20:00:21 <tdawson> #startmeeting EPEL (2021-11-03)
20:00:21 <zodbot> Meeting started Wed Nov  3 20:00:21 2021 UTC.
20:00:21 <zodbot> This meeting is logged and archived in a public location.
20:00:21 <zodbot> The chair is tdawson. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
20:00:21 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:00:21 <zodbot> The meeting name has been set to 'epel_(2021-11-03)'
20:00:22 <tdawson> #meetingname epel
20:00:22 <tdawson> #chair nirik tdawson bstinson pgreco carlwgeorge michel dcavalca
20:00:22 <tdawson> #topic aloha
20:00:22 <zodbot> The meeting name has been set to 'epel'
20:00:22 <zodbot> Current chairs: bstinson carlwgeorge dcavalca michel nirik pgreco tdawson
20:00:26 <pgreco> .hi
20:00:27 <zodbot> pgreco: pgreco 'Pablo Sebastian Greco' <pablo@fliagreco.com.ar>
20:00:27 <nirik> morning.
20:00:36 <michel> .hello salimma
20:00:37 <zodbot> michel: salimma 'Michel Alexandre Salim' <michel@michel-slm.name>
20:00:40 * michel back from parental leave
20:00:44 <tdawson> Hi pgreco
20:00:46 <tdawson> Hi nirik
20:00:53 <tdawson> Hi michel ... good to see you again
20:00:54 <carlwgeorge> .hi
20:00:54 <davide> .hello2 dcavalca
20:00:54 <zodbot> carlwgeorge: carlwgeorge 'Carl George' <carl@redhat.com>
20:00:54 <SSmoogen[m]> hello
20:00:56 <themayor> .hi
20:00:58 <zodbot> davide: Sorry, but user 'davide' does not exist
20:01:01 <zodbot> themayor: Sorry, but user 'themayor' does not exist
20:01:06 <SSmoogen[m]> hi Michel Alexandre Salim
20:01:18 <tdawson> Hi SSmoogen[m]
20:01:18 <davide> lol I guess the bot doesn't like my matrix nick
20:01:38 <tdawson> hi carlwgeorge
20:01:47 <tdawson> hi themayor
20:01:52 <themayor> hey troy
20:01:54 <tdawson> hi davide
20:01:57 <nirik> it doesn't like a number of nicks now...
20:01:59 <tdawson> :)
20:02:11 <themayor> davide: doesnt seem to be liking mine today
20:02:24 <tdawson> Wow ...everyone joined so quickly I had a hard time keeping up.
20:02:46 <michel> Davide Cavalca: did you have it registered in accounts.fp.o ? (I'm assuming yes)
20:02:52 <davide> I think so?
20:03:03 <nirik> it's broken right now.
20:03:04 <michel> .hello dcavalca
20:03:05 <zodbot> michel: dcavalca 'Davide Cavalca' <dcavalca@fb.com>
20:03:12 <davide> yeah it's there
20:03:27 <davide> oh it's .hello not .hello2
20:03:37 <nirik> well, for karama, I guess account searching still is fine.
20:03:47 * davide can never keep the bots straight
20:03:51 <nirik> .misc help hello
20:03:52 <zodbot> nirik: (hello <an alias, 1 argument>) -- Alias for "hellomynameis $1".
20:03:54 <nirik> .misc help hello2
20:03:55 <zodbot> nirik: (hello2 <an alias, 0 arguments>) -- Alias for "hellomynameis $nick".
20:04:03 <nirik> ^ different. ;)
20:04:13 <davide> TIL
20:04:16 <davide> thanks nirik
20:04:21 <michel> how's hello2 separate from hi?
20:04:23 <michel> .misc help hi
20:04:24 <zodbot> michel: (hi <an alias, 0 arguments>) -- Alias for "hello2 $*".
20:04:30 <michel> oh duh
20:05:07 <tdawson> #topic Old Business
20:05:42 <tdawson> Starting with willit ... no progress last week, other than it's running - https://tdawson.fedorapeople.org/epel/willit/status-overall.html
20:06:17 <tdawson> One we have active repos, I'll start working on will-it-build, and some updates to the will-it-install.   But that won't be till next week, possibly after next weeks meeting.
20:06:38 <michel> oh, nice name
20:06:50 <nirik> well, will-it-build is just koscheil right?
20:06:58 <tdawson> Thanks ... it's one of the few names that I like
20:07:09 <michel> what's the last column under install/binary? the one with the smaller numbers
20:07:10 <tdawson> nirik: Actually no
20:07:39 <tdawson> michel That's how many failed.
20:07:52 <davide> this is really cool
20:08:04 <davide> thanks for putting it together tdawson
20:08:11 <nirik> no? I mean I guess we would need something to file the FTBFS bugs, but doesn't koschei already try and build everything?
