18:01:35 #startmeeting EPEL (2019-02-13) 18:01:35 Meeting started Wed Feb 13 18:01:35 2019 UTC. 18:01:35 This meeting is logged and archived in a public location. 18:01:35 The chair is smooge. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:01:35 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:01:35 The meeting name has been set to 'epel_(2019-02-13)' 18:01:43 #meetingname epel 18:01:43 The meeting name has been set to 'epel' 18:01:51 #topic Chair and Introductions 18:02:02 #chair avij bstinson Evolution nirik smooge pgreco tdawson 18:02:02 Current chairs: Evolution avij bstinson nirik pgreco smooge tdawson 18:02:13 Hello everyone 18:02:20 Hello 18:02:27 hello 18:02:42 hey all, i'm in the middle of like 3 incidents. i'll try my hardest to pay attention for action items 18:04:16 sorry I'm late. 18:04:18 #action bstinson says he will take all the items 18:04:27 oh, just in time. I agree with this action 18:04:37 #topic Agenda 18:04:58 #info Smooge didn't send out an agenda (* AGAIN!!! *) 18:05:16 morning 18:05:26 #topic DevConf Summary 18:05:54 So Kevin and I went to Devconf and met with many people in the late January 18:06:21 there were a lot of meetings with modularity and other people on what will be needed for trying to get EPEL-8 possible 18:06:33 I am still writing up the summary and will send that out later today 18:06:58 The high level is that for EPEL-8-beta we will have only non-modular packages built 18:07:20 Will we be able to use modules for building? 18:07:33 not initially. 18:08:17 OK 18:08:20 depending on who we spoke with.. we can build with default modules only.. or we can fail a lot 18:08:24 but everything is subject to change... we need to make a beta and see how it all goes. 18:08:38 non-modular, but can they depend on modules as a build-req? 18:08:52 pgreco: yes. 18:08:56 default modules. 18:09:09 I'm specially thinking about something that Petr said at the dojo 18:09:20 there's a tooling change required to make this work. 18:09:37 again we will see when we set it up. for the beta the tooling changes may not be in place and working 18:09:43 after the beta we will know 18:10:44 in order to get that working we have to have a tool that takes the rhel8 modular repos and makes them back into distinct modules. 18:12:32 hi, sorry I'm late -- I have a double booking, I'm not at a proper computer at the moment 18:12:37 Evolution, bstinson nirik was there anything else high level I forgot to add? 18:12:42 hi avij 18:12:51 smooge: seems about right 18:12:54 Will our initial build repo's contain what is somtimes called the "CRB" ? 18:13:01 yes 18:13:10 cool 18:13:12 I have been told we won't compile much otherwise 18:13:17 like anything 18:13:44 smooge: You can ... I've tried ... but you can compile much much more with it. 18:14:03 The 3 repositories that EPEL-8-beta will a) not conflict with and b) will build with are: Base, Appstream and Code Ready Linux Builder. 18:14:53 It will not compile and may conflict with High Availability or Resilient because those are to be limited user channels 18:15:28 and their projects have requested we specifically don't use them 18:15:42 that is all I have from that. 18:15:56 #topic EPEL policy proposals 18:16:20 I sent out an email shortly before the meeting which will be up for approval/change/disapproval next meeting 18:16:43 #link https://fedoraproject.org/wiki/EPEL/Changes/Minor_release_based_composes 18:17:09 #link https://fedoraproject.org/wiki/EPEL/Changes/Release_based_package_lifetimes 18:17:18 (not sure if link works or not) 18:17:47 These are two policy changes which will affect EPEL-7 but will help make EPEL-8 a lot easier to produce 18:18:19 Please read, review, edit the wiki pages please 18:19:11 and thanks to everyone who has given tips on it before 18:19:21 so if everyoone is on board, we start that with the next 7.x? 18:19:33 I was going to start before then if possible 18:19:38 1 minor release is ~6-8 months 18:19:39 bi-annual ... that is meaning every 6 months, correct? not every two years? 18:19:58 correct. biennial is every two years 18:20:06 Ahh ... ok, thanks 18:20:14 I had to look that up several times 18:20:16 tdawson thanks for the question 18:20:24 it means twice a year 18:20:42 wrt to release based, I was saying that 13 months is closer to 2 minor releases 18:20:58 with our current schedule of ~6-8 months 18:21:55 so my main point is for the last major.minor release where 6.10 might be out there for over 13 months. I wasn't sure how to word that case 18:22:24 smooge, I meant for 7.x 18:22:35 no changes for 6.x sounds logic 18:22:44 yeah.. eventually there will be a 7.11 type release where it will fall into the same camp 18:23:09 but I could be trying to be too perfect and just go for good 18:23:25 when it comes up we deal with it? 18:23:31 versus code it in now 18:24:30 so I would like to test this out soon after it is approved for the current 7.6 release because of the Python34/36 changeover 18:25:17 but it may make sense to do the changeover and then do it on the next release 18:25:51 question wrt maintainers, what happens if a someone just stops doing it? without your guidelines I mean 18:25:53 So if you test it, then python34 would go into the archive, with python36 being the new one ... correct? 