20:00:12 <tdawson> #startmeeting EPEL (2021-07-07)
20:00:12 <zodbot> Meeting started Wed Jul  7 20:00:12 2021 UTC.
20:00:12 <zodbot> This meeting is logged and archived in a public location.
20:00:12 <zodbot> The chair is tdawson. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:12 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:00:12 <zodbot> The meeting name has been set to 'epel_(2021-07-07)'
20:00:13 <tdawson> #meetingname epel
20:00:13 <zodbot> The meeting name has been set to 'epel'
20:00:13 <tdawson> #chair nirik tdawson bstinson pgreco carlwgeorge michel dcavalca
20:00:13 <zodbot> Current chairs: bstinson carlwgeorge dcavalca michel nirik pgreco tdawson
20:00:13 <tdawson> #topic aloha
20:00:18 <pgreco> hello!!
20:00:28 <michel> .hello salimma
20:00:29 <zodbot> michel: salimma 'Michel Alexandre Salim' <michel@michel-slm.name>
20:00:40 <dcavalca> .hi
20:00:40 <carlwgeorge> howdy yall
20:00:40 <zodbot> dcavalca: dcavalca 'Davide Cavalca' <dcavalca@fb.com>
20:00:52 <tdawson> Hi michel
20:00:57 <tdawson> Hi dcavalca
20:01:01 <tdawson> Hi pgreco
20:01:06 <tdawson> Hi carlwgeorge
20:01:24 <tdawson> We have almost everyone here in the first minute.
20:01:45 <michel> \o/
20:02:06 <michel> at work we are trying to promote starting meetings at 5 past the hour, I wish it was 'ending meeting 5 before the hour' instead
20:02:15 <tdawson> Ha!
20:03:08 <tdawson> Well, I like the extra five minutes for everyone to gather ... but if you say the meetings start at 5 after the hour, then at 5 after the hour is when everyone starts ariving.
20:03:26 <michel> yeah, we'll see if that ended up being the case or not :p
20:03:28 <pgreco> michel: what about starting on time and ending at 5 past? :P
20:03:40 <michel> pgreco: ha, that's... actually what happened
20:04:00 * carlwgeorge waves at maxamillion
20:04:02 <maxamillion> .hello2
20:04:03 <zodbot> maxamillion: maxamillion 'Adam Miller' <maxamillion@gmail.com>
20:04:09 <tdawson> Hi maxamillion
20:04:10 <michel> back before remote work, office rooms have glass doors so what you do is hover ominously in front trying to intimidate people into evacuating
20:04:31 <tdawson> maxamillion: We haven't seen you here for a while, it's good to see you
20:04:34 <dcavalca> and/or start playing flappy bird on the room tool
20:04:46 <maxamillion> tdawson: it's good to be back :)
20:04:49 <tdawson> Does anyone know if nirik is going to make it?
20:05:28 <tdawson> Well, it's 5 after, so lets start
20:05:36 <tdawson> #topic Old Business
20:05:47 <tdawson> We'll start with the usual SIG
20:06:18 <tdawson> maxamillion: is there a specific agenda item you wanted to talk about?  Just want to make sure we leave enough time for whichever it is
20:06:39 <tdawson> Do we have any SIG issues and/or problems we need to talk about?
20:06:50 * michel need a minute
20:07:14 <maxamillion> tdawson: nope, just trying to slowly get involved again
20:07:44 <tdawson> maxamillion: Cool.  more hands the better, and it's always good to work with you again.
20:08:09 * nirik arrives late, reads up
20:08:28 <michel> sorry, baby issue. nothing much from me for EPEL Packagers this week, we had an extended break for 4th of July and I am still spring cleaning all my old tasks and unsubscribing from my old team's alerts and notifications :p
20:08:58 <dcavalca> my only update is that we got rust-below building in epel8
20:09:05 <dcavalca> which was less horrifying than I expected
20:09:12 <tdawson> :)
20:09:18 <tdawson> Did you use the vendor bundling?
20:09:23 <dcavalca> yep
20:09:36 <carlwgeorge> as one does in epel
20:09:36 <pgreco> dcavalca: rust an less horrifying in the same sentence?
