20:32:05 <tremble> #startmeeting EPEL
20:32:05 <zodbot> Meeting started Mon Nov 15 20:32:05 2010 UTC.  The chair is tremble. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:32:05 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:32:14 <tremble> chair nirik smooge
20:32:17 * nirik is around.
20:32:19 <tremble> chaira nirik smooge
20:32:24 * nirik hands tremble a #
20:32:24 <tremble> chairs nirik smooge
20:32:28 <tremble> #chairs nirik smooge
20:32:33 <tremble> #chair nirik smooge
20:32:33 <zodbot> Current chairs: nirik smooge tremble
20:32:44 <tremble> #topic Rolecall
20:32:49 * tremble is here
20:33:26 <tremble> smooge abadger1999 schlobinux_ ?
20:33:32 * abadger1999 here
20:33:37 * schlobinux_ is here
20:33:41 * nirik is here
20:34:01 * jsmith lurks
20:34:22 <smooge> sort of here
20:34:39 <smooge> monday madness and all that
20:35:03 <tremble> #topic Agenda items...
20:36:11 <tremble> RHEL6 is out!
20:36:12 <nirik> I have a FYI about the buildsystem and rhel6 final.
20:36:44 <tremble> Reminding packagers of unbuilt RHEL6 packages...
20:37:00 <tremble> Anything else?
20:37:31 <nirik> broken deps still.
20:37:59 <tremble> #topic Broken Deps
20:38:02 <tremble> nirik ?
20:38:16 <nirik> ok, there are still broken deps.
20:38:21 <nirik> epel5 stable isnt too bad.
20:38:36 <nirik> but epel5-testing, epel4-stable, and epel4-testing now need some work.
20:39:06 <nirik> I can try and poke at them over the next week or two... focusing on epel4-stable.
20:39:09 <nirik> but any help welcome
20:39:23 <tremble> Do we know what state epel6 is now in?
20:39:30 <tremble> for broken deps
20:40:03 <nirik> not off hand, perhaps we should ask stanmah to run a full slate of them.
20:40:24 <tremble> #info epel5 stable is now in a much better state
20:40:25 <nirik> thinking about it, I wonder how hard it would be to dump the output of each to a webpage.
20:40:36 <nirik> then run it like every 12 hours... so you can look at the web page for mostly up to date.
20:40:59 <tremble> #info nirik hopes to work on epel4 stable over the next couple of weeks.
20:41:31 <tremble> Personally I think that would be a fairly good thing to do.
20:42:22 <nirik> can ask
20:43:02 <tremble> #info nirik we ask around and see if we can get the script running regularly and updating a web page somewhere...
20:43:03 <nirik> anyhow, thats all I had. Would be good to know where epel6 is at.
20:43:17 <smooge> I think er put a list out saying " we are pulling the following from the EL-X on 12-01-2010"
20:43:20 <smooge> and do it
20:44:03 <nirik> the el4 stuff thats not easy to fix, I am fine with unpushing now.
20:44:13 <nirik> for el6, yeah, a warning might be nice.
20:45:12 * nirik has nothing more on deps.
20:45:30 <smooge> el4/el5
20:45:36 * tremble nods
20:45:47 <smooge> el6 I would say everything goes into epel-testing and we start pushing from there
20:46:08 <smooge> but I am a cranky old man
20:46:18 <tremble> 2 weeks notice that we're going to yank stuff that won't install anyway from EPEL4/5 doesn't seem too harsh
20:47:07 <nirik> well, we have warned about broken deps in el4/el5 for years. I don't think waiting more is worthwhile.
20:47:35 <smooge> I just want a date after thanksgiving so I can have enjoyed something before shit starts again
20:47:43 <smooge> :)
20:47:53 <nirik> in epel5 we should only have dogtag (they hope to fix in next 5.x update) qcairo-devel (a centos/repo issue, not us) and xulrunner-devel-unstable (a repo issue)
20:48:12 <nirik> so once dogtag is fixed in epel5, epel5 stable should be in perfect shape
20:48:23 <tremble> Proposal: We send out an email explaining that we're going to pull anything with broken deps, and just pull them one week later
20:48:43 * nirik thinks we already did.
