16:00:05 <geppetto> #startmeeting fpc
16:00:05 <zodbot> Meeting started Thu Jun 15 16:00:05 2023 UTC.
16:00:05 <zodbot> This meeting is logged and archived in a public location.
16:00:05 <zodbot> The chair is geppetto. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
16:00:05 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:05 <zodbot> The meeting name has been set to 'fpc'
16:00:05 <geppetto> #meetingname fpc
16:00:05 <geppetto> #topic Roll Call
16:00:05 <zodbot> The meeting name has been set to 'fpc'
16:00:11 <GwynCieslasheher> hi
16:00:24 <geppetto> #chair GwynCieslasheher
16:00:24 <zodbot> Current chairs: GwynCieslasheher geppetto
16:03:28 <carlwgeorge> .hi
16:03:29 <zodbot> carlwgeorge: carlwgeorge 'Carl George' <carl@redhat.com>
16:03:39 <geppetto> #chair carlwgeorge
16:03:39 <zodbot> Current chairs: GwynCieslasheher carlwgeorge geppetto
16:10:18 <geppetto> #topic Open Floor
16:10:28 <geppetto> Anything anyone wants to talk about with only the three of us?
16:10:48 <carlwgeorge> yes
16:11:03 <decathorpe> .hi
16:11:03 <carlwgeorge> i forget the syntax but here's the direct link
16:11:04 <carlwgeorge> https://pagure.io/packaging-committee/issue/1283
16:11:06 <decathorpe> sorry for being late
16:11:06 <zodbot> decathorpe: decathorpe 'Fabio Valentini' <decathorpe@gmail.com>
16:11:13 <geppetto> #chair decathorpe
16:11:13 <zodbot> Current chairs: GwynCieslasheher carlwgeorge decathorpe geppetto
16:11:23 <carlwgeorge> decathorpe: no worries, as you can see we haven't really started yet
16:11:47 * decathorpe juggling commenting on different GitHub tickets talking with people in two channels here
16:12:14 <decathorpe> re/ ticket: I'm not really happy with what's being proposed here
16:12:21 <carlwgeorge> same
16:12:34 <GwynCieslasheher> What a mess.
16:12:35 <carlwgeorge> decathorpe and tibbs already commented on the issue and echo points i had already made when this was discussed in a bugzilla
16:12:53 <decathorpe> is it really so hard to just ... update the package in EPEL? I know that soname bumps etc. are frowned upon but wouldn't that be the easiest solution?
16:13:11 <carlwgeorge> my stance is that packages shouldn't get to ignore update policies just because they have a special name like -latest
16:13:25 <decathorpe> agreed
16:13:31 <geppetto> Yeh
16:13:52 <carlwgeorge> updating this in epel would have to follow the incompat process, which they don't want to do it seems
16:14:10 <geppetto> Yeh, seems like it doesn't really wan tto be in EPEL
16:14:12 <carlwgeorge> they also don't want to do the alternative of shipping compat packages because they want to do this often, apparently
16:14:42 <GwynCieslasheher> -latest is usually only latest for a little while...
16:15:02 <carlwgeorge> i have suggested to them to use copr probably a half dozen times between this issue and the bugzilla, and they haven't replied about why that won't work
16:15:18 <decathorpe> are there so many dependent packages in epel?
16:15:24 <carlwgeorge> a few
16:15:30 <geppetto> Yeh,  assume it's because it's much easieron the user side to be in EPEL
16:15:48 <carlwgeorge> certainly not so many that coordinating a rebuild of them in one update would be impossible
16:16:05 <geppetto> I sympathize … but
16:16:23 <carlwgeorge> peak behind the curtain, this started as a customer request to go into rhel, and it got turned down
16:16:40 <carlwgeorge> go figure, they don't want to maintain something that changes soname so often
16:16:43 <geppetto> Yeh, rhel loves incompat. updates ;)
16:16:52 <decathorpe> still, I think going through the process of properly updating the package would be the cleanest way to handle it long-term
16:17:22 <geppetto> I'd just keep suggesting they have an old package in EPEL and new one in a COPR
16:17:33 <carlwgeorge> which is the proper amount of friction to ensure it only happens when absolutely necessary, not willy nilly
16:17:54 <carlwgeorge> can someone else write that up for the issue?  they're clearly not listening when i say it.
