19:34:06 #startmeeting epel 19:34:06 Meeting started Mon Oct 4 19:34:06 2010 UTC. The chair is tremble. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:34:06 Useful Commands: #action #agreed #halp #info #idea #link #topic. 19:34:40 #topic init process 19:34:59 So, who do we have this week? 19:35:14 me 19:35:26 * nirik is around 19:35:27 looks like nirik and smooge are partially around 19:35:37 * rsc is also partially around 19:35:43 #chair smooge stahnma nirik 19:35:43 Current chairs: nirik smooge stahnma tremble 19:36:37 agenda: Meeting Time finalization; Bugs Update; Broken Deps Update; EL-6 Branching 19:37:04 #topic EPEL Meeting Time 19:37:15 this time generally works ok for me. 19:37:17 Seems like the list has been rather uncaring about the times of these meetings 19:37:27 this time is good for me most often 19:37:28 So anybody *not* like the time ? 19:38:11 * tremble listens to the chirping of crickets. 19:38:16 I like the time. 19:38:26 sorry tremble thankyou for running htis from UK 19:38:38 .de these days actually 19:39:14 ok, I think we're on board now 19:39:19 meeting time finalzied? 19:39:21 +1 19:39:24 +1 19:40:09 ok everybody else said it was ok too 19:40:20 #topic epel bug update 19:40:26 #agreed New time slot 1930 UTC 19:40:41 bug count still below 200 :) 19:40:54 So they're still dropping, if not as fast as before. 19:40:55 #info 190 bugs --> see http://tr.im/epelbugs 19:41:20 already got some of the low lying fruit? 19:41:50 yeah, tremble has knocked out a bunch of it 19:41:57 yeah, good work on that. 19:41:58 I've got a few I think will clean up this week with pushes 19:42:13 I think most of the really low lying fruit's gone now, it's a case of going through the simple fixes rather than the out right closures. 19:42:14 goal is still 1 bug per week 19:43:20 * nirik nods. I will try and help where I can. 19:43:37 There's a number of "please update me" and "please branch me" bugs, I wonder if the please branch me for EL-4 ones should just get "Feel free to branch, if you can't sorry this is a WONTFIX" 19:44:16 I poked at the lua thing last week. 19:44:31 Success? 19:44:49 it oddly doesn't build for me from git. I am now wondering if it's patch got mangled by the git conversion. 19:45:04 jpo had an updated rpm that should work. 19:45:41 I'll get him to commit it or I will soon. 19:47:28 Next item? 19:47:34 sure. 19:47:44 #topic broken deps update 19:47:44 #topic EPEL broken deps update 19:47:46 eheh 19:48:01 ha 19:48:14 Not too much new here. I worked on the script a tiny bit last week. 19:48:20 guys are going to lunch. read summary later. 19:48:39 I was wondering how the emails should be separated 19:48:45 like one for EL-4 with all arches? 19:48:55 for stable and testing each? 19:49:16 that's 4 emails until el6 is live 19:49:21 yeah. 19:49:25 how much information do we want to send? 19:49:26 I think that makes sense to me. 19:50:06 ok 19:50:08 I can do that 19:50:16 hmm 19:50:21 there still probably won't be ppc 19:50:27 since I have to repoquery to run 19:50:35 s/have to/have no 19:50:51 deps are certainly looking better in the repos too 19:50:51 Do we want to un-push updates with broken deps from testing, as a start? Say at the end of the month? 19:50:52 right, thats the problem with running against centos 19:51:07 I'd rather unpush from stable 19:51:10 well, I would say we should get stable fixed first. 19:51:20 perhaps not even send the testing ones yet... until stable is fixed. 19:51:21 broken deps in testing may be waiting on a build override or something 19:51:52 stahnma: yeah, but shouldn't be pushed until 2 weeks after the whole lot is ready 19:52:13 mmm should 19:52:24 I know I have messed that up before ;) 19:52:54 And if we un-push them from testing then they have to wait for the deps to be ready too before pushing to stable. 