18:00:32 #startmeeting EPEL (2018-03-07) 18:00:32 Meeting started Wed Mar 7 18:00:32 2018 UTC. The chair is smooge. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:32 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:00:32 The meeting name has been set to 'epel_(2018-03-07)' 18:00:32 #meetingname EPEL 18:00:32 The meeting name has been set to 'epel' 18:00:32 #topic aloha 18:00:32 #chair avij bstinson Evolution nirik smooge 18:00:32 Current chairs: Evolution avij bstinson nirik smooge 18:00:47 hello 18:00:57 hello avij 18:01:05 looking to see if others can meet 18:01:17 o/ 18:02:52 #topic Agenda 18:02:52 #info Has anyone evaluated changes in upcoming 7.5? 18:02:52 #info Developertoolset 7 has been synced and turned on in koji 18:02:52 #info Other? 18:03:02 Any other items? 18:03:38 #topic EL-7.5 18:03:44 none here. 18:03:54 avij: have you poked this at all? 18:04:08 OK the EL-7.5 beta is out. I have not downloaded it yet but will do so later today to get an idea of what cruft is different from 7.4 18:04:18 well, I don't have access to RHEL 7.5, so I'm kind of limited in this regard 18:04:52 there are some rebases, but I'm unsure if they break EPEL stuff 18:05:12 it would be nice if they had some sort of you "Everything" ISO but oh well 18:05:13 I have read the beta release notes 18:05:23 avij, anything in the notes? 18:06:24 nothing that would strike me as "oh, this will need changes/rebuilds to EPEL packages". there's "GNOME Shell rebased to version 3.26" which may affect things. I don't know for certain. 18:07:26 there's apparently a new 4.14 kernel that'll be available as an option 18:07:34 wa? 18:07:37 but I don't think epel has anything kernel dependent in it 18:07:57 no.. we avoided that like the plague 18:07:58 oh, not for x86_64 though. 18:08:03 ha of course 18:08:05 kernel-alt, for arm, power and z 18:08:09 Evolution: I filed a C bug on this -- it was for s390x libvirt / docker matters 18:08:31 running docker on s390.. all 3 people 18:08:39 yeah. I'd like to see it for x86_64, but I'd also like to see a unicorn that dispenses whiskey 18:08:51 smooge: I have over 40k instances under s390x at aone customer 18:09:21 libreoffice, gimp, inkscape, webkitgtk4, qt5, nss, libreswan, dbus, cockpit, libvirt, samba, directory server, pki-core, binutils, rsync and others -- all rebased, but as I mentioned, I don't know if these require EPEL pkg rebuilds. 18:09:34 ok 4 sites.. and N customers 18:10:03 lots and lots of sites, the name of which is literally all across the country 18:10:15 out of scope here 18:10:18 yeah. 18:11:25 well at least I found someone admit they run s390x outside of IBM labs.. 18:11:50 * nirik sighs. sorry... here now 18:11:59 everything ok nirik ? 18:12:09 lost time in debugging something. 18:12:25 ah.. my usual problem 18:12:34 ok so the next one is all you. 18:12:38 most likely some rebuilds will be necessary for 7.5, I just don't know for sure which packages. I'll find out when 7.5's packages are available via CentOS continuous release (cr) repo. 18:13:24 OK thanks avij . Looks like we will need to do schedule some sort of flag day 18:13:46 #topic DTS-7 18:13:46 #info What to do with other SCL's? 18:13:46 #info What to do when dts-8 comes out? 18:13:51 Evolution: i almost caught a whiskey dispensing unicorn last night, but it turned out i was just dreaming... 18:13:54 thank you nirik 18:14:22 no problem... thank you and pbrobinson for doing all the heavy lifting so I could take credit. ;) 18:14:30 I got everything but aarch64 working but nirik found the right repo for that 18:14:46 there's already one issue... well, a bug in a scl... 18:14:52 LOL! 18:15:00 nirik: a bug in SCL? 18:15:01 NOWAI 18:15:02 of course there is 18:15:13 nirik: I'm sorry, I can't find my surprised face! 18:15:16 nirik: is it bleeding rpm deps? 18:15:18 there's a scl providing lz4-devel... and overriding the epel/rhel version 18:15:21 yep 18:15:35 fffffffffuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu 18:15:51 well its a good think nothing uses that 18:16:19 that should be easy to fix? 18:16:38 smooge: I think any of the dts releases should be parallel usable, so we could either use the latest, or just add 18:16:47 https://bugzilla.redhat.com/show_bug.cgi?id=1552431 18:17:21 they shouldnt be providing base things... 