16:00:38 <adamw> #startmeeting Fedora QA meeting
16:00:38 <zodbot> Meeting started Mon Dec 22 16:00:38 2014 UTC.  The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:38 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:43 <adamw> #meetingname fedora-qa
16:00:43 <zodbot> The meeting name has been set to 'fedora-qa'
16:00:45 <adamw> #topic Roll call
16:01:00 <adamw> ahoyhoy folks, anyone around for some QA? and have anything to talk about? :)
16:02:31 * tflink is here
16:03:11 <adamw> #chair tflink
16:03:11 <zodbot> Current chairs: adamw tflink
16:03:51 <adamw> smccann: ahoy there, are you the 'mccann2' whose name has been showing up on the nightly matrices?
16:04:19 * randomuser lurks
16:04:34 <smccann> yes that's me
16:04:45 <smccann> obviously have identity issues :-)
16:04:52 * satellit listening
16:05:46 <adamw> smccann: hehe :) thanks very much for helping out!
16:05:46 <adamw> #chair satellit
16:05:46 <zodbot> Current chairs: adamw satellit tflink
16:06:07 <smccann> adamw: np.  very excited to have found a bug :-)
16:06:16 * roshi_phone is here
16:06:34 <adamw> ahoy remote roshi
16:06:44 <adamw> try not to fall off anything
16:07:20 <roshi_phone> Doing my best :)
16:07:45 * satellit looking at new cinnamon live spin on bare metal
16:08:01 <adamw> satellit: wow, you sure try everything!
16:08:15 <adamw> okely dokely, well let's see what we've got
16:08:19 <adamw> #topic Previous meeting follow-up
16:08:27 <adamw> just one thing here, and i'm afraid it's an oops from me:
16:08:56 <adamw> #info "adamw to sync with cmurf and look at revising multiboot criteria" - sorry, didn't get the roundtuits for that one
16:09:01 <adamw> #action adamw to sync with cmurf and look at revising multiboot criteria
16:09:22 <adamw> we definitely should get any criteria changes nailed down ahead of alpha this time, though, so if anyone's thinking of some, time to get the proposals in :)
16:10:16 <adamw> #topic Nightly testing process
16:10:31 <adamw> so just wanted to check in and make sure this was working out OK for everyone...
16:10:44 <adamw> has it been clear what to do, OK finding images and so on?
16:10:52 <smccann> yep
16:11:17 <satellit> would be nice to list the approved koji links on test pages
16:12:01 <adamw> satellit: for the lives?
16:12:06 <satellit> yes'
16:12:10 <adamw> satellit: unfortunately it's tricky because the URL's completely unpredictable :(
16:12:20 <adamw> satellit: the search for recently built nightlies was the best I could come up with
16:12:39 <satellit> I knoe had to go to page 3&4 to find correct ones
16:12:52 <adamw> you could possibly do something really messy by scraping koji's web interface, or if you can search for the names of koji tasks, but i couldn't think of anything better...if anyone can it'd be great
16:13:08 <tflink> can it be done with the koji api?
16:13:22 <adamw> the boot.isos I can do because the location is predictable and based on the date, so that's just a bit of 'if this then that' in the template
16:13:28 <adamw> tflink: possibly? i haven't looked
16:13:35 * amita reading
16:13:38 * satellit flash the wireless workes on todays soas
16:13:50 <adamw> we'd have to have a thing whose only job was just finding nightly live locations and updating a template page i guess
16:13:58 <satellit> dd USb
16:14:17 <satellit> cand dates be searched?
16:14:30 <adamw> satellit: not in the web interface I don't think
16:14:40 <adamw> it doesn't seem to have a search that filters on the task name...unless i've missed it
16:14:57 <satellit> use different builder name than mash
16:14:59 <adamw> #info direct download links for the correct live images (or links to the exact correct set of koji tasks at least) would help, but currently hard to find a way to do it
16:15:10 <adamw> satellit: 'mash' is the builder that builds all the nightlies
16:15:29 <satellit> but pbrobinson and ausil can be searched
16:15:50 <roshi_phone> They only build the candidates
16:15:58 <roshi_phone> Not the nightly
16:15:59 <adamw> satellit: but the way the nightly thing works is, the nightlies just get built automatically, then we choose when to pick one for testing
16:16:07 <roshi_phone> iirc
16:16:08 <satellit> ok
16:17:18 <adamw> it'd be possible to put up a new page every day, but i thought the volume would drive folks nuts
16:17:34 <roshi_phone> That'd be hard to track
16:17:42 <adamw> (and then you'd really need testcase-stats or something to get any useful data out)
16:17:49 <roshi_phone> Imo Anyways
16:18:05 <adamw> and tcs would need a bit of changing too as there'd be just too many builds...