20:08:18 <nirik> yeah, very nice. :)
20:08:37 <tdawson> nirik:  will-it-build will look at the packages needed to build the package in epel9-next (which uses the cs9 buildroot) and determines if those same packages will be available for epel9.
20:08:58 <themayor> yeah, fantastic work on this tdawson
20:09:11 <tdawson> Thus, I need some packages in the epel9-next repo before I can start working on that.
20:10:12 <nirik> ah I see... it's 'will it buildrequire' or something
20:10:29 <tdawson> nirik Ya ... but ... will-it-build sounds better. :)
20:11:01 <tdawson> Anyway ... there won't be any updates (other than it runs once a day) this next week.
20:11:03 <nirik> well, you don't know it will build... but it will at least have a chance I guess
20:12:45 <tdawson> So, moving on ... pgreco how are the macros coming along?
20:13:11 <pgreco> so, the macros seem to be working, I haven't had much time to work on the packaging we discussed last week
20:13:31 <pgreco> I'm cautiously optimistic for next week :P
20:13:36 <tdawson> Cool
20:14:04 <tdawson> I think the dicussion moved to a bugzilla and/or an issue ... did it get resolved there?
20:14:17 <carlwgeorge> https://pagure.io/epel/issue/77
20:14:35 <carlwgeorge> well, revived an old issue, but yes
20:14:53 <tdawson> Thanks ... I missed putting that in my notes last week.
20:15:45 <tdawson> pgreco: Anything else we need to discuss about it?  Or it is just a matter of implementing it now?
20:16:01 <pgreco> seems like it is just a matter of implementing it
20:16:38 <tdawson> Cool ... I hope your cautios optimism is correct, and you get it done next week. :)
20:17:21 <pgreco> me too
20:17:34 <tdawson> I think that's it for old business, that hasn't moved on to it's own topic.
20:17:40 <tdawson> #topic EPEL-7
20:17:55 <tdawson> Do we have anything for EPEL 7 this week?
20:19:04 <tdawson> Sounds like a no, and I don't remember seeing anything
20:19:07 <tdawson> #topic EPEL-8
20:19:15 <tdawson> How about any epel8 stuff?
20:19:35 <michel> I hear there's a proposal to bring openssl 3.0 in epel?
20:20:04 <tdawson> michel There is, it is being led by davide
20:20:28 <tdawson> davide: Is there any progress on that?
20:20:31 <davide> I sent Michel Alexandre Salim links to the previous discussions on this out of band
20:21:00 <michel> I'm afraid I didn't get round to reading them all yet, but will do after this
20:21:03 <davide> I didn't make any progress, I think the thing to do here is to build a slightly less crappy PoC then the one I made in 5 min and then reach out to the maintainers
20:21:06 <tdawson> If I remember right we are at the "How do we properly do this" stage.
20:21:30 <davide> and see if they're open to refactoring the rawhide spec so it's saner, as otherwise we'd either need to add a million %if, or fork it altogether
20:21:37 <davide> and both are pretty undersirable outcomes
20:21:42 <michel> do we need to solve this generally? e.g. "we need a derived spec based on X, with name renamed from X to XVer"
20:22:04 <davide> IMO the only practical way to do this is if we can maintain it in the rawhide branch alongside the real one, so it's easy to keep up to date and in sync
20:22:17 <davide> but that requires maintainer collaboration
20:22:39 <nirik> if we could solve that generally we could ditch modularity and scls. ;)
20:23:01 <davide> one can dream :-)
20:23:11 <tdawson> :)
20:23:53 <tdawson> davide: Is there anything that we can do to help?
20:23:57 <davide> but yeah, I suspect the "general solution" to this is akin to "how to you parse latex" (you don't, you need latex to compile it for you)
20:24:05 <davide> so I'd probably focus on the specific one for now
20:24:35 <davide> tdawson: I'm mostly constrained by bandwidth, but now that Michel Alexandre Salim is back that's getting sorted out, so I expect we'll have some progress here
20:24:36 <tdawson> *laughs* sorry ... that's the first time I've heard that latex joke.
20:25:04 <davide> we could definitely use a review once we have a PoC of the package, and some support once we go talk to the maintainer
20:25:18 <davide> so this can hopefully come of as "EPEL would like to do this" vs "this rando would like to do this"
20:25:56 <michel> yeah, happy to help
20:26:21 <tdawson> Definatly ... reviews will be good, and I think there are many people that wouldn't mind doing a review, just many that don't want to touch it otherwise.
20:27:00 <tdawson> Anything for EPEL 8 before we move on?
20:27:18 <tdawson> Anything else I mean
20:28:05 <tdawson> #topic EPEL-9
20:28:12 <carlwgeorge> the fedora s390x builder migration is scheduled for next tuesday, which should unblock epel9-next
20:28:18 <tdawson> Ya!!
