QA/Meetings/20090422

From FedoraProject

Jump to: navigation, search

Contents

Attendees

  • Will Woods (wwoods)
  • Adam Williamson (adamw)
  • Jóhann Guðmundsson (viking_ice)
  • Jesse Keating (f13)

Previous meeting follow-up

  • [adamw] - discuss with mcepl and francois about radeon and intel F11Blocker bug status
    • overall, nouveau is looking quite good, radeon and intel have quite a few open bugs each, some of which are quite significant especially there's some major problems with old intel chipsets
    • i'm keeping an eye on those going forward, and appropriate bugs have been added to f11blocker / f11targert
    • jkeating asked about firefox font corruption issue (likely https://bugzilla.redhat.com/show_bug.cgi?id=495908)
  • [adamw] - post to fedora forums asking for feedback on pulseaudio issues
    • will do it this week instead.
  • [jlaska] - schedule autoqa discussion w/ wwoods and jkeating - is there any work we can do prior to F-11 GA?
    • Meeting held w/ jlaska+wwoods+jkeating to discuss current status and identify next steps
    • Wwoods mentioned slightly better logging in autoqa, post-repo-update test now does one test run for each repo (including one test per arch)
    • Expect to have automated test result mails prior to Fedora 11 GA
    • jkeating expressed a desire to add a post-repo-update test for package conflict detection
    • adamw recommended a blog post to increase autoqa awareness
  • [wwoods] - update on pulseaudio issues
    • I believe both intel-* bugs regarding snd_pcm_avail() exploding are closed
    • minor issues remain but AFAICT pulse is no longer exploding due to ALSA bugs

