2010-09-13 meeting minutes/summary

Kevin Fenzi kevin at scrye.com
Mon Sep 13 19:28:31 UTC 2010


=====================
#fedora-meeting: EPEL
=====================

Meeting started by nirik at 18:04:52 UTC. The full logs are available at
http://meetbot.fedoraproject.org/fedora-meeting/2010-09-13/epel.2010-09-13-18.04.log.html

Meeting summary
---------------
* Init process/agenda creation  (nirik, 18:05:07)

* Broken Dependencies  (nirik, 18:10:10)
  * ACTION: stahnma will break dep emails up into 2 (one for stable, one
    for testing)  (stahnma, 18:17:46)
  * ACTION: stahnma will mail maintainers of broken dep packages
    (nirik, 18:18:08)

* bugs  (nirik, 18:19:35)
  * we have 234 bugs currently.  (nirik, 18:21:34)
  * LINK: http://tr.im/epelbugs   (stahnma, 18:21:46)
  * ACTION: stahnma to write down whiteboard process for bugs.  (nirik,
    18:30:28)
  * ACTION: nirik to talk to bugzappers  (nirik, 18:30:34)
  * ACTION: all to work on bugs and try and close at least 1/week. ;)
    (nirik, 18:30:47)

* rhel6 outdated  (nirik, 18:32:00)
  * ruby might be out of date in 6, but they are working on it.  (nirik,
    18:33:12)

* rhel and epel version matching  (nirik, 18:36:37)
  * ACTION: tremble will generate a UPDATE-CAREFULLY file for packages
    that are only in RHEL6 for some branches that EPEL6 is carrying for
    the others.  (nirik, 18:45:25)
  * ACTION: nirik will commit that file to the indicated packages.
    (nirik, 18:45:34)

* Open Floor  (nirik, 18:49:18)
  * LINK:
    https://fedoraproject.org/wiki/Package_update_acceptance_criteria
    (stahnma, 18:49:58)
  * LINK:
    https://admin.fedoraproject.org/updates/rubygem-json-1.4.3-3.el5?_csrf_token=23055744d49c160c18dcef1b25f08cda668c4c10
    (stahnma, 18:51:48)
  * ACTION: stahnma will send an email to list about crazy BuildRoot
    error messages  (stahnma, 18:56:18)
  * will determine next weeks meeting time on list.  (nirik, 19:00:58)

Meeting ended at 19:01:02 UTC.




