17:03:50 <mitr> #startmeeting FESCO (2014-08-12)
17:03:50 <zodbot> Meeting started Wed Aug 13 17:03:50 2014 UTC.  The chair is mitr. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:03:50 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:03:52 <mitr> #meetingname fesco
17:03:52 <zodbot> The meeting name has been set to 'fesco'
17:03:54 <mitr> #chair dgilmore jwb kalev mattdm mitr mmaslano nirik sgallagh t8m thozza
17:03:54 <zodbot> Current chairs: dgilmore jwb kalev mattdm mitr mmaslano nirik sgallagh t8m thozza
17:03:54 <thozza> hey
17:03:56 <mitr> #topic init process
17:04:07 <nirik> morning
17:04:08 * jreznik is here for fesco needs but only for limited time today...
17:04:13 <jwb> hi, i'm here
17:04:26 <kalev> hello hello
17:06:26 <mitr> That’s 6, sgallagh and mattdm are travelling, so let’s start
17:06:33 <mitr> #topic #1322 F21 Changes - Progress on Changes Freeze
17:06:35 <mitr> .fesco 1322
17:06:36 <zodbot> mitr: #1322 (F21 Changes - Progress on Changes Freeze) – FESCo - https://fedorahosted.org/fesco/ticket/1322
17:06:47 <mitr> Does anyone have updates or concerns?
17:06:56 * number80 is here for a limited time if you need cloud WG input too
17:07:05 <jreznik> I don't have much updates, still catching up after Flock...
17:08:14 <mitr> Moving on…
17:08:16 <mitr> #topic #1178 Fedora 21 scheduling strategy
17:08:18 <mitr> .fesco 1178
17:08:20 <zodbot> mitr: #1178 (Fedora 21 scheduling strategy) – FESCo - https://fedorahosted.org/fesco/ticket/1178
17:08:27 <nirik> so, another week? ;(
17:09:06 <nirik> I think there was talk about doing that rebuild tomorrow. If so, could we still hopefully get a TC before next tuesday?
17:09:18 <nirik> (at the very least we should have a TC before entering freeze)
17:10:11 <kalev> realistically, we still don't have TC-s done yet and we didn't freeze yesterday so we're looking at a slip anyway
17:10:35 <jreznik> kalev: yeah, slip is, now the question is how long...
17:10:35 <nirik> I thought currently next tuesday was the scheduled enter freeze day?
17:10:36 <mitr> Does anyone own the mass rebuild?  If we are doing one tomorrow, devel@ should be told
17:10:56 <jreznik> mitr: dgilmore is going to do it
17:11:00 <mitr> nirik: http://fedoraproject.org/wiki/Releases/21/Schedule says yesterday
17:11:17 <jreznik> also sec arches guys would like to coordinate it with them - but it's not priority
17:11:34 <jreznik> so the question is not - do we slip but how long
17:11:58 <jreznik> if one week is enough to do mass rebuild and finish TC - we need info from dgilmore for it
17:11:59 <nirik> hum. I guess I was mentally off a week. ;(
17:12:42 <jreznik> nirik: flock did the same to me :)
17:12:44 <nirik> the rebuild should be pretty fast, even though it's 21 and 22...
17:12:51 <nirik> but it's also only archfull packages.
17:12:52 <thozza> I thought too there is a week after flock
17:13:03 <thozza> but the wiki is clear
17:13:07 <mitr> nirik: The f22 one could wait if the binding constraint were builder capacity
17:13:22 <nirik> theres no capacity problems, we have lots of builders.
17:13:23 <kalev> I don't think a mass rebuild is going to help with other stuff though, releng seems already stretched pretty thin
17:13:30 <nirik> and we don't want to break upgrade path/git
17:13:47 <jreznik> nirik: it should be fast, so my main concern is TC... also I heard jakub is hunting one gcc 4.9 bug with broken code generation but I'm not sure what's the number is - sharkcz, any idea?
17:14:04 <mitr> So, can FESCo do anything to help?  Or is it the case of “don’t even ask for status updates, you would be slowing the doers down?”
17:14:20 <nirik> so, I'd say lets push to next tuesday at least, and try and get status from dgilmore if we need more?
17:14:37 <t8m> nirik, +1
17:14:38 <mitr> jreznik: Sounds scary.  Would we consider doing a mass rebuild after Alpha?
17:14:56 <nirik> it would be a big pain if it was all packages...
