21:01:01 #startmeeting EPEL (2021-11-10) 21:01:01 Meeting started Wed Nov 10 21:01:01 2021 UTC. 21:01:01 This meeting is logged and archived in a public location. 21:01:01 The chair is tdawson. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 21:01:01 Useful Commands: #action #agreed #halp #info #idea #link #topic. 21:01:01 The meeting name has been set to 'epel_(2021-11-10)' 21:01:01 #meetingname epel 21:01:01 #chair nirik tdawson bstinson pgreco carlwgeorge michel dcavalca 21:01:01 #topic aloha 21:01:01 The meeting name has been set to 'epel' 21:01:01 Current chairs: bstinson carlwgeorge dcavalca michel nirik pgreco tdawson 21:01:05 morning 21:01:13 .hi salimma 21:01:14 michel: michel 'None' 21:01:31 .hi 21:01:32 dcavalca: dcavalca 'Davide Cavalca' 21:01:37 Hi nirik 21:01:39 Hi michel 21:01:43 Hi dcavalca 21:01:55 Hi SSmoogen[m] :) 21:02:18 gekki 21:02:30 * dcavalca is going to multitask with another meeting for the first half hour or so 21:02:31 .hi 21:02:32 carlwgeorge: carlwgeorge 'Carl George' 21:02:40 hi all 21:02:41 gekki? 21:02:46 Hi carlwgeorge 21:02:51 .hi 21:02:52 c4t3l: Sorry, but user 'c4t3l' does not exist 21:02:58 :( 21:03:06 hello with my fingers off by one it looks like 21:03:12 * michel finishing lunch, excuse any typo 21:03:13 Hi non existant c4t3l 21:03:24 .hello robert 21:03:25 rsc: robert 'Robert Scheck' 21:03:27 howdy 21:03:33 Hi rsc 21:03:46 .hi rcallicotte 21:03:47 c4t3l: Sorry, but user 'c4t3l' does not exist 21:05:16 c4t3l: I'll assume you are rcollicotte :) 21:05:23 yes 21:05:44 Welcome ... are you here for something specific, or just here to see what's up? 21:06:15 * carlwgeorge waves at c4t3l 21:06:37 carlwgeorge invited me 21:06:46 * c4t3l waves back 21:06:48 Ah, cool. 21:07:06 #topic Old Business 21:07:07 c4t3l mentioned to me in email that he's interested in getting involved in epel and fedora 21:07:25 More hands are always welcome 21:07:41 Gonna zip through old business rather fast. 21:07:43 c4t3l: welcome! 21:08:02 pgreco isn't able to make it. He made very little progress on the macros this past week. 21:08:15 michel: thank you 21:08:44 willit (will-it-install, will-it-build) didn't have any work happen to it this past week. 21:09:23 I believe the other old business (openssl3 and epel9-next) are in the other topics ... 21:09:23 I have an update for openssl3 but it can wait until the epel8 topic 21:09:40 Yep 21:10:33 Actually, I had it in the Packaging SIG topic ... but we can do it on #epel8 too ... it fits both. 21:10:40 #topic EPEL-7 21:11:13 Did we have any epel7 topics we needed to bring up? 21:12:19 I'll take the staggering silence for a no :) 21:12:25 #topic EPEL-8 21:12:41 Go for it michel 21:12:56 whee! 21:13:09 so I posted this to the mailing list yesterday (let me grab the link in a sec), but TL;DR 21:13:58 openssl3 is out, it's in Rawhide and CentOS Stream 9 but not in Fedora stable releases or CS8 21:15:15 dcavalca had an idea of just making the openssl srpm generate openssl3 packages on some branches, so we can maintain this in a single package. I did a simple POC and it seems to work, so I posted asking if this is ... 'kosher' or not. basically using `bcond_with` / `bcond_without` so it's easy to toggle, and if set defines a suffix variable that get expanded in %{name} 21:15:34 (I've not added conflicts with the base package since I wanted to check before we get that far) 21:15:58 Conan Kudo thinks this might be against policy and we need a separate SRPM, and that's my suspicion too but I've not found any written policy to that effect yet 21:16:35 #info proposal to generate openssl3 from openssl spec - https://lists.fedoraproject.org/archives/list/epel-devel@lists.fedoraproject.org/thread/TGPN4J334S4JNXX5I7A3IOXFMCBSD35V/ 21:16:48 commit showing how this is done: https://src.fedoraproject.org/fork/salimma/rpms/openssl/c/712fbe3ddd4ad02e55bdeca4aa914dcb147abcc5?branch=openssl3-for-epel 21:16:57 we can't use `openssl` SRPM name anyway, since it'll block all openssl RPMs from RHEL base 21:17:11 you mean a openssl.spec that makes openssl3 binary packages? 