Autoqa update

  • See autoqa update from previous meeting

Open discussion

Upcoming QA events

Action items

IRC Transcript

adamw | alright Apr 22 12:00
adamw | qa meeting time Apr 22 12:00
adamw | who's here? raise an extremity Apr 22 12:00
* | wwoods cough Apr 22 12:01
wwoods | is my clock slow, or is yours fast Apr 22 12:01
* | viking_ice drinks yogurt.. Apr 22 12:01
* | Sonar_Guy lifts a leg, whoops wrong reference ;) Apr 22 12:01
adamw | wwoods: mine's showing 9am Apr 22 12:01
adamw | wwoods: all of the five I can see from here :) Apr 22 12:01
wwoods | step time server 10.5.26.10 offset 209.648472 sec Apr 22 12:01
adamw | heh Apr 22 12:01
wwoods | yup, mine's slow Apr 22 12:01
wwoods | okay woo yeah! I am now prepared to be completely awesome Apr 22 12:01
adamw | f13: ping? Apr 22 12:01
wwoods | jlaska is in meetings and won't be joining us Apr 22 12:02
adamw | whew, because i stink Apr 22 12:02
* | f13 Apr 22 12:02
adamw | allrighty, i think we can proceed Apr 22 12:02
adamw | so, follow up from last week Apr 22 12:02
--- | itami is now known as itami_ Apr 22 12:03
wwoods | here's our agenda https://www.redhat.com/archives/fedora-test-list/2009-April/msg01378.html Apr 22 12:03
adamw | jlaska followed through on his email about test day live CD generation, so follow that up on the mailing list Apr 22 12:03
adamw | review of x.org bugs: ok, I went through all bugs open on rawhide in the three major x drivers this week Apr 22 12:03
adamw | overall, nouveau is looking quite good, radeon and intel have quite a few open bugs each, some of which are quite significant Apr 22 12:04
adamw | especially there's some major problems with old intel chipsets Apr 22 12:04
adamw | i'm keeping an eye on those going forward, and appropriate bugs have been added to f11blocker / f11targert Apr 22 12:04
adamw | anyone else have any thoughts on that? Apr 22 12:05
wwoods | can you identify a couple of the scarier bugs? Apr 22 12:05
wwoods | is there any forward momentum on 'em, or are they just kind of.. floating Apr 22 12:05
f13 | adamw: have you ran across a bug with intel driver that causes fonts in firefox to become corrupted? Apr 22 12:06
adamw | with old intel chipsets: on a default f11 boot currently you get no visible cursor Apr 22 12:06
adamw | f13: yes, there's two reports for that one, i'll dig them out for you in a sec Apr 22 12:06
wwoods | old being i8xx? Apr 22 12:06
adamw | yes Apr 22 12:06
adamw | the cursor works but you can't see it Apr 22 12:06
adamw | booting nomodeset fixes it Apr 22 12:06
adamw | but that's obviously highly visible out of the box Apr 22 12:06
wwoods | bug id? workaround? Apr 22 12:06
f13 | adamw: ah good, thanks. Apr 22 12:06
wwoods | bluh. is there a workaround other than "nomodeset"? Apr 22 12:06
adamw | bug id is 488980, workaround is nomodeset, i don't think there's another Apr 22 12:06
f13 | nomodeset is enough for Preview, but certainly not final Apr 22 12:07
adamw | (though I should probably get someone to check if the SWCursor / HWCursor stuff still does anything for that driver) Apr 22 12:07
adamw | the other major problem with those chipsets is significant opengl borkage Apr 22 12:07
adamw | glxgears renders strangely, and compiz just won't work Apr 22 12:07
adamw | (i8xx again) Apr 22 12:07
adamw | f13's firefox font corruption issue is likely https://bugzilla.redhat.com/show_bug.cgi?id=495908 Apr 22 12:08
buggbot | Bug 495908: medium, low, ---, ajax, ASSIGNED, glyph corruption Apr 22 12:08
wwoods | if the devs have acknowledged that i8xx is just boned and we aren't likely to get it fixed up anytime soon, can we disable modeset for those specific chips? Apr 22 12:08
adamw | hmm, assigned to the wrong guy, let me fix that Apr 22 12:08
adamw | wwoods: i believe that would be possible, yeah, but so far i don't have a definite comment that they won't be able to fix it before ga Apr 22 12:08
adamw | nomodeset is only a workaround for the cursor issue, i don't believe it fixes the opengl issue Apr 22 12:09
wwoods | gotcha. Apr 22 12:09
adamw | beyond that, mostly the issues i'm worried about are hardware-specific 'it just doesn't work' ones Apr 22 12:09
adamw | of which there's a few more than i'd *like* Apr 22 12:09
