2024-11-13 18:00:56 <@tdawson:fedora.im> !startmeeting EPEL (2024-11-13)
2024-11-13 18:00:57 <@meetbot:fedora.im> Meeting started at 2024-11-13 18:00:56 UTC
2024-11-13 18:00:57 <@meetbot:fedora.im> The Meeting name is 'EPEL (2024-11-13)'
2024-11-13 18:01:04 <@tdawson:fedora.im> !meetingname epel
2024-11-13 18:01:04 <@tdawson:fedora.im> !topic aloha
2024-11-13 18:01:05 <@meetbot:fedora.im> The Meeting Name is now epel
2024-11-13 18:01:07 <@davide:cavalca.name> !hi
2024-11-13 18:01:09 <@zodbot:fedora.im> Davide Cavalca (dcavalca) - he / him / his
2024-11-13 18:01:10 <@conan_kudo:matrix.org> !hi
2024-11-13 18:01:12 <@zodbot:fedora.im> Neal Gompa (ngompa) - he / him / his
2024-11-13 18:01:17 <@xavierb:bachelot.org> hi
2024-11-13 18:01:58 <@tdawson:fedora.im> Hi Davide Cavalca Conan Kudo and Xavier bachelot 
2024-11-13 18:02:01 <@salimma:fedora.im> !hi
2024-11-13 18:02:02 <@zodbot:fedora.im> Michel Lind (salimma) - he / him / his
2024-11-13 18:02:09 <@carlwgeorge:matrix.org> !hi
2024-11-13 18:02:10 <@zodbot:fedora.im> Carl George (carlwgeorge) - he / him / his
2024-11-13 18:02:23 <@tdawson:fedora.im> Hi Michel Lind 🎩 UTC-6 and Carl George 
2024-11-13 18:02:30 <@nhanlon:beeper.com> !hi
2024-11-13 18:02:32 <@zodbot:fedora.im> Neil Hanlon (neil) - he / him / his
2024-11-13 18:02:33 <@nirik:matrix.scrye.com> morning
2024-11-13 18:02:37 <@dherrera:fedora.im> !hi
2024-11-13 18:02:38 <@zodbot:fedora.im> Diego Herrera (dherrera) - he / him / his
2024-11-13 18:02:47 <@tdawson:fedora.im> Morning nirik 
2024-11-13 18:03:03 <@tdawson:fedora.im> Hi Neil Hanlon and Diego Herrera 
2024-11-13 18:03:17 <@zodbot:fedora.im> neil gave a cookie to ngompa. They now have 150 cookies, 3 of which were obtained in the Fedora 41 release cycle
2024-11-13 18:03:22 <@nhanlon:beeper.com> take a cookie it might help
2024-11-13 18:03:48 <@conan_kudo:matrix.org> mmm
2024-11-13 18:05:48 <@tdawson:fedora.im> !topic EPEL Issues https://pagure.io/epel/issues
2024-11-13 18:05:48 <@tdawson:fedora.im> !link https://pagure.io/epel/issues?tags=meeting&status=Open
2024-11-13 18:06:17 <@tdawson:fedora.im> So, the only issue marked with meeting there deals with epel10 .... so .... I'm going to change the topic before we move to it.
2024-11-13 18:06:24 <@tdawson:fedora.im> !topic EPEL 10
2024-11-13 18:06:29 <@tdawson:fedora.im> !epel 300
2024-11-13 18:06:31 <@zodbot:fedora.im> ● **Opened:** 3 weeks ago by carlwgeorge
2024-11-13 18:06:31 <@zodbot:fedora.im> ● **Assignee:** carlwgeorge
2024-11-13 18:06:31 <@zodbot:fedora.im> ● **Last Updated:** a day ago
2024-11-13 18:06:31 <@zodbot:fedora.im> **epel #300** (https://pagure.io/epel/issue/300):**EPEL 10 launch**
2024-11-13 18:06:31 <@zodbot:fedora.im> 
2024-11-13 18:07:12 <@tdawson:fedora.im> So much is done .... ya!!!
2024-11-13 18:07:59 <@tdawson:fedora.im> It looks like the only things left are documentation.
2024-11-13 18:08:03 <@carlwgeorge:matrix.org> yes i feel much better on the epel side since we have the testing repo and bodhi composes working
2024-11-13 18:08:19 <@carlwgeorge:matrix.org> docs are my focus this week as soon as i put out enough fires
2024-11-13 18:08:39 <@nirik:matrix.scrye.com> I have a dumb question (that lots of other people will ask, so I will go ahead): will epel10 work fine with rhel10 beta? or we expect it mostly to, but have a few things, or ...
