16:00:09 <geppetto> #startmeeting fpc
16:00:09 <zodbot> Meeting started Thu Jul 27 16:00:09 2023 UTC.
16:00:09 <zodbot> This meeting is logged and archived in a public location.
16:00:09 <zodbot> The chair is geppetto. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
16:00:09 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:09 <zodbot> The meeting name has been set to 'fpc'
16:00:09 <geppetto> #meetingname fpc
16:00:09 <zodbot> The meeting name has been set to 'fpc'
16:00:09 <geppetto> #topic Roll Call
16:00:22 <decathorpe> .hi
16:00:23 <zodbot> decathorpe: Something blew up, please try again
16:00:26 <zodbot> decathorpe: An error has occurred and has been logged. Please contact this bot's administrator for more information.
16:00:27 <Eighth_Doctor> .hello ngompa
16:00:33 <zodbot> Eighth_Doctor: Something blew up, please try again
16:00:36 <zodbot> Eighth_Doctor: An error has occurred and has been logged. Please contact this bot's administrator for more information.
16:00:43 <geppetto> lol
16:00:44 <decathorpe> oh no
16:00:52 <Eighth_Doctor> oh dear
16:00:54 <geppetto> Need to regenerate
16:00:59 <geppetto> #chair decathorpe
16:00:59 <zodbot> Current chairs: decathorpe geppetto
16:01:04 <geppetto> #chair Eighth_Doctor
16:01:04 <zodbot> Current chairs: Eighth_Doctor decathorpe geppetto
16:01:33 <Eighth_Doctor> #chair Son_Goku
16:01:33 <zodbot> Current chairs: Eighth_Doctor Son_Goku decathorpe geppetto
16:01:40 <Son_Goku> I have a feeling we're in for a bad time
16:01:56 <decathorpe> zodbot asleep at the wheel
16:03:22 <decathorpe> or whatever the correct English idiom is for that
16:03:42 <Son_Goku> that's pretty much the the correct idiom
16:05:01 <decathorpe> good to know
16:05:08 <geppetto> #unchair Eighth_Doctor
16:05:08 <zodbot> Current chairs: Son_Goku decathorpe geppetto
16:05:29 <Son_Goku> ehh, my other self is there, I just don't feel like dealing with the desync issues right now
16:05:42 <Son_Goku> and with zodbot slightly on fire, ehhh
16:05:56 <decathorpe> probably doesn't matter much with only the three of us
16:06:01 <Son_Goku> I'm particularly cursed in this room
16:06:13 <Son_Goku> there are sync issues that only I experience in f-m-1
16:07:04 <geppetto> weird
16:07:22 <geppetto> Technology is too magic
16:08:01 <decathorpe> .hi
16:08:02 <zodbot> decathorpe: decathorpe 'Fabio Valentini' <decathorpe@gmail.com>
16:08:16 <decathorpe> looks like nirik fixed it in fm3
16:10:11 * nirik nods
16:10:45 <geppetto> #topic Open Floor
16:11:16 <geppetto> Well, there's still only 3 of us (Son_Goku doesn't count twice ;)
16:11:23 <geppetto> Anyone have anything to talk about?
16:11:34 <Son_Goku> .hello ngompa
16:11:35 <zodbot> Son_Goku: ngompa 'Neal Gompa' <ngompa13@gmail.com>
16:11:46 <geppetto> I think we can just close 1279
16:11:46 <Eighth_Doctor> .hello ngompa
16:11:47 <zodbot> Eighth_Doctor: ngompa 'Neal Gompa' <ngompa13@gmail.com>
16:11:50 <Son_Goku> nah, I got nothing
16:11:50 <geppetto> .fpc 1279
16:11:51 <zodbot> geppetto: Issue #1279: Packaging repo files to external repositories? - packaging-committee - Pagure.io - https://pagure.io/packaging-committee/issue/1279
16:12:06 <Son_Goku> oh this :/
16:12:12 <Son_Goku> I really don't know if we want to do this
16:12:27 <decathorpe> I'd rather have FESCo weigh in on this
16:12:32 <decathorpe> than make guidelines yet
16:13:15 <Son_Goku> yeah, this also gets into uncomfortable territory about third party sources
16:13:33 <decathorpe> it's not technically against the rules to package repo files, but ... I feel like it's against the spirit of the third party sources guidelines
16:13:42 <geppetto> Yeh, I didn't think it was allowed … but tibbs seemed to think it was fine as long as they are disabled.
16:14:00 <decathorpe> yes, it would be good to clarify that
16:14:05 <tibbs> I thought distro policy allowed.
16:14:09 <tibbs> Sorry for being late.
16:14:13 <geppetto> #chair tibbs
16:14:13 <zodbot> Current chairs: Son_Goku decathorpe geppetto tibbs
16:14:39 <decathorpe> technically, I think so.
