22:00:15 #startmeeting EPEL (2021-03-12) 22:00:15 Meeting started Fri Mar 12 22:00:15 2021 UTC. 22:00:15 This meeting is logged and archived in a public location. 22:00:15 The chair is tdawson. Information about MeetBot at http://wiki.debian.org/MeetBot. 22:00:15 Useful Commands: #action #agreed #halp #info #idea #link #topic. 22:00:15 The meeting name has been set to 'epel_(2021-03-12)' 22:00:16 #meetingname epel 22:00:16 The meeting name has been set to 'epel' 22:00:18 #chair nirik tdawson bstinson pgreco carlwgeorge michel_slm 22:00:18 Current chairs: bstinson carlwgeorge michel_slm nirik pgreco tdawson 22:00:19 #topic aloha 22:00:24 * carlwgeorge waves 22:00:36 morning 22:00:55 Hi carlwgeorge 22:00:57 Hi nirik 22:01:32 hey, it's that time again! 22:01:33 hello everybody 22:01:46 Hi pgreco ... yes, it is that time again. 22:01:49 here 22:01:51 Hi dcavalca 22:01:52 .hello salimma 22:01:53 michel_slm: salimma 'Michel Alexandre Salim' 22:01:54 Hi smooge 22:02:02 Hi michel_slm 22:02:04 * michel_slm mostly lurking 22:03:18 Hi all 22:03:23 Hi rsc 22:03:25 .hello robert 22:03:26 rsc: robert 'Robert Scheck' 22:03:49 We've got a full house this week. 22:04:58 #topic Old Business 22:05:06 we are all getting old 22:05:12 *laughs* 22:05:22 Yes we are 22:05:34 Let's start with the EPEL Packaging SIG 22:05:56 dcavalca and/or michel_slm Do we have any news/progress on that? 22:06:36 That collaborator requesting branches patch was merged I think... just need a new release of the tool, which hopefully will be next week. 22:06:48 Ya!! 22:06:54 Yay 22:07:16 tdawson: michel_slm got the stronswan build for EPEL8 updated, which was nice 22:07:22 * strongswan 22:07:54 Good, cuz we don't need weak swans 22:08:08 *sigh* I think smooge is rubbing off on me ... just not as well. 22:08:12 tdawson: go to the corner, bad tdawson ;) 22:08:43 That package is not officially maintained by the sig yet, i wonder if we've managed to get in touch with the maintainers 22:09:13 dcavalca: michel_slm: Do you know if there are other packages that the SIG can/should pick up? Something we will be able to test out the new patch with? 22:09:37 wish we knew what was popular... ;) 22:09:55 I have a few open tickets that I hadn't checked in a while, one sec 22:10:10 bridge-utils: https://bugzilla.redhat.com/show_bug.cgi?id=1841942 22:10:19 python-tmuxp: https://bugzilla.redhat.com/show_bug.cgi?id=1908449 22:10:23 yeah, looking at requests that are stale would be good. 22:10:52 pretty sure I had more but bugzilla is failing me right now 22:10:59 The wiki has a search query for stale requests 22:11:15 (On my phone now so I can't easily fish it out) 22:11:20 Two should be good to start with. 22:11:43 erlang-rebar: https://bugzilla.redhat.com/show_bug.cgi?id=1749520 22:11:55 Plus, we still have to wait for the tool to be deployed. 22:12:03 elixir: https://bugzilla.redhat.com/show_bug.cgi?id=1752945 22:12:31 t1lib: https://bugzilla.redhat.com/show_bug.cgi?id=1891943 22:13:12 Elixir will be interesting. Is erlang already in? 22:13:32 I think the answer is yes, but do we have the steps documented for the SIG bringing in a package? 22:13:47 https://paste.centos.org/view/1da6a0b8 22:13:48 erlang-22.0.7-1.el8.src.rpm 22:13:53 fyi, if you file an epel request against rawhide, it will eventually be moved to a branch version and then later closed when that branch is eol 22:14:13 yeah, I've just noticed that 22:14:16 carlwgeorge: in some cases there is no epel8 in RHBZ ;-( 22:15:12 you can avoid that with the futurefeature keyword, or by using the fedora epel product (not always possible) 22:15:12 the two I linked were never in EPEL, so it doesn't show up at all 22:15:12 you can file against epel only if the package happens to also be in epel6/7 22:15:12 rsc: i know 22:15:12 Yeah, a problem the first time a package is ever requested for epel 22:15:20 Perhaps we could get a template and get people to use that. 22:15:25 that's what i'm saying, if it's never existed in epel before use rawhide but add the futurefeature keyword so it's excluded from that automation 22:15:29 python-typeshed: https://bugzilla.redhat.com/show_bug.cgi?id=1913141 22:15:49 I was wondering earlier this week on how to find a list of all packages which had EPEL branches 22:16:18 could query bz or src.fp.o I would guess... 