Action Items
------------
* stahnma will break dep emails up into 2 (one for stable, one for
  testing)
* stahnma will mail maintainers of broken dep packages
* stahnma to write down whiteboard process for bugs.
* nirik to talk to bugzappers
* all to work on bugs and try and close at least 1/week. ;)
* tremble will generate a UPDATE-CAREFULLY file for packages that are
  only in RHEL6 for some branches that EPEL6 is carrying for the others.
* nirik will commit that file to the indicated packages.
* stahnma will send an email to list about crazy BuildRoot error
  messages
--
18:04:52 <nirik> #startmeeting EPEL
18:04:52 <zodbot> Meeting started Mon Sep 13 18:04:52 2010 UTC.  The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:04:52 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
18:04:56 <nirik> #meetingname epel
18:04:56 <zodbot> The meeting name has been set to 'epel'
18:05:07 <nirik> #topic Init process/agenda creation
18:05:15 <nirik> #chair tremble stahnma smooge
18:05:15 <zodbot> Current chairs: nirik smooge stahnma tremble
18:05:24 <nirik> Who has items for the agenda?
18:05:47 <stahnma> I have deps list...the continued saga.  And Bug Count...moving the wrong direction.
18:06:05 <stahnma> and RHEL 6, outdated before it GAs
18:06:06 <tremble> Matching RHEL versions to EPEL versions...
18:06:07 <stahnma> :(
18:06:25 <smooge> ok I will be in a board meeting
18:06:32 <smooge> so continue you on me
18:06:46 <nirik> smooge: sorry for poor timing. ;(
18:06:58 <stahnma> smooge: is this a common conflicting time?
18:07:06 <stahnma> other topic idea:  meeting times
18:07:50 <nirik> ok, so I have:
18:07:52 <nirik> meeting times
18:07:52 <nirik> broken dependencies
18:07:52 <nirik> bugs
18:07:52 <nirik> rhel6 outdated
18:07:52 <nirik> rhel and epel version matching
18:07:56 <smooge> stahnma, it will be from now on
18:08:07 <stahnma> smooge: ok, we'll keep discussing itmes
18:08:09 <smooge> I am sorry about that we just changed also this week
18:08:10 <stahnma> times even
18:08:21 <smooge> an hour earlier and 2 hours later work ok for me
18:08:35 * stahnma votes to take times to list...again
18:08:42 * tremble laughs
18:08:51 <nirik> meeting times are always a pain. ;(
18:09:05 <stahnma> I really liked this one...but oh well
18:09:18 <stahnma> moving on?
18:09:31 <nirik> ok, any other agenda items? or shall we start on them?
18:09:47 <stahnma> that's all I can think of right now
18:09:55 <tremble> We can always aob at the end
18:10:02 <nirik> yep. ok.
18:10:10 <nirik> #topic Broken Dependencies
18:10:25 <nirik> was meaning to ask: was the list against just the stable updates?
18:10:32 <stahnma> I keep attempting to run a dep check at least once a week.
18:10:47 <stahnma> It had both, stable and then further down the email was stable+testing
18:11:11 <stahnma> We need to come up with a resolution plan for packages in stable with broken deps
18:11:22 <stahnma> either move them back to testing, or get the deps in and push karma ASAP
18:11:46 <nirik> I'm having trouble seeing where one ends and the other begins...
18:11:59 <nirik> I fixed my package that was listed.
18:12:07 <nirik> tremble fixed one I reported. ;)
18:12:11 <stahnma> I'd also love it if bodhi wouldn't push to stable with deps being met, but that may be more difficult
18:12:26 <nirik> we need autoqa for that I fear.
18:12:31 * tremble grins
18:12:43 <tremble> Can we get autoqa on the depsolve only?
18:12:53 <tremble> just as a starting point?
18:13:16 <tremble> Although having the rpmlint run would be nice too.
18:13:31 <stahnma> rpmlint gives a lot of false positives, at least on ruby stuff
18:13:43 <nirik> tremble: once it's ready, sure.
18:13:50 <stahnma> mostly due to any filename with a % in it is considered an unexpanded macro
18:14:07 <tremble> Yeah, you don't block but it's nice to have the rpmlint output mailed to you
18:14:08 <stahnma> nirik: in the future I could break that into two emails
18:14:27 <nirik> stahnma: might be good.
18:14:28 <stahnma> I'd also like to notify package maintainers directly.
18:14:37 <nirik> that also would be good.