20:49:18 <nirik> It's just a matter of doing it for the el4 ones that can't be easily fixed.
20:49:29 <smooge> we did. like nirik said we have given a lot of warning. we just need a date to do it
20:49:48 <tremble> Actually I'm referring to pulling all of them, not just the "hard to fix" ones.
20:49:49 <nirik> I'm happy to try and do it over the next week. ;)
20:50:11 <smooge> heck I don't mind if its all of them. its been 2 months of warning I think?
20:50:25 <nirik> tremble: all the el4-stable ones you mean?
20:50:34 <tremble> yes
20:50:50 <nirik> I'm happy to try and do it over the next week. ;)
20:51:20 <tremble> Anyone object?
20:51:51 * nirik listens to the silence.
20:52:19 * tremble laughs as his laptop starts playing "The sound of silence"
20:52:41 <nirik> ha
20:53:04 <tremble> #agreed All EPEL 4 packages with broken deps will be pulled over the next week or 2 as nirik has time.
20:53:43 <tremble> #topic RHEL-6
20:53:54 <tremble> #info It's finally reached GA!
20:54:47 <nirik> #info builders are updated with final bits.
20:55:11 <nirik> it would be good to now look at those packages that were only in workstation-optional and see if we have them...
20:55:18 * tremble nods
20:55:52 <nirik> that would be: fftw
20:56:03 <nirik> openoffice.org-sdk
20:56:21 <nirik> openoffice.org-opensymbol-fonts
20:57:19 <nirik> I think none of them are available in server.
20:57:45 <nirik> but I could be wrong.
20:58:46 <smooge> ugh
20:59:00 <nirik> the perl-Time-modules and perl-MIME-tools I think are available.
20:59:10 * tremble logs into a handy satellite server...
20:59:16 <nirik> humm... or not.
20:59:18 <smooge> let us queue discussion of EL-5 where we said "gee all these things won't build so do we wait for CentOS?"
20:59:56 <nirik> well, it's unclear to me that we have an issue, it would be good to confirm before we do anything.
20:59:58 <tremble> fftw is in an "optional" repo
21:00:11 <nirik> tremble: which one? ;)
21:00:27 <tremble> RHEL Server Optional (v. 6 64-bit x86_64)
21:00:32 <nirik> great.
21:00:36 <nirik> how about the openoffice* things.
21:00:49 <smooge> probably not.
21:01:03 <smooge> so lets branch libreoffice into EPEL-6 and use those :)
21:01:22 <nirik> ha.
21:01:28 <tremble> RHEL Server Optional (v. 6 64-bit x86_64)
21:01:45 <nirik> both of them? hurray!
21:02:24 <tremble> Yup
21:02:31 <nirik> so, I think thats the full list. I can update tickets and ask people to build/confirm.
21:03:03 <tremble> It looks like the decision may have been to put them all into a "we sort of support this" repo...
21:04:42 <nirik> yeah, which is just fine for us.
21:04:44 <tremble> Not quite sure how additional entitlements work on top of this lot, but it looks like life's a lot simpler than we feared,
21:05:00 <nirik> at least so far...
21:05:11 <nirik> until someone needs something only in workstation* again. ;)
21:05:13 * tremble laughs
21:05:56 <tremble> nirik Do you and tibbs have access to somewhere that it would even be possible for you to check before branching?
21:06:16 <nirik> well, the builders possibly, but they haven't been updated to final yet. ;(
21:06:42 <nirik> as soon as there is a centos6 we could look at that...
21:06:59 <nirik> or if someone gave me a rhel6 entitlement I could setup a virtual here.
21:07:11 <smooge> I will try to make a list of what is different between the two so we can have a "well that won't work" list
21:07:21 <nirik> smooge: that would be great.
