20:59:26 #startmeeting EPEL (2022-12-14) 20:59:26 Meeting started Wed Dec 14 20:59:26 2022 UTC. 20:59:26 This meeting is logged and archived in a public location. 20:59:26 The chair is smooge. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 20:59:26 Useful Commands: #action #agreed #halp #info #idea #link #topic. 20:59:26 The meeting name has been set to 'epel_(2022-12-14)' 20:59:32 #meetingname epel 20:59:32 The meeting name has been set to 'epel' 20:59:32 .hi 20:59:33 rcallicotte: rcallicotte 'Robby Callicotte' 20:59:35 .hi 20:59:37 jonathanspw: jonathanspw 'Jonathan Wright' 20:59:42 #chair nirik tdawson pgreco carlwgeorge salimma dcavalca dherrera gotmax[m] smooge 20:59:42 Current chairs: carlwgeorge dcavalca dherrera gotmax[m] nirik pgreco salimma smooge tdawson 20:59:48 #topic aloha 20:59:50 .hi 20:59:51 pgreco: pgreco 'Pablo Sebastian Greco' 21:00:13 I thought this was cancelled? 21:00:22 looks like smooge is in a hurry today starting a few seconds before 21:00:38 no. tdawson is sick so smooge was gonna run it 21:00:59 .hi 21:01:00 carlwgeorge: carlwgeorge 'Carl George' 21:01:03 tdawson, asked someone to run it because he was sick. 21:01:48 .hello yselkowitz 21:01:49 yselkowitz[m]: yselkowitz 'Yaakov Selkowitz' 21:02:07 hello everyone and I hope you are having a good day 21:02:20 this will be a fairly short meeting I think 21:02:26 #topic End Of Life (EOL) 21:02:27 i can fix that 21:02:34 RHEL 7 will go EOL on 2024-06-30 21:02:40 CentOS Stream 8 goes EOL in 2024-05-31 21:02:42 ruh roh 21:02:47 CentOS Stream 9 goes EOL in 2027-05-31 21:03:12 carl has to go and ruin everyone's day 21:03:19 #info RHEL 7 will go EOL on 2024-06-30 21:03:27 #info CentOS Stream 8 goes EOL in 2024-05-31 21:03:33 #info CentOS Stream 9 goes EOL in 2027-05-31 21:03:40 #topic EPEL Issues https://pagure.io/epel/issues 21:04:05 the only item I see for this meeting is https://pagure.io/epel/issue/212 21:04:53 jonathanspw, your ticket, your time 21:05:14 Hmm, first time in this process. Should I have something to say? 21:05:35 i think the ticket covers it 21:06:03 +1 from me, seems straightforward 21:06:05 is there anything outside of the ticket that came up 21:06:05 FWIW there's also a BZ requesting an update because the old version in EPEL7 is from the very early days of novnc and the old version is basically unusable anyway. 21:06:07 .hello ngompa 21:06:07 Eighth_Doctor: ngompa 'Neal Gompa' 21:06:21 +1 from me, super straightforward 21:06:27 old, prerelease version I might add. 21:06:42 are there any members in disagreement with jonathanspw's proposal? 21:07:25 gotmax[m], pgreco salimma 21:07:39 michel-slm ? 21:07:58 ok going once.. 21:08:08 going twice.. 21:08:35 approved by the old man in the chair 21:09:58 any other issues for this meeting? 21:10:01 i just opened an issue tagged as meeting for my thing, but we can save it for the open floor 21:10:08 #topic Old Business 21:10:32 any old business for this week? it was epel-10 last week 21:11:04 smooge: sorry work call got in the way, +1 21:12:01 no problem 21:12:04 #topic General Issues / Open Floor 21:12:21 ok anything for open floor ? 21:12:29 https://pagure.io/epel/issue/214 21:12:42 apologies for getting this in so late 21:13:31 i don't believe there is any kind of waiting period involved, so if possible it would be nice to vote on it today to put it behind us before the holidays 21:13:36 It makes the most sense for neither apptainer or singularity-ce to "Provides: singularity". apptainer has had 9+ months to rename (done improperly) and he's basically trying to squat on the "singularity" name forever. Let the users choose is my view. 21:14:11 i know jonathanspw is already familiar but i'll give everyone else a moment or two to read over the issue 21:14:27 For transparency, I'm