19:53:20 This at least prevents things from getting worse which is a starting point from which to improve things. 19:53:25 how about this: 19:53:42 do stable and mail people. unpush any that aren't fixed in a week. 19:53:48 do testing and mail the list. 19:53:54 revisit that after stable is fixed. 19:54:10 +1 19:54:12 +1 19:54:52 ok 19:55:20 #agreed Stahnma will mail out the list of currently broken "stable" packages 19:55:43 #agreed One week after that we will unpush any that are still broken. 19:55:51 cool 19:56:49 #agreed We will also run the dep-check on testing and revist what to do with them after we un-push the packages with broken deps from stable 19:57:09 ok 19:57:30 Moving swiftly on? 19:57:45 #topic EL-6 Branching 19:58:16 So we have a load of packages that didn't branch because they were in RHEL-6 beta 1, but have now gone. 19:58:56 do we have a list of them? 19:59:02 This has resulted in broken deps for a few packages and some people have been asking after branches for these. 19:59:30 I don't know if rubygem-rails was in the beta1, but's not in beta2 19:59:44 and it's holding out a decent sized portion of the ruby stuf 19:59:47 perl-Tk disappeared 20:00:39 We can generate a full list reasonably easily... 20:00:39 these are ones that are not at all in beta2 refresh? 20:00:45 yes 20:00:51 ie, not just the ones that are in workstation only? 20:00:54 Gone completely 20:01:06 ok 20:01:14 I wish we knew if they would be back. ;) 20:01:56 Hopefully we don't need to wait too much longer. 20:03:00 So do we... 20:03:08 a) Just branch them 20:03:36 b) Tell people to go through the usual branching process, basically waiting a week if there's no response 20:04:06 c) Just let people branch them but expect them to at least add the existing owner as an owner. 20:04:14 how many are there? 20:04:17 d) Wait for GA 20:04:46 * nirik guesses he doesn't care much... perhaps b seems easiest. 20:05:20 I'm aware of 1 or 2 for certain did find a list earlier... 20:06:31 I'm kind of in favor a D, but at the same time, I'd like to have latest packages ready to go in epel 20:08:01 So refresh... Removed package: dogtail 20:08:01 Removed package: mozldap 20:08:01 Removed package: perl-Mozilla-LDAP 20:08:01 Removed package: python-bugzilla 20:08:33 we already branched python-bugzilla I think. 20:08:42 .branches python-bugzilla 20:08:43 nirik: EL-5 EL-6 F-12 F-13 devel f14 20:09:10 hmm, I use perl-Mozilla-LDAP a lot 20:09:15 or, at least I use to 20:10:10 only those 4? 20:10:12 Beta 1->2 Removed package: amqp 20:10:12 Removed package: bitmap-fonts 20:10:12 Removed package: bluecurve-icon-theme 20:10:12 Removed package: dcbd 20:10:12 Removed package: eclipse-egit 20:10:12 Removed package: eclipse-jgit 20:10:14 Removed package: fastback 20:10:16 Removed package: fedora-bookmarks 20:10:20 Removed package: fedora-gnome-theme 20:10:22 Removed package: fedora-icon-theme 20:10:24 Removed package: htmlview 20:10:26 Removed package: jetty 20:10:28 Removed package: libmcrypt 20:10:30 Removed package: livecd-tools 20:10:32 Removed package: perl-Config-Tiny 20:10:34 Removed package: python-krbV 20:10:36 Removed package: qpid-tests 20:10:38 Removed package: rhpl 20:10:40 Removed package: rng-utils 20:10:42 Removed package: wxGTK 20:10:54 and like I said, I can't find perl-Tk any more... 20:10:59 some of those should die. ;) Like rhpl. 20:11:02 So there may be more. 20:11:30 The list isn't as big as I feared it might be. 20:11:32 so, why not just allow them to branch as per normal... as we find them and someone wants to maintain them. 20:11:40 Yeah 20:12:16 It was asked that we also discuss the Server/Workstation packages game... Personally I think we just wait for GA. 20:12:41 It's difficult to be prepared for an unknown target 20:13:07 sadly, I think so too. I know it's frustrating to wait while deps are broken... 