21:03:03 <nirik> #startmeeting
21:03:10 <nirik> #meetingtopic EPEL
21:03:12 * SmootherFrOgZ is
21:03:16 <nirik> #meetingname epel
21:03:23 <nirik> #topic Init process
21:03:42 * stahnma made it :)
21:03:47 <smooge> I am here
21:03:50 <smooge> hurray
21:03:52 <dgilmore> im kinda here
21:04:00 <dgilmore> doing the migration to koji
21:04:10 <smooge> yeah koji
21:04:11 * abadger1999 here
21:04:12 <stahnma> very nice
21:04:42 <nirik> ok, shall we go ahead and get started?
21:04:48 <smooge> yes please
21:04:51 <stahnma> good with me
21:05:08 <nirik> #topic Policy about packages moving from EPEL to RHEL
21:05:46 <nirik> so, this came up with the recent openjdk thing.
21:06:11 <stahnma> well, openjdk is really only the latest one
21:06:16 <stahnma> I believe there were others
21:06:20 <nirik> I thought the policy was 'when it appears in rhel, it is removed from epel', but I guess it's unclear.
21:06:29 <stahnma> well, I think that's the policy
21:06:44 <stahnma> unless RH keeps a lower EVR
21:06:56 <stahnma> or they only publish some of the binary packages rather than all of them
21:07:45 <nirik> well, that does open a can of worms.
21:08:14 <dgilmore> openjdk has been removed from epel
21:08:39 <smooge> or the RHEL doesn't have a 'needed' patch the EPEL one had for some reason
21:08:44 <stahnma> but the EVR in EPEL is higher than what's currently available in RHEL
21:09:16 <nirik> once it goes into rhel, we should report bugs there if any, etc... we shouldn't try and carry our own package.
21:09:25 <dgilmore> nirik: right
21:09:39 <nirik> stahnma: right, so those people who installed from epel still have the epel version. ;( Only rhel can fix this.
21:09:54 <stahnma> bug has been filed
21:10:00 <stahnma> it should  be fixed in 5.4
21:10:05 <dgilmore> they know it and will be making sure the next update is higher
21:10:12 <nirik> this gets back to needing better coordination.
21:10:18 <stahnma> agreed
21:10:41 <nirik> ok, so do we need to adjust our policy here? or just strive to enforce it moving forward better?
21:10:46 <stahnma> at the beta we should check new packages and EVR
21:10:54 <stahnma> but I also think RH should be doing the same
21:12:01 <smooge> at the beta we would need to do a tree check and then mark/push stuff that is obsoleted
21:12:25 <smooge> but it only helps so much.
21:12:58 <nirik> sometimes things come in as updates too...
21:13:02 <stahnma> well, wehn 5.4 beta hits (unless it already has) we should probably check
21:13:07 <nirik> we sometimes can't know what rhel is going to be doing.
21:13:12 <stahnma> true
21:13:22 <smooge> my guess is that the beta should be this fall/winter
21:13:53 <smooge> my frustration is that why are we constraining ourselves to a 1 way relationship.
21:14:06 <smooge> then I take a happy pill and go back to work
21:14:14 <stahnma> agreed
21:14:21 <stahnma> we need a bi-directional feedback process
21:14:30 <slipp3d_> agreed stahnma
21:14:32 <stahnma> but we need somebody close to releng for RHEL
21:14:35 * nirik nods.
21:14:48 <stahnma> and I really don't know who those people are
21:15:04 <smooge> stahnma, I am not sure anyone knows who those people are :P
21:15:47 <abadger1999> f13: Do you know who might fill that role?
21:15:58 <nirik> It was suggested to mail notting and/or spot and inquire.
21:16:01 <stahnma> I am asking here as well
21:16:07 <nirik> Perhaps someone could take that task?