20:28:25 <nirik> hope it goes smoothly.
20:28:33 <davide> yay
20:29:09 <michel> oh, that reminds me, I need to check if my %limit_build macro that's in Fedora can be included in RHEL 9
20:29:22 <carlwgeorge> i'm hoping to get with releng later this week to go ahead and enable it in bodhi so we can push epel-release and epel-rpm-macros through the pipeline and make sure it works
20:29:23 <tdawson> Let's say it goes smoothly, do we have an estimate of how long it would take?
20:30:19 <nirik> tdawson: a few minutes, it's just powering off on one box and bringing up on another from what I understand. That said. I am doing more things in that outage... updating servers, moving to new koji version.
20:31:04 <tdawson> michel Do you have a bugzilla for your macros?
20:31:30 <michel> for rhel9? not yet, I'm going to install a test system first
20:31:41 <tdawson> nirik So, it does have the potential to be smooth, as well as the potential ... to not
20:31:45 <nirik> https://pagure.io/fedora-infrastructure/issue/10302 is the outage ticket. I need someone(s) to review it to make sure I didn't make any horrible mistakes and will be sending it out likely later today to devel announce.
20:32:02 <michel> need to check if the documentation change for Fedora landed, that would make it easier to point the RHEL people to. esp since some packages like webkit2gtk3 are already using it
20:32:04 <nirik> the big question is networking... hopefully all that works, but we won't know until we try
20:33:19 <tdawson> nirik  I'm hoping it does ... and, they've hopefully been getting better and better at this with each one they do.
20:33:31 * nirik nods
20:34:17 <tdawson> carlwgeorge I sorta missed your comment ... ya, if we can get bodhi going and push those packages through ... I'd appreciate it.
20:34:37 <carlwgeorge> need to anyways to test the pipeline
20:35:15 <tdawson> Yep, I'd hate for all the s390x stuff to finally get done, and then fine out there was a different bug just waiting for us.
20:36:16 <tdawson> Cool ... it sounds like this time next week we'll (hopefully) have alot of progress made.
20:36:31 <tdawson> Or we could be in the middle of fire ... one or the other. :)
20:36:38 <nirik> 🤞
20:36:58 <tdawson> anything else for epel9 before we move on?
20:37:09 <michel> one thing
20:37:49 <michel> looks like tibbs is pushing `%constraint_build` to replace `%limit_build` - https://pagure.io/packaging-committee/pull-request/1130
20:37:50 <michel> currently it requires RPM 4.17 features but he said he'll backport it
20:39:19 <tdawson> Oh my ... that would be so good for a couple packages.
20:39:40 <davide> yeah I could have used that for below
20:40:26 <tdawson> Is that what you want to try to get into RHEL9?
20:40:38 <michel> the big improvement is it overrides `macros._smp_build_ncpus` cleanly, which is the part I didn't know how to do
20:41:06 <michel> yeah, I'll check with tibbs, he mentioned he'll backport it for EPEL and stable Fedora -- anything that works with F34 can go into RHEL9, I take it
20:42:00 <tdawson> michel Cool ... let us know next week how it is going.
20:42:44 <tdawson> I'm going to move on
20:42:51 <tdawson> #topic EPEL-Packaging-SIG
20:43:28 <tdawson> davide: Do we have any Packaging-SIG business this week?
20:43:32 <davide> I have a couple of things
20:43:43 <davide> below, resctl-demo and resctl-bench all made it to stable
20:43:43 * SSmoogen[m] is looking to install CS9 as his laptop OS this weekend once I get openvpn and some other things compiled for it
20:43:58 <SSmoogen[m]> oh and we are already getting systems asking for epel-9
20:44:08 <davide> I wouldn't call them "good examples", but they do provide a blueprint for getting non-trivial rust packages in EPEL with vendored crates
20:44:11 <tdawson> davide: Cool
20:44:17 <SSmoogen[m]> 3 RHEL-9 boxes and 1 Mageia
20:44:23 <davide> also, I was looking at the tracker ticket, and there's a bunch of stuck requests
20:44:38 <michel> Mageia using epel9 ?!
20:44:38 <davide> we should probably have a process to walk over these say once a month or something and triage them
20:44:39 <tdawson> SSmoogen[m]: That doesn't surprise me
20:44:45 <davide> e.g. https://bugzilla.redhat.com/show_bug.cgi?id=1990157
20:44:52 <tdawson> davide: That's a good idea.
20:44:55 <davide> which ended up starting a non-responsive maintainer process
20:44:57 <nirik> rhel10 or bust!