18:14:45 <stahnma> I need to look at the Fedora code that does that, rather than hack together my own
18:15:00 <stahnma> at least from a starting point
18:15:17 <tremble> stahnma : I've got most of that from the EPEL-6 nag script, can send it over if you'd like
18:15:23 <stahnma> sure
18:15:39 <stahnma> is there a method to remove a package from stable?
18:15:43 <stahnma> is the rel-eng ticket?
18:15:49 <stahnma> is that ^^
18:16:00 <nirik> sure, or if you get a list, I or dgilmore can do I think.
18:16:10 <stahnma> ok
18:16:22 <stahnma> I know of a couple that won't have deps met for a long time (possibly ever)
18:16:32 <nirik> yeah, there's some old ones in there. ;(
18:16:44 <tremble> There's no way the git-bugzilla one's getting fixed.
18:16:55 <stahnma> any other action required on deps?
18:17:28 <tremble> So what's the plan...
18:17:34 <nirik> might be worth setting up a time to get some of us together and work on them... ie, decide which can be fixed and fix those.
18:17:46 <stahnma> #action stahnma will break dep emails up into 2 (one for stable, one for testing)
18:18:08 <nirik> #action stahnma will mail maintainers of broken dep packages
18:18:18 <stahnma> right
18:18:24 <stahnma> I'll look into a few other ideas I have also
18:18:36 <stahnma> I think setting up a time is a good idea
18:18:47 <stahnma> that can transistion us into talking about bugs too :)
18:18:52 <nirik> ok. It might be nice to mail them, then unpush after a bit ... to give them time to deal with it.
18:18:58 <nirik> indeed.
18:19:11 <nirik> anything else on broken deps?
18:19:11 <stahnma> yeah, some more mature workflow will be required
18:19:22 <stahnma> not right now.  We might revisit in a week or two
18:19:31 <nirik> yeah, sounds good.
18:19:35 <nirik> #topic bugs
18:19:37 <tremble> Rather than mailing is it worth bugzilla entries
18:19:56 <tremble> Can then put a tracking bug in.
18:20:11 <stahnma> at the rate of our bug rate closure, I would bet not :)
18:20:23 * tremble laughs
18:20:26 <nirik> tremble: I think some of them are filed... that one you fixed with perl-Font-TTF I had already filed.
18:20:52 <tremble> Oh ok didn't spot that, I was going off stahnma's email
18:21:18 <nirik> I added karma to the update and noted the bug number.
18:21:34 <nirik> #info we have 234 bugs currently.
18:21:46 <stahnma> http://tr.im/epelbugs
18:21:54 <stahnma> still works...until tr.im is completely shut down
18:22:33 <nirik> so, what can we do here? there are probibly some that are low lying fruit.
18:22:38 <stahnma> Some bugs haven't had attention at all
18:22:47 <nirik> we could ask bugzappers to try and get some people triaging them if we think it would help.
18:22:49 <stahnma> some are requests for something really simple, but I always wonder if I should just fix it
18:23:03 <stahnma> others are requests for new versions...which can be problematic
18:23:03 <nirik> I am thinking we should start doing so...
18:23:11 <nirik> just step in and fix what can easily be fixed.
18:23:23 <stahnma> I mean, I am a proven packager, but I never know when I am overstepping my bounds
18:23:56 <stahnma> I'd like to have some sort of goal to close 1 bug a week at least
18:24:02 <stahnma> It's not much, but it's better than 0
18:24:05 * tremble nods
18:24:08 <nirik> yeah.
18:24:28 <stahnma> I mean, I guess that's only 3 bugs a week if we all do it.  But, maybe it will be contagious
18:24:34 <nirik> we could try another bug day, but the last one flopped. ;(
18:24:38 <stahnma> and I am in favor of working with bugzappers
18:24:57 * tremble isn't PP so can't do much with most of them.
18:25:18 <nirik> tremble: attach patch, wait a bit, and then ask a pp to commit I guess...
18:25:24 * tremble nods
18:25:27 <stahnma> tremble: you can triage at least, and let us know which ones to fix also
18:25:35 <tremble> Fair pont
18:25:43 <nirik> stahnma: I can ask them... what do we want them to do exactly ?
18:25:58 <stahnma> we have a fair number (40 or so I am guessing) of bugs asking for 'please branch XYZ into EPEL'
18:26:07 <stahnma> we need a list of willing co-maintainers for epel
18:26:19 * tremble is willing to do so.