16:18:02 <geppetto> :)
16:18:13 <decathorpe> either way, splitting dependent packages into "I use the API" and "I use the server" but they don't work together seems to worst solution
16:18:35 <carlwgeorge> indeed
16:18:41 <geppetto> Yeh, and we have no process for doing anything like that
16:19:24 <carlwgeorge> geppetto: are you active on internal slack yet, or still using gchat
16:20:09 <carlwgeorge> just going to send you the link to the bz.  i wish i could just make it public but it started with some mention of specific customers so i can't.
16:21:26 <geppetto> I'm only on sack for a while
16:21:30 <geppetto> slack, too
16:21:44 <geppetto> I commented/closed the FPC issue.
16:21:47 <carlwgeorge> thanks
16:22:16 <carlwgeorge> dovetailing from that, how do we feel about the current guidelines around package naming for multiples
16:22:19 <carlwgeorge> https://docs.fedoraproject.org/en-US/packaging-guidelines/Naming/#multiple
16:22:48 <carlwgeorge> it lightly discourages -latest without outright calling it "should not" or "must not"
16:24:13 <geppetto> Yeh, basename as latest seems best for Fedora … maybe for everything.
16:24:33 <decathorpe> yep, that's usually the way to do it
16:24:38 <carlwgeorge> it also presents the versioned suffix and the descriptive suffix as equal options, and i think there should be a strong preference for the former
16:25:08 <decathorpe> isn't this the reason we got rid of "compat-foo" packages? :)
16:25:43 * decathorpe remembers review request for compat-openssl, which luckily was then changed to openssl1.1 ...
16:26:29 <carlwgeorge> i can take a stab at rewriting this section as long as no one is fundamentally opposed to what i vaguely described here
16:26:30 <geppetto> Probably add some wording about how the descriptive suffix should be used if it's expected people will want to stay on that "branch" of the package for a long period of time
16:26:43 <geppetto> Although it looks a like mini-modules
16:27:47 <carlwgeorge> i'm trying to think of named suffix i would be ok with, and i'm struggling
16:28:00 <geppetto> I thought docker had one for a while
16:28:08 <carlwgeorge> i think -stable is almost as bad as -latest
16:28:17 <decathorpe> carlwgeorge: maybe "-lts"
16:28:38 <carlwgeorge> ok that's a good one, and probably a better example to use in the guidelines than -stable
16:28:44 <decathorpe> that has a specific meaning, and could be reused for other LTS version
16:29:00 <carlwgeorge> or firefox-esr
16:29:11 <decathorpe> that's one example that I was thinking of
16:29:26 <carlwgeorge> in rhel that changes versions but is always the designated esr (lts)
16:29:43 <carlwgeorge> ok i've got my project i think
16:29:53 <geppetto> 👍
16:30:09 <carlwgeorge> thanks for the discussion yall
16:30:36 <geppetto> no problem … anyone have anything else, if not I'll give you most of 30m back.
16:30:42 <decathorpe> I tagged another ticket with "meeting"
16:30:46 <GwynCieslasheher> Nope
16:31:22 <decathorpe> .fpc 1281
16:31:23 <zodbot> decathorpe: Issue #1281: RFC: Recommend using %bcond to control building docs and tests - packaging-committee - Pagure.io - https://pagure.io/packaging-committee/issue/1281
16:32:28 <decathorpe> probably easiest way forward would be to describe a current "best practice" for how to do it, but not do MUST USE THIS
16:32:53 <geppetto> Sounds good
16:33:41 <decathorpe> that way if RHEL maintainers want to add a "with docs" or "with tests" condition they can at least point at the guidelines and say "this is in line with best practices, please accept my humble PR"
16:34:53 <carlwgeorge> just yesterday i was in a discussion where someone asked me if rhel had guidelines about <thing>, and i told them for the most part rhel follows fedora guidelines (with exceptions that can be found in the epel guidelines)
16:35:34 * carlwgeorge wonders if it would be possible to backport %bcond to epel-rpm-macros since it's not in rhel currently
16:36:29 <decathorpe> ngompa indicated that it might be ... and I usually trust him on stuff like that ;)
16:36:46 <carlwgeorge> ah i see that now, should have finished reading the issue
16:36:47 <geppetto> Anything is possible =)
16:36:59 <carlwgeorge> possible != reasonable
16:37:17 <carlwgeorge> i should have said reasonable lol
16:39:56 <geppetto> On that note…
16:40:01 <geppetto> #endmeeting