19:00:21 <gundalow> #startmeeting Ansible Molecule Working Group 19:00:21 <zodbot> Meeting started Wed Apr 10 19:00:21 2019 UTC. 19:00:21 <zodbot> This meeting is logged and archived in a public location. 19:00:21 <zodbot> The chair is gundalow. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:00:21 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic. 19:00:21 <zodbot> The meeting name has been set to 'ansible_molecule_working_group' 19:00:39 * gundalow waves 19:00:41 <gundalow> who's around? 19:01:10 * themroc waves 19:01:27 <smyers> howdy 19:01:38 <gundalow> #chair themroc smyers 19:01:38 <zodbot> Current chairs: gundalow smyers themroc 19:02:06 <gundalow> #info agenda https://github.com/ansible/community/issues/427 19:02:16 <gundalow> #topic Next Molecule release 19:03:05 <gundalow> #info for reasons that are now fixed the current molecule release is in PyPi is `2.20.0` 19:03:10 <gundalow> Fixing that is a pain 19:03:35 <gundalow> #info Though if the next release will be 2.21 in PyPi and Quay.io 19:04:05 <gundalow> So are we OK with that, we could just release what we have as 2.21 and do promotion/publicity for that 19:06:51 <gundalow> Not sure how much it matters 19:06:59 <gundalow> themroc: smyers any thoughts? 19:07:00 <smyers> So no fixing 2.20, and no 2.20.post0? 19:07:08 <themroc> no thoughts against it 19:08:28 <gundalow> cool 19:08:46 <gundalow> #topic Open Floor 19:08:50 <smyers> Something occured to me, which is that it would be nice to fix 2.20, but there's no real need to couple the fixing of 2.20 with the release of 2.21. If we want to just release 2.21 "soon", 2.20 can be released (or not) independently 19:09:35 <gundalow> I think we can just add a changelog then release 2.21 19:09:50 <gundalow> though I'd like to get input from other users here before we do that 19:09:50 <smyers> Definitely works for me 19:10:46 <gundalow> #info We now have a proposal form https://github.com/ansible/molecule/issues/new/choose and some proposals https://github.com/ansible/molecule/labels/proposal 19:11:28 <gundalow> Would welcome people to add their thoughts on https://github.com/ansible/molecule/issues/1923 19:11:41 <gundalow> smyers: themroc anything from yourselfs? 19:12:06 <smyers> Pretty much anything I could've come up with I opened up as issues or PRs... 19:12:23 <fabianvf> Sorry was stuck in traffic, am now here 19:12:51 <smyers> There's an open discussion right now regarding the use of yapf, and conflicts between the latest yapf and flake8...lemme find that link for the minutes 19:13:14 <smyers> https://github.com/ansible/molecule/issues/1916 19:13:24 <smyers> Not sure if that should be #info'd or what 19:14:18 <themroc> from what i have seen looking at the functional tests, there is a lot of duplicate code, so probably some place for making it better 19:15:19 <themroc> the intrication of tox / pytest / molecule is not super easy to follow at first sight 19:15:49 <gundalow> #chair fabianvf 19:15:49 <zodbot> Current chairs: fabianvf gundalow smyers themroc 19:15:54 <gundalow> fabianvf: all good, welcome :) 19:16:13 <gundalow> URLs get auto added 19:18:01 <themroc> smyers, ah yes it seems there is some reccurent discussion on that topic 19:19:35 <themroc> as a python newbie i can't really take position, but i can say that i am pretty happy to have a linter explaining to me some common errors and avoiding to commit absurd things 19:20:34 <themroc> i would let the specialists choose the best tool, and follow its rules 19:20:53 <smyers> flake8 is definitely great for that. I'm not personally fond of tools like yapf making style changes for you without giving you an easy way to understand *why* it wants things one way or another. 