2024-11-13 18:08:41 <@carlwgeorge:matrix.org> that and writing the literal announcement post
2024-11-13 18:09:03 <@davide:cavalca.name> does rhel10 beta exist already?
2024-11-13 18:09:18 <@salimma:fedora.im> apparently it launched yesterday ish?
2024-11-13 18:09:25 <@salimma:fedora.im> I have not seen an announcement yet though
2024-11-13 18:09:54 <@carlwgeorge:matrix.org> i expect it mostly to, but i would not be surprised if some exceptions are found.  the beta is like a point release before zero, and we never explicitly targeted it, so it's certainly possible for there to be library problems
2024-11-13 18:10:05 <@carlwgeorge:matrix.org> in fact, i'm willing to bet that late perl rebase will cause issues
2024-11-13 18:10:57 <@nirik:matrix.scrye.com> right, so we might want to make sure and have docs/notes about that
2024-11-13 18:11:08 <@nirik:matrix.scrye.com> The release notes are up, but yeah, not sure there was a press release yet
2024-11-13 18:11:28 <@carlwgeorge:matrix.org> i'm also not finding a press release to 10 beta, just 9.5
2024-11-13 18:11:45 <@salimma:fedora.im> I suspect some -epel packages with a tight constraint on the original version-release might break
2024-11-13 18:11:55 <@tdawson:fedora.im> I haven't seen an announcement for it, but I was able to download it from the normal place .... so the beta is at least there.   Hmm ... I could easily do a test to see what installs and what doesn't.
2024-11-13 18:12:01 <@salimma:fedora.im> oh, 9.5 is out too? I might have some Rust packages that are unblocked now
2024-11-13 18:12:33 <@salimma:fedora.im> fwiw for tight constraints, Carl George and I were discussing this yesterday and just depending on the original version might be better, between c10s, rhel beta, and rebuilds
2024-11-13 18:12:37 <@nirik:matrix.scrye.com> yep. 9.5 is also out. There is a press release for that. ;)
2024-11-13 18:13:33 <@yselkowitz:fedora.im> how soon until epel9-build includes 9.5?
2024-11-13 18:13:47 <@carlwgeorge:matrix.org> i think it already does
2024-11-13 18:14:02 <@nirik:matrix.scrye.com> yeah, should already have it.
2024-11-13 18:14:30 <@carlwgeorge:matrix.org> i saw someone mention a build failure that looks to be a result of them building in the middle of the sync, they had a different selinux-policy between different arches, so i check and suggested they just build again
2024-11-13 18:15:18 <@carlwgeorge:matrix.org> my recommendation for -epel packages is to only require the matching version, not the matching release.  it's way too difficult to keep that in sync, and if you make a packaging mistake you can't bump the release and still match.
2024-11-13 18:15:20 <@nirik:matrix.scrye.com> Thats pretty odd. But ok
2024-11-13 18:15:49 <@conan_kudo:matrix.org> I usually set it to be greater than or equal to original evr
2024-11-13 18:15:54 <@carlwgeorge:matrix.org> i.e. `Requires: foo-%{version}`, not `Requires: foo-%{version}%{release}`
2024-11-13 18:16:20 <@conan_kudo:matrix.org> but yeah `Requires: foo = %{version}` can work too
2024-11-13 18:16:20 <@carlwgeorge:matrix.org> i.e. `Requires: foo-%{version}`, not `Requires: foo-%{version}-%{release}`
2024-11-13 18:16:41 <@conan_kudo:matrix.org> In this sense, I'd pair it with a `Conflicts: foo < %{version}-%{origrelease}`
2024-11-13 18:17:17 <@nirik:matrix.scrye.com> well, in general you should let rpm handle Requires...
2024-11-13 18:17:28 <@salimma:fedora.im> yeah I was thinking of using >= but that means you also have to add a < (version+1)
2024-11-13 18:17:52 <@carlwgeorge:matrix.org> it's more complexity than it's worth imo, just require a matching version
2024-11-13 18:18:10 <@salimma:fedora.im> RPM behaves in weird ways sometimes so I don't want to push my luck. Yesterday TIL that an unversioned provides is treated as if you're providing version-release
2024-11-13 18:18:16 <@carlwgeorge:matrix.org> ideally these are short term packages anyways and the real build is on the way to crb, but we all know that can get stalled out or even denied
2024-11-13 18:18:18 <@tdawson:fedora.im> So this doesn't turn into a "rpm packaging" meeting ... do we have any other EPEL 10 stuff?