21:16:21 <stahnma> I will
21:16:28 <stahnma> I will try to find out
21:16:55 <smooge> spot is on vacation/travel til I think late June
21:17:24 <smooge> notting might be a good choice.. or a desperate and forlorn plea on the rhel mailing list
21:17:37 <nirik> #action stahnma will try and find someone to interface with on the RHEL side of things.
21:17:53 <nirik> cool. Shall we move on? or anything else on this topic?
21:18:25 <smooge> not that I can think of. stahnma if I can help I will definately do so.
21:18:39 <stahnma> ok
21:18:40 <stahnma> next
21:19:05 <nirik> #topic Review Bug Day ideas
21:19:22 <nirik> ok, so this is scheduled now... what do we want to try and accomplish?
21:19:23 <stahnma> As I mentioned in #epel earlier, I'd like to classify bugs
21:19:31 <stahnma> so we can work with them easier
21:19:54 <stahnma> we basically have requests for branch, requests for update and actual bugs
21:20:17 <stahnma> it's a fairly even divide as far as numbers go(at a glance)
21:20:32 <slipp3d_> stahnma, how else would the grouping go?
21:20:35 <stahnma> I think bug day goal could be to touch/triage every bug
21:20:46 <nirik> well, we should be able to just process the branches and close those.
21:21:03 <nirik> request for updates could be trickier if they need a lot of other things.
21:21:04 <stahnma> nirik: well the branch requests are not from the maintainers
21:21:09 <smooge> stahnma, do we have a listing of every bug?
21:21:12 <stahnma> they are from end-users
21:21:17 <stahnma> http://tr.im/epelbugs
21:21:20 <nirik> ah, so requests to package?
21:21:39 <stahnma> requests for the maintainer to branch their fedora package to EPEL
21:22:19 <nirik> well, we can try and ping maintainers, branch and maintain with existing people, or the like
21:22:31 <stahnma> slipp3d_: not sure on groupings. Just trying to make a managable listing
21:22:53 <nirik> I think if we have enough folks it should be possible to touch on all of them... 135 right now.
21:23:25 <stahnma> I've  been trying to read through and triage some of them, or ask for updates
21:23:32 <stahnma> some are misfiled, and moved some
21:23:56 <nirik> cool.
21:24:03 <stahnma> the bugzapper guys were very interested in helping, but we probably need clear goals
21:24:25 <nirik> for the bugs we should be able to try and at least make sure there is sufficent info provided, etc.
21:24:45 <stahnma> true
21:24:57 <smooge> I think I had a couple of requests for branches that got closed for some reason
21:25:23 <stahnma> I think we should come up with some expectations and documentation for people who want to help
21:25:44 <smooge> yes that is a good plan.
21:25:59 <smooge> expectations: bug is looked at and updated with any info from tester
21:26:09 <smooge> requests for branches ?
21:26:14 <smooge> I am not sure how to handle those
21:26:25 <slipp3d_> i would think that would be a good idea stahnma ... otherwise someone like me who is new would be over whelmed ..
21:27:03 <nirik> well, ping maintainer and try and see if they want to maintain it. If not, come up with a list of them and ask on the epel-list for maintainers?
21:27:19 <stahnma> i'd like a way to classify them in bugzilla
21:27:26 <stahnma> so at a glance we can tell what we are working on
21:27:39 <stahnma> I am not that familiar with BZ, so I might need help there
21:27:40 <nirik> we could use the whiteboard field.
21:27:48 <nirik> you can put anything you like in there...
21:27:53 <stahnma> oh great
21:27:59 <stahnma> that should be perfect
21:28:14 <stahnma> can I report/sort on that field?
21:28:50 <smooge> nirik, a couple of the ones were ones I was wanting to maintain.. but its been a long while since I looked at what they were.
21:29:05 <nirik> yeah, I think you can query it ok
21:29:25 <smooge> nirik, cool. I didn't know about the whiteboard field
21:29:55 <stahnma> anything else on this topic?