21:07:27 <nirik> also there is client*
21:09:07 <smooge> client the bastard child of server+workstation
21:09:12 <tremble> #info it looks like most of the "workstation-optional" packages that were causing problems are available for server
21:10:38 <tremble> smooge: Does infrastructure have access to RHN?  Could we give nirik/tibbs access to somewhere they could run "sudo yum list <pkgname>"
21:10:43 <smooge> ok will get those lists asap
21:10:49 <skvidal> tremble: ?
21:10:51 <skvidal> huh?
21:10:56 <skvidal> why do you need access to rhn to run that?
21:11:14 <tremble> You need a box plugged into RHN/Satellite.
21:11:17 <smooge> I think they are wanting to see if a package is available to build against
21:11:25 <smooge> tremble, or a local repo :)
21:11:35 <skvidal> tremble: for what?
21:11:49 <nirik> well, we just need a rhel6 box that is up pointed at the final repos.
21:11:52 <skvidal> tremble: we have the pkgs available in repos named the same way for use in building pkgs
21:12:15 <nirik> skvidal: maintainer foo requests a epel6 branch for package bar... but we have no way of checking if bar is already in RHEL6.
21:12:51 <skvidal> nirik: I believe you have access to that data
21:13:05 <skvidal> nirik: but if you do not - I can make sure that folks in epel can
21:13:21 <nirik> skvidal: yeah. (Although are all 6 boxes pointed to the final repos now? or only the ones upgraded to 6 final)?
21:13:55 <nirik> anyhow, where are we...
21:14:11 <skvidal> most/all of our rhel6 boxes are on beta - but  a lot of those are publictest/staging boxes
21:14:32 <skvidal> nirik: but I can give you the urls you can use to repoquery things
21:14:43 <nirik> skvidal: ok, cool.
21:14:53 <tremble> #info Hopefully there should be a way (RSN) for tibbs/nirik to check the validity of EPEL branch requests.
21:15:51 <tremble> I think we previously agreed that we'd wait for CentOS 6 and then switch EPEL from its "beta" mode.
21:16:00 <nirik> yes, I think thats best
21:16:13 <tremble> #info  we previously agreed that we'd wait for CentOS 6 and then switch EPEL from its "beta" mode.
21:17:01 <tremble> Do we want another run of the "nag" script?  The following packages have not yet been built for EPEL6 ?
21:17:13 <nirik> yeah, I think that would be good.
21:17:25 <nirik> add a "RHEL6 is out now... " type thing to the top of the email.
21:18:09 <tremble> #action tremble Try to find the time to run the script to ask maintainers to build any unbuilt EPEL6 packages.
21:19:14 <tremble> #topic Bugs
21:19:30 <tremble> #info 200 EPEL bugs found ...
21:19:45 <nirik> back up to 200? ;(
21:19:51 <tremble> #info Hopefully with RHEL6 now GA we should be able to close off a load.
21:20:30 <tremble> I've not have the time/energy to do another triage run for the last week or two.  That will probably allow us to close a few.
21:20:50 <nirik> cool
21:21:23 <tremble> Any one have anything else to say on the subject?
21:22:09 <nirik> nope
21:22:12 <tremble> #topic Open Floor
21:23:12 <skvidal> umm
21:23:13 <skvidal> I do
21:23:18 <skvidal> so - I don't want this to be taken wrongly
21:23:22 <skvidal> but why are we waiting for centos?
21:23:55 <tremble> skvidal: Just because that gives us time to get everything tidied up now we know what's actually available.
21:24:04 <smooge> also most users of EPEL are centos systems.
21:24:16 <tremble> There was a faff with various dependencies not being available
21:24:32 <skvidal> smooge: really? we have stats on that?
21:25:11 <nirik> yeah, now that final is out, people can look at whats really in it and decide what they are landing for epel...
21:25:24 <nirik> I guess we don't need to wait for centos, but waiting a bit seemed like a good idea to me.