2024-11-13 18:18:32 <@carlwgeorge:matrix.org> i have one more thing
2024-11-13 18:18:37 <@tdawson:fedora.im> Yep
2024-11-13 18:18:39 <@salimma:fedora.im> e.g. look at the Fedora/c10s/c9s packaging for iptables where iptables-nft and iptables-legacy both provide iptables unversioned, but match if you say you need iptables = version-release
2024-11-13 18:18:48 <@salimma:fedora.im> sorry troy
2024-11-13 18:19:53 <@carlwgeorge:matrix.org> we originally were targeting 2024-11-19 for the announcement, but at this point my guess is we're going to slip it a week later to 2024-11-26.  things aren't ready on the centos side, and it's beneficial for us to wait for them.
2024-11-13 18:20:14 <@xavierb:bachelot.org> Troy Dawson anything that can be done to help with KDE ?
2024-11-13 18:20:25 <@davide:cavalca.name> 2024-11-26 is thanksgiving week fyi
2024-11-13 18:20:34 <@davide:cavalca.name> I'd expect a lot of people will be out and might miss it
2024-11-13 18:20:56 <@salimma:fedora.im> yeah. maybe delay to December?
2024-11-13 18:20:58 <@carlwgeorge:matrix.org> yeah i'm aware, we could alternatively slip into december, but many folks take that whole month off
2024-11-13 18:21:13 <@carlwgeorge:matrix.org> we have the red hat shut down and also people trying to use up their pto
2024-11-13 18:21:16 <@conan_kudo:matrix.org> yeah, I don't think we should slip that much
2024-11-13 18:21:37 <@davide:cavalca.name> yeah, you're right, probably not worth delaying further
2024-11-13 18:21:48 <@carlwgeorge:matrix.org> best case scenario is everything centos falls into place this week and we announce next tuesday, but that seems unlikely
2024-11-13 18:21:48 <@salimma:fedora.im> is 11-21/22 late enough?
2024-11-13 18:21:54 <@salimma:fedora.im> the Thu/Fri before Thanksgiving week
2024-11-13 18:22:31 <@tdawson:fedora.im> It might not look like it, but we're actually pretty good right now.  Just finished and updated plasma-workspace, and maliit-keyboard's depenendency has built, but still in a side-tage.
2024-11-13 18:22:49 <@carlwgeorge:matrix.org> that could work.  originally i was looking at tuesdays semi-arbitrarily, but it doesn't have to be on a specific day of the week
2024-11-13 18:22:54 <@tdawson:fedora.im> I think it's very possible ... depending on what you care about. ;)
2024-11-13 18:23:25 <@nhanlon:beeper.com> that's awesome news. do let us know if you need bodhi help or anything!
2024-11-13 18:23:46 <@carlwgeorge:matrix.org> besides the secureboot issue me and troy were talking about separately, i talked to shaun and he said he's not ready to announce either.  i need to touch base with him more about how much more time he needs.
2024-11-13 18:24:10 <@tdawson:fedora.im> Ahh ... ok.  I hadn't heard from him.
2024-11-13 18:24:33 <@carlwgeorge:matrix.org> either way, aside from docs and the announcement text, we're ready on the epel side, just want to align with centos.  that's the main relevant part for this meeting, we can move on i think.
2024-11-13 18:24:39 <@tdawson:fedora.im> I should say in short.  They stuff I'm in charge of ... it will be ready by the original date. :)
2024-11-13 18:25:06 <@tdawson:fedora.im> But that doesn't mean it's totally ready.
2024-11-13 18:25:37 <@tdawson:fedora.im> I should say in short.  They CentOS Stream 9 stuff I'm in charge of ... it will be ready by the original date. :)
2024-11-13 18:25:51 <@tdawson:fedora.im> I should say in short.  They CentOS Stream 10 stuff I'm in charge of ... it will be ready by the original date. :)
2024-11-13 18:26:00 <@nirik:matrix.scrye.com> I think at this point, just do the best you can and if we need to change it we can look at that next week?
2024-11-13 18:26:06 <@tdawson:fedora.im> I'll stop typing now ... I'm just digging myself deaper.
2024-11-13 18:26:45 <@tdawson:fedora.im> I agree with you both (nirik and carl)
2024-11-13 18:27:01 <@tdawson:fedora.im> Any other EPEL 10 stuff before we move on?
2024-11-13 18:27:20 <@salimma:fedora.im> I have some sort of epel 10 / old business update
2024-11-13 18:27:23 <@carlwgeorge:matrix.org> check if you have any fti bugs from me 😀
2024-11-13 18:27:30 <@salimma:fedora.im> but that can be old biz / open floor, not sure
2024-11-13 18:27:41 <@carlwgeorge:matrix.org> i think i have them filed for everything except kde related stuff, as requested
2024-11-13 18:27:49 <@tdawson:fedora.im> Michel Lind 🎩 UTC-6: Let's move to old business, then go for it.
