18:00:14 <felixfontein> #startmeeting Ansible Community Meeting
18:00:14 <zodbot> Meeting started Wed May 11 18:00:14 2022 UTC.
18:00:14 <zodbot> This meeting is logged and archived in a public location.
18:00:14 <zodbot> The chair is felixfontein. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
18:00:14 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
18:00:14 <zodbot> The meeting name has been set to 'ansible_community_meeting'
18:00:14 <felixfontein> #topic Agenda https://github.com/ansible/community/issues/645
18:00:14 <felixfontein> acozine andersson007_ baptistemm bcoca briantist cyberpear cybette dericcrago dmsimard felixfontein geerlingguy gundalow gwmngilfen ikhan_ jillr jtanner lmodemal misc nitzmahone resmo samccann tadeboro cidrblock thaumos zbr: ping!
18:00:17 <briantist> o/
18:00:18 <felixfontein> #info Agenda: https://github.com/ansible/community/issues/645 / Topics: https://github.com/ansible-community/community-topics
18:00:21 <felixfontein> #topic Updates
18:00:21 <andersson007_> o/
18:00:48 <cybette> o/
18:00:57 <felixfontein> #chair briantist andersson007_ cybette
18:00:57 <zodbot> Current chairs: andersson007_ briantist cybette felixfontein
18:02:06 <andersson007_> #info A kind reminder for the Steering Committee that the [ibm.spectrum_virtualize collection inclusion request](https://github.com/ansible-collections/ansible-inclusion/discussions/35) is waiting for the second review (more than a month passed since the first round was done). Could anyone from the Committee please review it? Thanks.
18:02:24 <dmsimard> \o
18:02:38 <felixfontein> #chair dmsimard
18:02:38 <zodbot> Current chairs: andersson007_ briantist cybette dmsimard felixfontein
18:03:06 <samccann> o/
18:03:49 <felixfontein> #chair samccann
18:03:49 <zodbot> Current chairs: andersson007_ briantist cybette dmsimard felixfontein samccann
18:05:13 <felixfontein> looks like not too many are around today :)
18:06:33 <felixfontein> #info The ansible-core 2.13 GA date has been changed from May 23th to May 16th
18:07:45 <acozine> o/
18:07:49 <andersson007_> #info The [vmware.vmware_rest](https://github.com/ansible-collections/vmware.vmware_rest) collection has been [included](https://github.com/ansible-community/ansible-build-data/pull/120) in Ansible 5. Thanks everyone!
18:07:49 <felixfontein> #chair acozine
18:07:49 <zodbot> Current chairs: acozine andersson007_ briantist cybette dmsimard felixfontein samccann
18:07:56 <cybette> #link https://groups.google.com/g/ansible-devel/c/UlHhhRDVUzc
18:08:03 <acozine> though I may disappear without warning - we're having thunderstorms and the internet is iffy
18:10:29 <felixfontein> i unfortunately forgot to prepare a PR for the proposed rules on dependencies (see https://github.com/ansible-community/community-topics/issues/94) we talked about last week...
18:12:00 <felixfontein> I have a quick question on https://github.com/ansible-community/community-topics/issues/98: does the general procedore we used for community.kubernetes -> kubernetes.core sound good? i.e. add new collection, replace old collection by set of deprecated redirects, then eventually remove old collection?
18:12:43 <felixfontein> if nobody objects or wants to have a discussion on this, I can create a voting issue for that proposal, then we'd have an 'official' procedure for this
18:12:46 <acozine> I think the kubernetes transfer went well, but then I don't really use that collection
18:13:06 <acozine> were there any Angry Issues opened about it at the time?
18:13:14 <samccann> it sounds like a logical progression to me
18:13:28 <felixfontein> I haven't heard anything... but I don't use them either
18:13:39 <acozine> I don't see anything linked to the community topic
18:13:40 <cybette> sounds good to me (also as a non-user) :P
18:15:00 <remindbot[m]> @cybette:ansible.im cyb-clock chimes every 15 minutes during the community meeting
18:15:03 <acozine> Professor Google doesn't find any rants immediately
18:15:10 <samccann> heh
18:15:25 <acozine> so either the transfer worked for most folks, or else they havne't noticed yet ;)
18:16:27 <felixfontein> maybe we'll notice once Ansible 6.0.0 gets released, since then community.kubernetes and its redirects is on longer there :)
18:16:30 <samccann> So theer were deprecation warnings in Ansible 5, and now it will be removed in Ansible 6, right?
18:16:38 <felixfontein> yes
18:16:50 <samccann> And does this show up in the Ansible 6 porting guide under.. erm.. breaking changes?