16:18:45 <satellit> but it is great that wireless is working again in rawhide
16:18:45 <adamw> so yeah...if anyone can come up with a viable method for grabbing the correct nightly build urls, that'd be a definite improvement
16:18:51 <adamw> satellit: ah, that got fixed? cool
16:19:05 <satellit> todays soas x86
16:19:13 <roshi_phone> Releng-dash should have them...
16:19:29 <roshi_phone> The right links, I mean
16:20:01 <satellit> but only for latest  not the weekly approved ones
16:20:05 <adamw> roshi: ralph showed me the code that does that
16:20:17 <adamw> roshi: it's the one with the 'let me be clear. i don't know what this is.' comment in it :)
16:20:37 <roshi_phone> I figured you grabbed the weekly on the first day of the week or something
16:20:49 <roshi_phone> Haha, yeah
16:21:00 <adamw> roshi: so far it's just been kind of random - the first one i did when relval's nightly support started working
16:21:11 <adamw> the second one i did when i added --needed and testing it and noticed another was 'needed'
16:21:18 <adamw> so so far it's sort of tied to relval feature additions :P
16:21:24 <roshi_phone> Then it's just a weekly cron job, I didn't know there was more to it
16:21:35 <roshi_phone> Lol
16:21:41 <roshi_phone> Makes sense
16:21:53 <adamw> the idea was to have it run on every new compose
16:21:57 <roshi_phone> Phone typing is slow...
16:22:02 <adamw> though in practice probably just running it nightly as a cron job would probably work fine
16:22:07 <adamw> probably.
16:22:20 <satellit> lots of pages...
16:22:36 <adamw> satellit: with that i meant with the --needed parameter
16:22:57 <adamw> which would actually only nominate a new build at most every 3 days, and only if important packages had changed
16:23:19 <adamw> aaaanyhoo
16:24:00 <adamw> it seems like mostly it's working out? as long as that's the case i'll make sure new builds keep getting nominated and see if i can figure anything better out for the live links
16:24:10 <satellit> did kernel change in todays koji?  (wireless)
16:25:16 <adamw> satellit: haven't looked, to be honest - i wake up 15 minutes before this meeting starts, hit the head, make a coffee and some toast, and that's pretty much it :P
16:25:30 <roshi_phone> Lol
16:25:34 <satellit> : P  same here
16:25:50 <adamw> the last kernel build looks to be 2014-12-19. could be an NM change or something?
16:26:02 <satellit> k
16:26:43 <adamw> so, anyone have any other topics?
16:27:05 <adamw> ooh, here's one
16:27:07 <roshi_phone> I don't
16:27:09 <adamw> #topic Retrospective check-in
16:27:29 <adamw> (adamw: making up the agenda as he goes along since 2010)
16:27:30 * roshi_phone needs to add stuff to that
16:27:36 <adamw> i did just take a quick look at https://fedoraproject.org/wiki/Fedora_21_QA_Retrospective and there's a good point from kparal on it
16:27:58 <adamw> "Kparal (talk) - I wish there was a different process than the blocker bug process for product working groups' requirements, like desktop icons polish. It doesn't really make sense to block all products indefinitely just because of some polish needs. I think a separate process with clearly defined restrictions could be better"
16:28:13 <roshi_phone> Yeah
16:28:23 <adamw> there's a lot of detail to kick around, but there's certainly something similar for Server, in the server role requirements
16:28:39 <tflink> then would we release one product before another?
