20:00:07 <tdawson> #startmeeting EPEL (2022-04-06)
20:00:07 <zodbot> Meeting started Wed Apr  6 20:00:07 2022 UTC.
20:00:07 <zodbot> This meeting is logged and archived in a public location.
20:00:07 <zodbot> The chair is tdawson. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
20:00:07 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:00:07 <zodbot> The meeting name has been set to 'epel_(2022-04-06)'
20:00:07 <tdawson> #meetingname epel
20:00:07 <tdawson> #chair nirik tdawson pgreco carlwgeorge salimma dcavalca
20:00:07 <tdawson> #topic aloha
20:00:07 <zodbot> The meeting name has been set to 'epel'
20:00:07 <zodbot> Current chairs: carlwgeorge dcavalca nirik pgreco salimma tdawson
20:00:41 <dcavalca> .hi
20:00:41 <SSmoogen> here
20:00:41 <zodbot> dcavalca: dcavalca 'Davide Cavalca' <dcavalca@fb.com>
20:01:00 <tdawson> Hi dcavalca
20:01:02 <tdawson> Hi SSmoogen
20:01:14 <rcallicotte> .hi
20:01:15 <zodbot> rcallicotte: rcallicotte 'Robby Callicotte' <rcallicotte@mailbox.org>
20:01:41 <michel> Hi folks. Going to be slightly late
20:01:42 <michel> .hi salimma
20:01:42 <davide> Michel Alexandre Salim: I'm not seeing your messages on the IRC side of the bridge
20:01:42 <zodbot> michel: michel 'None' <m.wahlheim@t-online.de>
20:01:42 <michel> .hello salimma
20:01:44 <tdawson> Hi rcallicotte
20:01:45 <zodbot> michel: salimma 'Michel Alexandre Salim' <michel@michel-slm.name>
20:01:54 <rsc> .hello robert
20:01:55 <zodbot> rsc: robert 'Robert Scheck' <redhat@linuxnetz.de>
20:02:04 <michel> Yeah zodbot doesn't hear me either. Sigh
20:02:07 <carlwgeorge> .hi
20:02:08 <zodbot> carlwgeorge: carlwgeorge 'Carl George' <carl@redhat.com>
20:02:17 <michel> dcavalca: looks like it's fixed now?
20:02:27 <tdawson> Hi michel
20:02:29 <tdawson> Hi rsc
20:02:31 <dherrera> .hi
20:02:32 <zodbot> dherrera: dherrera 'None' <dherrera@redhat.com>
20:02:34 <tdawson> Hi carlwgeorge
20:02:38 <tdawson> Hi dherrera
20:02:54 <dcavalca> michel: yup, I can see you now
20:03:08 <nirik> morning
20:03:20 <tdawson> Hi nirik
20:05:10 <SSmoogen> what's next?
20:05:11 <tdawson> And then it got quiet ...  too quiet ...
20:05:17 <nirik> heh
20:05:18 <tdawson> #topic EPEL Issues https://pagure.io/epel/issues
20:05:19 <tdawson> https://pagure.io/epel/issues?tags=meeting&status=Open
20:05:27 <nirik> call me conversation killer.
20:05:33 <tdawson> .epel 135
20:05:34 <zodbot> tdawson: Issue #135: Modular content for EPEL9 - epel - Pagure.io - https://pagure.io/epel/issue/135
20:05:46 <tdawson> Let's start with the modular content
20:06:07 <tdawson> carlwgeorge: How was the response to the emails?
20:06:40 <carlwgeorge> only one more reply since last week, nodejs maintainer would like to have epel9 modular
20:06:43 <tdawson> Is this a little wierd to others, since we talked about this on the open office hours?
20:07:20 <tdawson> I can understand nodejs wanting it ... modules seem like they are made for nodejs.
20:07:29 * nirik still would prefer not to provide any modular support for epel9
20:07:54 <carlwgeorge> i'm with nirik, but i don't want to block it if someone that cares about it is willing to drive the work
20:08:01 * salimma back from dealing with baby and cats
20:08:08 * sgallagh is here
20:08:20 * carlwgeorge waves at sgallagh
20:08:26 <carlwgeorge> your ears must have been burning
20:08:30 <tdawson> Hi sgallagh
20:09:05 <nirik> ha.