adamw | in particular there's an i945 variation which has cropped up in two mac systems which just doesn't work Apr 22 12:10
adamw | that's https://bugzilla.redhat.com/show_bug.cgi?id=494390 Apr 22 12:10
buggbot | Bug 494390: medium, low, ---, krh, ASSIGNED, install hangs at graphics: pci 8086:27A2 Intel 945 Apple Macintosh mini Apr 22 12:10
adamw | so far i'm getting good feedback / interaction from reporters but i haven't seen a lot of intel bugs get closed; let me take a look at the numbers on that Apr 22 12:11
adamw | well, 9 rawhide bugs in intel closed this month Apr 22 12:11
adamw | so work is definitely happening Apr 22 12:11
adamw | just not sure whether kristian will get to all the important issues before ga :\ Apr 22 12:12
adamw | this is one situation where i'd like to have severity / priority in place :) ah well Apr 22 12:12
f13 | well and you'd like maintainers to be responding to such things (: Apr 22 12:13
viking_ice | +1 Apr 22 12:13
adamw | well, even without that - at least i'd have a handy one-glance way to track the most important bugs myself Apr 22 12:13
f13 | F11Blocker? (: Apr 22 12:14
adamw | i was saying to jlaska that i've yet to find a way to convince bugzilla to accept a search for, say, bugs in xorg-x11-drv-intel that block F11Blocker Apr 22 12:14
adamw | it looks like it should be possible with the boolean fields, but it doesn't seem to work, no matter what type of match i select or whether i put 'F11Blocker' or the bug number in the field Apr 22 12:14
adamw | always gets zarro boogs Apr 22 12:14
adamw | if anyone knows that it'd be appreciated =) Apr 22 12:15
adamw | anyhow, that's that topic...any more on it before we move on? Apr 22 12:16
adamw | alright Apr 22 12:16
adamw | next action item was for me again: post to forums for feedback on pulse. er, i forgot. sorry! will do it this week instead. Apr 22 12:17
adamw | next was for jlaska, but wwoods and f13 can fill us in: did you guys have an autoqa discussion? Apr 22 12:17
f13 | we did Apr 22 12:18
adamw | excellent...what did you conclude? Apr 22 12:18
f13 | and we had some action items fall out of it, which have yet to be completed Apr 22 12:18
wwoods | a couple things got fixed up Apr 22 12:18
f13 | we've got one or two things to fix up and then we'll turn on the automated part of the autoqa Apr 22 12:18
f13 | with results being reported to autoqa-results@lists.fedorahosted.org Apr 22 12:19
wwoods | slightly better logging in autoqa, post-repo-update test now does one test run for each repo (including one test per arch) Apr 22 12:19
wwoods | made sure we had a plan for updating/maintaining the autoqa code on the test system Apr 22 12:20
adamw | so for the run-up to ga we will have actual useful testing being done by autoqa? Apr 22 12:20
wwoods | if you consider repoclosure / verifytree "useful" then yes Apr 22 12:21
wwoods | if you're expecting automated installer testing.. that may not happen before GA Apr 22 12:21
adamw | awesome Apr 22 12:21
f13 | I'd like to hook conflict checking into that as soon as possible though Apr 22 12:21
wwoods | file conflicts or other? Apr 22 12:22
wwoods | should be pretty trivial if there's code already written to check it Apr 22 12:22
wwoods | and.. not so bad otherwise Apr 22 12:22
f13 | wwoods: any. Apr 22 12:22
f13 | yeah, i think seth already has code for this Apr 22 12:22
wwoods | righteous. I assume that's a post-repo-update test? Apr 22 12:23
f13 | yep Apr 22 12:23
wwoods | rad. okay, that goes on the list as the Next Test To Write Apr 22 12:24
adamw | ok, sounds awesome...on a pr note, maybe you could write up a blog post about this stuff will? Apr 22 12:25
adamw | i think some people have the idea autoqa is still just a big ball of gas, not something we actually have written and working right now Apr 22 12:25
wwoods | hm, sure - wonder if the autoqa-results list is open-access Apr 22 12:25
wwoods | it's not production-ready, but yeah, there's code and it's generating results Apr 22 12:25
viking_ice | perhaps pr for f12 Apr 22 12:26
wwoods | https://fedorahosted.org/pipermail/autoqa-results/ Apr 22 12:26
wwoods | whee Apr 22 12:26
adamw | yeah, works for me and i'm not in any vpns or anything Apr 22 12:26