21:17:23 yes, thats a no no as Eighth_Doctor says. 21:18:09 You can't just do a branch of openssl, it has to be a seperate dist-git repo called openssl3 ... and, for the reasons everyone has already said. 21:18:10 we can make the openssl spec have conditionals to do this properly, so that it's a trivial sync, but the filename and the package names have to be adjusted 21:19:21 yeah, I think this is a no, but for clarity - where's this documented? I take it it's a "spec name must match srpm" 21:19:41 Koji IIRC fails if spec file name != srpm name. I tried that years ago myself. 21:20:47 ah ok. I got it working earlier because I generated the srpm for Fedora then rebuild for EPEL, so the spec + srpm are openssl and it only became openssl3 when rebuilt 21:20:56 let me try real quick but I suspect everyone is right 21:21:00 koji operates on source packages for a lot of things... if there's foo in several repos it uses foo from one repo and ignores foo (and everything it produces as a binary package) in others. 21:21:55 ok, next question: since this is for EPEL, should I take the cs9 spec as "upstream" or the Rawhide spec? I'll just add it as a new Git remote and periodically merge, so it's as trivial to use one or the other 21:22:27 so with openssl3 SRPM containing openssl.spec, mock is happy but yeah I would not want to risk having weird breakages in Koji 21:22:55 michel: take careful note of usage of `%{name}`, changing the Name field changes that macro and can have unintended side effects 21:23:27 i think using the cs9 one would be more sustainable long term 21:23:33 carlwgeorge: yup, I had to change the `%setup` line to not use `%{name}` 21:23:39 If it were me, I'd use the cs9 spec. I suspect it will change less, thus have less updates. 21:23:45 carlwgeorge: +1 21:23:55 but if we're maintaining this in a new package I'll just change the name in a separate commit 21:24:36 tdawson: that's what I figure too, make sense. so that way, we might as well carry the name change downstream first, and only try to get it changed in CS9 if merging is too painful 21:25:21 alright, that's it then. I have a plan to go forward. Will also check in some documentation on how to update this package when we have a repo. ah, anyone here interested in reviewing this? 21:26:13 * tdawson notices everyone take a step back. :) 21:26:28 I'll help out obviously 21:26:41 more reviewers would definitely be appreciated :) 21:26:42 where does one review it? 21:27:31 probably as PRs in dist-git? or just pastebin 21:27:36 michel: what works best for you? 21:27:38 * SSmoogen[m] got back from the dentist so a bit woozy on details at the moment 21:27:38 SSmoogen: I don't have anything ready to review yet, but I'll ping the epel IRC channel and post to epel-devel when done 21:28:16 oh, to review before I do an official review request? yeah since this is going to be a new repo anyway... let me see, can I do a fork in gitlab? 21:28:35 yep, gitlab has forks 21:28:53 yeah, that's easiest then 21:29:09 Depending on how busy I am when you are ready for a review, I can look at it too. 21:29:25 tdawson: i've also got an epel8 thing when we're ready to move off openssl3 but before epel9 21:29:51 carlwgeorge ok. I was just about to ask if anyone had any more epel8 stuff 21:29:57 I just checked, yeah this works. I'll put up changes in https://gitlab.com/michel-slm/openssl 21:30:06 michel are you all set? Care if we move on? 21:30:12 oh, I should rename it top openssl3 21:30:14 yeah, let's move on 21:30:18 thanks all 21:30:21 carlwgeorge go for it 21:31:47 openldap-servers was removed from rhel8 a while back, and there has been discussion about doing an "extracted subpackage" for epel8 21:32:15 but good news, it won't be necessary, openldap-servers is getting added back to rhel8. it's already back in c8s. https://bugzilla.redhat.com/show_bug.cgi?id=1972601 21:32:31 Oh, nice 21:32:56 well that'll make things easier 21:33:05 are we going to have it in c9s as well? 21:34:04 and will it get added in 8.5 or do we have to wait for 8.6? 