20:00:00 <jborean93> #startmeeting Ansible Windows Working Group
20:00:00 <zodbot> Meeting started Tue May 16 20:00:00 2023 UTC.
20:00:00 <zodbot> This meeting is logged and archived in a public location.
20:00:00 <zodbot> The chair is jborean93. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
20:00:00 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:00:00 <zodbot> The meeting name has been set to 'ansible_windows_working_group'
20:00:10 <nitzmahone> o/
20:00:11 <jborean93> #topic Agenda https://github.com/ansible/community/issues/682
20:00:17 <jborean93> #chair nitzmahone
20:00:17 <zodbot> Current chairs: jborean93 nitzmahone
20:00:26 <jborean93> Nothing on the agenda so straight into our favourite
20:00:29 * nitzmahone lurks while eating sad lunch of re-heated corn dogs
20:00:30 <jborean93> #topic Open Floor
20:01:32 <nitzmahone> Looking at that thing up above around the bad EE dep metadata
20:01:32 <jborean93> Nothing too interesting to share. With the impending doom of Server 2012 and 2012 R2 in CI, I'm finally jumping the gun and bumping the Windows collections to 2.x.y and removing some of the deprecated stuff from years ago
20:02:12 <nitzmahone> Are the 2.x collections "hard" PS 5.1+ or just soft "we're not testing less"?
20:02:12 <jborean93> Yea that's not too suprising, I don't know enough about EE to really understand the problem. All I know is I could build the container with ansible-builder with the file currently checked in
20:02:19 <jborean93> the latter
20:02:24 <nitzmahone> noice
20:03:12 <jborean93> https://www.irccloud.com/pastebin/R1VtqMXL/
20:03:38 <jborean93> I did have that originally but then builder wouldn't work because `python39-devel` didn't work on whatever default image it used
20:03:46 <nitzmahone> Yeah, this kind of stuff is going to be an endless battle with builder
20:04:12 <jborean93> IIRC I just copied what the networking team had in their bindep files
20:04:46 <nitzmahone> We need to come up with some better standards and recommendations there, as well as the ability to mask deps from busted collections so you can still build without hacking the collection itself
20:04:59 <nitzmahone> It's on the fire for builder 3.1
20:05:04 * jborean93 Just need to implement a pure Python Kerberos library so we don't need to compile anything :)
20:05:28 <nitzmahone> You'll question your life choices there after the first production segfault reports ;)
20:05:56 <jborean93> heh, shouldn't segfault, it's pure Python :)
20:06:01 <jborean93> But yes I value my sanity here
20:06:35 <jborean93> I will admit I hate it how there's no `async` method available to get the TGT as that could block the loop if you don't put it in a thread
20:07:24 <nitzmahone> Yeah, the only way I can think of to handle the version-dependent Python packages is to add a bunch more custom environment markers (I forget what bindep calls them, "configs" or something?), but that still only solves a small part of the problem in the face of "works on any RHELish base distro"
20:07:45 <jborean93> sure but this is a very common case
20:07:57 <jborean93> Even if it's ugly, the current way it works just isn't really viable
20:09:44 <jborean93> In any case I've got nothing else, will close the meeting in a few minutes unless something else pops up
20:10:45 <nitzmahone> Yeah, there's no obvious easy solution to that one that keeps everyone happy
20:12:28 <nitzmahone> (downstream stuff is all "we're not compiling stuff, we don't need this stinking headers package!"
20:15:20 <jborean93> so much fun, sucks the docs for this is a bit minimal oh well will investigate it further if support end up reaching me about it
20:16:41 <jborean93> cool, will close the meeting, enjoy!
20:16:46 <nitzmahone> thanks for driving
20:16:47 <jborean93> #endmeeting