20:09:08 <sgallagh> Or I have a trigger on "nodejs" ;-)
20:09:16 <carlwgeorge> same difference
20:10:06 <sgallagh> Indeed
20:10:09 <SSmoogen> personally I am going to expect that this will need as much work as it did the last time to get modularity 'working' for EPEL-8
20:10:30 <SSmoogen> that took a specific inititiative with CPE to get the parts working and things kickstarted.
20:10:54 <SSmoogen> I would prefer to say 'no sorry we do not have the capacity'
20:10:54 <sgallagh> So, there are valid needs for a Node.js module in EPEL9, in large part because the RHEL9 non-modular Node.js RPMs do not provide the development packages needed for things like R-v8
20:11:30 <nirik> couldn't that just be normal epel versions?
20:11:33 <carlwgeorge> can those be requested for crb?
20:11:40 <nirik> or the overlap is too much?
20:11:45 <sgallagh> If it comes down to it, I can find a workaround for Node.js
20:11:46 <salimma> can't we just get the development packages shipped?
20:12:29 <carlwgeorge> i'd prefer the only use case for epel modularity be shipping alternate versions, not shipping missing subpackages
20:12:32 <SSmoogen> going from how things went last time, what is shipped in 9.0 CRB will probably be it for a while as they wait to see what is 'really' needed versus 'wanted' and what can be handled
20:13:00 * salimma has a backlog project to rewrite the bitwarden cli in rust because ugh, it's currently nodejs
20:13:04 <sgallagh> If Node.js isn't built expecting to be linked against (as it is in RHEL) we have to rebuild the whole thing the "right way"
20:13:17 <sgallagh> carlwgeorge: That's essentially what it's doing in EPEL 8 right now, except that it's also shipping an alternate version of the same release that RHEL is (named `nodejs:16-epel`)
20:14:19 <sgallagh> As I said above: if it's only Node.js that is asking for Modules, I can live with finding a workaround.
20:15:11 <tdawson> We do have two others (zabbix and one more I belive) that would like to build on modules.
20:15:39 <carlwgeorge> 389-ds-base and nextcloud
20:16:04 <carlwgeorge> but i'm skeptical of nextcloud because the maintainer doesn't seem to be aware that the current epel8 modules don't even install
20:16:20 <tdawson> So, there is a desire for moudles, although it is low.  We just don't have someone to champion getting the modules in.
20:16:24 <salimma> isn't 389-ds.. a Red Hat project?
20:16:30 <SSmoogen> *cough*
20:16:51 <rcallicotte> hmm
20:16:59 * salimma : asking awkward questions since before the millenium
20:17:48 <carlwgeorge> looking at the epel8 389 streams, they have next/stable/testing, which i think may be a better fit for copr repos
20:18:13 <tdawson> sgallagh: hmm ... just how much do you want your nodejs modules?  Would you be willing to help get modularity into epel9?
20:18:46 <carlwgeorge> yeah what we're lacking at this point is someone to own and drive the process, coordinate with releng and mbs developers, etc.
20:19:03 <sgallagh> Not that badly
20:19:08 <tdawson> :)
20:19:27 <carlwgeorge> works for us, no one on the steering committee wants to own it either
20:19:28 * nirik again points out it's not _just_ that
20:19:31 <tdawson> I think at this point, everyone has looked at the work and taken one step back.
20:19:32 <sgallagh> I'd recommend talking to ppisar who is largely maintaining that stack nowadays for Fedora and RHEL
20:20:01 <carlwgeorge> nirik: it's the first step i'd say
20:20:21 <carlwgeorge> lets call gathering the info on which maintainers want it step 0
20:20:28 <nirik> sure, but 6 months after someone drives it in... they disappear and I get to maintain it.
20:20:49 <SSmoogen> hey now.. I disappeared way before EPEL modularity got done
20:20:57 <tdawson> *laughs*
20:20:58 <sgallagh> But I tend to agree: it's a LOT of effort to bootstrap modularity, and if interest is that low, it may not be worth it.