18:16:56 <felixfontein> there already was an announcement in Ansible 4.2 or so (but no deprecation warnings)
18:17:22 <samccann> so announcement in 4, deprecation warning in 5, removal in 6. Sounds reasonable
18:17:33 <felixfontein> samccann: removed_features, actually: https://github.com/ansible-community/ansible-build-data/blob/main/6/changelog.yaml#L16
18:18:26 <felixfontein> I guess for community.sap we could add the new collection in 6.x (whenever we decide on it), replace the old with a redirect placeholder in 7.0, and remove the old one in 8.0
18:19:11 <samccann> felixfontein: ok that shows up in the porting guide so +1 from docs;-)
18:19:35 <samccann> and yeah that sounds like a good plan
18:20:02 <felixfontein> sounds good, then I'll write this out in a bit more detail and create a vote for it
18:20:29 <andersson007_> +1
18:20:36 <felixfontein> since nobody started a discussion in the discussion issue I guess it won't have much (if any) opposition
18:20:43 <acozine> sounds great, thanks felixfontein
18:21:12 <felixfontein> so does anyone have topics to discuss about today?
18:21:32 <felixfontein> I have one (from #ansible-diversity), on naming of new features in collections
18:21:46 <felixfontein> if nobody has something else, let's talk a bit about that
18:22:15 <samccann> I had one question about Ansible5
18:22:48 <felixfontein> do you mean
18:22:48 <felixfontein> 18:40 < samccann> Since the question came up - will we have a 5.x release after 6 releases? We did that with Ansible 4 but i think that was a one-off and won't happen again?
18:22:51 <samccann> Ansible 4 had one release that happened at or after Ansible 5 released. Do we anticipate something similar for Ansible 6 (one release of 5 right at or after the 6 release?)
18:22:52 <felixfontein> ?
18:22:58 <samccann> yep that one
18:24:03 <dmsimard> yes
18:24:11 <dmsimard> I'm expecting a 6.10
18:24:21 <dmsimard> er, 5.10
18:24:30 <felixfontein> good question. right now the 6.0.0 release is planned for June 21st, and the next 4.x.0 releases would be 5.8.0 on May 17th, 5.9.0 on June 7th, and potentially 5.10.0 on 28th
18:24:48 <samccann> ok. We need to find a way to make this clearer in the docs
18:24:56 <felixfontein> so I guess 5.10.0 would be similar to the last 4.x.0 release we had which was one (or two?) week(s) after 5.0.0
18:25:10 <dmsimard> felixfontein: right
18:25:11 <samccann> because right now we say 5 goes EOL when 6 comes out and that's not quite true. There is a final release that is immediately EOL
18:25:38 <samccann> ok thanks
18:25:53 <felixfontein> dmsimard: why would you want to release 5.10.0?
18:25:57 <acozine> samccann: yeah, we changed our minds, but I guess we never updated the docs
18:26:12 <felixfontein> (I'm not opposed, I'm mainly trying to figure out a rule for when to do it and when not in the future :) )
18:26:40 <samccann> acozine: it's a bit hard to word in one tiny cell of a table to say 'EOLs shortly after Ansible 6' or something
18:26:50 <dmsimard> mostly because the release is close enough
18:26:58 <samccann> it's also not easy to understand 5.10 is EOL on arrival
18:26:59 <felixfontein> should we have something like "there should be at least one X.y.0 release on the day or after the day of the (X+1).0.0 release"? or only under more restrictive circumstances?
18:27:28 <felixfontein> dmsimard: how close is close enough? :)
18:27:31 <andersson007_> + the explanation why would be nice to have in the doc
18:27:36 <andersson007_> short
18:27:53 <felixfontein> or let me ask differently: if 5.9.0 would have been released a week before 6.0.0, would you also do a 5.10.0 release?
18:28:15 <samccann> I can probably add a note that says 'Each Ansible EOL release may include one final version to capture the most recent changes.'
18:28:28 <felixfontein> (right now it is two weeks before the 6.0.0 release)
18:29:15 <dmsimard> I don't personally need a 5.10, it's something we can do because it is inexpensive and not a lot of work to flush out any remaining updates given we don't have something like an LTS
18:29:24 <acozine> I'd flip that, samccann :  "Each Ansible EOL version may issue one final release at or shortly after the first release of the next version"
18:29:32 <dmsimard> "one last maintenance relsaa
18:29:57 <dmsimard> release* after a major release could be good enough
18:29:58 <felixfontein> acozine: +1
18:30:01 <remindbot[m]> @cybette:ansible.im cyb-clock chimes every 15 minutes during the community meeting
18:30:02 <acozine> heh, and each of us makes the note a little longer
18:30:07 <acozine> sorry samccann
18:30:14 <felixfontein> in 15 minutes it will be half a page
18:30:23 <acozine> heh, yep
18:30:27 <felixfontein> and in two weeks you can buy a book ;)
18:30:51 <cybette> lol
18:30:53 <acozine> that's documentation by committee for you
18:31:04 <felixfontein> :+1:
18:31:07 <samccann> hehe
18:31:22 <samccann> I'll spin up a pr with this shortly
18:31:26 <samccann> thanks
18:31:27 <briantist> I had expected there to be one more version...