wwoods | rad. Apr 22 12:26
viking_ice | it's open Apr 22 12:26
adamw | viking_ice: sure, we don't want to do a full-court press now, but just a note in will's blog (and hence planet) to say that we've actually got code running and generating results now Apr 22 12:26
viking_ice | I can see all the unresolved deps.. to what purpose usefull to me I really do not know.. Apr 22 12:27
adamw | viking_ice: this is mostly for maintainers and releng Apr 22 12:27
viking_ice | yup kind new that.. Apr 22 12:27
wwoods | viking_ice: well, the more interesting tests will come soon, but e.g. the verifytree test will tell you whether or not there are rawhide images Apr 22 12:27
wwoods | (that's post-tree-compose, not post-repo-update) Apr 22 12:28
adamw | autoqa is a big evil robot that will make all of us fleshy qa units obsolete, not help us ;) Apr 22 12:28
viking_ice | sweet .. Apr 22 12:28
wwoods | and obviously we want to get a better reporting mechanism than "email on a list" Apr 22 12:28
wwoods | but.. baby steps Apr 22 12:28
adamw | ok, awesome, autoqa is rockin' Apr 22 12:28
adamw | final action item from last week was me again - priority/severity stuff Apr 22 12:29
adamw | which as you can see I did, and is now generating a robust and enthusiastic discussion Apr 22 12:29
adamw | =) Apr 22 12:29
wwoods | hah. well said Apr 22 12:29
adamw | if anyone wants to chat about that here we can do too, but hopefully not too much, don't want to eat too much meeting time with it Apr 22 12:30
wwoods | I think any discussion here would be redundant Apr 22 12:30
adamw | ok Apr 22 12:30
adamw | one last thing from last week then i guess - anything further on pulseaudio from you, will? Apr 22 12:31
wwoods | so let's move on (plus I have a meeting about Fedora 11 Preview readiness in 30min) Apr 22 12:31
wwoods | I believe both intel-* bugs regarding snd_pcm_avail() exploding are closed Apr 22 12:31
wwoods | minor issues remain but AFAICT pulse is no longer exploding due to ALSA bugs Apr 22 12:31
adamw | awesome...so theoretically we shouldn't have any big bugs that make pulse just not work for large numbers of people left? Apr 22 12:32
wwoods | volume control issues are the remaining major class of bugs that I know of Apr 22 12:32
wwoods | there might still be problems with snd_pcm_avail() on less common audio chipsets (emu*?) Apr 22 12:32
viking_ice | Perhaps the default on pcm should be 80% percent... Apr 22 12:32
wwoods | viking_ice: that'd be reasonable if we could ask alsa which volume control was "PCM" Apr 22 12:33
adamw | fwiw I have a ice1724 and no problem with pulse, so that one's good. dunno about the emus. i'll definitely get that message on the forums. Apr 22 12:33
wwoods | but apparently it's not that simple Apr 22 12:33
viking_ice | frack.. Apr 22 12:33
adamw | mandriva defaults every mixer control to 80% and then has a big list of exceptions for ones that *shouldn't* be set to 80%. that's ugly. i'm not aware of any decent non-ugly solution to the default volume issue :\ Apr 22 12:33
wwoods | see http://people.redhat.com/alexl/files/why-alsa-sucks.png Apr 22 12:33
f13 | you said alsa, nothing is going to be simple. Apr 22 12:33
wwoods | and no, that is not a forged screenshot Apr 22 12:34
--- | warren_ is now known as warren Apr 22 12:34
viking_ice | Well let's fill forums release notes and the fedora channel with "alsamixer -c0" Apr 22 12:34
adamw | basically, i really think we need an old-skool graphical mixer in f11, and i will fight for that at higher levels if necessary. but aside from that, pulse is looking good i think Apr 22 12:34
wwoods | IIRC dude had to hack the X server or GTK to get the coordinate space large enough to actually show *all* the controls in one screenshot Apr 22 12:34
wwoods | it's ridiculous Apr 22 12:34
wwoods | eventually it'll get sorted such that we don't need to resort to alsamixer -c0 to turn the hidden knobs Apr 22 12:35
adamw | yeah, and we'll all get ponies :) Apr 22 12:35
wwoods | in the meantime, though, that's gonna remain a FAQ Apr 22 12:35
wwoods | and a PITA Apr 22 12:35
wwoods | and various other TLAs Apr 22 12:35
* | nirik can beat that even. If you have ever looked at the alsamixer output from a openmoko freerunner. ;) Apr 22 12:36