22:16:27 I still have a template on my plate of things to do, but I never seem to get down to doing it. 22:17:25 I'm hoping next week things will be calm enough for me to do it. 22:18:14 here's hoping for quiet. :) 22:18:38 It should be .. after Tuesday 22:18:55 Anything else for the SIG? 22:19:03 * michel_slm is on diaper duty, it's never quiet 24/7 22:19:13 *laughs* 22:20:17 OK, let's move on to epel8-next 22:20:31 carlwgeorge: Any progress on this? 22:20:41 https://pagure.io/fedscm-admin/pull-request/63 22:21:13 i decided to move this forward i'm going to skip the pkg_cs8.json file check for now, since it's not clear how to structure that 22:21:46 cool. hopefully we can get that in next week's release. 22:21:51 we can always add it later, and in the meantime the pkg_el8.json file covers the vast majority of the things we need to block 22:22:22 yeah when mohan is back next week i'm going to see if i can schedule something with him to start making the koji/bodhi changes in prod 22:22:31 .hello ngompa 22:22:31 Eighth_Doctor: ngompa 'Neal Gompa' 22:22:45 i'm doing a few more touch ups on fedpkg which will be added onto https://pagure.io/fedpkg/pull-request/429 22:22:54 Hi Eighth_Doctor and King_InuYasha 22:23:02 which is also easier by not checking for pkg_cs8.json 22:24:27 do we have an agenda item for this "epel for eln" concept? if so i'll wait till then to bikeshed about epel-next naming. otherwise, that's what i'd like to talk about. 22:24:56 Nope, do you mind if we wait for open floor? 22:26:27 works for me 22:26:55 OK, I've put it on the open floor area, you'll be first when we get there. 22:27:05 Anything else for -next ? 22:28:10 not from me 22:28:22 Next on agenda is documentation 22:29:02 At least from my side, there hasn't been any progress, but I'm leaving it on the agenda for at least one more week. 22:29:31 i haven't made any progress on rewriting that wiki page to the fedora format 22:30:04 it's not hurting or blocking anything, so if you'd like to remove it from the agenda until i can bring it back up, that's fine by me 22:30:56 And I think that brings us to the topic that I think most people are here for, a new time for the meeting. 22:32:20 just in time for the dst hellscape 22:32:25 I wish there was a clear "everyone likes this time" But we do have several times with 4 votes, that also correspond with only a couple negative votes. 22:32:36 :) Yep 22:32:43 do you have the link handy for that tool? 22:32:53 https://framadate.org/yNIptJFnWuUR2OME 22:33:09 meeting times are so difficult with lots of people. ;( 22:33:36 Right before the meeting, I realized I hadn't changed anything into any US time zones 22:34:45 tdawson: you have nothing checked? 22:35:12 nirik: I have some negatives 22:35:41 But I looked at my schedule and was like ... anytime but these times work for me. 22:35:42 pgreco: no way you could do wed at 18? 22:36:10 nirik: every other week I have a company meeting 22:36:38 so I can do half the time I guess, and half at 50% ;) 22:36:47 if that's the consensus I'll manage 22:39:53 It's looking like the 2100 is the most popular ... for both Tues, Wed and Thur 22:39:53 well, also 21UTC on tuesday/wed/thru... except for smooge 22:39:54 And King_InuYasha 22:39:54 yeah. 22:39:54 Although, anytime but this time is bad for King_InuYasha 22:39:54 smooge: can you do anything around 21utc? or too late? 22:39:55 i think King_InuYasha is a lost cause, his only possible spot is on a friday, and even that's a maybe 22:39:55 2100 is ok 22:39:55 Eighth_Doctor: you should do something about that King_InuYasha guy 22:39:55 So if we do 2100, which day would people prefer? 