18:26:21 does anybody have the power to update something? in case of a CVE or something like it? 18:26:32 pgreco, proven packagers do 18:26:36 sure, provenpackages can... 18:26:59 can, yes, but my question goes more to the "should"? 18:27:13 etiquette and all.... 18:27:17 pgreco, I don't think we have anyway to do must 18:27:26 there's a page on that. :) 18:27:39 ah goo 18:27:41 d 18:27:50 * pgreco would like to read that... :) 18:27:55 I should have looked at that 18:28:13 https://fedoraproject.org/wiki/Who_is_allowed_to_modify_which_packages 18:28:53 tdawson, my next subject would go into my python36 idea 18:29:17 One thing to avoid IMHO is packages where some provenpackagers are keeping it updated/alive, but aren't the maintainer... then people file bugs and no one reads them. ;( 18:30:41 nirik, well I figured that was the case for various 'low-maintenance' packages already :/ 18:30:59 nirik: And I don't think that is an EPEL-only problem. 18:31:18 quite right on both counts. ;) 18:31:18 it is hard to tell if they are there because they are low-maintenance or that no one is checking the bugzilla queue etc :) 18:32:17 ok if nothing else on this.. please read, comment and fix. We will vote up/down next week 18:32:22 And on the flip side, if you were the original maintainer, are no longer with a project where you added a package ... and you have no idea if anyone is still using the package or not ... but ... I guess that's a different topic. 18:32:37 smooge: Will do 18:32:48 #topic Python34 to Python36 rebuild and migration 18:34:11 So the python people have done a bunch of rebuilds and most packages 'built' without changes. the broken packages will need provenpackager fixes and then we also need a bunch of testing 18:34:59 looked like a number of the failures were deps that need to be added/built, so there might be ordering issues too... 18:35:42 yeah. I would like to get ideas on where we want to go next? 18:36:33 well, at some point we just need to schedule it and do it and clean up the broken parts afterward. 18:37:36 if there is a clear order on that, it is going to be useful when I rebuild the armhfp 18:38:00 make 'fixes' to the copr to get them all to rebuild and then put up for testing.. schedule a proven packager day and force through the changes? 18:38:40 I doubt we will get to 100%.... but yeah 18:38:48 there is a suggestion from 2 testers to not use obsoletes and only conflict on the main python36 package. The upgrade process seems easier if only those conflict 18:39:36 My plan after this meeting is write up the EPEL-8 items, and set up a EL7 box to test a python34 and htne try using that copr to upgrade 18:39:55 tdawson and nirik I believe are our provenpackagers 18:40:36 I'd be happpy to review pr's/patches for python36 18:40:58 so I was thinking when we do it we should make sure it is a time frame you are available 18:41:20 versus neck deep in F30 crunch time 18:41:45 thank you nirik 18:42:24 yeah, lets schedule a day. Unfortunately I am moving at some point and not sure when that is. Should be back to normal after that 18:42:31 smooge: Other than the end of march, most of my stuff can be worked around, so, as long as you miss the last couple weeks of march, I'm ok. 18:42:36 ok that is all I have to say on this. [That was the other thing] 18:42:41 might also be good to go with a time when fedora is frozen or not too busy 18:43:22 beta freeze starts tuesday 2018-03-05 18:43:36 how does 2019-03-07 sound 18:44:19 that should have been 2019-03-05 (yes the checks I wrote today are still the wrong year) 18:44:48 so that would be the first week of March. Fedora Infra should be in a freeze and we should have 'ample' cycles 18:44:55 could work 18:45:03 2019-03-07 sounds good to me 18:45:38 #info Python34 to Python36 build time is scheduled for 2019-03-07 . plan to be posted to list after meeting 18:46:25 if someone could send a calendar invite so we don't forget, that would be cool. 18:46:31 doing it 18:47:22 #topic Open Floor 18:47:32 that was all I have for this meeting with no agenda 18:48:59 It covered everything I was wondering about. 18:50:39 I have added the meeting to the epel calendar 18:50:44 nothing from me 18:50:51 ok that is all and I think we are done for the day. 18:50:54 thank you all for coming 18:51:04 Thanks 18:51:08 tip the waiter.. he has to put up with me all the time while you get to go home 18:51:14 #endmeeting