20:09:41 <dcavalca> https://src.fedoraproject.org/rpms/rust-below/blob/rawhide/f/update-vendor-tarball.sh
20:09:46 <dcavalca> in case you need it for another package
20:10:03 <tdawson> Glad it worked out ...
20:10:59 <tdawson> Oh ... you vendor bundled it in rawhide too ... well ... good luck with that.
20:11:14 <dcavalca> tdawson: nope, rawhide is done sanely
20:11:25 <dcavalca> but I don't like diverging branches, so the shellscript is there too
20:11:30 <dcavalca> and there's a few %if in the specfile
20:11:37 <tdawson> Oh .. ok.
20:11:50 <tdawson> And, now I see the %if's ...yep
20:12:17 <tdawson> OK, if the sig is doing good, I'm going to move onto epel-next, cuz I bet it's going to take a bit of time.
20:12:18 * nirik was gonna look at rust-below the other day...
20:12:21 <dcavalca> I'm also unretiring xcompmgr, because it turns out the Chromium OSS tests depend on it...
20:12:52 <tdawson> oh ... cool.
20:13:07 <tdawson> carlwgeorge: Anything for epel-next ?
20:13:42 <carlwgeorge> a few of us met earlier this week to plan out epel9-next at a high level
20:13:51 <carlwgeorge> https://hackmd.io/@carlwgeorge/BJV_Rez6O
20:14:14 <carlwgeorge> two things are kinda uncertain
20:14:31 <carlwgeorge> 1. how much of epel9 do we go ahead and set up as prep for epel9-next
20:15:03 <carlwgeorge> 2. what is our strategy for when rhel9 goes ga, such as doing a mass branch/rebuild from epel9-next to epel9
20:15:27 <carlwgeorge> iirc mboddu was going to start a mailing list thread to discuss 2
20:15:34 <nirik> 1. only as much as you need to (keys?)
20:15:52 <nirik> because if you make more people will start trying to use it/ask about when it's live/etc :)
20:16:02 <carlwgeorge> nirik: we were thinking of doing all the tags too, just to avoid changing the inheritance later
20:16:18 <dcavalca> yeah, I'd say if there's common components, you can spin both up, but otherwise I'd wait to do epel9 until GA
20:16:19 <carlwgeorge> tags with no build target would be fine, i think
20:16:43 <nirik> carlwgeorge: well, people will see them and ask about them.
20:16:55 <nirik> but perhaps it's fine
20:17:00 <tdawson> I agree with nirik for 1.  Just the minimum.  Let maintainers do what packages they need once it's released, but we should just get it up with as few packages as possible.
20:17:15 <dcavalca> what's the plan for signing? iirc c9s content is still unsigned at the moment
20:17:46 <carlwgeorge> the plan is whatever would happen for epel9 eventually anyways, a new key that we use for epel9 and epel9-next
20:18:28 <pgreco> epel would be signed even if stream isn't then?
20:18:38 <carlwgeorge> sure why not, they're independent things
20:18:50 <pgreco> not a problem, just understanding
20:18:52 <dcavalca> we should probably say somewhere that we're building from unsigned content then
20:18:58 <carlwgeorge> also i keep hearing that cs9 signing is coming soon(tm)
20:19:02 <dcavalca> at least until that's sorted out on the c9s side
20:19:23 * dcavalca can't do any work on c9s until it's signed :(
20:19:32 <nirik> yeah, we likely don't want to setup signing later... would be more hassle...
20:20:04 <carlwgeorge> well we plan to use the cs9 buildroot, which will stay unsigned i think
20:20:46 <michel> when we go from 9-next to 9 it will involve rebuilding anyway, right?
20:21:01 <carlwgeorge> not necessarily
20:21:19 <michel> ah. so... if we start 9-next while cs9 is still unsigned.. should we mass rebuild?
20:21:39 <carlwgeorge> i don't see the point
20:22:18 <pgreco> the point is in how trust worthy is what is built already
20:22:21 <nirik> I'd say no mass rebuilds, but might just be me.
