20:01:37 <smooge> #startmeeting EPEL
20:01:37 <zodbot> Meeting started Fri Oct 31 20:01:37 2014 UTC.  The chair is smooge. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:01:37 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:01:47 <smooge> #meetingname epel
20:01:47 <zodbot> The meeting name has been set to 'epel'
20:02:05 <smooge> #chairs nirik bstinson dgilmore Evolution
20:02:15 <smooge> #chair nirik bstinson dgilmore Evolution
20:02:15 <zodbot> Current chairs: Evolution bstinson dgilmore nirik smooge
20:02:30 <smooge> #topic Meats and Greats
20:02:39 <bstinson> hi all
20:02:40 <smooge> Hello everyone to the 2014-10-31 meeting.
20:02:57 <nirik> happy halloween
20:03:24 <smooge> happy halloween to all and to all a spooky night
20:04:07 <smooge> so ok everyone is here but dgilmore
20:04:30 <smooge> #topic Old Business
20:05:14 <smooge> Ok so we have a lot of outstanding tickets.. several of them waiting on feedback
20:05:43 <Evolution> I think several of us spent a chunk of last friday's meeting time putting feedback on them.
20:06:15 <Evolution> for example the epel-latest meta-package
20:07:00 <nirik> meta-package? I thought that was just going to be a link?
20:07:00 <Evolution> smooge: should we prioritize them or what needs to happen to move forward?
20:07:28 <smooge> I thought it was meant to be a link also.
20:07:44 <Evolution> right, poor wording on my part
20:08:05 <Evolution> https://fedorahosted.org/epel/ticket/8 is what I mean
20:08:06 <nirik> just needs round tuits. ;)
20:08:09 <smooge> so I think in this case it is mainly me writing a cron job to do this per day.
20:08:15 <smooge> I think I can get that out by Tuesday
20:09:12 <Evolution> nirik: fwiw (and to throw him firmly under the bus) bstinson was volunteering in #centos-devel to apprentice if it would help free up time to get stuff done.
20:09:12 <smooge> though ktdreyer's idea may be a good place to look also..
20:10:13 <nirik> more help the merriier. ;)
20:10:18 <bstinson> having mash handle those looks like an interesting idea
20:11:17 <smooge> nirik, do we still use mash? I have mostly left that part fo the releng as dark arts practiced by dgilmore
20:11:26 <nirik> yep.
20:11:35 <nirik> bodhi calls mash to create the repos.
20:12:58 <smooge> so putting the code into mash would fix a lot of people's issues then.
20:13:17 <smooge> and make it so its not manual or a cron job
20:13:34 <nirik> yeah, but it might have to be a nice patch for dgilmore to take it. ;)
20:13:55 <smooge> .whoowns mash
20:13:58 <zodbot> smooge: ausil
20:14:29 <smooge> ok so I make a vegemite sandwich
20:14:47 <bstinson> should we start with a cronjob while we're exploring mash options? or can this wait
20:15:06 <nirik> well, it's waited this long, I don't know that it's super urgent
20:17:36 <smooge> ok I see something went wrong yesterday and the ticket I had for the agenda of this meeting didn't get finished.
20:18:23 <smooge> bstinson, I would be happy to work with you on both sides of the thing.
20:19:04 <smooge> I think the cron job might be actually harder than having mash do it after I looked at it for a while as you need to figure out what is the latest epel-release in a tree to link correctly to it :)
20:19:55 <smooge> but I may have been overthinking it.. anyway.. after meeting item as its technical versus policy
20:20:03 <smooge> #topic ticket 2
20:20:53 <nirik> I can look at a proposed patch here..
20:21:03 <nirik> it sounds like we are all ok with trying it in epel-release at least for now?
20:21:08 <smooge> so I am for putting it in epel-release
20:21:56 <nirik> ok, I can try and whip up something this weekend.
20:22:08 <smooge> #topic ticket 4
20:22:38 <smooge> this one is about unretirement criteria.
20:23:10 <smooge> First question I have is: How hard would it be for people with proven-packager to dedicate a day to help review any outstanding packages we need to unretire?
20:23:11 <nirik> yeah, so we could either do: re-review required, or do: if you say you are going to maintain it you can unretire it.
20:23:26 <smooge> The main issue has been no one is doing the reviews
20:23:29 <nirik> well, it would only need people in packager group to review.
20:23:54 <nirik> I did review a few today... well, 1 and added a bunch of comments on another.
20:24:00 <Evolution> there are a few folks on the centos side with that status who mentioned they'd be willing to help.
20:24:10 <Evolution> I think mostly it's just generating awareness
20:24:13 <smooge> I think the main issue is that most of us who have 'packager' would need someone to help us go through one so we can say we know what the f we are doing in a review.