20:45:15 <michel> we do have a saved search on our wiki IIRC
20:45:31 <michel> looks at requests > 2 wks old, since those can go to non-responsive maintainer
20:46:00 <davide> oh nice, I'd missed that
20:46:00 <michel> https://docs.fedoraproject.org/en-US/epel/epel-packagers-sig/
20:46:04 <michel> oh joy the docs transition happened
20:46:05 <davide> I was just looking at dependents of https://bugzilla.redhat.com/show_bug.cgi?id=1914423
20:46:18 <davide> tdawson: gets all the credit for that
20:46:26 <michel> yeah, I think the idea is we look at bugzilla and add the ones we think are important to the tracking bug
20:46:35 <tdawson> Actually, for epel it's a "stalled request" ... because the fedora maintainer doesn't have to do epel stuff.
20:46:51 <davide> sure, my point was we also need a process to go over the ones we added to the bug
20:47:06 <michel> right. uh, do we have a doc for stalled requests like the one for non-responsive maintainer?
20:47:12 <tdawson> Yep, otherwise the bug just sits there.
20:47:13 <michel> I always feel awkward using the non-responsive flow
20:47:21 <davide> tdawson: yeah, I wasn't sure but doing a non-maintainer seems a bit heavyweight here
20:47:26 <tdawson> michel Yes we do ... just a sec
20:47:27 <davide> * non-responsive maintainer
20:47:29 <michel> https://docs.fedoraproject.org/en-US/epel/epel-policy/#stalled_epel_requests
20:47:36 <michel> ^ doesn't have a text template, the non-responsive one does
20:47:49 <tdawson> https://docs.fedoraproject.org/en-US/epel/epel-policy/#stalled_epel_requests
20:48:31 <tdawson> Ah ... well, I've got my standard that I do, and thus far, one one person has been offended by it. :)
20:48:43 <michel> hehe
20:48:43 * carlwgeorge snorts
20:48:55 <michel> if I can steal your text, tdawson, do point me to it
20:49:05 <davide> meh, that's not our problem
20:49:18 <michel> I need to dig up that tool I was working on for doing recursive branch requests
20:49:28 <davide> yeah, let's get this up on a docs page so we can point people to it if/when they complain
20:49:49 <michel> yep. "it's not me, it's the template!" - then they dig up the blame history and found it's you
20:49:50 <michel> haha
20:50:13 <tdawson> OK ... all of mine have gone through, so I don't have any sitting in my waiting list ... I'll get that done for next week ...
20:50:31 <davide> dude I've successfully used this technique at work so many times
20:51:33 <tdawson> Well, the one that was offended ... I don't really think I could have done anything different ... anyway, ya, I'll get that for next week.
20:51:56 <michel> no problem in CS can't be solved by an additional layer of indirection
20:52:18 <tdawson> What do you think about once a month go through the bugzilla list and post it to epel-devel, see how many people there want to take on some of the packages.
20:52:40 <davide> that sounds great, thanks
20:52:57 <michel> +1
20:53:42 <tdawson> Anything else from EPEL-Packaging-SIG ?
20:54:22 <davide> that's it from me
20:54:30 <tdawson> #topic General Issues / Open Floor
20:54:53 <tdawson> Is there anything that didn't fall into the other topics?
20:55:48 <carlwgeorge> https://github.com/rpm-software-management/mock/pull/793
20:55:59 <carlwgeorge> mock upstream is asking us to trim up what we install into chroots
20:56:26 <michel> huh
20:56:30 <carlwgeorge> based on my understanding so far, i think we can get away with just epel-rpm-macros in the mock configs, and have the koji groups install fedpkg-minimal
20:57:16 <michel> doesn't epel-release contain the repo definitions though?
20:57:18 <nirik> I had a quick announcemen....
20:57:35 <tdawson> michel It does, but, do you really need those in mock?
20:57:46 <tdawson> nirik Go for it
20:57:47 <pgreco> tdawson was faster...
20:57:58 <michel> oh true, the repos are hardcoded in the cfg files
20:58:06 <nirik> ansible announced finally when 2.9.x goes EOL... https://groups.google.com/g/ansible-announce/c/kegIH5_okmg/
20:58:38 <nirik> so I will be EOLing the epel ones then barring any changes.
20:59:27 <tdawson> nirik  Thank you for letting us know.
20:59:46 <tdawson> nirik I assume your writting up an email for epel-devel and/or announce
21:00:05 <nirik> yeah, it's on my (now quite long) list of todos
21:00:12 <tdawson> :)
21:01:06 <tdawson> Looks like our time is up.
21:01:33 <tdawson> Thanks everyone for coming, for the good discussions, and for all that you do for the EPEL community.
21:01:40 <tdawson> I'll talk to you next week if not sooner.
21:01:52 <pgreco> see you guys next week!
21:01:52 <tdawson> #endmeeting