18:01:13 #startmeeting EPEL (2019-06-19) 18:01:13 Meeting started Wed Jun 19 18:01:13 2019 UTC. 18:01:13 This meeting is logged and archived in a public location. 18:01:13 The chair is smooge. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:01:13 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:01:13 The meeting name has been set to 'epel_(2019-06-19)' 18:01:13 #meetingname epel 18:01:13 The meeting name has been set to 'epel' 18:01:13 #topic Chair and Introductions 18:01:13 #chair bstinson Evolution nirik smooge pgreco tdawson 18:01:13 Current chairs: Evolution bstinson nirik pgreco smooge tdawson 18:01:23 * tdawson is here. 18:01:23 OK now I am doing this meeting right this time 18:01:36 * bstinson waves 18:01:40 morning 18:01:42 hello 18:02:26 hello all again\ 18:02:51 So this will be an open floor meeting as I pretty much have been bogged down on other things 18:03:34 ok, I have 1 thing in mind 18:03:34 The topics from last meeting were sgallagh's EPEL rawhide/wagonwheel concept #topic EPEL Rawhide (take 2) 18:03:34 #info https://lists.fedoraproject.org/archives/list/epel-devel@lists.fedoraproject.org/message/FYOOSAJOTDNCE4RLF32ORFO6WL2M5WD5/ 18:03:51 and EPEL-8 status 18:03:59 pgreco, you can go first 18:04:05 #topic Open Floorish 18:04:15 just note it for later, 7.7beta/python 3.6 18:04:26 let's start with epel-8 18:04:28 .hello2 18:04:30 sgallagh: sgallagh 'Stephen Gallagher' 18:05:47 #topic EPEL-8 status 18:06:29 from last week, We have grobisplitter splitting out repos daily and staging koji pointed to it 18:07:36 mboddu, got all the items for building working in staging koji and we were able to build a couple of pakcages 18:08:07 tdawson, discovered a problem where koji was kicking out some packages from the buildroot due to mergerepo 18:08:48 mboddu, reset the rhel-8 to be seen as a bare mode versus other merge modes, then discovered a problem in koji where it was applying blockers and failing if we were in bare mode. 18:09:12 mikem then worked out a patch and we need to apply the patch to koji and see if that fixes the issue 18:09:23 nirik, does that sound about right? 18:09:27 oh, fun days..... 18:09:59 yep. sounds right. I am building that patched koji now. 18:10:29 we will be working through these for the next couple of days and hope to get a set of test builds done and then a 'compose' out the door 18:10:52 well, a compose tested... not sure we want to push the stg composes anywhere public 18:12:14 yeah sorry.. we won't be pushing it but I will be pointing out the builds on list so people can 'test' things out of koji.stg.fedoraproject.org 18:13:01 smooge: That sounds about right, more fun stuff to do :D 18:14:14 So in the past with many different ways we branch and compose.. we used to put up a web page where developers would put up what packages they wanted branched. 18:15:11 With the current ways we branch etc.. is that still needed or is there a better mechanism to follow? 18:15:56 yeah, not sure. I think the normal process won't scale well to people branching a lot... 18:16:09 but I guess we could do them as releng tickets? or we could do a wiki page I suppose 18:16:49 I am ok with whatever works better for the releng team? 18:17:28 smooge: Question (lack of historical knowledge): Is it for maintainers who wants a epel-8 branch in dist-git? 18:17:43 From looking at the old pages this was for maintainers 18:17:49 Not for end users 18:18:08 mboddu: yes. 18:18:16 it was so a large mass branch could be done by releng vs 1 by 1 by 1 18:18:38 so, say I want to branch the Xfce desktop... I could file 25 tickets one for each package, or I could put them all in a list and just say 'do all these' 18:18:52 I was wondering if a master ticket in releng would work where we say if you want packages processed on YYYY-MM-DD please add your package to this ticket. 18:19:16 smooge: Okay, if we know all of them or bunch of them, then wiki is okay with me or if its 1 by 1 then releng ticket, and I can update the wiki once I processed the tickets for infomational purposes 18:19:17 Then on YYYY-MM-DD all of those are done and we close the ticket and do another one 18:19:21 we do want a record and make sure it's authenticated so we know the maintainer(s) really wanted it 18:19:40 Would the ticket allow for authentication? 18:19:51 nirik: I guess one ticket per request, doesn't matter if they had 1 or 100 18:19:52 yes, they would have to login to comment 18:19:52 The wiki last time had a bunch of people using it as a wishlist 18:20:11 mboddu: well, for maintainers it kinda does... 18:20:20 I don't want to spend an hour filing tickets... 