15:00:49 <sdoran> #startmeeting Ansible Core Public IRC Meeting https://github.com/ansible/community/issues/568
15:00:49 <zodbot> Meeting started Thu Oct 29 15:00:49 2020 UTC.
15:00:49 <zodbot> This meeting is logged and archived in a public location.
15:00:49 <zodbot> The chair is sdoran. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:49 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:49 <zodbot> The meeting name has been set to 'ansible_core_public_irc_meeting_https://github.com/ansible/community/issues/568'
15:01:32 <sdoran> Happy Thursday, everyone.
15:01:54 <sdoran> #topic https://github.com/ansible/ansible/pull/71734
15:02:18 <cyberpear> o/
15:02:34 <sdoran> felixfontein: Do we need to discuss anything more on that one or does it just need reviews?
15:02:49 <sdoran> #chair cyberpear bcoca
15:02:49 <zodbot> Current chairs: bcoca cyberpear sdoran
15:04:04 <sdoran> Ok, moving on.
15:04:19 <sdoran> #topic opinions on `fallible` or `ansiboom` for the canary branch formerly known as `ansible_chaos`, last call for other ideas?
15:04:44 <sdoran> Of those two, I prefer `fallible`.
15:05:17 <bcoca> fearable ...
15:05:45 <cyberpear> first 4 chars collide w/ fallocate, though
15:05:52 * agaffney votes for the bike shed to be painted blue
15:06:02 <cyberpear> obviously ansiboom would collide w/ ansible itself
15:06:28 <bcoca> asciicow
15:06:38 <sivel> oh, hey, meeting or something
15:07:08 <cyberpear> fearible++
15:07:22 <sdoran> I suppose we could  go with `'–__//\fallible` to prevent prefix collisions. 😉
15:07:23 <sivel> I don't think partial naming collisions is something we're going to concern ourselves with
15:07:38 <sivel> we'll be debating this until the end of time if we do that
15:07:54 <bcoca> sivel: did you just define bikesheding?
15:08:03 * bcoca ducks
15:08:12 <agaffney> what is this thing we're debating the naming of, anyway? I have no idea what ansible_chaos is
15:08:35 <sivel> agaffney: ansible-chaos will be a packaging of ansible-base that is used for experimental features we are looking for feedback on
15:08:37 <bcoca> agaffney: lab branch/package for experiments that might make it into main ansible , but distributed so people can test
15:08:49 <sivel> they may or may not ever be included in ansible or the devel branch
15:09:15 <sivel> we'll have a branch named the same thing, where that development will happen
15:09:57 <sivel> basically an easier way to get people to test out and provide feedback for "blessed" experiments, usually from the core team, that are pretty large in scope
15:10:20 <sivel> such as rewriting the process model, moving more work such as loops pre-fork, etc...
15:10:59 <sivel> as a separate package, it will also be installable beside of ansible, using a different package name, and diffent CLI tool names
15:11:22 <agaffney> doesn't having a single branch/package for that cause a bottleneck/conflicts for testing these things?
15:11:59 <sivel> agaffney: quite possibly, but it lowers the bar for testing, in that we are packging this on a regular basis, and it will be available at minimum on pypi also
15:12:03 <agaffney> I'm not sure what the separate package with separate tool names really gains over just pip-installing from a git branch into a virtualenv
15:12:23 <agaffney> lowering the bar is good, I guess
15:12:32 <sivel> Your proposed workflow has the issue of requiring someone more familiar with python development practices
15:12:43 <sivel> that's not necessarily the person we're targeting
15:12:52 <sivel> but more targeting the average community user
15:13:00 <agaffney> fair enough
15:13:23 <sivel> As of not the core team is largely in favor of the name "fallible"
15:13:32 <sivel> now*
15:13:35 <bcoca> s/not/now/
15:13:44 <sdoran> Any more votes for `fallible` or `ansiboom`?
15:13:56 <sdoran> Seems like we're going to go with `fallible`.
15:14:08 <cyberpear> I like "fearible" since there's no name collision, even on the first 2 chars, (in a minimal container)
15:14:35 <bcoca> fable ... you can read in many ways
15:14:39 <bcoca> also short
15:15:13 <felixfontein> fallible has 4 letters of tab completion collision with fallocate
15:15:18 <sivel> in any case, I think with the current core vote, which is easily 5 people, we'd need some other name to get more votes than what we already have for the current name.
15:15:39 <sivel> felixfontein: yeah, that was stated, and not really something we're going to worry with
15:15:41 <felixfontein> ansiboom is even worse with that regard
15:16:04 <cyberpear> fable also has no collisions in the first 3 chars (in the fedora package set)
15:16:11 <agaffney> imo, the name should be recognizable as being related to ansible. 'ansiboom' fits that more than 'fallible' does, but I have no strong opinion either way
15:16:15 <cyberpear> but doesn't sound like "ansible"
15:17:57 <sivel> sounds like we have our other potential ideas. we can probably move on
15:18:07 <sdoran> #topic open floor
15:18:50 * bcoca hugs ansiball
15:20:56 <felixfontein> antsible?
15:21:23 <sdoran> What about it?
15:21:31 <cyberpear> collision w/ the build tool, and doesn't say "caution!"?
15:21:53 * bcoca goes to the ansibar
15:22:29 <felixfontein> ansibleu :)
15:22:44 <felixfontein> sdoran: nothing serious
15:22:59 <sdoran> Ok. Just wanted to make sure I wasn't missing. something. :)
15:23:29 * felixfontein is trying to stop thinking about random words sounding similar to 'ansible' and try to get back to work :D
15:24:12 <cyberpear> re the `pip install --upgrade` issue... if you try to `pip install ~/path/to/ansible-base` devel version, with `--use-feature=2020-resolver`, the ansible package gets downgraded from 2.10.1 to 2.9
15:24:14 <bcoca> felixfontein: lost cause .... once it gets deep in there, very hard to weed out
15:24:34 <cyberpear> it "works" without that, but you get a warning that when the new resolver is default, it will break
15:25:23 <cyberpear> but I guess it's separate issue from the 2.9 to 2.10 upgrade issue
15:26:53 <bcoca> fun part will be collections that require a specific ansilbe version
15:27:09 <sivel> and even more fun when we rename ansible-base again ;)
15:27:13 <cyberpear> but maybe it's a pip resolver bug and "not our problem"
15:28:28 <cyberpear> pip should probably error out and refuse to downgrade without an extra flag
15:28:29 <bcoca> cyberpear: runtime evaluation of required version makes it our problem also, not just a packaging issue
15:29:22 <sivel> It's largely a lost cause trying to address the conflicts with pip and python packaging
15:29:30 <sivel> we've done about as best as we can
15:30:10 <sivel> we'll be in a similar situation as docker vs docker-py for a while
15:32:41 <sivel> effectively, going forward for a few releases pip upgrades are going to be problematic
15:33:35 <felixfontein> bcoca: ansible.posix doesn't like ansible-base 2.11.0.dev
15:34:08 <sivel> I doubt that team has concerned themselves with it yet
15:34:08 <bcoca> im aware
15:34:13 <felixfontein> bcoca: fortunately ansible just warns and ignores it (AFAIK)
15:37:17 <cyberpear> `feasible` kind of like `fearible` but it's an actual word
15:37:46 <bcoca> fungible
15:38:06 <bcoca> funable
15:44:09 <sdoran> If there are no other topics,  I will end the meeting in five minutes.
15:50:08 <sdoran> #endmeeting