20:24:19 <nirik> there's 11 open epel tickets for reviews: http://fedoraproject.org/PackageReviewStatus/EPEL.html
20:24:43 <nirik> a lot of those look bogus too.
20:25:11 <Evolution> yeah, I'm not sure how accurate that is.
20:25:29 * nirik goes to close some of them.
20:26:44 <Evolution> if there's an accurate list, I could probably get a few of the proven-packager folks in the centos community to have a monthly bug-day or something similar.
20:26:48 <bstinson> i wonder if we could do a review "workshop" of sorts, to help reviewers who are feeling a little shaky
20:26:58 <smooge> I would like that.
20:27:01 <nirik> sure.
20:27:08 <nirik> fedora-review also helps do a lot of the grunt work
20:27:09 <bstinson> +1 for bug day too
20:27:40 <SommitRealWeird> How is that list generated?
20:27:40 <smooge> I haven't reviewed a package in a couple of years and when I did I got feedback that I had done it wrong :)
20:27:56 <SommitRealWeird> Because I'm sure there's a Pound 2.7d that should be in the review list.
20:28:00 <nirik> SommitRealWeird: there's a script that makes it... all package reviews with epel versions
20:28:17 <buggbot> kevin@scrye.com changed the Status on bug 1148327 from NEW --- to CLOSED ERRATA.
20:28:18 <buggbot> Bug https://bugzilla.redhat.com/show_bug.cgi?id=1148327 Package Review, unspecified, unspecified, ---, nobody, CLOSED ERRATA, EPEL-7 branch: icewm
20:28:44 <nirik> oh yeah, sorry for spamming the channel.
20:28:48 <buggbot> kevin@scrye.com changed the Status on bug 1133548 from NEW --- to CLOSED ERRATA.
20:28:48 <buggbot> Bug https://bugzilla.redhat.com/show_bug.cgi?id=1133548 Package Review, unspecified, unspecified, ---, nobody, CLOSED ERRATA, libmodplug unretirement
20:29:05 <SommitRealWeird> https://admin.fedoraproject.org/updates/Pound-2.7-0.4.d.el7.1 <-- it's been pushed to testing, apparently - does that not need a review.
20:30:01 <SommitRealWeird> (or have I missed something in the process, which is obviously possible)
20:30:38 <nirik> SommitRealWeird: review is for entering the collection at all. If it's in testing it's past review
20:30:39 <buggbot> kevin@scrye.com changed the Status on bug 1128077 from NEW --- to CLOSED ERRATA.
20:30:40 <buggbot> Bug https://bugzilla.redhat.com/show_bug.cgi?id=1128077 Package Review, unspecified, unspecified, ---, nobody, CLOSED ERRATA, EPEL-7 branch: dbench
20:30:52 <SommitRealWeird> ah - cool.
20:31:10 <buggbot> kevin@scrye.com changed the Status on bug 1128076 from NEW --- to CLOSED ERRATA.
20:31:10 <buggbot> Bug https://bugzilla.redhat.com/show_bug.cgi?id=1128076 Package Review, unspecified, unspecified, ---, nobody, CLOSED ERRATA, EPEL-7 branch: fs_mark
20:31:46 <nirik> there may be some mixed in with the main review queue if they didn't set an epel version on them. ;(
20:32:11 <buggbot> kevin@scrye.com changed the Status on bug 1122749 from NEW --- to CLOSED NEXTRELEASE.
20:32:12 <buggbot> Bug https://bugzilla.redhat.com/show_bug.cgi?id=1122749 Package Review, unspecified, unspecified, ---, nobody, CLOSED NEXTRELEASE, EPEL7 branch erlang-oauth erlang-mochiweb erlang-snappy
20:32:32 <buggbot> kevin@scrye.com changed the Status on bug 1122747 from NEW --- to CLOSED ERRATA.
20:32:32 <buggbot> Bug https://bugzilla.redhat.com/show_bug.cgi?id=1122747 Package Review, unspecified, unspecified, ---, nobody, CLOSED ERRATA, EPEL7 branch autoconf-archive
20:32:36 <Evolution> so what do we need to do/decide for this one?
20:32:36 <smooge> ok so I think if someone walks me through how to properly review things I can review all the things which we needed
20:33:06 <Evolution> smooge: try to organize a bug-killin' day and review/refresher type stuff?
20:33:18 <smooge> a) I volunteer to be a reviewer who will do reviews of needed packages on Friday afternoons after I have been retrained
20:33:21 <nirik> yeah, so the answer to this ticket then would be: "yes, we want a re-review" ?
20:33:33 <smooge> b) We have a bug-killing day once a month where CentOS packagers help
20:33:50 <nirik> sounds lovely to me
20:33:51 <Evolution> nirik: I think so. if it's dead long enough it may be out of spec for current stuff.
20:34:04 <bstinson> i can take a review every now and then too
20:34:04 <nirik> yeah, and if it's bad we want to know before it enters the collection.