the reviewer of singularity-ce and carl and I have been emailing off-list with the apptainer and singularity-ce maintainers for a couple of weeks now. 21:14:29 i thought that items need a week 21:14:31 if this isn't EPEL specific then shouldn't it go to FPC? 21:14:56 the apptainer maintainer agreed to do this change for fedora branches, but doesn't want to for epel 21:15:22 my original feeling was that it would result in a fesco escalation, but he specifically asked for the epel steering committee to vote on it 21:15:35 if we want to pass the buck i'm ok starting a fesco issue instead 21:16:25 ok my understanding from fpc and fesco, things which need a vote need at least a couple of days before they can be voted on 21:17:51 I'm trying to remember if we had specific rules for a waiting period 21:18:08 we do for incompatible upgrades 21:18:49 this to me is an incompatible upgrade inside of an enigma wrapped inside of other things 21:19:04 yeah it's messy 21:19:28 I would like to get tdawson and nirik to look at it but they are both out 21:19:34 as long as we don't cancel next week for the holidays I'm ok waiting a week to vote 21:20:01 I don't think there is a problem with voting for/against it in the ticket and getting a majority there 21:20:28 of course what is a majority for this group I don't know 21:20:50 so I have some questions on this 21:20:58 shoot 21:21:15 is the removal of the provides for just EPEL7 or all releases or all releases after EPEL7? 21:21:29 shoot 21:22:07 I would like to see all releases, and the maintainer has agreed to do it in fedora releases 21:22:18 does the lack of provides potentially break anything in scientific linux 21:23:00 this is more of a 'smooge can answer this at 0520 when CERN people go OMG' 21:23:05 the only thing that no longer works would be `dnf install singularity`, which would be intentional 21:23:45 ugh this is such a mess 21:24:00 but yeah, I agree not having either be "singularity" is a good outcome 21:24:03 `dnf search singularity` will show both options 21:24:03 understood but I would like to make sure that a super collider doesn't explode because it can't find a singularity 21:24:22 my read of the situation is that both solutions are incompatible with the package known as "singularity" 21:24:26 smooge: the upgrade path is taken care of by a (proper) obsoletes that will remain in apptainer 21:24:28 ... cymbal ... 21:24:38 so even worse situations would come up if someone upgraded into it 21:25:26 singularity-ce is actually more compatible than apptainer 21:25:54 but they don't want to fight over the provides 21:26:38 Eighth_Doctor: we're already there with users upset over the apptainer replacement 21:26:47 ffff 21:27:20 I don't think apptainer dropping the obsoletes would be so bad either at this point. What's done is done for a lot of folk but why continue letting the breaking change exist against policy? 21:27:52 I feel like if this had followed the incompat process we would have declined the obsoletes 21:28:10 I think we should request the obsoletes+provides be removed entirely 21:28:16 stem the bleeding 21:28:54 yep that's what makes the most sense to me 21:29:00 ok could i get a clear proposal in the ticket of what is being asked for? Add it below the main wall of text 21:29:46 the goal of the ticket is to remove the provides, but i can add on the stretch goal of removing the obsoletes too. be aware that would mean people that haven't upgraded yet stay on the singularity package unmaintained. 