16:00:03 #startmeeting Fedora QA meeting 16:00:03 Meeting started Mon Feb 15 16:00:03 2016 UTC. The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:03 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:00:03 The meeting name has been set to 'fedora_qa_meeting' 16:00:06 #meetingname fedora-qa 16:00:06 The meeting name has been set to 'fedora-qa' 16:00:09 #topic Roll call 16:00:22 * kparal is here 16:00:31 ahoyhoy folks, who's been pining for all the raw knuckle excitement of a QA meeting so long they've been camped at the keyboard all day? 16:00:47 I KNOW I HAVE 16:01:11 am I going to be alone in here with this lunatic? 16:01:26 it puts the lotion in the basket 16:01:33 kparal: yer on yer own ;) 16:01:48 * danofsatx sits back to watch 16:01:49 adamw: didn't your day just start? 16:01:55 tflink: minor details 16:02:30 * satellit_e listening 16:02:44 morning satellit_e 16:02:59 * pschindl is here 16:04:01 alllrighty 16:04:13 i have a feeling we have lots to talk about (though i can't remember what half of it is) 16:04:17 so let's get rolling 16:04:21 #chair kparal danofsatx 16:04:21 Current chairs: adamw danofsatx kparal 16:04:29 #topic Previous meeting follow-up 16:04:36 we had a few things at the last meeting... 16:05:35 #info "adamw to get blockerbugs non-media blocker support merged" - I talked to tflink about it in brno and updated the diff - https://phab.qadevel.cloud.fedoraproject.org/D689 . tflink didn't merge it yet 16:06:03 #action tflink to merge D689 already, damnit flink 16:06:14 :) 16:06:18 fair? fair. :P 16:06:37 jeez, this is a one-man show, all these action items are mine. how does that keep happening? 16:07:12 #info "adamw to look into the reported FESCo vote to stop blocking on i686 16:07:12 media and take necessary steps" - I sent out some mails about this and updated the criteria preamble. there's an active test@ thread about the best way to adjust the matrices 16:07:21 damnit, line wrap 16:07:22 #undo 16:07:22 Removing item from minutes: INFO by adamw at 16:07:12 : "adamw to look into the reported FESCo vote to stop blocking on i686 16:07:43 #info "adamw to look into the reported FESCo vote to stop blocking on i686 media and take necessary steps" - I sent out some mails about this and updated the criteria preamble. there's an active test@ thread about the best way to adjust the matrices 16:08:20 #action adamw to complete matrix changes for i686 no longer blocking media 16:08:35 sorry, what's the test thread called? 16:08:39 * adamw looking 16:08:49 +1 all action items for adamw 16:09:20 #info ml thread for 32-bit matrix discussion is "Fedora 24: i686 images no longer 'release blocking'" 16:09:30 you posted in it 16:09:30 :P 16:09:43 anyone have further thoughts on that here, or should we carry on with it on-list? 16:10:02 I see it now 16:10:02 * linuxmodder listens and attempts to learn form back row 16:10:04 that one, or the "Non-image blocker process change proposal" one? 16:10:12 it doesn't seem to be much active lately 16:10:22 do i686 bugs go to FE? 16:10:22 linuxmodder: that's right next to the lotion basket 16:10:34 adamw, not cool 16:10:36 :) 16:11:07 danofsatx: the "non-active image blocker process change proposal" thread is about how we handle things like upgrade bugs that should "block the release" in some sense but aren't related to the media 16:11:10 surely hope no lotion is needed -- 16:11:12 kparal is still leading that afaik 16:11:34 satellit: yeah, i think, more or less; think of i686 images now as being like SoaS or Xfce or LXDE images 16:11:40 k, it was the first one I found. 16:11:41 k 16:11:42 (when they're still shipped) 16:11:42 adamw, like a secondary arch or a dependent packages issue? 16:11:46 adamw: I'm still waiting for your response I think :) 16:11:51 * danofsatx doesn't have his glasses on, missed the F24 thread 16:12:04 kparal: damnit, lay off a guy who's on vacation :P 16:12:16 we can throw a topic at that in a minute 16:13:05 I'll try to revive that thread, but nobody is keen on dealing with that stuff 16:13:09 so adamw am I understanding that right its seen as a non critical if its i686 , soas , or other non main spin ? for media blocking purposes? 