21:30:01 <stahnma> I'll try to classify bugs
21:30:18 <nirik> not off hand
21:30:29 <nirik> thanks stahnma
21:30:48 <nirik> anything else?
21:30:55 <smooge> not me
21:31:02 <nirik> next up...
21:31:08 <nirik> #topic FAQ updating (posting even)
21:31:31 <stahnma> I looked at hte FAQ (very quickly) and wondered if we should updat eit
21:31:40 <nirik> whats the link?
21:32:28 <smooge> yes we should update/rewrite it.
21:32:37 <stahnma> I don't remember :)
21:32:40 <smooge> its been pretty static since EPEL just founded
21:32:56 <stahnma> looking
21:32:58 <smooge> http://fedoraproject.org/wiki/EPEL/FAQ
21:33:02 <stahnma> yeah
21:33:04 <stahnma> that one
21:33:39 <nirik> yeah, we should go over it and update it...
21:33:54 <nirik> how best to do that? just do it in the wiki and discuss changes?
21:34:01 <nirik> or mock up a new one and move it when we are happy with it?
21:34:15 <smooge> I would say mock up a new one.
21:34:26 <smooge> what are the top 10 questions we need to answer regularly.
21:34:27 <slipp3d_> I would think that a mock would be the better option
21:34:34 <smooge> and then the rest
21:35:05 <nirik> ok. I know quaid was interested in updating the wiki pages, perhaps he could lead this? ;)
21:35:31 <stahnma> repotags
21:35:32 <stahnma> updating policy
21:35:32 <stahnma> how often we push
21:35:39 <stahnma> replacing core
21:36:04 <stahnma> compatibility with downstreams ELs
21:36:20 <smooge> mixing repos
21:36:42 <stahnma> how  we handle stuff thats in EL on one arch but not another
21:37:05 <nirik> section for maintainers would be good...
21:38:08 <nirik> so, who wants to lead this? or should we ask for help on the list?
21:39:34 <smooge> nirik, when do we want a first stage mockup? next week/week after? I can do it for week-after... next-week would be pushing it.
21:39:49 <nirik> I don't think it's urgent... just needs to get done sometime.
21:40:07 <smooge> ok I will lead. put my deadline for 2 weeks
21:40:24 <nirik> cool. thanks.
21:40:34 <smooge> also I was wondering.. do we want a trac instance etc?
21:40:35 <nirik> #action smooge will mock up a replacement/updated FAQ page for us.
21:40:51 <nirik> well, for tag requests and such we can just use the existing rel-eng one.
21:41:07 <nirik> not sure what else we need one for, but if there is a good reason we can request one.
21:42:12 <nirik> ok, anything else on faq?
21:42:19 <smooge> duh... I remember this being put in last meeting for using the rel-eng one.. sorry
21:42:34 <smooge> no I will have a mockup on list soon
21:42:41 <nirik> excellent. thanks.
21:42:45 <nirik> #topic Stable Push & Koji/Bodhi progress report
21:42:51 <nirik> dgilmore: you around? how do things look?
21:42:59 <nirik> and thanks for working on it!
21:43:00 <dgilmore> nirik: push done
21:43:06 <dgilmore> im syncing it now
21:43:15 <dgilmore> then ill import rpms into koji
21:43:50 <nirik> cool. :) How does the bodhi stuff look? or we will see when we try and use it?
21:44:02 <dgilmore> we will see when we try
21:44:10 <dgilmore> lmacken is looking over the patches
21:44:26 <dgilmore> most of it is config
21:44:27 * smooge makes sure to keep babies away from bodhi when you try.
21:44:34 <nirik> excellent. I'm happy to try pushes or anything whenever needed.
21:45:00 <nirik> ok, so thats all I had on the agenda.
21:45:05 <nirik> #topic Open Floor
21:45:12 <nirik> any other items?
21:45:19 <stahnma> I'd also like training on the pushes using the newer systems.