18:20:38 would 1 ticket per set of requests work? 18:20:52 if you have 30 packages just put them in one ticket? 18:20:54 nirik: Right, they can put 100 packages in a request or 1 package in a request, thats what I meant 18:21:00 Wouldit make more sense to default to branching, then let people dead.package it if they don't want it? 18:21:11 no. 18:21:20 mboddu: the fedscm tool will kick all those out? 18:21:51 sgallagh: branching from what? epel7? 5 years ago you wanted to maintain this package... thats a long ass time to assume 18:22:05 nirik: I think we can update fedscm-admin to handle epel-8 branch requests, which might be much easier for people 18:22:17 ok 18:22:33 mboddu: but it can only do one per ticket right? 18:22:39 nirik: yes 18:23:04 so, that doesn't solve the problem of maintainers that want to branch 100 of their packages... 18:23:06 ok sounds like a wiki page still sounds like the right thing to do again 18:23:35 or putting tickets in EPEL ticket tracker and we feed that to releng 18:23:45 nirik: We can provide a simple script to pass a one-per-line text file that calls fedscm 18:23:53 nirik: I guess people can write a shell script which will clone and request the branches, but not sure if everyone can do that 18:24:15 yeah, just so long as we have a simple answer... 18:24:24 does fedscm do the branching or does it just request to releng to do the branching 18:24:24 * nirik doesn't much care what it is. 18:24:39 aka it just files a ticket 18:24:48 smooge: fedpkg does the requesting and fedscm-admin processes them 18:25:01 * nirik wonders how far we are from replacing the tickets setup, but likely further than we can wait for 18:25:35 so what needs to be made so that fedpkg requests would work 18:25:45 right, ok. So a simple script to call fedpkg doesn't seem too hard 18:25:57 smooge: General workflow is "fedpkg request-branch epel-8" which files a ticket at pagure.io/releng/fedscm-requests and then fedscm-admin tool is used to process the requests by using "fedscm-admin process all" which will process all the open tickets 18:26:19 smooge, sgallagh : right 18:27:07 for i in xfce*;do pushd $i;fedpkg request-branch epel-8;done 18:27:17 smooge: I can create the script for you 18:27:20 of course there is a popd missing ;) 18:27:26 well, of course they aren't all named xfce* ;) 18:27:32 pgreco: Not exactly, what if the repo isn't cloned 18:27:37 cd EPEL-sources 18:28:12 nirik, well, we were using xfce as the multi-package example ;) 18:28:46 ok I think we are ratholing here.. I think we have a couple of solutions we can deal with on the list? 18:29:09 yes, but my point is that the list of packages I would branch for xfce isn't named like that. ;) xfdesktop, xfce4-session, Thunar, they have all kinds of names. 18:29:18 yeah, we don't need to decide this today 18:30:42 OK next up 18:30:49 #topic RHEL-7.7 changes 18:30:57 #info RHEL-7.7 beta is out 18:31:15 #info RHEL-7.7 ships with python36 18:31:29 #info wheeeeeeeeeeeeeeee 18:31:41 pgreco, does that cover it :) 18:31:58 yeap, exactly that 18:32:11 Will we have to rebuild all the python36 based packages? Or should they "just work" ? 18:32:34 my guess is that it will be more of a remove than a rebuild job 18:32:45 I expect most should just work after we remove the conflicts 18:32:59 but it would be good to coordinate the version they will release 18:33:03 OK 18:33:15 so epel doesn't have a bigger ENVR 18:33:39 thankfully the same developer is doing both builds... 18:34:47 that simplifies things a lot :D 18:35:06 so if they do... they only have themselves to blame :P 18:35:37 pgreco, did you see anything else? I have not had time to blow up 7.7b to see if there are other changes 18:36:14 I don't think so 18:36:21 that was more than enough 18:36:26 yeah :) 18:36:50 ok so anything else for this meeting or should I let you find people have 20 minutes before the next meeting or 2-3 for the day? 18:36:57 glad epel could proof of concept a python3 stack for rhel. ;) 18:37:10 * smooge puts that as another win for EPEL :P 18:37:14 yeap 18:38:58 Was there anything people wanted to discuss on the sgallagh proposal? Should I tap it on the mailing list for review again and we 'vote' on it next week? 18:40:20 ok I will tap it on the list and we will try to vote it up/down/concensus next week 18:40:25 #topic open floor 18:41:24 thank you all for coming... see you next week 18:41:27 #endmeeting