16:13:15 kparal: i think there's enough in the thread for me to pick an approach (for the i686-no-longer-blocking changes), i'll probably send another email for any further thoughts then go ahead and do something later this week if nothing comes up 16:13:39 linuxmodder: yes. there's actually a list of "blocking" and "non blocking" images: https://fedoraproject.org/wiki/Fedora_Program_Management/ReleaseBlocking/Fedora24 16:13:54 linuxmodder: bugs specific to any of the "non-blocking" images cannot block release (highest they can go is FE) 16:14:18 also a quick note on this topic: FESCo voted 6-0 that 32-bit upgrade issues are also non-blocking from F24 on 16:14:41 * kparal likes that 16:14:42 #info FESCo voted 6-0 that 32-bit upgrade issues are also non-blocking from F24 on: https://fedorahosted.org/fesco/ticket/1539#comment:14 16:15:03 some things simply need to die in a fire 16:15:06 i kinda think it's breaking a promise to people who installed f22/f23 i686, but ah well, i'm not the one who has to fix stuff :P 16:16:08 okay, where were we... 16:16:45 #info "adamw to mail list about co-ordinating a QA 'year in review' post for 2015" - I sent that mail out, haven't gotten any further in the process yet thanks to devconf and vacation. if anyone has content for a year in review please send it to the list! i'll probably set up a trac ticket or something 16:17:01 that's all the follow-up items on my list...anything else? 16:17:32 nope. 16:17:35 motion to close. 16:18:34 alrighty, onto new topics 16:18:36 er, let's see 16:18:46 #topic Non-media blocker process 16:19:11 * kparal tries to remember 16:19:15 yeah, me too 16:19:34 so, I posed two major questions 16:19:40 == Question #1: Do we slip always? == 16:19:44 == Question #2: For how long do we slip? == 16:19:49 right, and didn't get any direct replies 16:19:54 I received no relevant feedback 16:19:58 right 16:20:40 so, your answers were "yes" and "for a week unless we can fudge go/no-go by a day" 16:20:46 I can't blame people, I wouldn't read such boring emails either 16:20:48 given the position of releng i think that's reasonable 16:20:58 kparal: bet you never knew how non-fun it was to write them :P 16:21:18 I could have never imagine 16:21:38 so yes, the answers seem to be yes, and a week 16:21:42 so my answer is +1 to adjust the relevant policies to say updates must be stable for all non-media blockers before we sign off, and no change to the policy on what happens if we're not go 16:22:04 i suspect we might wind up with an interesting go/no-go meeting if this becomes a live issue, but i think that's at least where we should start 16:22:24 any other thoughts? 16:22:39 the mail kparal sent is http://lists.fedoraproject.org/archives/list/test@lists.fedoraproject.org/message/3NC74B5QTKFUMO5TMVLMOMGRDR3DY7II/ 16:22:52 (man, i wish hyperkitty did shorter links) 16:22:56 we should also define a process how releng are going to execute the new mm2-emergency-purge (or similar name) script, to make sure the latest updates are available to our users 16:23:09 is that our problem or releng's? 16:23:15 e.g. we can ask for that in the same ticket we use for composes, or in a different ticket 16:23:28 Is the plan to not post broken i686 or nonblocking images for release day or put up some sort of warning or not do anything special? 16:23:31 adamw: the problem is 'how releng will know about it?' 16:23:35 * adamw would rather things didn't work by us specifically asking for them and releng manually doing them 16:24:45 brunowolff: it's not QA's responsibility so i don't know for sure, but I believe the idea is to either not show 32-bit images on getfedora or all or to show them at a clearly lower level like other non-blocking images 16:24:47 By the way this isn't completely hypothetical as i686 is broken now, and getting it fixed, looks like it could take a while. 16:24:56 brunowolff: i'm aware :P 16:25:24 kparal: i guess this is kinda part of the whole problem of improving how we get from the blocker bug list to an RC compose 16:25:26 adamw: well, I guess they won't magically know that a day ago a critical update was pushed to F23 updates repo and the emergency script needs to be run (in case we're aiming for release announcement soon) 16:25:41 kparal: they really *should* :) 16:25:54 bz 1302071 16:26:09 adamw: do you have some process improvements ideas up your sleeve? 16:26:11 but OK, in a non-ideal world, fair point, i guess doing it in the TC/RC request ticket would be OK 16:26:16 kparal: i got all your ideas right here... 16:26:21 ideas are cheap! 