19:21:23 <nwilburn> For what its worth I am all for autoformatters if they work. I think caring about how your code looks is silly, as long as it follows the style guide. Having an autoformatter relieves you of that burden 19:21:56 <smyers> I also want to briefly mention, as in the issue, that I'm only trivially frustrated with yapf and am mostly interested in having tools that work together. Right now, yapf is generating code that flake8 doesn't like, and so it's a tricky problem. 19:22:18 <gundalow> For gh/ansible/ansible the coding standard is "What's enforced by CI at the time your PR is about to be merged", anything else is off-topic 19:22:43 <gundalow> unless it really isn't readable 19:23:14 <gundalow> I haven't used many of these auto formatters 19:23:16 <fabianvf> I like autoformatters as well, I just contribute to too many projects to remember all the different guidelines. If no autoformatter then what gundalow said 19:23:18 <smyers> That's generally been working for molecule, too 19:23:49 <gundalow> If there are autoformatters that are integrated into GitHub's `suggestion` system that could be interesting 19:24:10 <smyers> The only issue is that if you use the latest yapf, format-check passes but lint doesn't, or vice versa, which is sad. There is an open issue on yapf now about this, so a potential option is to keep yapf pinned until it gets fixed 19:24:17 <gundalow> otherwise I think we should just stick with "what's enforced by CI" as we have other things to do 19:24:42 <gundalow> #agreed whatever formatting (yapf, etc) tooling we use should be version pinned 19:24:47 <gundalow> smyers: Excellent point 19:25:17 <Not-e6c6> [molecule] webknjaz opened pull request #1929: Upgrade RTD config to use v2 schema - https://git.io/fjqCB 19:25:25 <Not-e6c6> [molecule] webknjaz closed pull request #1929: Upgrade RTD config to use v2 schema - https://git.io/fjqCB 19:26:21 <smyers> This problem provides an opportunity to review molecule's usage of yapf, but there's no imminent need to make a decision about it. 19:26:48 <smyers> I'll take some of these points and put them up in the issue, since I'm not sure I was clear about it there 19:26:48 <gundalow> nod 19:27:05 <gundalow> smyers: it's a difficult topic, thank you for taking the time 19:27:45 <Not-e6c6> [molecule] webknjaz opened pull request #1930: Add pre-commit's config to codeowners - https://git.io/fjqCw 19:27:52 <Not-e6c6> [molecule] webknjaz closed pull request #1930: Add pre-commit's config to codeowners - https://git.io/fjqCw 19:28:20 <pabelanger> o/ So, a few weeks back I said I find some time to run a zuul jobs against ansible/molecule, I think I might be ready to do some of that testing later this week. Is that something we are still okay with doing? 19:28:37 <gundalow> pabelanger: Sure, that would be great 19:29:14 <pabelanger> cool, I'll likely raise a PR in the next few days and likey see some comments from ansible-zuul user 19:30:22 <smyers> I've been wanting to get into zuul, so I'm personally pretty excited about it 19:31:03 <themroc> me too :) 19:31:20 <pabelanger> https://dashboard.zuul.ansible.com is how you can follow along on jobs that we are currently running. 19:31:32 <pabelanger> just working with the ansible-runner team now, to get them migrated first 19:31:54 <pabelanger> it will pretty much look like this, https://github.com/ansible/ansible-runner/pull/255 but against molecule 19:34:43 <gundalow> #info After Ansible 2.8 has been released we will look at using `ansible-core-ci` (the interface that Ansible uses to spin up AWS, Azure, Hetzner, etc, etc instances) for testing Molecule 19:35:04 <gundalow> pabelanger: Thanks for the update :) 19:35:14 <gundalow> If nothing else I'll close in a few minutes 19:35:31 * smyers has nothing else 19:39:53 <gundalow> #chair pabelanger 19:39:53 <zodbot> Current chairs: fabianvf gundalow pabelanger smyers themroc 19:39:57 <gundalow> (just for the logs) 19:40:00 <gundalow> Thanks everybody! 19:40:02 <gundalow> #endmeeting