2024-11-13 18:28:03 <@tdawson:fedora.im> Carl George: Thank you.
2024-11-13 18:28:26 <@conan_kudo:matrix.org> I've got a question for open floor as well
2024-11-13 18:28:28 <@smooge:fedora.im> hey I am here
2024-11-13 18:28:36 <@salimma:fedora.im> woah!
2024-11-13 18:28:36 <@tdawson:fedora.im> !topic Old Business
2024-11-13 18:28:48 <@smooge:fedora.im> nearly missed it again.. but Old Business is what I am
2024-11-13 18:28:55 <@tdawson:fedora.im> Hi Stephen J Smoogen  ... I see you came just in time.
2024-11-13 18:28:59 <@zodbot:fedora.im> salimma gave a cookie to smooge. They now have 263 cookies, 1 of which were obtained in the Fedora 41 release cycle
2024-11-13 18:29:09 <@zodbot:fedora.im> jrichardson gave a cookie to smooge. They now have 264 cookies, 2 of which were obtained in the Fedora 41 release cycle
2024-11-13 18:29:20 <@tdawson:fedora.im> Michel Lind 🎩 UTC-6: you had some old business ... go for it.
2024-11-13 18:29:54 <@salimma:fedora.im> 
2024-11-13 18:29:54 <@salimma:fedora.im> And a similar guide for releng on how to grant access, since sometimes they are inconsistent
2024-11-13 18:29:54 <@salimma:fedora.im> 
2024-11-13 18:29:54 <@salimma:fedora.im> I got snowed under with work last week, but I'm finally working on the documentation PR to provide better guidance for package maintainers to grant collaborator access
2024-11-13 18:30:24 <@tdawson:fedora.im> Michel Lind 🎩 UTC-6: That is great.
2024-11-13 18:30:33 <@salimma:fedora.im> and today it came up that some maintainers don't know how to deal with branch requests, so that PR will probably update the templates too to point to the new section I'm writing "dealing with requests for package maintainers"
2024-11-13 18:30:42 <@salimma:fedora.im> so ... expect a PR, and a mail to the mailing list, later today
2024-11-13 18:30:53 <@salimma:fedora.im> I yield to Nirik
2024-11-13 18:31:37 <@carlwgeorge:matrix.org> gonna be a busy open floor today (i have a thing too)
2024-11-13 18:32:45 <@nirik:matrix.scrye.com> So, first (smaller) one. Just FYI, https://bugzilla.redhat.com/show_bug.cgi?id=2308951 is requesting ansible for epel10... but... I don't think it's a good idea and neither does my comaintainer. So, instead folks should get specific collections they need in. I hope to look over my list of those and branch some this weekend if things are not on fire.
2024-11-13 18:32:51 <@tdawson:fedora.im> Thank you Michel Lind 🎩 UTC-6 
2024-11-13 18:33:29 <@tdawson:fedora.im> !topic General Issues / Open Floor
2024-11-13 18:33:37 <@salimma:fedora.im> ah. yeah personally I'm just missing one resource (flatpak) so happy to help just getting the smallest collection containing it branched
2024-11-13 18:34:10 <@tdawson:fedora.im> nirik: continue, I was just changing topics officially.
2024-11-13 18:34:24 <@carlwgeorge:matrix.org> ah this is that whole "ansible-core is ansible, ansible is a meta package of collections" thing isn't it
2024-11-13 18:34:38 <@zodbot:fedora.im> ngompa gave a cookie to smooge. They now have 265 cookies, 3 of which were obtained in the Fedora 41 release cycle
2024-11-13 18:35:04 <@nirik:matrix.scrye.com> second one is a bit complex to explain, but let me try: In the past when we had a epel package that other epel packages build against and it got moved into rhel, we needed to unblock it in the build tag in order for the rhel one to get used. But... this is a pretty rare corner case and I don't think we should just blindly do it all the time anymore... would folks be ok with making this a manual thing that the sig notices and files tickets on if it happens at rhel minor releases and we can drop automation doing it all the time?
2024-11-13 18:35:29 <@nirik:matrix.scrye.com> Carl George: right. 'ansible' is like 100+ collections... ansible-core is the engine part.