20:22:37 <pgreco> but if we're always going to build from unsigned, then yeah
20:22:47 <nirik> well, koji doesn't check that.
20:23:08 <carlwgeorge> mboddu suggested the mass branch/rebuild strategy, and was fairly convincing, i'll let him comment if he likes (either now or on list)
20:23:08 <nirik> it's up to us to make sure the buildroot contents we use are valid/correct.
20:23:34 <pgreco> right, but koji is the same server
20:24:00 <tdawson> carlwgeorge: Note:  When you say "mass branch/rebuild" you are talking about going from epel9-next to epel9, not epel8 -> epel9-next
20:24:03 <pgreco> but same network/cage should be pretty much the same level
20:24:11 <nirik> well, I mean epel7/8 we sync rhel content down, but koji doesn't check that they are signed.
20:24:17 <carlwgeorge> the basic idea is making the assumption that if a maintainer requested an epel9-next branch, they intend to maintain the package in epel9 too
20:24:30 <carlwgeorge> tdawson: yes exactly, sorry if that wasn't clear
20:25:01 <tdawson> To me, it was, but it sounded like michel was thinking a mass rebuild of epel8 -> epel9-next
20:25:07 <nirik> as long as we are clear with that expectation I guess.
20:25:36 <michel> tdawson: oh, no, sorry, I mean rebuilding everything in 9-next that was initially built pre-signing
20:25:55 <carlwgeorge> originally for 8 we discussed enforcing epelX existing if epelX-next is requested, but we never implemented it
20:26:00 <tdawson> Ah, ok
20:26:13 <michel> mass-rebuilding probably won't fly, we still want epel9-next to be opt-in, right? hopefully I can focus on getting my bulk branch request tool done soon, so it will be less painful than the epel8 bringup
20:26:30 <carlwgeorge> but since we never implemented it, we can use that as the opportunity to mass create epel9 branches from epel9-next branches and be able to populate epel9 really early
20:27:36 <carlwgeorge> michel: it will still be opt-in, but the idea is you can't only opt in to epel9-next
20:27:40 <nirik> we need to be 100% clear in the epel9-next announcement that that is the expectation... if we do that I am ok with that.
20:27:58 <carlwgeorge> right, doing epel9 only is ok, doing epel9-next only is not
20:28:20 <carlwgeorge> i would be surprised if anyone actually wanted the later anyways
20:28:42 <michel> carlwgeorge: right. signing up for 9-next means signing up for 9
20:28:43 <pgreco> the mass branching from epel9-next to epel9 should happen before 9.1 starts landing in stream
20:29:07 <nirik> or... as soon as 9 lands.
20:29:34 <michel> so mass-branching from X-next to X will be the default for X >= 9, basically
20:29:47 <carlwgeorge> pgreco: the mass branching could happen then, but not the building, because rhel9 wont' be available to set up the build target yet
20:30:18 <carlwgeorge> michel: only at the start of epelX, one time thing
20:30:32 <pgreco> yeah, my worry (maybe unfounded) is that the branched epel9 won't build because it may depend on something newer
20:30:56 <carlwgeorge> there will definitely be build failures to sort out, just like after a fedora mass branch/rebuild
20:31:33 <carlwgeorge> i'm sure we'll have more than a few related to unshipped devel packages
20:31:36 <tdawson> mohan (who seemed like he volunteered to do the mass branch/build) would be using the rawhide branch/build script(s).  One of the benefits this would bring, other than being very well tested, is that if people want to do the builds manually there is already a way for them to do that (say I have 300 packages I want done in a specific order and already have my own script)
20:31:39 <michel> <carlwgeorge "Michel Alexandre Salim: only at "> yup, I did mean that, sorry for not explicitly saying that
20:32:00 * carlwgeorge nods
20:32:39 <carlwgeorge> we don't need to hash out all the details yet, we can move on for the meeting and wait for mboddu's list email to discuss further
20:33:08 <tdawson> carlwgeorge: Good point ... and yep, that's still a ways off ... moving on.