20:13:24 but adding them in now and then having to remove them later would be a pain... and it would also cause issues for rhel 20:13:32 well, at the same time, most orgs don't adopt over night ;) 20:13:55 It wouldn't be too hard to yank them once the package list is public. 20:14:46 We could also implement the suggestion of putting in a repo priority, however that could catch out people on i686 and ppc64... 20:15:00 * nirik is not liking it because it's us basically bypassing the workstation/server split. Ie, would paying rhel folks get something from us when they normally would have paid for access to another channel. 20:15:56 dgilmore: I don't suppose you have heard anything on this recently? 20:16:23 nirik: The other side is that they wouldn't be getting the support from RH for those packages. 20:16:52 sure. 20:17:10 I don't have much of an opinion on that, other than sometimes servers get things classified as workstation packages 20:17:19 nirik: i dont expect to have any update until rhel is ga's at this point 20:17:23 and sometimes workstations get server packages 20:17:29 I really don't see why they seperate them at all 20:17:36 The split is generally rather arbitary 20:17:38 but, I'm not a sales/marketing person 20:17:42 * nirik doesn't either personally, but thats just us I guess. 20:17:50 Workstation is sold at a lower price. 20:18:42 true 20:18:56 but I don't see how when I pay the higher price, I am missing things from the lower price 20:19:01 and that I can understand, but the libs from Workstation that can't be found in Server makes EPEL's life awkward. 20:19:08 right. 20:19:19 alas, we will not solve that problem today 20:19:37 So how many of these do we know of? 20:19:47 I think I know of 2 that are being asked about a lot by maintainers... 20:20:10 we could add those 2 as a special exception now, and pull them later... 20:20:16 or just wait. ;) 20:20:27 I think we should just wait 20:20:32 +1 20:20:39 (on waiting) 20:21:36 * nirik shrugs. 20:21:52 I think it's worth people considering the options, but I think actually doing anything should wait. 20:22:49 +1 to wait 20:23:20 ok 20:23:58 #agreed We're not going to branch the Workstation only packages (beyond anything already branches) until at least after GA. 20:24:29 #agreed For packages that got dropped between betas people should just put in branch requests as per normal. 20:25:31 #topic Open Floor 20:25:46 just a reminder to try to get the latest version of your package in to epel6 20:26:08 there was some discussion on list about system management tools. It got no real resolution, but I liked the discussion 20:26:17 * tremble nods 20:27:17 Oh, packages we're building to satisfy some arches... 20:27:25 * stahnma has to go to next meeting 20:27:27 thanks! 20:28:11 * nirik vuagely thinks he had something, but can't recall what it was. 20:28:48 My suggestion is as follows : file "README-PACKAGER" in the EL-6 branch 20:29:00 Warning: 20:29:00 This package is available for *some* architectures in RHEL6. 20:29:00 The EPEL policy is that where possible we should simply rebuild the 20:29:01 RHEL6 source packages, so that we can avoid overriding files/RPMs. 20:29:01 As such, be very careful when updating this package. 20:29:09 oh yeah, that. 20:29:16 sounds good. Make it so! :) 20:29:34 Okay I shall start dropping that file into Repos. 20:30:10 I said "where possible" because I actually found a package that doesn't build in koji due to a missing dep on perl-CGI. 20:30:22 ;( 20:30:45 #action tremble to start putting a warning message into EL-6 branches where we're rebuilding RHEL SRPMS. 20:31:23 Any Other Business? 20:31:32 * nirik can't think of his. 20:32:03 Closing in 1 minute if nobody comes up with anything... 20:32:46 thanks for running the meeting tremble! 20:33:04 #endmeeting