17:14:57 <jreznik> mitr: I'd say no, please, no
17:14:59 <mitr> nirik: Given the little information we have, +1
17:15:17 <nirik> yeah, need more data. ;(
17:15:17 <thozza> nirik: +1 from my side, I'm for one more week and then revisit where we are
17:15:19 <mitr> jreznik: That would mean waiting with the mass rebuild until Jakub is certain enough, then
17:15:44 <nirik> https://fedorahosted.org/rel-eng/ticket/5962 is the releng ticket BTW
17:15:57 <jreznik> mitr: I don't have more details :( sharkcz told me on Monday
17:16:21 <kalev> as much as I am against slipping, I don't think there's any way to avoid at least one week slip at this point :(
17:16:25 <nirik> there's some more info in the releng ticket as to exact bugs, etc.
17:16:37 <jreznik> mitr: based on ticket, it's probably fixed
17:16:49 <jreznik> nirik: thanks for link
17:17:03 <mitr> (added to FESCo ticket for tracking)
17:17:07 <nirik> thanks
17:17:11 <kalev> regarding the rebuild, I tend to trust the toolchain guys -- if Jakub says to do a mass rebuild, we should do one (I assume he did, since everyone is talking about that?)
17:17:26 <nirik> yeah.
17:17:43 <jreznik> so it looks good
17:17:48 <mitr> More votes?
17:18:26 <mitr> (at +4 from mitr, nirik, t8jm, thozza)
17:18:53 <kalev> what was the vote, to slip one week until the 19th?
17:19:06 <mitr> nirik:  so, I'd say lets push to next tuesday at least, and try and get status from dgilmore if we need more?
17:19:16 <mitr> kalev: ^^
17:19:20 <kalev> +1 from me too then
17:19:25 <mitr> Thanks
17:19:41 <mitr> #agreed Slip freeze to next Tuesday at least (+5)
17:19:53 <mitr> #topic #1326 change to fesco replacement process?
17:19:56 <mitr> .fesco 1326
17:19:58 <zodbot> mitr: #1326 (change to fesco replacement process?) – FESCo - https://fedorahosted.org/fesco/ticket/1326
17:20:00 <nirik> so I assume that means all other milestones move as well, right?
17:20:11 <mitr> #undo
17:20:11 <zodbot> Removing item from minutes: <MeetBot.items.Topic object at 0x6d584d0>
17:20:28 <mitr> I guess so.  jreznik?
17:20:37 <jreznik> mitr: thanks, I'll do announcement... just I'm going to be on PTO until next Wed so please, if FESCo could watch it :)
17:20:43 <jreznik> nirik: yes
17:21:07 <nirik> ok, just confirming. thanks.
17:21:55 <mitr> #info jreznik will not be available until next Wednesday; FESCo members, please follow status of the mass rebuild.
17:22:03 <mitr> #topic #1326 change to fesco replacement process?
17:22:05 <mitr> .fesco 1326
17:22:06 <zodbot> mitr: #1326 (change to fesco replacement process?) – FESCo - https://fedorahosted.org/fesco/ticket/1326
17:22:11 <mitr> So we ended up not discussing this in any formal Flock session.
17:22:26 <nirik> yeah. ;(
17:23:37 <nirik> personally, I would like to avoid: too many more elections and also complex process.
17:23:44 <t8m> nirik, +1
17:23:50 <thozza> I think it is not clear from the ticket what proposal we want ot make
17:24:12 <nirik> yeah, we don't have a clear proposal I don't think
17:24:21 <mitr> I would like to hope that mid-term replacements will be infrequent in the future; we just ended up having many of them _this_ term (due to its length and I guess other factors)
17:24:31 <number80> may I give a suggestion as a board member ?
17:24:37 <mitr> Sure
17:25:19 <nirik> I guess abadger1999 did make concrete proposals in comment 17
17:25:23 <number80> if there are only few seats to be replaced, I think that appointing people is fine, but if too many people leaves, you may want an election
17:25:46 <jreznik> yeah, I really think it's not going to happen more than once for a while
17:26:00 <jreznik> it's some work, I'm not saying it is not
17:27:27 <mitr> Proposal: defer for a complete proposal to be added to the ticket and discussed there, drop meeting keyword until then (i.e. let’s not keep thinking about this only during the meetings)
17:27:35 <jwb> yes
17:27:50 <t8m> mitr, that's a good suggestion
17:28:00 <thozza> mitr: +1
17:28:02 <dgilmore> hey all sorry was distracted
17:28:11 <mitr> comment:17 is an option (at least it gives FESCo the _freedom_ to run elections :) ) but I don’t feel quite comfortable approving it between the 6=>7 of us today)
17:28:12 <nirik> mitr: +1
17:28:26 <abadger1999> <nod>
17:28:44 <abadger1999> And please, consider 17 a strawman if someone would like to continue modifying it from there.
17:28:55 <mitr> #agreed defer for a complete proposal to be added to the ticket and discussed there, drop meeting keyword until then (i.e. let’s not keep thinking about this only during the meetings) (+5)
17:28:58 <abadger1999> as I won't have time to continue working on it personally.