22:39:55 I can do it on tues/wed/thurs. 22:40:08 * nirik is fine with any of them 22:40:09 all 3 days are meetings from end to end for me so I don't care 22:40:12 2100 on t/w/t all work for me 22:40:30 wed is my most meeting day... so that might be best. 22:40:35 so it's just one more. ;) 22:41:02 I'm fine with Wedensday, is everyone else? 22:41:39 Wed should work for me on most weeks 22:41:46 Wed 21:00 UTC, which in fact it will end up being 20:00 UTC after DST ? 22:41:56 :) 22:42:09 yeah. 22:42:13 death to DST! 22:42:50 So, are we ready to vote. Wedesnday at 2100 22:43:13 (which will be 20:00 UTC next week) 22:43:17 +1 22:43:26 +1 22:43:27 +1 22:43:32 +1 22:45:05 dcavalca: michel_slm: smooge: Ya, Na ? 22:45:13 +1 22:45:37 +1 22:45:38 michel_slm, is probably dealing with baby 22:46:04 true, i'll put him down for absent 22:46:30 #info New meeting time passed - Wednesdays at 2100 (moving to 2000 next week) 6 agree 0 disagree 1 absent agree 22:47:05 I'll send out an email, and work on getting the Fedora calendar setup for the change. 22:47:29 Anything for EPEL 7 or 8 before we move to Open Floor? 22:48:28 #topic General Issues / Open Floor 22:48:48 carlwgeorge: you wanted to talk about "epel for eln" 22:49:54 i don't have all the details (perhaps dcavalca or michel_slm would like to summarize) but as i understand it they want to create an "epel for eln" to ensure more fedora packages are ready for the next major version of rhel. that has some pretty clear naming confusion with epel-next. 22:50:11 i think i even saw mention of an epeln tag 22:50:25 there was some discussion about this on the devel list... but I didn't think there was a concrete proposal out of it yet? 22:50:32 Is that in an email somewhere? I haven't caught up on my fedora-devel email. 22:50:46 that's interesting, a lot of work, but interesting 22:50:48 nirik: it was discussed a bunch during the ELN meeting this morning 22:50:52 there's details in the minutes 22:50:58 are there any plans to make eln a "deliverable" product? 22:51:04 carlwgeorge's summary is correct 22:51:11 Wait ... there's an ELN meeting? 22:51:18 eln sig, yes 22:51:23 yeah, I saw that in passing, but... someone should make a real proposal out of it so we can see if it's going to work out/could be made better. 22:51:40 when i proposed epel-next i was adamant about not calling it epel-stream. this is mainly because after cl8 is gone, and centos stream is the only centos, i'd like to eventually champion for dropping the stream. that will be harder if we go with epel-stream. 22:52:36 but if eln is going to cross over into epel content, then i don't think i have a choice but to call it epel-stream. if that's the case i can adjust fedpkg/fedscm-admin code now pretty easily, and we haven't done anything in prod koji yet. 22:53:06 regardless of the ELN thing, IMHO epel-stream is a lot less confusing than epel-next with the currents status quo 22:53:10 call it epel-playground for even less work 22:53:15 I personally don't see epel eln going anywhere, for starters, it's already building for RHEL10 ... so I wouldn't worry about it too much. 22:53:38 i don't see the point of epeln, as one could just use rawhide if they want those packages 22:53:46 tdawson: yes, epel eln would target whatever eln targets (which is 10 right now) 22:53:49 so, this would allow maintainers to build their rawhide package for epel10 before it exists to test things? or ? 22:54:18 nirik: that, plus allow people that want to deploy ELN to test things out to use (a subset of) EPEL 22:54:36 So ... they are planning on maintaining them for 3 years before maintaining them for 10 years? 