20:34:45 <smooge> c) We expect all packages that have been removed from the collection over 90 days to be re-reviewed.
20:35:02 <Evolution> +1 here
20:35:11 <smooge> or is there a different time frame already agreed to in Fesco?
20:35:21 <nirik> I think it's 2 weeks for fedora ones.
20:35:29 <smooge> c) We expect all packages that have been removed from the collection over 14 days to be re-reviewed.
20:35:35 <nirik> +1
20:35:40 <Evolution> (amended) +1
20:35:44 <smooge> +1
20:35:45 <bstinson> +1
20:36:23 <smooge> #topic ticket 7
20:36:36 <smooge> I am going to put this in the ticket so someone talk about 7 for a few secs
20:36:57 <nirik> 7 was a pretty good movie. Kinda creepy tho.
20:38:15 <Evolution> the scripts mentioned in the ticket. are those running?
20:38:46 <smooge> https://fedorahosted.org/epel/ticket/7
20:39:25 <smooge> Evolution, there is a set of scripts which email every week about how long packages have been orphaned
20:40:26 <Evolution> well, the list for el7 seems fairly minor, but that's probably expected.
20:40:30 <smooge> from: opensource@till.name SubjecT: [EPEL-devel] Orphaned packages in epel6
20:40:31 <Evolution> the list for 5 is.... substantial.
20:40:36 <smooge> yeah..
20:42:12 <smooge> I think 5 has reached the "We have a lot of people who want stuff in it.. but no one and I mean no one wants to help fix it."
20:42:34 <smooge> Which may mean we should look at retiring 5
20:43:09 <nirik> well, I think it will slowly die into the sunset.
20:43:12 <Evolution> to some extent I'm fine with going full BOFH on it. keeping it as-is is degrading security at least.
20:43:27 <smooge> nirik, it is still 1/2 of our EPEL usage
20:43:34 <nirik> fun
20:43:53 <nirik> well, I guess we may get some screams of anguish when/if we retire things.
20:43:56 <Evolution> just because a package is orphaned doesn't mean it's out of date. if it's orphaned AND has outstanding CVEs, I'd vote kill it.
20:44:26 <smooge> Evolution, well here is the issue. if it is orphaned it means no one is checking to see if it has CVEs except us.
20:44:37 <nirik> well, orphaned is kinda bad too... even if no security issues there could be other issues... and no one is listening. ;(
20:44:59 <Evolution> quite a few of those packages have multiple names attached to them. do those people get notifications?
20:45:06 <nirik> yep.
20:45:13 <nirik> emails go to the people listed too.
20:45:26 * nirik removed himself from several he was on that he no longer cares about at all.
20:46:00 <smooge> Evolution, and htose people could have removed the orphan status by taking the package if they wanted to.
20:46:17 <smooge> which basically says 'we arent really maintaining it either.'
20:46:21 <bstinson> i've tried to poke a couple people on packages i depend on, but i haven't heard back from anyone
20:47:34 <Evolution> so, just as a CYA, lets escalate. have this group triage/email those people with a notice to take the package else it will be orphaned. then 30d of no action, kill it.
20:47:47 <Evolution> that way NO ONE can say they didn't know, or weren't told.
20:47:55 <smooge> okie dokie
20:48:01 <smooge> I am ok with that.
20:48:13 <bstinson> +1
20:48:36 <nirik> +1
20:48:40 <smooge> +1
20:49:33 <smooge> actually i would like this a bit more specific. 'those people' is defined by co-maintainers of packages listed as orphaned?
20:49:46 <Evolution> smooge: yes.
20:49:57 <smooge> in the case of packages without comaintainers do we remove now or wait for the 30 day period also?
20:50:22 <bstinson> is there a way to automatically distinguish as of now?
20:50:33 <smooge> I don't think so.
20:50:54 <Evolution> special announcement to the list instead of the automated nag emails that could be in danger of being ignored?
20:50:57 <nirik> the emails list them...
20:51:06 <Evolution> not that those wouldn't be ignored also, but still.
20:51:34 <nirik> yeah, at some point there is only so much we can do. ;)
20:51:49 <nirik> and removing them might get some folks to realize they need them and show up and unretire them. ;)
20:51:50 <Evolution> right
20:51:53 <smooge> OK next question.. what do we need to do with stuff that because we removed orphaned stuff breaks the repos?
20:52:31 <nirik> the emails also list the things broken by removing those things, so we just keep going until we have 0 packages or 0 broken things. ;)
20:52:35 <smooge> as in packageA requires packageB which we remove because it was orphaned
20:52:50 <smooge> do we do it all in one swoop?
20:53:03 <nirik> I think so... or we could do packageB's in the next pass.
20:53:15 <nirik> but if it depends on packageA thats gone, it won't work very well.