16:29:00 <adamw> i think in practice it's a problem that would 'solve itself' to some degree as i don't see two Go/No-Go Meetings in a row voting no-go over desktop icon polish, but it'd be nice to nail down something properly
16:29:04 <roshi_phone> We'd release "unpolished"
16:29:14 <adamw> yeah, that sounds more likely
16:29:34 <adamw> i think the conclusion on the split release cycle thing was it'd need a bunch of planning
16:30:48 <adamw> i guess kamil will come up with a more in-depth proposal there, but if anyone else wants to think about it too, more ideas always good...
16:31:12 <roshi> I'll take a look at add stuff to the retrospective today...
16:31:36 <adamw> #agreed kparal's note on Product-specific release blockers is a good one, we'll all think about ways to refine his ideas
16:31:40 <adamw> roshi: thanks!
16:32:01 <adamw> there's another point on there that's yet another variation on the perennial 'no freeze on comps or spin-kickstarts'
16:32:13 <adamw> nirik: around, by any chance?
16:34:16 <adamw> dgilmore: or you?
16:35:02 <adamw> oh, well :/ we should really talk about that with releng though.
16:35:11 <adamw> it does seem to always cause some kind of problem.
16:36:29 <adamw> welp, that's about all I had in the bag of tricks...
16:36:30 <adamw> #topic Open floor
16:37:00 * roshi has nothing
16:37:08 <adamw> anything else, from anyone else? testing? taskotron? test days? favourite colour of M&M? favourite spelling of favorite? or color?
16:37:28 * satellit nothing else   have a nice holiday
16:37:41 * roshi is doing stuff with things
16:37:53 <roshi> testCloud for disposable clients
16:38:09 <smccann> did M&M come out with a white choc variety or did I just dream that?
16:38:25 * satellit_e testing cinnamon live here...
16:38:29 <roshi> I thikn that was a dream
16:38:42 <smccann> (sigh)
16:38:49 <adamw> roshi: stuff with things, you say? that's terrible. you should be doing things with stuff.
16:39:04 <adamw> smccann: it sounds like a wonderful dream.
16:39:07 <tflink> looks like there's another package affected by a somewhat obscure depcheck bug that hasn't been figured out yet
16:39:10 <tflink> https://phab.qadevel.cloud.fedoraproject.org/T372
16:39:20 <adamw> tflink: was that the one that someone asked about on devel@?
16:39:26 <tflink> I think so, yeah
16:39:28 <adamw> i happened to see that at the top of my giant pile of unread devel@ mail...
16:39:37 * tflink hasn't responded to that yet
16:40:07 <roshi> I'll have to take a look at swapping those two - sometime during the time at the place
16:41:37 <adamw> roshi: the surgical precision of your planning is astonishing
16:41:59 <adamw> oh, for anyone who isn't following G+, one of SUSE's QA folks got tired of waiting for us to get automated install testing working and is doing it for us...:)
16:42:25 <tflink> is that the person who was asking in #fedora-qa over the weekend?
16:42:28 <roshi> same guy who we were talking to in fedora-qa?
16:42:29 <satellit> saw that on IRC
16:42:31 <adamw> i dunno, i saw it in g+
16:42:32 <adamw> https://plus.google.com/u/0/101804214464023770367/posts/Esxh59nVjik
16:42:44 <roshi> yeha, same guy
16:43:16 <adamw> i met him at lfnw i think, nice guy
16:43:35 <adamw> i'll take free testing from anyone!
16:44:58 <roshi> yeah, he seemed like a nice fellow :)
16:46:06 <adamw> #agreed many thanks to richard brown for the free automated install testing, in exchange we offer this box of shiny beads
16:46:12 <adamw> oooh, so shiny.
16:46:25 * tflink still needs to look deeper into openqa
16:46:33 <tflink> it's been years since I peered into the guts of it
16:46:34 <roshi> aw, I didn't know we had shiny beads!
16:46:45 <roshi> I thought all we had were shiny baubles
16:46:49 <tflink> so much to do, so little time :(
16:48:58 <amita> same here, need to learn about openqa
16:49:44 <amita> even shiny beads and baubles are new to me
16:49:58 <adamw> hehe
16:50:09 <amita> :)
16:50:20 <adamw> alllrighty, anything else from anyone? if not i'll set the fuse
16:50:23 <adamw> thanks a lot for coming out, folks
16:51:25 <roshi> np - thanks for running it :)
16:52:44 <satellit> afk
16:55:04 <adamw> #endmeeting