2024-11-13 18:36:05 <@conan_kudo:matrix.org> nirik: I'm not sure I'd want to make this manual now that `epel10` tracks centos stream
2024-11-13 18:36:17 <@carlwgeorge:matrix.org> a lovely naming scheme when `/usr/bin/ansible` is in ansible-core (not your fault, just often a cause of the confusion)
2024-11-13 18:36:28 <@zodbot:fedora.im> dherrera gave a cookie to smooge. They now have 266 cookies, 4 of which were obtained in the Fedora 41 release cycle
2024-11-13 18:36:34 <@tdawson:fedora.im> nirik: Are you asking to make it part of the whole "new release, time to expire the packages that are in RHEL" workflow?
2024-11-13 18:36:43 <@smooge:fedora.im> ok lets keep talking about the first item and then move to the second?
2024-11-13 18:36:53 <@nirik:matrix.scrye.com> yes, but only in the case where they are buildrequires for other epel packages.
2024-11-13 18:37:16 <@nirik:matrix.scrye.com> sorry, yeah... the ansible thing was a FYI mostly, unless someone has some further thought on it
2024-11-13 18:37:31 <@yselkowitz:fedora.im> wouldn't that be most of the time though?
2024-11-13 18:37:37 <@carlwgeorge:matrix.org> nope, makes sense, thanks for the heads up
2024-11-13 18:37:44 <@tdawson:fedora.im> I honestly didn't even know that was something that needed to be done.  I thought they were all just swirlled in a pot and the highest NVR won.
2024-11-13 18:37:48 <@nirik:matrix.scrye.com> Conan Kudo: thats a good point... I'm not sure how much this is happening tho.
2024-11-13 18:38:01 <@smooge:fedora.im> I think the 'no make ansible for EPEL-10' makes sense
2024-11-13 18:38:22 <@nirik:matrix.scrye.com> I could find 0 cases in rhel9. It used to happen back in the 6/7 days a fair bit.
2024-11-13 18:38:27 <@smooge:fedora.im> Troy Dawson: the problem in the past is that the EPEL one generally got the higher NVR
2024-11-13 18:38:37 <@nirik:matrix.scrye.com> koji cares nothing for nvrs
2024-11-13 18:38:59 <@nirik:matrix.scrye.com> package foobar is blocked. Therefore you should not gather / merge any packages from the foobar source package
2024-11-13 18:39:20 <@smooge:fedora.im> well there was something in koji which would weigh the newer EPEL package over the older but official RHEL one..
2024-11-13 18:39:28 <@tdawson:fedora.im> But, when we remove the epel2rhel packages at the beginning of the release, shouldn't that be enough.   It will no longer be available.
2024-11-13 18:39:28 <@carlwgeorge:matrix.org> for builds, i know nvrs are ignored and most recently tagged wins, but how does that comparison work with a build versus an external repo?
2024-11-13 18:39:44 <@conan_kudo:matrix.org> Stephen J Smoogen: iirc that's only if they don't have matching source package names
2024-11-13 18:40:03 <@conan_kudo:matrix.org> if they have matching source package names, koji filters them out unequivocally
2024-11-13 18:40:12 <@nirik:matrix.scrye.com> foobar is blocked in epel10 say, the epel10-base tag with the external repos is inheriting from that and the epel10-build tag inherits from that.
2024-11-13 18:41:02 <@nirik:matrix.scrye.com> if we unblock it in epel10-build it can come in from rhel, but not epel...
2024-11-13 18:41:15 <@nirik:matrix.scrye.com> I guess we could just always keep doing it. Seems a waste.
2024-11-13 18:41:37 <@tdawson:fedora.im> Ohhh ... which reminds me ... RHEL 9.5 just came out, we need to do an epel snapshot ... is the ticket created?
2024-11-13 18:41:45 <@carlwgeorge:matrix.org> i be ok dropping the automation to see what that looks like without it
2024-11-13 18:42:24 <@carlwgeorge:matrix.org> https://pagure.io/releng/issue/12456
2024-11-13 18:42:26 <@nirik:matrix.scrye.com> I would too, but it might cause some confusion when people hit it... as long as we are aware
2024-11-13 18:42:36 <@carlwgeorge:matrix.org> i'd be ok dropping the automation to see what that looks like without it
2024-11-13 18:43:03 <@zodbot:fedora.im> tdawson gave a cookie to carlwgeorge. They now have 69 cookies, 1 of which were obtained in the Fedora 41 release cycle
2024-11-13 18:43:35 <@smooge:fedora.im> i don't have a horse in this race. I remember that we had a lot of problems with EL5/6/7 BUT everything from koji/mock/how repos are pulled in have changed in many different ways 
2024-11-13 18:43:55 <@nirik:matrix.scrye.com> https://pagure.io/releng/issue/12440 has more details if folks want to read later and chime in.