18:26:31 <nirik> well, some of those, the answer is: foo is too old, sorry.
18:26:35 <stahnma> nirik: I am not sure yet.  Probably classify the bug into categories...new branch, real bug, update foo, blah
18:26:49 <nirik> we probibly need a page/list/checklist then...
18:26:49 <stahnma> maybe even teach the bugzilla API
18:26:51 <stahnma> :)
18:27:14 <tremble> Use the whiteboard?
18:27:20 <stahnma> Along the lines of co-maintainership; I am willing to do anything ruby related. :)
18:27:30 <nirik> tremble: yeah, thats the typical way.
18:27:55 <tremble> nirik: Or a set of tracker tickets? AIUI
18:28:08 <nirik> yeah, thats possible too... but seems overkill.
18:28:14 * tremble nods
18:28:28 <tremble> See how we get on with whiteboarding?
18:28:28 <stahnma> I tried using the whiteboard a year or two ago for epel bugs
18:28:44 <stahnma> I bet I could reverse engineer what I did then for the bugzappers or any volunteer
18:28:57 <nirik> that would be lovely if you are able.
18:29:33 <stahnma> I mean, it was a manual set of steps.  Review bug, put label "foo", "bar" or "baz" in whiteboard
18:29:39 <stahnma> I just don't remember what the labels were
18:29:48 <stahnma> but since most of the bugs are still open, I can find out
18:30:03 <nirik> yeah, if it's something we can write down I can talk to bugzappers about having some folks do it.
18:30:15 <stahnma> sounds good
18:30:28 <nirik> #action stahnma to write down whiteboard process for bugs.
18:30:34 <nirik> #action nirik to talk to bugzappers
18:30:36 <stahnma> yes
18:30:37 <tremble> I'm likely to have some time to do a bit of triaging over the next couple of weeks.
18:30:42 <stahnma> super
18:30:47 <nirik> #action all to work on bugs and try and close at least 1/week. ;)
18:30:53 <nirik> tremble: excellent.
18:31:01 <stahnma> are dep issues more important than long-standing bugs?
18:31:02 <nirik> anything more on bugs?
18:31:11 <nirik> stahnma: I would say yes.
18:31:14 * stahnma thinks dep issues are huge
18:31:21 * tremble nogs
18:31:25 <stahnma> ok
18:31:28 <stahnma> that's all on bugs
18:31:30 <stahnma> from me
18:31:36 <tremble> And also closable in a reasonable timeframe.
18:31:42 <nirik> fixing deps may close some bugs too. ;)
18:31:46 <stahnma> true
18:31:49 * tremble nods
18:31:51 <nirik> ok, moving along then...
18:32:00 <nirik> #topic rhel6 outdated
18:32:08 <nirik> stahnma: care to expand on this?
18:32:14 <stahnma> I think I was wrong on the one thing I really cared about for this
18:32:35 <stahnma> I was under the impression that RHEL6 had 1.8.6 which would hurt a lot
18:32:36 <tremble> There's a couple of ruby pkgs that were just behind the useful release...
18:32:48 <stahnma> but, they are moving it to 1.8.7 :)
18:32:55 <tremble> Cool :)
18:32:55 <nirik> ah ha.
18:32:58 <stahnma> also, the version of rubygems might be too old
18:33:03 <stahnma> for some packages
18:33:12 <nirik> #info ruby might be out of date in 6, but they are working on it.
18:33:34 <llaumgui_zhukov> .fas llaumgui
18:33:35 <zodbot> llaumgui_zhukov: llaumgui 'Guillaume Kulakowski' <llaumgui at gmail.com>
18:33:36 <stahnma> nirik: I think it actually is resolved.  But, I don't know how to verify yes.
18:33:40 * MrTom asks French meeting attendees to go to #fedora-meeting-1 not to disturb EPEL meeting exceptionaly
18:33:53 <nirik> MrTom: oops. are we cutting into your meeting time? :(
18:34:05 <nirik> MrTom: sorry if so... we could move.
18:34:12 <MrTom> nirik, go on, we move to #fedora-meeting-1
18:34:24 <MrTom> no, we move, it will be hell with meeting bot if you move :)
18:34:34 <stahnma> MrTom: thanks. We'll be better about scheduling next time.
18:34:37 <nirik> MrTom: ok. sorry about that. ;(
18:34:39 <MrTom> #fedora-meeting-1 is free
18:34:42 <MrTom> don't worry
18:34:50 <MrTom> it's all fixable :)
18:35:02 <nirik> MrTom: we will not overlap next week... ;)
18:35:15 * tremble thinks MrTom for being understanding.
18:35:20 <nirik> stahnma: yeah, do you know if they plan a beta3?
18:35:21 * tremble thanks MrTom for being understanding.
18:35:43 <tremble> nirik There was a post on the beta list suggesting not.