17:29:11 * nirik nods
17:29:45 <mitr> #topic #1328 glfw maintainer continuously breaking ABI
17:29:48 <mitr> .fesco 1328
17:29:49 <zodbot> mitr: #1328 (glfw maintainer continuously breaking ABI) – FESCo - https://fedorahosted.org/fesco/ticket/1328
17:30:48 <nirik> so, no response from them in bug.
17:31:08 <nirik> well, in the last 5 months
17:31:16 <number80> last news, I had from the maintainer, he was in a business trip
17:31:35 <thozza> I'm for having stable versions in fedora
17:32:01 <mitr> Also, this ticket was filed _in response_ to an unresponsive maintainer call on devel@
17:32:05 <thozza> but it would be nice to hear the maintainer's  reasons
17:33:14 <mitr> ... otoh that call didn’t explicitly follow the nonresponsive maintainer policy it seems (reusing another bug)
17:33:33 <mitr> number80: Do you know what would be the appropriate time to wait, then?
17:34:12 <number80> mitr: we're looking for his phone number
17:34:16 <jwb> why can't we have a proven packager just do a stable release in the stable fedora branches and the maintainer can justify later when they return why that isn't possible
17:34:28 <number80> I think that it's safe to fix the package
17:34:44 <nirik> the reporter of the fesco ticket has offered to do that if given privs.
17:35:19 <number80> but this is my personal opinion, I disagree to consider the maintainer as inactive
17:35:44 <thozza> can it be worse with the stable version?
17:35:52 <mitr> jwb: FWIW the latest ABI breakages happened 10 months ago or even earlier, the package is really more unmaintained than breaking at this moment
17:36:04 <nirik> proposal: grant acls to mpreisle for the package so they can push stable releases to fedora stable branches and work with the maintainer to meet guidelines moving forward.
17:36:14 <jwb> nirik, +1
17:36:30 <t8m> nirik, +1
17:36:34 * number80 agree with nirik proposal
17:36:45 <jwb> mitr, to be dealt with separately.  fix the problem we can fix now, fix the maintainer issues as we can.
17:36:47 <thozza> nirik: +1
17:36:59 <kalev> +1
17:37:08 <nirik> yeah.
17:37:34 <mitr> nirik: Can we add an explicit “Pushing unstable upstream snapshots to released versions of Fedora as used to happen in glfw is contrary to the updates policy"?  Your proposal implies it but doesn’t actually say it.
17:37:46 <mitr> nirik: +1 anyway
17:37:49 <number80> We'll try to contact him so that he could answer, but that shouldn't prevent taking action for glfw
17:38:21 <nirik> mitr: fine with me. I think it should be a given that we want people  to follow the updates policy tho...
17:38:26 <t8m> mitr, I would be +1 to that as well
17:38:36 <mitr> #agreed grant acls to mpreisle for the package so they can push stable releases to fedora stable branches and work with the maintainer to meet guidelines moving forward.(+6)
17:39:00 <mitr> nirik: That’s a fair point.
17:39:17 <mitr> #topic #1329 systemd-sysusers in F21?
17:39:19 <mitr> .fesco 1329
17:39:20 <zodbot> mitr: #1329 (systemd-sysusers in F21?) – FESCo - https://fedorahosted.org/fesco/ticket/1329
17:40:47 <nirik> not much new info here.
17:41:38 <mitr> Going by the book, the change to user account packaging needs to go through FPC (as is already planned for F22 in https://fedoraproject.org/wiki/Changes/SystemdSysusers ), and in the meantime, use of this mechanism is IMHO technically violating the existing guideline.
17:41:56 <jwb> nothing is using it in the distribution
17:42:09 <jwb> so i don't see how it's violating anything
17:42:12 <mitr> jwb: the entire issue is systemd using it for itself
17:42:35 <mitr> Hence my real question: are there any negative effects from systemd doing this?
17:42:42 <thozza> sounds like journal-remote-gateway is using it
17:43:23 <mitr> This was motivated by some real breakage from that mechanism, which has AFAIK been fixed in the meantime, so it’s either “there might be more bugs” or “not using the official scripts breaks $this” (or perhaps “we insist on rules being followed Just Because”)
17:44:19 <mitr> I can certainly imagine bordering-on-outlandish systems that would be broken by this (e.g. reimplementing fedora-usermgmt by replacing the shadow-utils package) but that’s probably out of scope of setups we care about.
17:44:27 * nirik wasn't clear on what the orig breakage was, good to hear it's fixed.
17:46:30 <mitr> I’m tempted to just redirect this to FPC since they own the current guideline and know the full motivation and corner cases (and they will have to review the F22 packaging _anyway_)
17:47:13 <jwb> fine with me
17:48:02 <nirik> we could.