18:31:55 <briantist> I think I asked about this a little while ago since I was busy trying to get one more collection release out beforehand
18:32:16 <briantist> or am I reading this wrong? sorry I got pulled away and am trying to skim and catch up
18:32:29 <felixfontein> briantist: I remember your question, but I didn't thought you meant a release after 6.0.0, just whether there are more 5.x.0 releases
18:32:34 <felixfontein> #chair briantist
18:32:34 <zodbot> Current chairs: acozine andersson007_ briantist cybette dmsimard felixfontein samccann
18:32:38 <briantist> oh no not after 6
18:32:53 <briantist> you're right
18:33:07 <felixfontein> 5.8.0 and 5.9.0 will definitely happen before 6.0.0
18:33:21 <briantist> works for my purposes 😅
18:33:23 <felixfontein> (in one resp. four weeks)
18:33:41 <felixfontein> for me too :)
18:34:17 <felixfontein> c.g 5.0.0 will get out next week. that will be interesting, it's the first huge collection (ok there are only two of them anyway ;) ) which stops using symlinks for flatmapping
18:34:39 <felixfontein> I'm wondering how the devel docs build will do the day after :)
18:36:00 <felixfontein> ok, let's talk about some word choices...
18:36:27 <felixfontein> #topic How to handle offensive terminology for new collection content
18:37:02 <andersson007_> how are we gonna detect the terminology?
18:37:21 <felixfontein> this got inspired by a PR in community.general which wants to extend the 'alternatives' module (a front-end for the update-alternatives tool, which originated in Debian but is now used in many distros) to support that tool's master/slave feature
18:38:00 <felixfontein> someone opened a ticket in that tool many years ago about changing the terminology to something less offensive, but that issue seems to have had no impact
18:38:20 <felixfontein> there are no aliases, or anything public on whether this might be changed eventually in the future (or not)
18:38:56 <felixfontein> I brought this up in #ansible-diversity yesterday, and the general consensus was that we should try to avoid adding new features using such names
18:39:03 <briantist> in #ansible-diversity I mentioned this GitHub App that can be used to detect that stuff andersson007_ : https://github.com/apps/in-solidarity
18:39:03 <briantist> It's in use on the  `lowlydba.sqlserver` : https://github.com/lowlydba/lowlydba.sqlserver
18:39:03 <briantist> (but this is just an addendum, felixfontein 's question is not about detection)
18:39:05 <felixfontein> and that we should discuss this in a larger venue, like this one (and the SC)
18:39:52 <andersson007_> briantist: cool, thanks! I wonder if we should add something about it in the collection requirements
18:40:10 <samccann> so the first level questoin - what do we do about this PR that tries to add 'slave' terminology to support what is called slave in the underlying tech.
18:40:17 <felixfontein> I personally find this a lot easier if the project the collection's part is about (in this case: the dpkg repository, which contains update-alternatives) already discussed that question and selected potential alternatives they want to use (maybe even just somewhen in the future), then we could use these terms (and potentially have the others as aliases)
18:40:28 <samccann> aka it's a new introduction of problematic language
18:40:28 <briantist> I think we should consider it, maybe enable it on `ansbile-collections` and let's see how it goes, probably takes some tweaking
18:40:35 <felixfontein> but in this case we have nothing, and we could only pick something ourselves with no guarantee that the upstream project will use the same term (or will ever use other terms)
18:41:01 <briantist> (but again I'm trying not to distract from felixfontein 's core question, which really is tricky)
18:41:14 <felixfontein> :)
18:41:16 <samccann> felixfontein: agreed. If/when they update terminology, we'd have to change the collection again to match imo
18:41:53 <acozine> felixfontein: do you have a link handy to the issue raised on the dpkg / update-alternatives repo?
18:41:54 <felixfontein> personally I (at least now) think that wouldn't be too bad, we'd just have another set of aliases. except that it could be even more confusing for users of our collection(s)
18:42:13 <felixfontein> #link https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=884368
18:42:27 <samccann> felixfontein: can we have more than one alias and can we leave them there forever?
18:42:30 <felixfontein> #link https://github.com/ansible-collections/community.general/pull/4654
18:43:04 <samccann> like we decide to call it 'secondary' and the downstream eventually calls it 'alternate'. Can we have both in the collection when that happens?