20:33:08 <pgreco> do we have a way to list the packages used from the buildroot? so we can ask RHEL developers to make sure that those are shipped
20:33:25 <pgreco> since CBR is supposed to be enough to build epel?
20:33:41 <pgreco> even though it isn't in practice
20:33:52 <pgreco> maybe we could start clean in 9
20:34:00 <tdawson> pgreco: Hmm ... I'm betting we could script that.
20:34:26 <tdawson> Grab the latest root logs from all the epel builds, and compare.
20:34:49 <pgreco> yeah, I was thinking about something more "formal", hehe
20:34:53 <pgreco> or even httpd logs
20:35:11 <tdawson> I'm going to move on, or I'll get sucked back into the -next subject, but ya, it's very do-able.
20:35:45 <tdawson> badges and logos
20:36:18 <tdawson> Badges: https://pagure.io/fedora-badges/issue/829
20:36:26 <tdawson> Logos: https://pagure.io/design/issue/770
20:36:44 <tdawson> I haven't seen any progress on the badges
20:36:55 <carlwgeorge> i don't know what happens next on the badges, i guess we're waiting on feedback from the design team?
20:37:16 <tdawson> I guess ... I've never done one, so I don't know.
20:38:08 <tdawson> The logo, I created an SVG of the purple drop with the e ... though it looks like my browser doesn't like it.
20:38:47 <carlwgeorge> image attachments in pagure issues never load consistently for me
20:39:37 <nirik> yeah, I think it's waiting for feedback
20:41:15 <tdawson> anyway, there is progress on the logo at least.
20:41:55 <tdawson> Do we think we need to keep bringing these up in this meeting?  Or should we just wait until we have good progress and are needing a final approval?
20:42:40 <michel> I think it's fine to have a quick status check?
20:42:41 <carlwgeorge> the later i think
20:43:07 <tdawson> OK, then at most, I'll do a quick status check.
20:43:28 <tdawson> But, especially the logo, please feel free to make comments on the issues.
20:43:45 <tdawson> So we don't run out of time, I'm moving on.
20:43:51 <tdawson> #topic EPEL-7
20:44:29 <tdawson> I didn't notice anything dealing with epel7 this week, anyone see anything?
20:44:48 <nirik> nope, all quiet
20:45:19 <tdawson> #topic EPEL-8
20:45:53 <carlwgeorge> i got a thing
20:46:26 <carlwgeorge> epel8 currently has tesseract 4.  i noticed the other day that tesseract v3 is planned for rhel 8.5.
20:47:04 <carlwgeorge> https://bugzilla.redhat.com/show_bug.cgi?id=1826085 (mostly private comments, don't bother if you're not @redhat.com)
20:47:18 <tdawson> ugg
20:47:29 <nirik> thats... unfortunate. ;(
20:47:53 <michel> ouch. so... we're going to be regressing?
20:48:07 <michel> can we rename to tesseract4 and make it parallel isntallable?
20:48:32 <carlwgeorge> the maintainer can
20:48:32 <michel> or convince the RHEL team to pick v4
20:49:24 <carlwgeorge> but the rename would be a new srpm and dist-git repo, no matter what the current tesseract has to be retired
20:50:26 <nirik> yep
20:50:55 <carlwgeorge> shipping 4 instead of 3 was discussed and declined in the private comments
20:51:55 <pgreco> do we have a list of the affected packages?
20:52:51 <pgreco> and if v3 is api compatible with v4? I'm not familiar with tesseract
20:52:52 <carlwgeorge> i don't see anything that buildrequires tesseract-devel
20:52:59 * mboddu is here
20:53:01 <carlwgeorge> it's a new soname
20:53:06 <tdawson> vapoursynth-plugins
20:53:22 <pgreco> :(
20:53:50 <nirik> I don't see anyone noting the epel package in that bug...
20:53:52 <carlwgeorge> ah, `pkgconfig(tesseract)`
20:54:05 <nirik> might mention it and see if they would reconsider 4 vs 3?
20:54:35 <tdawson> Ya, according to the new "add a new package to RHEL" they are supposed to check if a package is already in EPEL, and if it is, open a bug about it.