17:48:05 <mitr> I guess that would be an implicit decision that the current state of F21 is not a blocking problem
17:48:15 <thozza> mitr: +1 for redirecting to FPC
17:48:59 <mitr> Proposal: File a FPC ticker placeholder for the sysusers feature, tell them that F21 already uses this, and do nothing else with F21 unless FPC comes back to us.
17:49:13 <thozza> mitr: +1
17:49:41 <nirik> well, f21 'uses' it only for systemd right?
17:49:57 <t8m> mitr, +1
17:50:03 <kalev> I wouldn't want to micromanage systemd here, they know what they are doing
17:50:03 <mitr> nirik: s/F21/systemd in F21/ , yes
17:50:22 <kalev> we could say that we are looking into applying this systemwide for F22 and that needs FPC to help get guidelines done
17:50:27 <nirik> sure, +1. If they find some serious issues with that, we can re-evaluate.
17:50:36 <mitr> kalev: As a rule we don’t outsource the packaging guidelines to systemd, and systemd has rather a history of not dealing with packaging guidelines soon enough I'm afraid.
17:52:03 <mitr> kalev: … and to complete the picture, FPC goes much smoother if there is someone actively writing drafts instead of just telling FPC “here solve this problem”
17:52:28 <kalev> fair enough
17:53:45 <t8m> mitr, +!
17:53:47 <t8m> mitr, +1
17:53:47 <mitr> kalev: (and this does actually let systemd do what they want in F21 unless someone comes back with a specific objection)
17:54:03 <mitr> At +4; more votes?
17:54:18 <kalev> +1, the proposal makes sense to me
17:54:35 <mitr> File a FPC ticker placeholder for the sysusers feature, tell them that systemd in F21 already uses this, and do nothing else with F21 unless FPC comes back to us. (+5)
17:54:42 <mitr> #topic Next week's chair
17:54:45 <thozza> mitr I think we have just 4
17:54:46 <mitr> Anyone?
17:54:47 <thozza> no?
17:54:55 <thozza> t8m: voted twice
17:55:09 <jwb> +1
17:55:13 <thozza> ok :)
17:55:27 <t8m> thozza, the second thing was just plussing what mitr said in regards to systemd
17:55:33 <t8m> not a vote
17:55:36 <t8m> :)
17:55:46 <mitr> thozza: I counted myself without voting formally, sorry
17:56:00 <thozza> mitr: OK, makes sense
17:56:57 <thozza> what's next?
17:57:08 <mitr> Anyone wants to chair the next meeting?
17:57:09 <thozza> open floor?
17:57:28 <thozza> mitr: I can if you tell me what I have to do (not sure precisely)
17:57:49 <mitr> thozza: Great, thanks.
17:58:09 <mitr> Please follow https://fedoraproject.org/wiki/FESCo_meeting_process (and if anything in there is wrong or unclear, please do ask and update the page)
17:58:19 <mitr> #info thozza to chair next week’s meeting
17:58:24 <thozza> mitr: OK thanks
17:58:26 <nirik> cool. ;)
17:58:36 <mitr> #topic Open Floor
17:58:42 <mitr> Anything for open floor?
17:58:48 <kalev> dgilmore: are you around now by any chance? did you see the scheduling discussion above, is this realistic?
17:58:49 <nirik> dgilmore: what was the current status on tc?
17:58:51 <nirik> :)
17:59:12 <dgilmore> I am here
17:59:18 <thozza> I have to leave, so if there's nothing really important, you'll have to do open floor without me
17:59:24 <dgilmore> still sorting small issues out
17:59:28 * thozza will read the log later
17:59:45 <dgilmore> ive been somewhat wreaked this week
18:00:00 <nirik> dgilmore: we agreed to move freeze to the 19th. With the mass rebuild tomorrow, what are the chances we could have a TC before the 19th? (next tue)
18:00:06 <dgilmore> and pulled in a million directions by different people wanting my attention
18:00:14 <nirik> yeah. ;(
18:00:16 <dgilmore> nirik: we wills ee
18:00:20 <dgilmore> we will see
18:00:28 <thozza> ok, bye guys.... sorry....
18:00:29 <dgilmore> I need to rewrite the massrebuild script today
18:00:45 <dgilmore> someone is free to help with that
18:01:06 <dgilmore> I can tell you what we need to have it do
18:01:26 <mitr> #info Help wanted rewriting massrebuild script totday, ask dgilmore for details.
18:04:51 <mitr> Anything else for open floor?
18:05:11 <mitr> If not, I’ll close the meeting in 2 minutes
18:05:21 * dgilmore has nothing
18:07:00 <mitr> Thanks everyone!
18:07:03 <mitr> #endmeeting