2024-11-13 18:43:57 <@carlwgeorge:matrix.org> exactly, let's scream test it to make sure things work the way we think they do
2024-11-13 18:44:02 <@smooge:fedora.im> i think simplifying releng's life is a better idea 
2024-11-13 18:44:08 <@tdawson:fedora.im> nirik: Can you write this up in an email, or issue.   ..... ha ... like that. :)
2024-11-13 18:44:18 <@carlwgeorge:matrix.org> considering the failure mode is just "my build failed", it's not bad at all
2024-11-13 18:44:59 <@tdawson:fedora.im> Stephen J Smoogen: I think you said you had something for open floor?
2024-11-13 18:45:13 <@smooge:fedora.im> no I just said I was here for it
2024-11-13 18:45:19 <@tdawson:fedora.im> Ahh ... ok.
2024-11-13 18:45:39 <@smooge:fedora.im> my apolgoies for the confusion
2024-11-13 18:45:47 <@tdawson:fedora.im> Not a problem.
2024-11-13 18:46:02 <@tdawson:fedora.im> Carl George: I think you were the other one that said they had something for open floor.
2024-11-13 18:46:51 <@tdawson:fedora.im> I think ... sometimes when there is alot of conversation, I'm just looking at the color of the name.
2024-11-13 18:47:05 <@carlwgeorge:matrix.org> yup, i wanted to bring up the epel-packagers-sig.  i expect this will be controversial, but bear with me to make my point so we can have a discussion around it.
2024-11-13 18:47:34 <@tdawson:fedora.im> Go for it.
2024-11-13 18:47:49 <@zodbot:fedora.im> salimma gave a cookie to carlwgeorge. They now have 70 cookies, 2 of which were obtained in the Fedora 41 release cycle
2024-11-13 18:48:16 <@carlwgeorge:matrix.org> the sig has grown to have permissions on a lot of packages.  so many that the notifications are overwhelming, including to the point of some people wanting to leave the sig.  i think the sig is also being abused for branch permissions with no intention of ongoing maintenance.
2024-11-13 18:48:49 <@carlwgeorge:matrix.org> i think the recent mailing list post was related to this https://lists.fedoraproject.org/archives/list/epel-devel@lists.fedoraproject.org/thread/OGUFKOWNUD5TXKGDTFMA23GW4ORX6QMY/
2024-11-13 18:50:03 <@carlwgeorge:matrix.org> there was also a recent epel10 branch request where the requestor wanted to only do the epel10 build and not become the default epel bz assignee, but used the default template asking to be a full co-maintainer.
2024-11-13 18:50:31 <@carlwgeorge:matrix.org> i think we're in a situation now where epel requests are putting a bad taste in the mouth of fedora maintainers, and that worries me.
2024-11-13 18:50:56 <@conan_kudo:matrix.org> this has always been kind of a problem, even before epel-packager-sig
2024-11-13 18:51:35 <@salimma:fedora.im> in our discussion this, Carl suggested having more focused SIGs and it's not a bad idea
2024-11-13 18:51:39 <@conan_kudo:matrix.org> I experienced this plenty of times when I was working on getting fedora infra apps in EPEL
2024-11-13 18:51:40 <@carlwgeorge:matrix.org> i know it would be a lot of work to unwind having the sig on so many packages, but it's not something that would need to happen overnight.  if we agree to go in that direction, we could start by removing mention of it from the branch request, i.e. if you ask for a branch you yourself must be agreeing to be a co-maintainer.
2024-11-13 18:51:48 <@salimma:fedora.im> we can have a moratorium on adding new packages to epel-packagers-sig for now?
2024-11-13 18:52:13 <@carlwgeorge:matrix.org> i know it would be a lot of work to unwind having the sig on so many packages, but it's not something that would need to happen overnight.  if we agree to go in that direction, we could start by removing mention of it from the branch request, i.e. if you ask for a branch you yourself must be agreeing to be a co-maintainer and the default epel bz assignee.
2024-11-13 18:52:27 <@davide:cavalca.name> can we quantify the problem?
2024-11-13 18:52:45 <@carlwgeorge:matrix.org> not without directly calling people out by name, and i'd prefer not to do that
2024-11-13 18:53:02 <@conan_kudo:matrix.org> also, is it *new*?
2024-11-13 18:53:08 <@carlwgeorge:matrix.org> it's getting worse
2024-11-13 18:53:15 <@salimma:fedora.im> does it matter whether it's new or not?