20:53:23 <smooge> nah I would rather do it all in one official pass.
20:53:33 <nirik> yeah, me too.
20:53:48 <nirik> people who want packageB back can unretire a and b and maintain them both.
20:54:17 <bstinson> agreed, i think we pick a day to start the clock and say "any package mentioned in this email will be retired in...."
20:54:40 * nirik nods
20:55:41 <smooge> OK additional proposal: On December 10, 2014, we are going to have a clean out day. All orphaned packages and packages requiring those orphaned packages will be removed from EPEL repositories.
20:56:20 <smooge> That makes it a wednesday so should miss any Fedora release date :)
20:56:47 <smooge> and gives us over 30 days to set all the balls in motion
20:56:58 <nirik> smooge: we will be at the FAD in RDU...
20:57:05 <nirik> but otherwise fine.
20:57:19 <smooge> December 17th?
20:57:43 <Evolution> "merry christmas, we deleted your packages"
20:57:57 <smooge> well it is that or January
20:58:06 <Evolution> no, I'm totally fine with doing it.
20:58:12 <Evolution> I'm just going to joke about it too
20:58:25 <smooge> OK amended:
20:58:39 * nirik is fine whenever really.
20:58:47 <nirik> we are coming up on the hour tho
20:59:02 <bstinson> i'm +1 with either date
20:59:40 <smooge> On December 17, 2014, EpSCO will be cleaning out the EPEL repositories of all orphaned packages. Packages that are orphaned will be removed and all packages depending on those orphaned packages will be removed from the repository. Packagers requiring packages need to take over orphaned packages.
20:59:44 <smooge> +1
20:59:48 <Evolution> yeah. +1.
20:59:56 <Evolution> and we should make some noise about that.
21:00:01 <smooge> Oh I am going to...
21:00:15 <bstinson> maybe we could colocate a bug day?
21:00:24 <smooge> sounds good
21:00:32 <smooge> I will add to my task list that
21:01:17 <smooge> nirik, one last +1/-1 and then I will go to Open Flood and close out
21:01:22 <nirik> +1
21:01:46 <smooge> #agreed On December 17, 2014, EpSCO will be cleaning out the EPEL repositories of all orphaned packages. Packages that are orphaned will be removed and all packages depending on those orphaned packages will be removed from the repository. Packagers requiring packages need to take over orphaned packages.
21:01:58 <smooge> #topic Open Flood
21:02:19 <smooge> Meetings from now on will just have 2 topics in them.. I think that is about all we can get to :)
21:02:23 <Evolution> only thing I've got to ask here is about how updates get submitted for karma
21:02:52 <Evolution> things like the thunderbird update sat for about a month because it wasn't in bodhi properly and didn't get feedback
21:03:16 <Evolution> https://lists.fedoraproject.org/pipermail/epel-devel/2014-October/010372.html
21:03:17 <Evolution> that.
21:03:40 <smooge> It looks like a bug in bodhi that they are looking at.
21:04:03 <Evolution> okay cool. so long as it's being addressed.
21:04:03 <smooge> nirik, do you know anything on it?
21:04:15 <nirik> hum. not sure... were those pulled into rhel?
21:04:55 <Evolution> nirik: tb was yanked out between 6 and 7, I've not heard anything for putting it back.
21:05:20 <smooge> yeah.. only Evolution is the mail client for 7
21:05:49 <nirik> not sure then... so they didn't have updates? sounds like the maintainers built them and didn't submit? possibly because it was between beta and non beta?
21:06:01 <nirik> ie, they thought they would just go out, but we switched on bodhi
21:06:18 <Evolution> nirik: could be.
21:06:48 <Evolution> is there a way to see what else could be in that state?
21:06:51 <nirik> yeah, thats my theory anyhow.
21:07:03 <Evolution> or maybe send an email to maintainers reminding them to submit?
21:07:14 <smooge> OK I think we need to send a reminder to maintainers that bodhi is on and that they need to submit
21:07:35 <nirik> Evolution: possibly, but not easily.
21:07:47 <Evolution> then let's do it smooge's way and email.
21:07:58 <Evolution> that's cheaper/easier on required manhours.
21:08:40 * nirik nods
21:08:45 <smooge> okie dokie
21:08:51 <smooge> anything else for Flood?
21:09:10 <smooge> closing at 21:10
21:09:17 * nirik has nothing.
21:09:28 <Evolution> that's it for me
21:09:36 <bstinson> thanks everyone!
21:09:59 <smooge> ok dokie thanks for coming and such.
21:10:04 <nirik> thanks smooge
21:10:14 <smooge> btw I am going to ask what time works better for you guys on Friday?
21:11:21 <smooge> Actually.. next week is end of DST in US. Should I make the meeting 16:00 UTC still or 17:00 UTC [nirik, bstinson, Evolution ?]
21:11:32 <smooge> #endmeeting