15:02:10 <bcoca> #startmeeting ansible core public irc meeting
15:02:10 <zodbot> Meeting started Thu Feb 17 15:02:10 2022 UTC.
15:02:10 <zodbot> This meeting is logged and archived in a public location.
15:02:10 <zodbot> The chair is bcoca. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
15:02:10 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:02:10 <zodbot> The meeting name has been set to 'ansible_core_public_irc_meeting'
15:03:14 <bcoca> #topic Type hinting in core
15:03:29 <bcoca> going to assume we still don't have the doc detaliing our plans on this?
15:04:08 <bcoca> #topic extend validate-modules to all plugins
15:04:58 <sivel> sorry, was lost elsewhere
15:05:05 <sivel> we do have a proposal for type hinting
15:06:54 <bcoca> ah, nice, link?
15:06:59 <bcoca> #topic Type hinting in core
15:08:41 <felixfontein> o/
15:10:03 <bcoca> bueler?
15:12:14 <bcoca> k, skipping topic again, thinki broke the sivel
15:12:26 <bcoca> # topic  extend validate-modules to all plugins
15:12:33 <felixfontein> \o/
15:12:33 <sivel> sorry, lost in gdb right now :(
15:12:34 <bcoca> not sure what we are waiting on this
15:12:40 <bcoca> dammit .. you are doing that on purpose!
15:12:41 <felixfontein> me neither :)
15:13:03 <sivel> https://github.com/ansible/proposals/issues/202
15:13:12 <bcoca> felixfontein: actually, right now, i would wait for test/filters to also be available, but other than that i gave my +1 long ago
15:13:22 <felixfontein> a bit of more work might be necessary once the sidecar PR is merged though, since it adds new documentable plugin types
15:13:33 <felixfontein> yep exactly :)
15:13:34 <bcoca> jinx
15:13:54 <bcoca> k, since rest are PST i'll try to ping later to see what hold up is
15:13:57 <bcoca> #topic Type hinting in core
15:14:04 <bcoca> #link  https://github.com/ansible/proposals/issues/202
15:14:05 <felixfontein> thanks!
15:15:11 <bcoca> #topic Type hinting in core
15:15:17 <bcoca> #link  https://github.com/ansible/proposals/issues/202
15:15:35 <bcoca> k, leaving that for consideration for htat topic
15:16:06 <bcoca> #topic should we normalize passing lists to lookups
15:16:19 <bcoca> #link https://github.com/ansible/ansible/pull/76388
15:16:39 <bcoca> ^ PR above does this for ONE lookup, thoughts were 'normalize this' across lookups since most are 'supposed to' handle lists as input
15:17:00 <bcoca> and eventually add CI rule (validate-modules seems good place for htis)
15:17:39 <bcoca> thoughts? pros/cons?
15:18:02 <bcoca> also avoids having to teach people the *mylist pythonism
15:18:39 * bcoca searches for criket nest to wake em from hybernation
15:20:39 <bcoca> #topic add backup_dir to backup options
15:20:47 <bcoca> @rabin ?
15:20:59 <felixfontein> bcoca: I just started reading on the *list PR
15:21:23 <felixfontein> I don't think this should be done globally, some lookup plugins don't want their data modified that way
15:21:55 <bcoca> like? they  are expecte to work with with_ which does take lists as input
15:22:25 <bcoca> jsut as they are expected/semi required to output lists as output
15:23:07 <felixfontein> I'd have to look for concrete instances, but lookups are used for many more things than for loops
15:23:10 <bcoca> bad enforcement has made it so not all behave correctly, tried to unify at least the way input is handled via 'config system' but most still have their own input parsing
15:23:33 <bcoca> i know, but they still have to be 'loop compatible'
15:24:19 <felixfontein> the cartesian lookup would be broken by this
15:24:34 <felixfontein> it really needs the lists you pass in to arrive as lists
15:24:59 <bcoca> list of lists
15:25:00 <felixfontein> if you flatten its input, you get a wrong result (except in a very special case :) )
15:25:29 <felixfontein> this would be anothe rlevel of magic, and magic always comes back to bite you
15:25:30 <bcoca> not flattening, also cartesian should be removed, its is redundant in many ways (i wrote it when unaware of 'toghether')
15:25:32 <felixfontein> ALWAYS :)
15:25:56 <bcoca> i dont disagree, but there is sooo much magic in lookups, this is more of making it consistent
15:26:11 <bcoca> since the maigc is already present in many
15:30:26 <felixfontein> I think you first need to write down how exactly the mechanism should work
15:30:45 <felixfontein> at least right now I don't have enough brain capacity left to understand how it works and whether it's bad or good magic ;)
15:31:18 <bcoca> mostly as the PR shows, but of course making it 'smarter' in cases wher eyou are looking for 'lists of lists' like nested/toghther/cartesian
15:47:17 <bcoca> k, seems we are out of gas, closing for now, most issues will still need revisit next meeting since we also lack a meaningful quorum
15:47:21 <bcoca> #endmeeting