16:14:39 <decathorpe> but I think it's complicated
16:16:45 <tibbs> I don't know, the policy seems pretty clear.
16:16:49 <tibbs> https://docs.fedoraproject.org/en-US/fesco/Third_Party_Repository_Policy/
16:17:55 <tibbs> The only fesco involvement is for repositories which are enabled by default.
16:18:18 <Son_Goku> I think the main concern I have is that the Intel repositories could break Fedora
16:18:30 <tibbs> Or am I misunderstanding?  This all goes way beyond packaging committee stuff.
16:18:38 <Son_Goku> we've generally tried to avoid shipping repo files that contain content that could lead to system breakage
16:18:58 <Son_Goku> at least with Workstation that's been the case
16:19:17 <tibbs> It's really just so easy to add external repositories that I don't see all that much benefit from packaging them, but....
16:20:02 <geppetto> tibbs: I think the big thing is the first key requirement
16:20:15 <Son_Goku> Workstation has generally had two requirements:
16:20:17 <tibbs> I think that our involvement shouldn't be much more than maybe adding a pointer to this policy.
16:20:19 <Son_Goku> 1. AppStream repodata
16:20:29 <Son_Goku> 2. Not obviously potentially break the distro
16:20:59 <decathorpe> yes, the policy is pretty clear that you need FESCo approval to ship something as enabled - but shipping something as diabled doesn't really provide anything on top of copying the repo file into /etc/yum.repos.d manually
16:21:19 <tibbs> Right
16:21:32 <Son_Goku> some historical tickets for Workstation:
16:21:37 <Son_Goku> #link https://pagure.io/fedora-workstation/issue/291
16:21:46 <Son_Goku> #link https://pagure.io/fedora-workstation/issue/283
16:21:56 <Son_Goku> #link https://pagure.io/fedora-workstation/issue/356
16:22:00 <tibbs> But I honestly don't think we have much to say here.  It's just that someone wants to package something, so the question is going to come to us.
16:22:05 <geppetto> Again, the first key requirement is that anything on the hosted repo (even if disabled) not break any of the legal requirements for Fedora packages.
16:22:14 <Son_Goku> right
16:22:24 <tibbs> We could maybe help with that by pointing to the policy somewhere, but none of the other questions seem to be ours to answer.
16:23:02 * geppetto nods
16:23:27 <Son_Goku> Heck, I would have added Slack's yum repo to fedora if it made sense and worked in a discoverable fashion
16:23:47 <tibbs> I think the distro policy is pretty clear here, and it's also clear that questions about that policy should go to FESCo.
16:23:50 <decathorpe> https://docs.fedoraproject.org/en-US/fesco/Third_Party_Repository_Policy/#_key_requirements_for_third_party_repositories
16:24:14 <decathorpe> these "key requirements" apply whether the repo is shipped enabled or not, so I really think FPC has not much to add here
16:24:25 <tibbs> So the remaining question, I guess, is whether we need to add something to the guidelines, and if so, what.
16:24:32 <Son_Goku> yeah, and it's probably worth asking FESCo to add some technical color
16:24:48 <Son_Goku> because it seems Workstation's rules are "unwritten", which is probably bad
16:25:11 <decathorpe> "put files into the only place where they work"
16:25:17 <decathorpe> but then ... I think that's usually implied
16:26:26 <decathorpe> the "we need appdata for the repositories" seem to be https://docs.fedoraproject.org/en-US/fesco/Third_Party_Repository_Policy/#_software_labeling_and_metadata
16:26:33 <decathorpe> (though the wording is unnecessarily obtuse IMO)
16:27:01 <Son_Goku> I think it's worth asking for it to be made clearer
16:28:15 <tibbs> I guess anyone is welcome to do that.
16:28:34 <decathorpe> either way, shipping third party repositories already does require WG or FESCo approval, whether enabled by default or not
16:28:41 <decathorpe> so I don't think we need to say anything as FPC here
16:28:56 <geppetto> 👍
16:29:00 <Son_Goku> we can at least punt it to FESCo :)
16:29:13 <Son_Goku> some of the questions we've discussed in here belong there anyway
16:29:26 <geppetto> .fpc 1276
16:29:29 <zodbot> geppetto: Issue #1276: Should /sbin/ldconfig be /usr/sbin/ldconfig? - packaging-committee - Pagure.io - https://pagure.io/packaging-committee/issue/1276
16:29:30 <geppetto> This one annoys me
16:30:06 <tibbs> I have sort of lost track of just what the new dnf is going to break.
16:30:14 <decathorpe> DNF5 annoys me to be honest, but that's a separate topic
16:30:31 <tibbs> But given the recent announcement that it's being put off until f40 or f41, there's certainly no urgency.