18:43:09 <felixfontein> samccann: we can have a list of aliases. there's no limit I think, though practically you shouldn't have too many
18:43:27 <acozine> wow, that's been sitting there since 2017
18:44:04 <felixfontein> maybe they had long discussions on this somewhere, but I don't know how the debian process works, so maybe it was just not at a public place, or I just didn't find it
18:44:10 <samccann> whatever we decide here sets a precedent for sure
18:45:00 <remindbot[m]> @cybette:ansible.im cyb-clock chimes every 15 minutes during the community meeting
18:46:04 <felixfontein> one potential complication is also that this tool ships outside of dpkg with other distributions, who might or might not consider dpkg upstream (maybe they made a copy a long time ago and independently developed it further)
18:47:05 <samccann> felixfontein: so potentially debian could rename it to one thing, and say fedora to something else (if fedora made a copy etc)?
18:47:38 <felixfontein> yes
18:47:55 <felixfontein> and both could end up with different names than us ;)
18:48:27 <briantist> 😩
18:48:46 <acozine> I would hope those communities would be open to the change if it were made upstream . . . though maybe we should do a little research - it's possible some of the downstream distros have already made a choice of different terminology and we could copy what they have done
18:49:35 <andersson007_> adding aliases should be an exception. I would suggest adding one alias and explaining matching things of the underlying service in argument's doc. primary / secondary for master / slave seems suitable imo
18:49:42 <felixfontein> does anyone know which fedora/RHEL packages provide update-alternatives?
18:50:02 <felixfontein> andersson007_: parent / child would also work
18:51:17 <andersson007_> yep
18:51:37 <andersson007_> maybe
18:52:12 <andersson007_> though i imagine people
18:53:14 <samccann> fwiw the RH suggested replacements are:... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/7c1c177d1cbc4663f052dcb3026b52977dd6f817)
18:53:22 <samccann> aka parent/child aren't on that list
18:53:23 <acozine> from my slight experience with update-alternatives, they aren't really parent/child relationships
18:53:44 <samccann> from https://source.redhat.com/aboutredhat/weareredhat/inclusion/diversity__inclusion_wiki/conscious_language_project_faq
18:54:34 <andersson007_> https://www.irccloud.com/pastebin/2Y48VtlY/
18:54:38 <acozine> it's been a while since I wrestled with update-alternatives, but my recollection is that the settings are "use Setting A and if you can't find A then use Setting B" or something along those lines - does anybody have more recent experience with the tool?
18:55:04 <felixfontein> samccann: I can't see that page without logging in
18:55:12 <acozine> it feels really weird that they would have used `master` and `slave` for that, though
18:55:48 <samccann> felixfontein: woopsie!
18:56:20 <felixfontein> https://github.com/fedora-sysv/chkconfig seem to be the sources for fedora's package that contains update-alternatives, and the alternatives.c header says "Copyright 1997-2009 Red Hat, Inc."
18:56:53 <cybette_> This is the Conscsious language project page on github https://github.com/conscious-lang
18:57:04 <samccann> ok this list is pointed to by RH as well - https://inclusivenaming.org/word-lists/tier-1/
18:57:46 <andersson007_> i've just found several cases of slave in c.network
18:58:25 <felixfontein> it looks a bit like the Fedora and Debian implementations are independent implementations
18:58:52 <felixfontein> andersson007_: which doesn't surprise me, since it's a common term in networking... c.g's netcli module also uses it
18:59:10 <samccann> andersson007_: yeah the first question is - what do we do for NEW items in collections we control.
18:59:24 <samccann> the 2nd question - what do we do for EXISTING items in collections we control
19:00:19 <andersson007_> samccann: do we have these questions in a corresponding community topic? sorry if i missed it
19:00:27 <felixfontein> npt yet
19:00:31 <felixfontein> s/npt/not/
19:00:54 <felixfontein> I guess I'll create one, since time is now up
19:00:55 <acozine> I'd vote for filing issues in the collections
19:01:11 <felixfontein> acozine: you mean for existing stuff, or for new stuff?
19:01:14 <samccann> yeah one thing we have to be clear about in the community topic is that this will be a Steering Committee decision.
19:01:28 <samccann> these are topics that can easily blow up in bad bad ways
19:01:51 <cybette> +1 samccann
19:01:51 <andersson007_> a though about a corresponding sanity doc check  has just come to mind
19:01:52 <felixfontein> indeed. one reason why I didn't do it yet ;)
19:03:16 <felixfontein> anyway, time's up!
19:03:27 <felixfontein> thanks all for the discussions today :)
19:03:30 <felixfontein> #endmeeting