18:35:43 <stahnma> nirik: I don't. I haven't had as much exposure to EL6 as I would have liked, mostly due to $DAYJOB changes.
18:35:55 <nirik> ok.
18:35:58 <stahnma> ruby-sig just mentioned that it was moved to 1.8.7 :)
18:36:11 <nirik> ok, anything further on this topic?
18:36:18 <stahnma> I don't think so.
18:36:33 <nirik> ok, moving on...
18:36:33 * stahnma </rant style=happier>
18:36:37 <nirik> #topic rhel and epel version matching
18:36:44 <nirik> tremble: this was your topic?
18:37:17 <tremble> Yeah, I've cleaned up most of the perl packages with version mismatches, what do we want  to do about the rest.
18:37:30 <tremble> cairomm was the significantly notable one...
18:38:03 <tremble> You suggested just untagging it since it seems to be available for all arches...
18:38:47 <nirik> what was the case on that one again?
18:38:58 <nirik> in beta1, dropped beta2, added back in refresh?
18:39:19 <tremble> On server it's available for all arches and EPEL is 1.8.4, RHEL 1.8.1
18:39:41 <tremble> Something like that I believe
18:39:54 <nirik> yeah, so I think we should just block it and ask the maintainer to dead.package it.
18:40:23 <tremble> We also need to decide on a standard way to put a warning flag on the packages
18:40:51 <nirik> perhaps a README.rhel file?
18:41:14 <tremble> Yeah but making it obvious that that file is intended for the package maintainers.
18:41:36 <nirik> 00-MAINTAINER-README-NOW ?
18:41:38 <nirik> :)
18:41:48 <tremble> Soomething like that :)
18:41:59 * nirik notes we had a case or cases of this for epel4 long ago.
18:42:23 * nirik looks for the packages.
18:42:50 <tremble> In this case probably wants to state which RHEL flavours/arches it IS available for.
18:42:57 <nirik> got it
18:42:59 <nirik> UPDATE-CAREFULLY
18:43:09 * tremble laughs
18:43:10 <nirik> ie, for yum in el4:
18:43:19 <nirik> This package exists in CentOS-4 base.
18:43:19 <nirik> Please be careful when updating that the EVR here remains lower than in CentOS-4.
18:43:28 * tremble nods
18:43:44 <nirik> so, I think UPDATE-CAREFULLY would be good to keep using.
18:43:55 * tremble nods
18:43:56 <nirik> just need content about this issue, and a list of packages to commit it to.
18:44:10 <nirik> tremble: can you whip up content and list and I can commit it?
18:44:25 <tremble> Sure, want me to email it to you.
18:44:32 <nirik> works for me.
18:45:25 <nirik> #action tremble will generate a UPDATE-CAREFULLY file for packages that are only in RHEL6 for some branches that EPEL6 is carrying for the others.
18:45:34 <nirik> #action nirik will commit that file to the indicated packages.
18:45:39 <nirik> anything more on this topic?
18:45:40 <tremble> With regards the pkgs that need untagging, I'd suggest that we only untag for now and wait for GA before blocking/dead.pkg -ing
18:46:12 <nirik> yeah, they could change before release for sure...
18:46:44 <tremble> I seem to recall  that un-blocking in koji can get messy.
18:47:31 <nirik> well, sure...
18:47:41 <nirik> is cariomm the only one currently in that state?
18:48:00 <tremble> There's a few, again shall I just email you?
18:48:22 <nirik> sure, works for me. I will need to check with dgilmore on how best to block or untag them... might cc him on it as well.
18:48:31 <tremble> Ok
18:49:00 <stahnma> I have a couple of open discussion items if we have any time left
18:49:18 <tremble> 10 mins...
18:49:18 <nirik> #topic Open Floor
18:49:22 <nirik> take it away. ;)
18:49:37 <stahnma> The Acceptance Criteria page on the wiki does not mention epel
18:49:43 <stahnma> yet, I am pretty sure our rules are different
18:49:52 <tremble> link?
18:49:58 <stahnma> https://fedoraproject.org/wiki/Package_update_acceptance_criteria
18:49:59 <nirik> which acceptance criteria?
18:50:07 <nirik> yes, thats fedora specific.
18:50:12 <lmacken> bodhi does not enforce that for EPEL (yet)
18:50:35 <stahnma> well, bodhi at the top of the page tells me it does.  Then when I try to push, it says I don't meet the criteria.
18:50:43 <stahnma> Then I just wonder why, because according to the page, I do.
18:51:02 <lmacken> stahnma: the top of the page says 'for all fedora releases'