16:30:53 <decathorpe> I don't understand what's going to break anymore
16:31:00 <tibbs> Really the question is whether we need to react to https://bugzilla.redhat.com/show_bug.cgi?id=2180842 in any way.
16:31:02 <geppetto> Makybe they'll fix it given the extra time
16:31:45 <tibbs> The fundamental issue was that they were going to stop downloading the whole file lists and break file dependencies outside of three directories.
16:32:14 <tibbs> Months ago I said that this kind of breakage really needs to go through the change process, but I was ignored.
16:32:24 <geppetto> Do the guidelines now say no package can make a file requirement?
16:32:47 <Son_Goku> nope
16:32:48 <tibbs> Well presumably a guidelines change would be part of the change process that they weren't going to do.
16:33:06 <tibbs> We discourage certain file dependencies but do not ban them.
16:33:12 <tibbs> And the list of directories seems to be different.
16:33:47 <tibbs> Oh, no, it's not.
16:33:51 <tibbs> https://docs.fedoraproject.org/en-US/packaging-guidelines/#_file_and_directory_dependencies
16:34:09 <tibbs> We say "SHOULD NOT They also MAY depend on paths listed in an explicit Provides:. They SHOULD NOT include dependencies on other paths as that requires additional repository metadata to be downloaded."
16:34:58 <tibbs> I pasted an extra two words at the beginning there, oops.
16:35:29 <tibbs> Anyway, we're fairly in sync except that we would have to change "SHOULD NOT" to "MUST" because otherwise the dependencies would simply not work.
16:36:02 <geppetto> Yeh, so not sure what to say "Please don't break the package installer to not follow the guidelines, or make a change request" is tempting, but will probably cause minor fires
16:36:28 <decathorpe> I still think DNF5 should fix this - they originally planned to download file lists on demand, but that's apparently on the back burner
16:37:05 <tibbs> Given the delay, they will have more time to get to back burner things.
16:37:16 <geppetto> Yeh
16:37:21 <decathorpe> I hope so
16:37:41 <tibbs> I think the only remaining question is whether there is an issue with using /bin/ldconfig in our examples.
16:38:10 <Son_Goku> there should not be
16:38:12 <tibbs> Someone says that it's confusing because it creates dependencies on /bin/ldconfig which contradicts this guideline.
16:38:27 <Son_Goku> do we even have a /bin/ldconfig?
16:38:32 <Son_Goku> I thought we only have /sbin/ldconfig
16:38:35 <geppetto> I'd vote no atm. … that's like /bin/sh … it should still be used even if it's not quite true.
16:38:44 <tibbs> Sorry, of course I mean /sbin/ldconfig.
16:39:09 <Son_Goku> I'm with geppetto on this one
16:39:11 <geppetto> I thought there was an explicit provides for it anyway.
16:39:20 <tibbs> Since glibc added an explicit Provides: /sbin/ldconfig, I think the issue went away and only the confusion remains.
16:39:33 * geppetto nods
16:39:40 <tibbs> It originally didn't have one and that's why so many packages were set to be broken.
16:39:47 <geppetto> I guess just close it and hope everything gets better
16:39:52 <Son_Goku> yeah
16:39:55 <Son_Goku> this is a non-issue
16:40:14 <tibbs> Only one of the packages that would be broken has yet to be fixed.
16:40:23 <tibbs> https://bugzilla.redhat.com/show_bug.cgi?id=1731700 if anyone cares.
16:40:39 <geppetto> Okay, I closed it.
16:40:46 <decathorpe> +1 for closing
16:40:54 <tibbs> Yeah, I think we're good.
16:40:55 <geppetto> Anything else, or do you want 20ish minutes back
16:41:21 <tibbs> I was trying to clean some stuff up last night.
16:41:31 <geppetto> tibbs++
16:41:31 <zodbot> geppetto: Karma for tibbs changed to 2 (for the release cycle f38):  https://badges.fedoraproject.org/tags/cookie/any
16:41:39 <Son_Goku> tibbs++
16:41:42 <zodbot> Son_Goku: Karma for tibbs changed to 3 (for the release cycle f38):  https://badges.fedoraproject.org/tags/cookie/any
16:41:54 <tibbs> We really do have too many open issues.
16:42:13 <tibbs> Now pagure is being super slow....
16:43:01 <tibbs> There's certainly nothing urgent.  When pagure comes back I will just go through a few more tickets and PRs and see if I can clean anything up.
16:43:24 <decathorpe> apparently there's intermittent network issues somewhere in the sphere (radius to be determined) around the pagure datacenter
16:44:16 <tibbs> Ah, well, not a big deal.  Anything I was working on is by definition not urgent anyway.
16:44:26 * geppetto gives some karma to the network engineer that's trying to fix that.
16:44:50 <geppetto> Ok, enjoy the rest of your day
16:45:04 <geppetto> #endmeeting