21:25:33 <skvidal> so waiting for a bit is fine
21:25:36 <smooge> not really. but since a large number of systems come from clouds which don't offer RHEL .. I make an assumption
21:25:41 <skvidal> if that's why we're waiting
21:25:55 <skvidal> nirik: but that we're 'waiting for centos' feels a bit odd to me
21:26:21 <skvidal> and also feels kinda crappy to rhel customers who want to use epel
21:26:28 <nirik> well, people who use centos are consumers of ours... but of course going out of beta wouldn't hurt them any.
21:26:32 <smooge> well the other issue is that many of the packagers say they can't wont be able to test until CentOS
21:26:37 <nirik> well, they can use epel fine right now.
21:26:44 <skvidal> nirik: right 'beta'
21:26:54 <skvidal> which a lot of the risk-averse folks using an EL-distro are not going to touch
21:27:00 <smooge> skvidal, I don't think EPEL ever gets out of beta
21:27:13 <nirik> oh, I almost forgot: https://bugzilla.redhat.com/show_bug.cgi?id=653146 reviewers welcome.
21:27:39 <nirik> skvidal: yeah, true. Although many risk averse EL people aren't going to move to EL6 until they have pounded on it for a while.
21:27:51 <skvidal> nirik: also true - but some of that pounding includes epel
21:27:58 <skvidal> anyway - it's fine for us to wait for a bit
21:28:05 <skvidal> and even to wait for centos to have a release
21:28:09 <nirik> perhaps we should ask for more input from the list as well?
21:28:10 <skvidal> but I wanted to make sure we had a reason for it
21:28:16 <skvidal> nirik: a good idea
21:28:52 <tremble> skvidal: We have reasons for waiting for "a while", CentOS 6 just seems like a milestone.
21:28:54 <nirik> I can post to the list about it.
21:29:35 <skvidal> tremble: again - I'm fine with it - I just worried that it feels a little crappy to say "yes, we know these are pkgs for rhel, but we're going to wait for the rhel-rebuild to call it stable"
21:29:45 <tremble> #action nirik to post the EPEL-devel about when we want to come out of beta.
21:30:11 <tremble> skvidal: Yeah, I can see your point.
21:30:39 <tremble> We could just as easily say "1 month post GA"
21:30:52 <skvidal> which seems reasonable to me
21:31:15 <skvidal> the other reason  why 'waiting for centos' bugs me is that it means our release cycle is completely outside of fedora control
21:31:27 <skvidal> s/fedora/fedora epel/
21:31:33 <tremble> Let's see what the view of the list is...
21:31:38 <skvidal> nod
21:31:40 <skvidal> agreed
21:31:48 <nirik> yeah, perhaps 1 month would make more sense, but will see.
21:31:53 <gholms> Does it matter that people have to wait for centos to do mock builds?
21:32:52 <skvidal> gholms: they don't
21:32:56 <skvidal> gholms: they can do them in koji
21:33:07 <skvidal> and afaik, those will be against rhel6
21:33:08 <gholms> Sure, if they have access to a koji server with RHEL packages.
21:33:09 <smooge> skvidal, but can they test them?
21:33:29 <skvidal> gholms: if they have an account in fedora and are a packager - they have access to scratch builds
21:33:32 <skvidal> which is a koji with rhel pkgs
21:33:54 <nirik> yeah, testing is harder.
21:34:01 <skvidal> smooge: maybe you  and I should talk to someone about getting packagers in epel access to something
21:34:02 <nirik> you can test with beta2.
21:34:04 <skvidal> smooge: or an entitlement
21:34:12 <nirik> there was talk about that long ago...
21:34:12 <skvidal> smooge: to let them setup their own instances
21:34:15 <nirik> but it didn't go anywhere. ;(
21:35:15 <smooge> skvidal, ok off line
21:35:38 <skvidal> smooge: sounds fine w/me
21:36:30 <tremble> Shall we call it a day and leave smooge and skvidal to see what they can arrange?
21:36:34 <skvidal> nod
21:36:37 <skvidal> I think so
21:37:10 <tremble> closing in 1 minute...
21:38:06 <tremble> #endmeeting