16:00:00 #startmeeting EPEL (2012-07-27) 16:00:00 Meeting started Fri Jul 27 16:00:00 2012 UTC. The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:00 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:00:00 #meetingname epel 16:00:00 #topic init process/agenda 16:00:00 #chair smooge tremble dgilmore 16:00:00 EPEL meeting ping abadger1999 rsc stahnma tremble dgilmore smooge nb maxamillion tremble Jeff_S HackMan 16:00:00 The meeting name has been set to 'epel' 16:00:00 Current chairs: dgilmore nirik smooge tremble 16:00:08 any folks around for an epel meeting? 16:00:09 here 16:00:12 * abadger1999 here 16:00:54 * pyther24 is just a fly on the wall 16:00:59 anyone have items for agenda? 16:01:12 * nirik was going to go over the broken deps script which ran last weekend. 16:01:16 * nich0s is with pyther24 16:01:20 Oh, oh, I'm curious about the policy for packages that exist in both EPEL and RHEL 16:01:33 jokajak: so are we all. ;) 16:01:56 I meant to write up a new proposal this week, but didn't get to it again sadly. 16:02:53 we can discuss some more in meeting today if folks like. 16:03:02 anyhow, anything else for agenda? or shall we start in? 16:03:39 #topic Broken deps reporting script 16:03:49 * skvidal is here... just lurking 16:03:52 so, we have a broken deps script again and it ran this last weekend. 16:04:00 keeps readig that as Broken Derps 16:04:09 Strangely, I haven't heard any complaints about it... ;) 16:04:47 there's a few enhancements it needs... better text explaining when it's talking about epel vs epel+epel-testing 16:05:01 and it would be nice if it sent to the mailing list as well. 16:05:22 finally it would be great if it could also have a whitelist of some corner cases that aren't really broken... 16:05:37 If anyone wants to work on that, I'd be happy to point you to the code. ;) 16:05:50 Do you have a link to the output of the script? 16:05:56 I may have D'd it 16:06:08 it mails each owner about each broken dep 16:06:31 nirik: sure, I might could use it for $job 16:06:45 for example, here's one I get (which turns out to be kind of a false positive): 16:06:48 collectd has broken dependencies in the epel-5 tree: 16:06:49 On x86_64: 16:06:49 collectd-4.10.3-1.el5.i386 requires libpython2.4.so.1.0 16:06:49 Please resolve this as soon as possible. 16:07:13 (this is due to python not being multilib in rhel5) 16:07:37 ah 16:08:35 jokajak: I can get the scripts pushed to the rel-eng git tree and point you at them... 16:08:49 nirik: yes, I'm more or less around 16:09:22 nirik: regarding the broken dependencies...I think the optional channel is missing for these checks. At least it was for mine IIRC 16:09:45 rsc: oh? odd. it should have been in there... 16:12:11 rsc, what was the package that was missing 16:12:35 anyhow, I will get the scripts pushed out. 16:13:43 ok, anything else on broken deps? 16:14:21 #topic Overlapping packages part 12 16:14:35 As noted I meant to make a new proposal, but keep getting sidetracked. ;( 16:14:36 DUH-DUH-DUUUUUUUUUUUH 16:14:42 I'll really work on it this next week... 16:14:47 but basically it will be: 16:14:48 it's a 12 part series, how exciting 16:15:08 jokajak: just that this discussion has been going on for a long time now sadly. ;( 16:17:21 epel6 will not overlap with rhel6 in the channels that match up to advanced platform, except where they are needed for limited arch support. 16:17:37 I need to try and figure out what channels match up to AP. 16:17:41 and see if that will work out 16:18:02 what about when packages are needed for limited arch support, will they still be built for the supported arches? 16:18:22 http://fedoraproject.org/wiki/EPEL:Packaging#Limited_Arch_Packages 16:18:30 * nirik thinks that policy is pretty set. 16:19:00 so, yes, if rhel6 ships foo only on x86_64, we can ship foo-0.x for all arches we support. 16:19:17 ok, that is quite clear :) thanks 16:19:40 the big problem has been figuring out which channels we will look at for overlaps. 16:19:43 and if a package violates that policy, what is the recourse? 16:20:14 file bug? ask maintainer to fix it. escalate to sig or rel-eng to fix it if they don't. 16:20:47 when I accidentally got into this meeting, I would point out one personal example of EPEL vs. RHEL 16:20:48 great :) 16:20:51 https://bugzilla.redhat.com/show_bug.cgi?id=741324#c5 16:21:55 (the comment in the link was not intentional) 16:21:56 jpokorny: right. 16:22:17 the problem/question is if we care about ha channel or not. various of our proposals would have had us doing so. 16:22:56 thanks for the good example jpokorny 16:23:22 I think we need some sort of ven diagram.. if we stick to X how much of EPEL goes away. if we stick to X+X0 , etc etc. 16:23:26 Don't almost all of our proposals include hs? 16:23:29 *ha 16:23:38 abadger1999: many of them yeah. 16:24:08 the last one did, but inode0 asked why ha and lb... and we didn't really have much reason, just 'we added them a while back' 16:25:23 The kind of de facto right now is.. if they're in the buildsystem's definition of rhel 16:25:27 anyhow, I will try and look over the AP mappings and see if they help us. 16:25:30 16:25:43 I can't seem to find the email with them, but will keep looking. 16:26:40 #action nirik to work on a new policy draft. help welcome. 16:26:45 anything else on this? 16:27:53 #topic Open Floor 16:27:59 any items for open floor? 16:29:10 * nirik will wait for a few minutes, then close out if nothing comes up 16:31:29 nothing for open floor from me 16:32:11 ok, thanks for coming everyon. 16:32:16 everyone even 16:32:17 #endmeeting