20:21:00 <sgallagh> ha
20:21:01 <carlwgeorge> i won't lose any sleep if we want to label this as "rejected by releng"
20:21:08 <carlwgeorge> or "rejected by epel steering committee"
20:21:34 <nirik> how about rejected for now, will reconsider if new information/people/requests ?
20:21:34 <SSmoogen> Formal vote? -1
20:21:41 <tdawson> I believe nirik has the best point of all.
20:22:02 <carlwgeorge> so who wants the job of saying no formally in the issue?
20:22:13 <rcallicotte> agreed
20:22:15 <tdawson> I'll say no formally
20:22:17 <SSmoogen> done
20:22:25 <salimma> I'm happy to lend a -1
20:22:28 <sgallagh> Please at least be loud about the rejection, so if there are others who REALLY need it but haven't noticed that it's in jeopardy, they'll have a chance to notice?
20:22:31 <carlwgeorge> would you like me to summarize the maintainer email responses first?
20:22:52 <tdawson> carlwgeorge: I think that would be a good idea.
20:22:53 <carlwgeorge> on the issue i mean
20:23:04 <tdawson> carlwgeorge: Yep, that's what I thought you meant.
20:23:10 <carlwgeorge> will do, and will ping you once that's done for you to deliver the crushing blow
20:23:16 <sgallagh> Though... I do wonder if we could arrange something with CentOS Stream to allow us to build EPEL modules in their MBS infra
20:23:33 <sgallagh> Which would also solve the "RHEL dependencies" problem, in theory.
20:23:42 <tdawson> carlwgeorge: I believe sgallagh has a point, we will need to announce this on epel-devel.
20:23:43 <carlwgeorge> they wouldn't be guaranteed to work on rhel
20:23:59 <SSmoogen> ticket is close
20:24:01 <carlwgeorge> there is already an epel-devel thread we can reply to
20:24:03 <sgallagh> carlwgeorge: Nothing in EPEL is guaranteed...
20:24:10 <tdawson> Ha!   SSmoogen beat me to it. :)
20:24:47 <tdawson> I'll still write up an email
20:24:48 <SSmoogen> I will let carlwgeorge and tdawson to announce this on the list
20:24:48 <carlwgeorge> sgallagh: how about "we know that will cause multi-month gaps where a library bump in stream will cause a module to not be installable
20:24:50 <carlwgeorge> "
20:25:07 <SSmoogen> actually it would be worse than that
20:25:27 <sgallagh> Do we know that? I'm not sure.
20:25:34 <sgallagh> But it's not worth debating here.
20:25:35 <SSmoogen> the version numbers built in CentOS Stream will not match the ones in RHEL and dependencies between modules will break
20:26:06 * SSmoogen tried a hybrid in the past to get around a 'missing thing'
20:26:08 <tdawson> We still have a few more things, so I'm going to move on.
20:26:11 <SSmoogen> move along
20:26:12 <carlwgeorge> sgallagh: we definitely know it's possible, which is why epel-next exists
20:26:33 <sgallagh> ok
20:26:46 * sgallagh departs
20:26:57 <tdawson> salimma: Do you mind if we put off the CVE issue until it's one month mark?
20:27:18 <salimma> tdawson: not at all
20:27:33 <salimma> haven't had time to work on it anyway so I have no update on that front
20:27:33 <tdawson> Thanks.
20:27:40 <tdawson> #topic Old Business
20:27:52 <tdawson> Depending on HA or outside packages
20:28:10 <tdawson> Carl has a pull request to to documentation on this ...
20:28:27 <tdawson> https://pagure.io/epel/pull-request/168
20:29:17 <tdawson> Although  he's going to re-word one part that nirik asked about, I was wondering how everyone felt about it.
20:29:58 <dcavalca> seems reasonable
20:30:01 <rsc> Wasn't an exception list mentioned, too?
20:30:06 <SSmoogen> seemed reasonable
20:30:31 <tdawson> rsc: Yes, he has that in the last paragraph
20:30:49 <rsc> tdawson: ah, right. Sorry!
20:31:44 <tdawson> I guess the question is, is anyone against it?
