard | dynaconf would normally grab ARA_* variables | 19:25 |
| dmsimard | so those lines are /mostly/ no-op, dynaconf will still correctly "overwrite" them if they are defined and the config file will have precedence if the env vars aren't set | 19:25 |
| apollo13 | dmsimard: I am under the impression that that would still work if you import dynaconf in, manage.py / wsgi.py | 19:26 |
| apollo13 | from https://github.com/rochacbruno/dynaconf/issues/89#issuecomment-425691017 | 19:26 |
| dmsimard | That's how it is set up right now | 19:26 |
| apollo13 | so why do you need that os.environ stuff? | 19:26 |
| dmsimard | In my testing, there seemed to be a sort of circular run around. I forget the specifics. | 19:27 |
| dmsimard | Something like the database being set to the defaults until dynaconf took over and that was sufficient to be off | 19:28 |
| dmsimard | So those lines ensures that it is consistent with or without env vars | 19:28 |
| apollo13 | and you also need it for ARA_BASE_DIR so the default for DATABASE_NAME also works? | 19:29 |
| dmsimard | Yeah, and at the end of settings.py we write out a configuration template file. I'm not convinced that this specific part needs to live in settings.py outright, though, I've considered writing a django-admin command for it | 19:31 |
| apollo13 | +1 on the admin-command | 19:31 |
| dmsimard | I'd probably have the admin-command be able to write a configuration template in different formats (yaml, ini, toml, etc.) | 19:32 |
| apollo13 | would that really be such a big deal? it is just a config file, don't overcomplicate it :D | 19:32 |
| dmsimard | ¯\_(ツ)_/¯ | 19:33 |
| dmsimard | for the database thing, I looked at what pulp did and they sort of cheat in the sense that they don't ship a working config: https://github.com/pulp/pulp/blob/a3aaa5e80ee8640c25d3850c7d4a5c12fc6e8faa/pulpcore/app/settings.py#L171-L182 | 19:34 |
| dmsimard | I want to ship a working config :) | 19:34 |
| apollo13 | https://github.com/rochacbruno/dynaconf/blob/master/dynaconf/contrib/django_dynaconf/dynaconf_django_conf.py#L3 is most likely triggering django to go through databases and whatnot | 19:40 |
| apollo13 | imo they should directly import DJANGO_SETTINGS_MODULE there | 19:41 |
| apollo13 | although that will probably miss the global settings | 19:42 |
| apollo13 | but then this will always trigger django evaluation which tends to break things as can be seen :D | 19:42 |
| dmsimard | yeah... it's never going to be able to 100% prevent settings from loading before setting it's own settings :p | 19:49 |
| apollo13 | actually it wouldn't need to, it just overcomplicates it | 19:52 |
| apollo13 | anyways, gotta fight soap for a bit | 19:53 |
| dmsimard | Xaroth: btw there's no license or anything for https://github.com/Xaroth/ansible-uwsgi -- did you want to contribute it to the ara role as-is ? I don't want to unnecessarily copy paste things without attribution or credit of some kind :p | 20:29 |
| *** PaddyF has joined #ara | 21:10 | |
| ara-slack | yooynas: @yooynas has joined the channel | 21:16 |
| *** weshay is now known as weshay_pto | 21:44 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!