adamw | alright, i think we all know the basic status there, though. we can't do much about it in this meeting, and time's tight, so - moving on again Apr 22 12:36
wwoods | agreed Apr 22 12:36
adamw | next topic is autoqa review, which i think we covered already, right wwoods / f13? Apr 22 12:36
wwoods | pretty much, yeah. if anyone has CODE for simple tests we could run after a repo is updated, or a new tree is built Apr 22 12:37
wwoods | please let us know about it Apr 22 12:37
adamw | ...or forever hold your peace Apr 22 12:37
adamw | (or just send it to the mailing list) Apr 22 12:37
wwoods | indeed. Apr 22 12:38
viking_ice | btw to you have task list for project autoqa somewhere accessible online Apr 22 12:38
viking_ice | encase any wants to provide code.. Apr 22 12:38
wwoods | no. Apr 22 12:38
adamw | ah, yeah, good question - is this list of tests to write online anywhere, for e.g. Apr 22 12:38
wwoods | we've got a trac instance so I could start filing tickets there. Apr 22 12:38
f13 | That photo crashed my box Apr 22 12:38
adamw | heh Apr 22 12:38
adamw | wwoods: that sounds like a good idea, and a way maybe to get da community involved in autoqa (i remember you were having trouble with that before) Apr 22 12:39
f13 | https://fedoraproject.org/wiki/Automated_QA_Testing_Project as well Apr 22 12:39
f13 | we wanted to wait until we had a well functioning harness before opening up for test submission Apr 22 12:39
wwoods | right Apr 22 12:39
f13 | we used the two existing tests as examples to get all the other bits working right Apr 22 12:39
f13 | so we're not quite ther eyet Apr 22 12:39
adamw | ah, i see Apr 22 12:39
wwoods | but as far as I can see, there's no harm in tracking the current status in trac tickets Apr 22 12:39
--- | fab__ is now known as fab Apr 22 12:40
wwoods | it'll help me keep track of tasks and milestones Apr 22 12:40
wwoods | and give some transparency into the process Apr 22 12:40
adamw | yeah, as long as it doesn't add too much overhead it's probably a win Apr 22 12:40
viking_ice | Fedora QA track instance could also be used I think.. Apr 22 12:40
adamw | what trac are you using currently? Apr 22 12:41
wwoods | anyway, let's move on, since we've got some big things to discuss with Preview looming Apr 22 12:41
f13 | I'll try to slap something together today Apr 22 12:41
adamw | OK, move on sounds good Apr 22 12:41
adamw | next topic's a biggie - f11 preview preparation Apr 22 12:41
wwoods | I'm not. I'm using 'vim ~/autoqa-todos.txt' Apr 22 12:42
adamw | jlaska split this into three bits for us Apr 22 12:42
adamw | blocker bug, schedule, install results Apr 22 12:42
wwoods | so moving that to trac is probably a good idea. but yes. blocker bug! Apr 22 12:42
adamw | so, for blocker bug...looks pretty manageable Apr 22 12:43
adamw | 14 open bugs, a lot of which are MODIFIED or ON_QA Apr 22 12:43
wwoods | f13: when do we need to start pushing the Preview bits out? Apr 22 12:43
f13 | we need a lot of verification of those though Apr 22 12:43
f13 | today would be a very good day to compose the bits Apr 22 12:43
f13 | tomorrow less so Apr 22 12:43
--- | thomasj_ is now known as thomasj Apr 22 12:44
f13 | frelling deltarpms have been making rawhide composes take too fucking long though and eating into our testing time Apr 22 12:44
wwoods | yikes. okay. Apr 22 12:44
f13 | this is why I asked for a matrix and help testing things last week Apr 22 12:44
adamw | f13: what are the real biggies in your view? Apr 22 12:45
wwoods | I've gotta get some upgrade test cases into the install matrix Apr 22 12:45
f13 | it's hard to say since there hasn't been much testing Apr 22 12:45
f13 | verifying all the MODIFIED stuff is big Apr 22 12:45
f13 | and correcting any of the big Beta issues that we shipped with Apr 22 12:46
f13 | I'm going to make up some isos today, once rawhide lands Apr 22 12:46
wwoods | unfortunately I lost a test system and one of my other test systems is EFI, which is.. not great Apr 22 12:46
f13 | just waiting on ppc to finish buildinstallinating Apr 22 12:47
f13 | ppc is one major thing too Apr 22 12:47
f13 | since ppc was DOA for beta Apr 22 12:47