20:32:11 <salimma> I agree with it with nirik's note about just making it a meeting issue
20:32:32 <carlwgeorge> I think I've accommodated/addressed all previous concerns, save the last bit about meeting issues
20:33:11 <carlwgeorge> can non-maintainers add their own issue tags?
20:33:26 <tdawson> I'm pretty sure they can
20:33:55 <tdawson> I believe they have to have an account on pagure.io ... but other than that I think it's open.
20:34:28 <SSmoogen> yep
20:35:34 <tdawson> So, I'm going to say this has passed, when Carl get's the meeting issues thing written.
20:36:01 <tdawson> carlwgeorge: Thank you for writing that up.
20:36:15 <SSmoogen> +1
20:36:37 <tdawson> The other thing in Old Business is changing the "Stalled Package" time
20:37:02 <tdawson> Ugg ... I don't have the email thread handy ...
20:38:08 <carlwgeorge> i do if i can find the tab
20:38:09 <tdawson> https://lists.fedoraproject.org/archives/list/epel-devel@lists.fedoraproject.org/thread/FBTH7FGVYFVBFJX6QKAPLC6FCV66OY6D/
20:39:05 <salimma> I feel like this is in danger of turning into bikeshedding, so I'm throwing my weight behind tdawson 's option D but if it does not work we can always reevaluate
20:39:28 <salimma> so - file request, wait a week, needinfo and wait two weeks before releng
20:39:52 * nirik is ok with that.
20:40:02 <carlwgeorge> so needsinfo after a week?
20:40:16 <nirik> needinfo is pretty harsh, but I suppose people who aren't responsive won't care about that much more than the bug itself.
20:40:29 <SSmoogen> needinfo after a week
20:40:44 <dcavalca> +1 for needinfo after a week as well
20:40:45 <carlwgeorge> i think last week there was a suggestion of an immediate needsinfo, which i disagree with, but i can live with needsinfo after 1 week
20:40:54 <nirik> I guess that means needinfo on the asignee ?
20:40:56 <salimma> yeah, immediate needinfo feels rude
20:41:01 <SSmoogen> I offered the immediate and withdraw it
20:41:25 <tdawson> nirik Correct, I guess we need to be specific.
20:41:26 <salimma> if we still get pushback we can always adjust it to either 2 weeks / 1 week or 2 weeks / 2 weeks
20:41:49 <salimma> ideally we can do needinfo on /any/ maintainer with admin access, but that's not possible :p
20:41:51 <carlwgeorge> yeah even if we tweak it again later this feels like a step in the right direction
20:41:56 <SSmoogen> I think that the second notification needs to be stronger to show 'yes we tried to get in touch with you'
20:42:34 <carlwgeorge> salimma: you can set multiple needsinfos, you just have to specify each person, no easy button for all
20:42:34 <tdawson> Yep
20:43:06 <tdawson> So, it's sounding like Proposal D is the winner?
20:44:00 <carlwgeorge> works for me
20:44:04 <nirik> +1
20:44:28 <dcavalca> +1
20:44:59 <carlwgeorge> i can offer up the pr to implement
20:45:38 <tdawson> #info Stalled Package Proposal D passed - 6 for, 0 against
20:45:53 <tdawson> carlwgeorge: Thanks, that was going to be my next question.
20:46:18 <tdawson> Moving on.
20:46:29 <tdawson> #topic EPEL-7
20:46:30 <tdawson> CentOS 7 will go EOL on 30 June, 2024
20:46:43 <tdawson> Anything for EPEL 7 specifically?
20:47:42 <tdawson> #topic EPEL-8
20:47:42 <tdawson> CentOS Stream 8 goes EOL in 2024-05-31
20:48:06 <tdawson> Ha ... I just noticed I have the date formats in two different formats. :)
20:48:10 <tdawson> Anything for EPEL 8?
20:48:29 <nirik> YYYY-MM-DD is the only way to specify dates. ;)
20:49:04 <carlwgeorge> nirik: +1
20:49:08 <SSmoogen> nothing el8 from me
20:49:09 * nirik notes rhel 8.6 beta came out
20:49:18 <SSmoogen> good catch
20:49:36 <SSmoogen> we should probably put in a ticket to have a snapshot of 8.5 archived
20:49:38 <carlwgeorge> i've got a couple of epel8 packages on my radar to retire when 8.6 drops
20:49:46 <tdawson> Is that public?  Or just for certain partners?