21:30:29 Proposal: apptainer is to remove the Provides: singularity from its EPEL packages. or something else. 21:30:46 that is what it currently says, yes 21:30:55 I would prefer that if there is a stretch goal it is for EPEL and Fedora and not just EPEL 21:31:11 I believe this is already resolved for Fedora isn't it? 21:31:15 yes i'm not keen on doing a different solution in different branches 21:31:20 the maintainer just wants us to force the matter for EPEL 21:31:37 (he won't do it unless we say he should) 21:31:44 Eighth_Doctor: he agreed to drop the provides, not the obsoletes for fedora. 21:31:44 Eighth_Doctor: not merged yet https://src.fedoraproject.org/rpms/apptainer/pull-request/1 21:32:13 oh goody, new commits, i need to rebase 21:32:14 * Eighth_Doctor sighs 21:33:03 is everyone ok with leaving the discussion in the issue and not introducing it on list, as only steering committee maintainers votes count for this? 21:33:13 oh wow, this thing is a messy spec file 21:33:14 yes. I am 21:33:19 yes 21:33:28 The apptainer spec violates about 10 packaging guidelines. 21:33:36 * Eighth_Doctor headdesks 21:33:47 we should really document our exact voting timeline 21:34:04 what does fesco do, a week with no -1 and at least x number of +1s? 21:34:26 I think that sounds right 21:34:34 no wait i think that's for becoming a package sponsor 21:34:36 at least 2 +1s 21:34:41 and 0 -1s 21:34:44 after one week 21:34:48 ^ 21:34:57 if there's only 1 +1, extend another week 21:35:11 I would like to say no -1's and 3 +1, as we do not vote for members like FESCO 21:35:12 if it stays there after another week, accept 21:35:22 ok so we technically don't need next week's meeting then, just vote in the ticket 21:35:40 obviously if anyone -1s it then we can discuss further or outright decline it 21:35:44 if there are -1's in the ticket on next Wednesday, it would be moved ot the next meeting 21:36:09 i'm saving most of my pto for january, so i'll be around next week 21:36:13 otherwise it is considered approved and will be 'formal' when the next meeting occurs 21:37:35 is that ok for everyone? 21:37:38 should i file a separate issue for the obsoletes removal? or tack it onto this one? 21:37:59 i kinda like the idea of separate votes to make everyone's intentions clear 21:38:21 I would prefer seperate votes AND I would prefer if it was clear why we were differing from FESCO 21:38:52 are we differing from FESCo? 21:39:01 it wasn't a problem in Fedora 21:39:03 nobody brought it up to FESCo, the packager agreed on his own for Fedora 21:39:05 just not EPEL 21:39:08 ah ok 21:39:14 the maintainer is willing to make the change in fedora branches, but doesn't want to in epel branches 21:39:33 any justification for not doing it in epel? 21:39:37 this is purely to recommend the maintainer to do the same in EPEL that he's doing in Fedora 21:39:59 "If you wanted to do it only on Fedora, that's fine with me; it's EPEL that I care the most about." 21:40:23 so no actual justification to diverge 21:40:23 ha finally a real packager 21:40:32 blech 21:40:34 ruuuude 21:40:38 hmm 21:40:56 i'll avoid commenting further on that part to not taint the vote 21:41:32 anyway.. I am going to ask that people read, and possibly explain why they think it is a good idea for this in the ticket 21:42:03 sounds like a plan 21:42:24 thanks everyone for your input 21:42:32 beyond that is there any other New Business? 21:43:54 #topic Next Meeting? 21:44:28 OK we can have the next meeting on 2022-12-21 or 2023-01-04 21:45:08 i'll be around on both those dates 21:45:11 same 21:45:31 yeap, same here 21:45:44 not the "middle" week though 21:45:48 ok I am going to say next meeting for 2022-12-21 but it may be a very short one 21:46:07 i can also run the meeting if troy or smooge aren't here 21:46:09 #info Next meeting will be 2022-12-21 but may be limited 21:46:12 sounds good to me 21:46:34 #info carlwgeorge may run if troy is not available 21:47:06 ok thank you all for coming to this weeks meeting. 21:47:14 #endmeeting