21:45:36 <stahnma> I was wanting to gather some metrics on EPEL
21:46:09 <dgilmore> id like to keep the people who know the passphrase for the epel key small
21:46:09 <smooge> no the main things I was wondering was how we might be able to use Fedora Community for ourselves
21:46:09 <abadger1999> Policy on when pushes will be performed?
21:46:17 <stahnma> basically, are we losing contributors in EPEL?
21:46:22 <nirik> yeah, we should do them more often... weekly? daily for testing?
21:46:27 * smooge realizes that made no sense.
21:46:28 <stahnma> also, why are certain fedora maintianers not interested in EPEL?
21:46:59 <abadger1999> testing daily +1
21:47:01 <smooge> I have an off-the wall idea I would like to propose
21:47:04 <stahnma> also, what are the primary reasons for people not using EPEL and using repo_x?
21:47:06 <nirik> stahnma: the two reasons I think I hear most are: 1) involved in another rhel addon repo, or 2) have no time/interest in rhel, only use fedora.
21:47:19 <smooge> nirik, agreed
21:47:55 <smooge> stahnma, the reasons for choosing which repo are almost the same as choosing KDE vs GNOME vs LXDE
21:48:01 <nirik> abadger1999: I'm fine just adding them to my daily schedule once we have all the push procedures down. I think weekly might be better for stable tho, so people know when to expect them to appear. Or even bi-weekly.
21:48:31 <abadger1999> <nod>... the only thing there is that testing (at least in Fedora) doesn't go to the buildroot.
21:48:36 <smooge> my proposal is to change the name of the repos location from stable to 'current'.
21:49:15 <nirik> abadger1999: yeah, we should do the same for epel I think.
21:49:32 <smooge> I would like to see how 'current', 'next-week', 'next-month' would look
21:49:46 * nirik shrugs. Thats fine if you think 'stable' is too much of a promise I guess.
21:49:54 <nirik> more repos==more pain.
21:50:20 <smooge> nirik I think it is too much of a promise especially when we aren't really testing stability to the level we once hoped for.
21:50:54 <smooge> well actually next-week/next-month are more like staging/testing. but I can understand having too many repos
21:51:13 <smooge> actually staging/testing would be a better name
21:51:24 <nirik> well, we only call it 'stable' in docs right, there is no 'stable' directory.
21:51:44 <smooge> .... has been believing the docs too much
21:52:16 <nirik> it's just 'testing' and then the 5 or 4 repo.
21:52:17 <smooge> stahnma, what kind of statistics did you want to look for and how would we measure them? Mirror maanger
21:52:35 <smooge> nirik, ok so it would be more of a change in documentation.
21:52:45 <stahnma> I'm out.
21:52:46 <stahnma> thanks
21:52:48 <nirik> yeah, I think so. I agree we shouldn't promise 'stable'
21:52:58 <nirik> thanks for coming stahnma
21:53:32 <nirik> so, everyone ok with trying for daily pushes to testing and weekly (tuesday?) pushes to stable?
21:53:49 <smooge> thanks stahnma
21:53:50 <smooge> thanks
21:54:00 <smooge> nirik +1
21:54:26 <smooge> maybe we can shield ourselves from black tuesday
21:54:54 <SmootherFrOgZ> nirik: +1
21:55:24 <nirik> well, I think it's good to have a known day when they show up. Friday is bad as people might head out for the weekend and they appear... monday's are kinda hectic...
21:55:32 <nirik> I guess wed or thursday would be fine too, IMHO.
21:57:20 <nirik> ok, anything else? or shall we close out the meeting?
21:57:38 <smooge> I actually like Tuesday
21:57:51 <smooge> Microsoft picked it for similar reasons.
21:58:12 <smooge> I am ok with closing.
21:58:34 <nirik> ok, will close in 60sec if nothing else appears.
21:59:34 <nirik> #endmeeting