22:54:50 a package being in eln isn't a guarantee of it making it into the next rhel, so why not just add more packages to eln? 22:55:37 carlwgeorge that would solve the problem too, yes 22:55:45 carlwgeorge: yes, that would also work 22:55:53 Anyway ... why would epel9-next be confused with epel-eln ? 22:56:09 carlwgeorge: so could all packages be added to eln? 22:56:17 no 22:56:23 tdawson: isn't eln short for enterprise linux next? 22:56:31 rsc: very likely not 22:56:33 rsc: no idea, i'm not involved in eln 22:56:35 Actually, no. It's sort for ELN 22:56:45 epel-eln would be unversioned in fact 22:56:48 well, they could, but that would cause more problems too. 22:56:52 and even for EPEL the discussion so far would be to just do this for a subset, likely opt in, at least to begin with 22:57:21 fwiw I don't think anything around this is set in stone at this point 22:57:34 and I'd encourage folks to read the minutes from this morning if you're interested 22:57:36 dcavalca: if there is still no guarantee that eln packages make it into next RHEL...why? Or is ELN still a Enterprise Linux Next, even if "declined"? 22:57:40 no one wants all of rawhide dumped into eln, but i think it's logical to include future epel candidates into the eln package set 22:57:53 I'd prefer if everyone involved discusses a plan and comes up with one. 22:58:08 OK, something just clicked in my brain. So you would have the automatic builds of EPEL packages, just like you do ELN packages. 22:58:09 * nirik also again notes "epel" is not a specific set of packages. 22:58:17 so my main issue with this is that the ELN builds and triggered rebuilds cause enough havok with the regular Fedora builds. I would like to see a formal request and an idea of resources to make it happen 22:58:38 And, as nirik just noted ... EPEL isn't a specific set of packages, while ELN is. 22:58:43 yep, this is definitely contingent on sorting out resources 22:58:55 and yes, I expect an actual proposal will come out once things are fleshed out a bit more 22:58:59 we don't need to solve it today, i just want to think about it before we push forward with epel-next stuff and have to do it over as epel-stream 22:59:25 my biggiest concern here is annoying maintainers actually... but eln already has that problem somewhat. Hopefully we can solve that if we ever re-write FMW 23:00:08 tdawson: "But the interpretation of the name evolved, and now you can think about it as Enterprise Linux Next if you like." 23:00:14 https://docs.fedoraproject.org/en-US/eln/#_name 23:00:17 My opinion, stay with epel-next, and if needs be, incorporate ELN into epel-next ... epel8-next epel9-next epel-eln-next 23:00:43 Oh ... El Nino still lives. :) 23:00:50 nirik nirik nirik, we don't do that, this is fedora we only add new brand shiney new things and releases ontop of the old rotting infrastructure. 23:01:09 that would be another option... make epel10-next now, builds against eln. 23:01:26 that...actually makes sense 23:01:46 nirik: that's similar to what I proposed in the list, yeah 23:01:52 Ya ... that does make sense 23:01:57 yeah I could see that working 23:01:59 but branching/maintaining your package there does mean you have to do so for 3 more years... but for some things people may not care about that 23:02:08 and then switch epel18-next to cs10 when available 23:02:32 but also... yeah, what carlwgeorge said (assuing you meant epel10-next) 23:02:49 i have no idea how i typed 18 there, whoops 23:03:00 must be friday! 23:03:15 and 5 oclock somewhere (here!) 23:03:34 Yep, I let us go a little long cuz the conversation was reaching a conclusion 23:03:49 Thank you all for being here today, and for the good conversation. 23:03:51 there's a pandemic, i have nowhere to be :D 23:03:58 I'll talk to you all this wedensday 23:04:11 #endmeeting