18:17:26 only used when they are enabled. 18:17:32 AFAIK 18:18:10 nirik, can we blacklist that package somehow? 18:18:23 possibly yeah 18:19:15 will have to make sure you can block in external repos... 18:20:35 I can try to see if a reposync without it kills the stuff we need for dts. 18:21:20 ok so I guess this is hurting our current builds? 18:21:28 well, the maintainer worked around it and filed that bug, so I don't know how urgent it is. 18:22:19 ok 18:22:25 so I had one more thing... 18:22:32 #info Try this for EL-6? 18:23:11 we could I suppose. I am not sure how much interest there is, but sure... 18:23:21 smooge: personally I wouldn't bother for el6 unless it's explicitly requested 18:23:54 so EL-6 is our largest user 18:24:05 sure 18:24:26 and it has been requested for EL-6 multiple times in the past.. we just didn't at that time 18:24:47 what's the level of effort required? 18:24:48 because we didn't do anything with SCL's until approved by FESCO/FPC 18:25:30 on my part.. just work out the @$%#%$ repos and reposync them. on releng part getting them into koji, on developer part.. finding out what leaks we just caused 18:25:33 the main driver here was chromium, and I am not sure it was ever built on el6... 18:25:34 spot is crazy, but not that crazy. :) 18:25:52 for the record, the ath10k wireless SUXXORS. 18:26:22 but its free and opensource.. you can fix that microcode 18:26:34 * smooge should not try to be funny. 18:26:50 Personally I am fine with devtoolset as it's not use impacting at all... but anything that is I am not really in favor of as it will confuse and annoy users. 18:27:09 well, the firmware seems to be the issue. Trying to find one that works well with the current driver. :) 18:27:12 I thought it was built on chromium, but stopped 18:27:26 then again, I believe there was a glibc issue that I can't imagine an scl would fix. 18:28:18 yeah.. it requires statically linking a newer version 18:28:50 however other software was looking at it in the past.. but I will put it down to "if someone still needs it they can ask" 18:29:05 after I put a post up about it 18:29:36 yeah, blog post would be good... you can look at chromium to see how it's setup... 18:29:52 (but chromium still hasn't built yet due to a patch failure) 18:29:58 we could ask spot (from a safe distance) if he thinks its feasible 18:30:10 ok anyone have any other items on this. my cough medicine is giving up so I am needing to end this in a bit 18:30:30 nope. 18:30:38 nope 18:30:53 #topic Open Floor 18:31:14 This is the section where we look for items for the next meeting 18:32:19 sorry i missed most of this meeting. had a guy from the city show up wanting to look at my house 18:32:59 "youse got a nice house here.. be a pity if someting were to happen to it." type of guy from the city? 18:33:00 https://src.fedoraproject.org/rpms/chromium/blob/epel7/f/chromium.spec#_1067 is most of the magic (and adding BuildRequires: for devtoolset packages) 18:35:05 I have to say if you want to see an rpm spec file that puts TeX or OpenOffice to shame.. this is it 18:36:22 texlive still wins... it has a c program to make the .spec. :) 18:36:42 Spot needs to work harder.. 18:37:06 he needs to make a haskell program that writes a c program that makes the spec 18:37:29 ok I am going to call it a meeting in 30 18:37:33 seconds that si 18:37:39 Thank you all for coming 18:37:46 thanks again for chairing smooge 18:37:54 yup. thanks. 18:38:00 #endmeeting