2024-11-13 18:53:26 <@conan_kudo:matrix.org> it somewhat does
2024-11-13 18:53:38 <@salimma:fedora.im> FWIW I'm the person who brought the current SIG into existence and I'm inclined to agree with Carl here
2024-11-13 18:53:41 <@conan_kudo:matrix.org> because on some level, people are already not happy to acknowledge they don't actually own packages
2024-11-13 18:53:49 <@carlwgeorge:matrix.org> what michel alluded to was an idea we were brainstorming on, instead of a mega sig for everything epel, more smaller topic focused sigs
2024-11-13 18:53:59 <@smooge:fedora.im> it isn't new. it is the major reason various packagers have wanted EPEL to be moved under CentOS or out of Fedora 
2024-11-13 18:54:24 <@conan_kudo:matrix.org> and on another level, this repeatedly forces that point that there are no owners of packages because they can be bypassed
2024-11-13 18:54:28 <@nirik:matrix.scrye.com> The notifications thing was why I dropped out and I think Troy Dawson as well.
2024-11-13 18:54:28 <@salimma:fedora.im> I think we need to draw a distinction between people who want to just own packages, and reasonable people like Miro and Fabio who thinks if you're not the main maintainer or in the ACL you should at least put up a PR and wait for an official response
2024-11-13 18:54:48 <@smooge:fedora.im> and me
2024-11-13 18:54:48 <@tdawson:fedora.im> Yep.
2024-11-13 18:55:03 <@salimma:fedora.im> before we digress too much, Carl I think you should mention what you think should replace this
2024-11-13 18:55:10 <@conan_kudo:matrix.org> I just disabled the notifications for the sig for me, since it already overlaps with all the sigs I'm already part of
2024-11-13 18:55:20 <@carlwgeorge:matrix.org> yes, there is a timeline where epel-packagers-sig works as intended and fedora maintainers don't feel bypassed and everything goes through pr and everyone answers their bugs, but we don't live in that timeline
2024-11-13 18:56:16 <@carlwgeorge:matrix.org> instead of a mega sig for everything epel, more smaller topic focused sigs would work better i think
2024-11-13 18:56:28 <@nirik:matrix.scrye.com> I think there's also several parts to this, it's not one thing... people not wanting to be bypassed, but also people wanting someone who builds/pushes something to stay and maintain it and answer bugs, update it, etc
2024-11-13 18:56:41 <@tdawson:fedora.im> Like an epel-packagers-rust-sig ?
2024-11-13 18:57:20 <@davide:cavalca.name> so I think it's worth drawing a distinction between well-maintained and closely-followed ecosystems like rust and python and random packages
2024-11-13 18:57:37 <@carlwgeorge:matrix.org> a sig of 2-5 people who all care about a topic is a much better ownership model than a sig of ~20 with very unrelated things
2024-11-13 18:57:41 <@davide:cavalca.name> for the former, I totally agree we should coordinate more and better, and that the sig in its current form is likely to big of a hammer
2024-11-13 18:58:23 <@conan_kudo:matrix.org> if we're going into topic based sigs, then they should join the relevant sigs for those packages
2024-11-13 18:58:23 <@davide:cavalca.name> but for the latter, IMO we will always need some kind of catchall; tbh a lot of times when I end up branching something because I need it, I end up having to fix stuff in Rawhide too because the package is so old it barely even builds
2024-11-13 18:58:28 <@carlwgeorge:matrix.org> yeah that's a good way to think of it, the current set of epel-packager-sig stuff is too random
2024-11-13 18:58:28 <@conan_kudo:matrix.org> and if they don't exist, work with them to make it
2024-11-13 18:58:32 <@salimma:fedora.im> with Rust I think you should just join the Rust SIG. same with Golang
2024-11-13 18:58:48 <@conan_kudo:matrix.org> for example, almost every desktop in fedora branched into epel has a sig
2024-11-13 18:58:55 <@nhanlon:beeper.com> I believe I brought this idea up some time ago and it was turned down for some reason, but I can't recall the circumstances and can't find my notes. In any case, I'm in support of it.
2024-11-13 18:58:55 <@salimma:fedora.im> with Python just being in the SIG is not enough, some packagers prefer you also sign up to own EPEL bugs if you branch it
2024-11-13 18:58:57 <@conan_kudo:matrix.org> there is no reason people working on those desktops aren't members of those sigs
2024-11-13 18:59:08 <@salimma:fedora.im> and yeah, KDE EPEL works happen as part of the KDE SIG
2024-11-13 18:59:11 <@carlwgeorge:matrix.org> michel can you summarize the thing from fesco the other day?
2024-11-13 18:59:21 <@tdawson:fedora.im> Yep, I agree with that, same with KDE, and multi-media.
2024-11-13 18:59:29 <@carlwgeorge:matrix.org> about porting the stalled process to all of fedora, sort of?