20:54:39 <carlwgeorge> yeah the rationale given is the abi breakage, but it's currently buildroot only so i wouldn't think that applies
20:54:54 <tdawson> Looks like that step was skipped.
20:55:05 <carlwgeorge> this is my surprised face
20:55:11 <nirik> yeah, shocking.
20:55:19 <pgreco> I'm gonna go on a limb and say ignored
20:56:02 <carlwgeorge> i can fight this battle in the bz, unless someone else feels passionately about it and wants to drive it
20:56:07 <tdawson> carlwgeorge: Were you going to add a comment about epel in the bug?  Or would you like me to?
20:56:20 <carlwgeorge> i would love to pass it off to you if you'll have it
20:56:43 <tdawson> carlwgeorge: Let me take it, since I have a bookmark for the proceedure, which includes checking epel.
20:56:51 <carlwgeorge> god bless you
20:57:02 <carlwgeorge> tdawson++
20:57:10 <pgreco> tdawson++
20:57:10 <zodbot> pgreco: Karma for tdawson changed to 3 (for the current release cycle):  https://badges.fedoraproject.org/tags/cookie/any
20:57:13 <tdawson> Honestly, I haven't used tesseract, but now that I see it, it sounds like something I've been looking for for a while.
20:57:41 <carlwgeorge> i noticed it because i track the rhel pungi changes.  it's not in the 8.5 nightly compose yet, but i imagine it will be soon.
20:58:57 <tdawson> OK, I"ve writen it down and it's on my to-do list.  I'll report back next week.
20:59:09 <tdawson> Anything else for EPEL8?
20:59:24 <tdawson> Oh man, it's almost time ..
20:59:30 <tdawson> #topic General Issues / Open Floor
20:59:41 <tdawson> Anything for the general florr?
20:59:45 <tdawson> floor
21:00:26 <mboddu> Sorry I am late here, I changed the notifications location and they were lost totally
21:00:44 <Eighth_Doctor> tdawson: how are we on kde?
21:00:45 <maxamillion> mboddu: something something, computers may have been a mistake
21:00:56 <carlwgeorge> can we talk about not sending updates-testing reports to the main epel-devel list?
21:00:57 <tdawson> mboddu: Not a real problem ... we talked about your proposal and need you to send an email.
21:01:08 <mboddu> I will send an email about epel9 mass rebuild to epel-devel@ list
21:01:09 <carlwgeorge> a dedicated report list would be swell
21:01:19 <Eighth_Doctor> wait, epel9 what?
21:01:25 <tdawson> Eighth_Doctor: Really good actually.  I opened bugs on the rest of the packages, and half are already branched.
21:01:35 <mboddu> Eighth_Doctor: Wait for my email :D
21:01:35 <Eighth_Doctor> tdawson: sweet!
21:01:53 <nirik> carlwgeorge: well, the idea is that it gets someone testing them...
21:01:54 <tdawson> I'm through qt5, kf5, and just finished plasma.  Will start on the apps and everyrthing else later tonight.
21:02:02 <nirik> (dunno how well it works tho)
21:02:02 <Eighth_Doctor> tdawson: I'm looking forward to building a Hyperscale Workstation with KDE Plasma on it :D
21:02:08 <tdawson> Cool
21:02:10 <carlwgeorge> i'm pretty sure every just setups up filters to hide those
21:02:21 * Eighth_Doctor reads them
21:02:27 <carlwgeorge> *most people
21:02:35 <carlwgeorge> Eighth_Doctor is superhuman
21:02:38 <Eighth_Doctor> 😆
21:02:58 <Eighth_Doctor> at some point, c9s will become interesting again
21:03:15 <Eighth_Doctor> hopefully signed packages and signed repos are coming soon
21:03:20 <tdawson> Looks like our time is up, sorry for not having enough time for open floor
21:03:31 <tdawson> Thanks to everyone for coming and participating.
21:03:35 <tdawson> Talk to you next week.
21:03:42 <dcavalca> thanks tdawson
21:03:47 <tdawson> #endmeeting