18:51:03 <stahnma> I thought we were saying 2 weeks instead of 1 week like Fedora
18:51:13 <tremble> It's probably worth having it documented somewhere why RHEL tends to have more of the parallel installable pkgs to.
18:51:20 <nirik> stahnma: http://fedoraproject.org/wiki/EPEL/GuidelinesAndPolicies
18:51:21 <lmacken> stahnma: yeah, it may enforce the 2 week thing... but that ways always enforce, by hand anyway, for epel
18:51:29 <lmacken> s/ways/was/
18:51:30 <stahnma> lmacken: fair enough.  It still says I don't meet it.
18:51:38 <nirik> stahnma: for what update?
18:51:42 <stahnma> and it's been 15 days, from what I can see
18:51:44 <lmacken> stahnma: I'll try and clarify that message
18:51:48 <stahnma> https://admin.fedoraproject.org/updates/rubygem-json-1.4.3-3.el5?_csrf_token=23055744d49c160c18dcef1b25f08cda668c4c10
18:51:50 <stahnma> rubygem-json
18:51:53 <stahnma> for el5
18:51:54 <nirik> it's 2 weeks in testing unless it's security or gets enough karma
18:51:59 <stahnma> right
18:52:25 <nirik> it's not quite been 2 weeks...
18:52:31 <stahnma> I'm saying anybody is doing anything wrong.  I just think it could be clearer.
18:52:37 <lmacken> In [3]: PackageUpdate.byTitle('rubygem-json-1.4.3-3.el5').days_in_testing
18:52:40 <lmacken> Out[3]: 14
18:52:43 <stahnma> 29th was a sunday
18:52:46 <lmacken> bodhi should approve within the next 6 hours
18:52:47 <nirik> humm... or it has now.
18:52:49 <stahnma> today is a monday
18:52:56 <stahnma> so it's 15 days
18:53:10 <nirik> stahnma: it counts from when it was actually in testing...
18:53:13 <nirik> not when it was submitted.
18:53:17 <stahnma> I guess (> 14) != (>= 14)
18:53:20 <stahnma> ah
18:53:26 * stahnma learns
18:53:36 <lmacken> bodhi has a job that runs every 6 hours ( iirc ) that will approve updates
18:53:50 <nirik> it might be nice to fix http://fedoraproject.org/wiki/EPEL/GuidelinesAndPolicies to be clear about our policy...
18:53:59 <nirik> or make a new page with it spelled out better.
18:53:59 <stahnma> seems like 2 weeks is a long time to wait fix a broken dep in stable...
18:54:22 <nirik> if we only had 1 more tester it could have gone when they added the last karma.
18:54:28 <lmacken> stahnma: yeah, well fedora updates are vulnerable to the same type of problem.  We'll have to fix it there first.
18:54:35 <stahnma> lmacken: ok.
18:55:19 * nirik notes fedora-easy-karma works on rhel5/centos5...
18:55:30 <stahnma> Also, just as an FYI to anybody working in EPEL.  If you see a messsage "debug-info.sh find invalid predicate" or something, you need to add a BuildRoot back into the spec
18:55:37 <stahnma> That bit me for hours this weekend.
18:55:45 <nirik> ah, nasty. ;(
18:55:46 * tremble too
18:56:00 <nirik> anyone interested/able to fix wiki pages for updates policy? ;)
18:56:18 <stahnma> #action stahnma will send an email to list about crazy BuildRoot error messages
18:56:29 <stahnma> nirik: I would, but I doubt I will get to it this week.
18:56:53 <nirik> yeah, I don't think I have time for it...
18:57:14 <nirik> I'd suggest a special session sometime to fix all our wiki pages, but not sure when that would be...
18:57:35 <stahnma> we really need a few more individuals who care about epel
18:57:50 <stahnma> and are willing to put in some amount of time on it
18:58:34 <nirik> yes.
18:58:41 <nirik> how can we find them though? ;(
18:58:52 <stahnma> teh twitters ? ;)
18:59:30 <tremble> Trouble is that on the packaging side there's no easy way to get EPEL only maintaners in.
18:59:51 <tremble> Those people who
19:00:05 <tremble> 're asking for pkgs to be ported could possibly be tapped...
19:00:06 <stahnma> I agree
19:00:22 <nirik> yeah.
19:00:37 <stahnma> time to close
19:00:42 * tremble nods
19:00:42 <stahnma> until next week :)
19:00:43 <nirik> yep.
19:00:50 <nirik> sounds good.
19:00:58 <nirik> #info will determine next weeks meeting time on list.
19:01:02 <nirik> #endmeeting
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/epel-devel-list/attachments/20100913/a0354f0b/attachment.sig>


More information about the epel-devel-list mailing list