2024-11-13 18:59:38 <@smooge:fedora.im> in the past many of those sigs had no interest in EPEL (and were actively hostile to it)
2024-11-13 18:59:42 <@yselkowitz:fedora.im> so limit epel-packagers-sig to epel-specific packages?  or even to just epel-release etc.?
2024-11-13 18:59:47 <@salimma:fedora.im> Carl also mentioned that for the "we need to be able to fix packages if nobody touches EPEL" - the normal provenpackager is better
2024-11-13 18:59:49 <@nirik:matrix.scrye.com> side note: it would be cool if we could make the 'would you please branch this for epel' bugs more useful. Perhaps have a tool to make them.
2024-11-13 19:00:09 <@salimma:fedora.im> ebranch does, I hope to make a new release that's more useful soon
2024-11-13 19:00:23 <@salimma:fedora.im> we actually landed in this mess /because/ someone mass-fire bugs with ebranch :(
2024-11-13 19:00:30 <@carlwgeorge:matrix.org> we're out of time, i'm happy to talk about this more in the main channel, or table it until next week's meeting
2024-11-13 19:00:31 <@conan_kudo:matrix.org> epel-packagers-sig should be an rcm override, similar to fedora provenpackagers
2024-11-13 19:00:49 <@smooge:fedora.im> Carl George: could you write up a full proposal from the talk
2024-11-13 19:00:52 <@salimma:fedora.im> so Conan Kudo I initially thought so too but Carl pointed out why not use provenpackager then
2024-11-13 19:00:57 <@conan_kudo:matrix.org> and we can keep that group small and eliminate maintainership from epel-packagers-sig
2024-11-13 19:01:03 <@conan_kudo:matrix.org> Michel Lind 🎩 UTC-6: maybe we should?
2024-11-13 19:01:19 <@tdawson:fedora.im> Just so people know, as much as I like this conversation, our time is up ... but I'm going to extend another 5 minutes so we can wrap it up.
2024-11-13 19:01:20 <@davide:cavalca.name> provenpackagers might be too big of a hammer
2024-11-13 19:01:26 <@conan_kudo:matrix.org> the original reason was that provenpackagers cannot branch packages
2024-11-13 19:01:30 <@conan_kudo:matrix.org> that's why we didn't use it
2024-11-13 19:01:52 <@carlwgeorge:matrix.org> and for good reason, pp also used to get abused for branch and forget drive-bys
2024-11-13 19:01:58 <@conan_kudo:matrix.org> but if we are breaking things back up into topic sigs, then the justification changes
2024-11-13 19:02:25 <@conan_kudo:matrix.org> I'm not saying it wasn't, but a big part of the problem is we don't do much to push people working on epel back into fedora
2024-11-13 19:02:41 <@rcallicotte:fedora.im> !hi
2024-11-13 19:02:41 <@zodbot:fedora.im> Robby Callicotte (rcallicotte) - he / him / his
2024-11-13 19:02:44 <@conan_kudo:matrix.org> not saying it's any of us, but I know of people who are like this
2024-11-13 19:02:50 <@carlwgeorge:matrix.org> i'll work up a thread on discussions.fpo
2024-11-13 19:02:51 <@salimma:fedora.im> Carl write full proposal with alternatives (e.g. use provenpackager vs make epel-packagers-sig a build override SIG like provenpackager but without branching)
2024-11-13 19:02:51 <@salimma:fedora.im> so let's do this
2024-11-13 19:02:51 <@salimma:fedora.im> 
2024-11-13 19:03:11 <@salimma:fedora.im> and for my docs PR I can put a note saying please don't use epel-packagers-sig for now in requests until this is sorted out
2024-11-13 19:03:12 <@tdawson:fedora.im> Hi Robby Callicotte  ... just in time for us to end. ;) ... we're going over for 5 minutes.
2024-11-13 19:03:21 <@rcallicotte:fedora.im> neat
2024-11-13 19:04:26 <@tdawson:fedora.im> Sounds good.   And this that, let's close things up.
2024-11-13 19:05:16 <@tdawson:fedora.im> Thank you all for the very good conversations.  And especially thank you all for not "going after people" and instead kept it a nice civil discussion about what to do.
2024-11-13 19:05:31 <@salimma:fedora.im> the world could use more civility
2024-11-13 19:05:35 <@tdawson:fedora.im> Yep
2024-11-13 19:05:45 <@tdawson:fedora.im> I'll talk to ya'll next week, if not sooner.
2024-11-13 19:05:51 <@jrichardson:matrix.org> Thanks as always Troy Dawson 
2024-11-13 19:05:59 <@tdawson:fedora.im> !endmeeting