20:50:06 <nirik> did we archive 8.4? or did we/I drop the ball.
20:50:06 <carlwgeorge> swtpm and libtpms
20:50:17 <tdawson> nirik It got archived
20:50:33 <nirik> ok, cool.
20:50:53 <tdawson> https://archives.fedoraproject.org/pub/archive/epel/8.4/
20:51:08 <tdawson> #topic EPEL-9
20:52:16 <tdawson> Sorry SSmoogen, I missed that.  But, isn't it a bit early to get a snapshot made?
20:52:23 <tdawson> And, anything for epel9?
20:52:57 <nirik> oh, one thing we did a while back I can mention....
20:53:05 <SSmoogen> tdawson, I was thinking we can do a snapshot now and do one later when .6 drops
20:53:23 <SSmoogen> better to do it early than forget later
20:53:24 <tdawson> SSmoogen: Oh ... well, if we can do that ... sure
20:53:33 <nirik> we made a c9s instance "el9-test.fedorainfracloud.org" for package maintainers (see https://fedoraproject.org/wiki/Test_Machine_Resources_For_Package_Maintainers )
20:53:56 <SSmoogen> it is basically rsync -avSHP /pub/epel/8/ /pub/archive/epel/8.6/
20:54:43 <tdawson> nirik That's great.
20:55:07 <nirik> oh... also another epel9 thing...
20:55:45 <SSmoogen> waits for the Steve Jobs reveal
20:55:52 <nirik> salimma had filed a bug to request python-django for epel9... https://bugzilla.redhat.com/show_bug.cgi?id=2033064
20:56:14 <nirik> maintainers had balked, but I think if we just keep the LTS version and go from one to another they would be ok with it...
20:56:16 <salimma> nirik: nice
20:57:00 <nirik> I assume the primary case here is mailman?
20:57:31 <nirik> anyhow, anyone with thoughts welcome to add them to the bug. ;)
20:58:09 <tdawson> Thanks
20:58:31 <tdawson> We're getting close to time, I'm going to lump the final two issues together
20:58:43 <tdawson> #topic EPEL-Packaging-SIG / General Issues / Open Floor
20:59:24 <carlwgeorge> i thought we agreed to just merge those anyways, or do away with the epel packaging sig dedicated section
20:59:38 <tdawson> Ah ... I'm good with that.
20:59:56 <tdawson> Maybe that was the weeks I was gone.
21:00:20 <salimma> nirik: I think we need to use LTS anyway, Conan Kudo mentioned that hyperktity does not support django 4 yet anyway
21:00:35 <salimma> yeah, I think we don't need to make packaging SIG a special thing
21:00:39 * nirik nods
21:01:04 <tdawson> The EPEL Committee Badges had some progress - https://pagure.io/fedora-badges/issue/829   I think I'm going to tell them to use the latest graphic and finish things up.
21:01:20 * Eighth_Doctor waves
21:01:21 <Eighth_Doctor> .hello ngompa
21:01:22 <zodbot> Eighth_Doctor: ngompa 'Neal Gompa' <ngompa13@gmail.com>
21:01:22 <salimma> one epel9 update - once xml2dict is branched for epel9 I can follow up on asking the jc maintainer to branch it too. it's a super spiffy tool
21:01:43 <tdawson> Hi Eighth_Doctor
21:01:50 <salimma> convert many tools' output into json so you can use jq etc. to splice the data. it's like the best bits of Powershell :)
21:01:51 <tdawson> Just in time for the meeting to end :)
21:01:59 <salimma> tdawson: more badge :D
21:02:05 <tdawson> salimma: Cool.
21:02:27 <tdawson> If nobody has anything else, I'm going to close ... looks like we are over time.
21:02:36 <SSmoogen> hit the button tdawson
21:02:41 <tdawson> Thanks ya'll for coming, and for all the  work you do on EPEL
21:02:48 <tdawson> #endmeeting