16:27:58 I think the RC ticket is a good place, it's reasonable to assume we won't post a release announcement until that ticket was closed 16:29:07 or a different ticket that will be used for non-compose tasks 16:29:18 well, i guess if no-one has objections to kparal's proposal we should go ahead and implement it 16:29:22 it's been there for a month 16:30:19 * danofsatx has family activities to tend to. 16:30:22 #action kparal to implement changes to the process pages to specify that fixes for non-media blockers must be pushed stable for the appropriate release(s) for QA to vote Go 16:30:30 carry on, I've got ZNC finally working. 16:30:50 danofsatx: get them testing fedora! 16:31:05 somehow I know that editing those wiki pages will be the same amount of fun as writing up those emails 16:31:09 they are, without knowing it - they're my Cinnamon testers ;) 16:31:27 kparal: ooh, just acres of fun. remember to use lots of templates! templates are the best 16:31:37 everyone loves curly braces 16:31:47 {{{{{{{{{mediawiki}}}}}}}}}}}}} 16:32:06 adamw: you're not helping! 16:32:21 * kparal will have nightmares today 16:32:43 muahahaha 16:32:54 alright, so - what other topics do we have? i know there's...stuff... 16:33:20 danofsatx: excellent, carry on soldier 16:33:58 oh, i should mention: 16:34:03 #topic Wayland plans 16:34:34 so some of us talked to mclasen about Wayland last week. he told us the current status is that we're working on F24 with the assumption that Wayland-by-default will go ahead 16:35:00 iow: we will break your desktop once again, and this time even without lennart's help 16:35:08 but at some point we'll get together and make an evaluation about whether it's good enough, with the possibility to switch back if not 16:35:23 kparal: was it Alpha freeze when we were gonna decide? or Beta freeze? /me forgot 16:35:46 we said early march, before alpha 16:35:47 I'm not sure either 16:35:52 ok 16:36:00 mclasen___: hello 16:37:30 there are a few wiki pages useful when comparing differences or debugging wayland 16:37:37 https://fedoraproject.org/wiki/Wayland_features 16:37:45 https://fedoraproject.org/wiki/How_to_debug_Wayland_problems 16:37:47 hi mclasen 16:38:03 and it also contains links to wayland-related bugzilla trackers 16:38:04 hi 16:38:07 ok, so i think that was Alpha freeze 16:38:12 please use them if you find something important 16:38:32 (I follow them, at some wayland devs as well) 16:38:42 #info for now F24 work proceeds on the plan that Wayland-by-default will be important: we will get together with desktop SIG before Alpha freeze (2016-03-08) and decide whether to pull the plug 16:38:59 #info if Wayland-by-default goes ahead, we will organize a Test Day for it between Alpha and Beta 16:39:37 * satellit_e opt out on boot? 16:39:38 note that by-default does not mean "there's no X", or anything like that. it'll be as it is now, you can pick X instead. we know there are several cases where Wayland is not yet suitable 16:40:09 satellit: yeah, AIUI the 'opt-out' options will be pretty much as they are now (you can pick the X session in GDM instead, and if you need GDM to run on X there's gdm's config file) 16:40:19 if there are bigger plans mclasen can mention them 16:40:24 k 16:41:03 pretty much, yes. there's no plan to take the X session away as an easily available option 16:44:57 alrighy, thanks mclasen 16:45:06 anyone have any other topics? i feel like there's other things but my brain is mush 16:45:32 the pungi4 stuff is moving along, but still waiting on releng to pull the big switch 16:45:44 i know dgilmore was making progress with getting more of the images built last week 16:50:22 #topic Open floor 16:50:29 we've got a quick blocker review meeting at the top of the hour too 16:54:41 welp, sounds like everyone's done 16:54:48 so, meet back in #fedora-blocker-review in five minutes :) 16:54:51 thanks for coming everyone!@ 16:55:24 I'm working on trying to confirm that binutils 2.26 is what started breaking kernels. Live images from just before the change are still available and I am planning on using them to specifically test the binutils change on f24. (I can't do it with current rawhide, because binutils 2.25 no longer works in rawhide.) 16:55:53 brunowolff: yep, saw the mail - thanks for working on that 16:56:52 #endmeeting