21:36:18 dcavalca: not sure about 9 yet, the justification for adding it back to 8 was that it being there before and being removed caused upgrade issues with the clients package, because the rhel cdn doesn't remove stuff. seems easy for them to avoid that in 9 by never publishing it. 21:36:34 michel: looks to be coming to 8.5, but it's not there yet 21:36:54 ah, so I guess we might end up having to do the split package in EPEL9 after all :) 21:37:04 if it won't be in cs9 at all we'll just need to have an openldap-epel package I guess 21:38:06 sure, but doing it in just epel9 is easier than both epel8 and epel9 :D 21:38:43 yup, good news 21:39:09 anything else for EPEL8 before we move on to EPEL9 ? 21:39:30 not from me 21:39:47 #topic EPEL-9 21:40:00 nirik carlwgeorge How is epel9-next looking ? 21:40:27 nirik do you want to talk about the s390x thing first? 21:40:32 same as ever. Sadly the s390x move didn't go well, so we had to revert... going to try again next tuesday and hopefully things will be better. 21:41:12 What was the issue? 21:41:13 :( 21:41:42 in the meantime, we did get epel9-next added to bodhi, and pushed epel-release/epel-rpm-macros (noarch packages) through to testing to make sure it works. someone from releng unpushed it though, i'm not sure why yet. 21:42:06 rsc: https://pagure.io/fedora-infrastructure/issue/10302#comment-761270 21:42:12 The z/vm lpar wouldn't work because the z15 mainframe needed some errata applied. The kvm lpar moved just fine, but we couldn't get the network to come up. 21:43:13 * dcavalca hadn't realized linux on s390x runs on top of z/vm 21:43:19 it can 21:43:21 or kvm 21:44:48 carlwgeorge: huh, should we ask the person who unpushed it? 21:45:36 already did, waiting to hear back. mboddu said to go ahead and push it again though. 21:45:46 well, submit it to be pushed again 21:47:18 Well, it's sad that we have to wait another week ... but that's also why we said our goal is Dec. 1st ... cuz these things happen. 21:47:36 Anything else for EPEL9 before we move on? 21:47:44 hopefully we can be done in time to be part of the c9s launch announcement 21:48:13 nope, we can move on 21:48:33 Yep, I hope we are able to get in with that annoucement ... and ... moving on. 21:48:51 #topic EPEL-Packaging-SIG 21:49:37 So, I was suppose to go through my bugs and develop a template for requesting packages, as well as stalled package tickets 21:49:46 https://paste.centos.org/view/f56665df 21:50:15 I haven't got it in a documentation pull request yet ... but what do ya'll thing of that ^^ 21:50:23 looks nice but you probably want to add the requester's FAS to the template? 21:50:44 thanks I need to do this for perl-Net-Telnet and perl-Config-Grammar for the smokeping 21:50:54 michel Instead of "I" 21:51:02 Oh, and I see a typo already ... 21:51:12 also maybe we need to be specific about what access we need. e.g. collaborator is actually enough but in which case we should probably also request they add epel-packagers-sig as collaborator on epel* branches 21:52:35 Very good point. I actually had that in my original ones, but for the kde-sig ... I can add that fairly easy. 21:53:17 at some point the old package database let people just request access -- any idea if we can get that rewritten for Pagure? 21:53:49 I've had a branch request / comaintainer offer get a reply of "how do I add someone again?" - apparently it's not obvious 21:54:34 I don;t think anyone who can do that has cycles to work on that... at least in short term 21:54:49 we could add to the note about how to add someone? 21:54:58 It's not as obvious as it was before ... but ya, what nirik said. 21:56:10 That's good too ... I've put those three things on my notes. 21:56:25 I'd update the pastebin, but I can't type that fast. 21:57:42 Looks like we are running out of time ... I'll leave that subject for now. 21:57:54 #topic General Issues / Open Floor 21:58:38 Anything else that needs to be brought up? 22:00:09 Thank you all for coming. And thank you for all your work on EPEL. I appreciate all you do, and I know others do as well. 22:00:21 Talk to you next week, if not sooner. 22:00:28 thanks tdawson ! 22:00:29 thanks tdawson 22:00:34 thank you ! 22:00:34 #endmeeting