adamw | there's what looks like a showstopper ppc bug in the list, 496173 Apr 22 12:48
wwoods | true. I'll spin up some tests on ppc64 if it helps - are there images somewhere? Apr 22 12:48
adamw | status MODIFIED as of april 17th, looks like it needs a re-test Apr 22 12:48
f13 | I've just confirmed it Apr 22 12:48
wwoods | anaconda .46 is in rawhide, so yeah, should be fixed Apr 22 12:48
f13 | wwoods: there is yesterday's rawhide, and there will be today's rawhide Apr 22 12:49
f13 | well I confirmed it yesterday and forgot to close the bug Apr 22 12:49
wwoods | oh, anaconda .47 lands today? nice, maybe EFI will work Apr 22 12:49
f13 | and mdraid Apr 22 12:50
f13 | mdraid is a bit shakey Apr 22 12:50
f13 | so overall anaconda storage is a concern Apr 22 12:50
wwoods | mdraid has always been bunk Apr 22 12:51
adamw | color me surprised :\ Apr 22 12:51
wwoods | I hate it with hammers Apr 22 12:51
f13 | erm, no it hasn't Apr 22 12:51
f13 | perhaps you're thinking of dmraid Apr 22 12:51
wwoods | oh! Apr 22 12:51
wwoods | yes. Apr 22 12:51
f13 | oddly enough, dmraid is solid, mdraid is not Apr 22 12:51
adamw | and of course that's the trickiest thing to test, because you need appropriate hardware and an up-to-date anaconda and to run an actual install Apr 22 12:51
wwoods | I apologize. Please redirect my hammer-wielding hatred. Apr 22 12:52
wwoods | I have a preupgrade bug regarding an mdraid install which is probably this Apr 22 12:52
f13 | rawhide is rsyncing Apr 22 12:52
wwoods | the reporter says cute things like "Apparently this new version of the OS is dropping support for upgrading Apr 22 12:53
wwoods | software RAID volumes." Apr 22 12:53
adamw | i'm just poking through the bugs on the list now to see what needs to happen to get them tested / fixed... Apr 22 12:53
wwoods | yes, apparently this bug indicates that we are intentionally removing support for your configuration. never ascribe malice etc. Apr 22 12:53
f13 | anything else? I've got a meeting and need to input some food first Apr 22 12:55
adamw | yeah, the other two bits Apr 22 12:55
wwoods | no, I think we're in well-worn territory here. standard day-before-push panic. Apr 22 12:55
adamw | schedule and install test results Apr 22 12:56
adamw | i'm not sure exactly what jlaska wanted to talk about there Apr 22 12:56
adamw | presumably, do we think we're ok for the schedule? will we make it? Apr 22 12:56
wwoods | schedule: release is apr. 28. if we want to hit that, we need to push bits today or tomorrow Apr 22 12:56
wwoods | I'd feel more comfortable answering that question once we've had a chance to test with today's rawhide bits Apr 22 12:57
adamw | ok Apr 22 12:58
adamw | so, conclusion: we need to push some bits, build some images and test 'em Apr 22 12:58
wwoods | right. most of the stuff on the blocker list needs retesting, we need today's bits for some of it. Apr 22 12:58
wwoods | f13: any ETA on the rawhide compose? Apr 22 12:59
f13 | it's rsyncing to the master now Apr 22 12:59
wwoods | I'll see what I can do with current rawhide - get some ppc64 testing in, etc Apr 22 12:59
wwoods | nice. Apr 22 12:59
adamw | several of the bugs on the blocker list are claimed fixed last month but haven't been confirmed, i've just sent out pings on most of those Apr 22 12:59
wwoods | alright - f13 and I have another meeting to hit Apr 22 13:00
adamw | ok, thanks all Apr 22 13:00
f13 | it's first stage rsync though, not hte second stage with repodata Apr 22 13:00
f13 | later Apr 22 13:00
wwoods | I think we'll discuss it further in that meeting and afterward we'll have bits Apr 22 13:00
f13 | wwoods: I'll be on that meeting in a few minutes Apr 22 13:00
wwoods | we can begin retesting in earnest Apr 22 13:00
adamw | does anyone have anything else to bring up, which doesn't involve f13 or wwoods? :) Apr 22 13:00
wwoods | and have a much better picture of how boned we are in a few hours Apr 22 13:00
adamw | update the boneometer! Apr 22 13:01
viking_ice | what'ś that tied to the gaydar it's the boneometer.. Apr 22 13:01
viking_ice | available for just 9billiion swahili dollars.. Apr 22 13:02
adamw | aha Apr 22 13:03
adamw | ok, meeting adjourned! thanks all Apr 22 13:03

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!