16:03:28 #startmeeting 16:03:28 Meeting started Wed Feb 29 16:03:28 2012 UTC. The chair is slangasek. Information about MeetBot at http://wiki.ubuntu.com/meetingology. 16:03:28 16:03:28 Available commands: #accept #accepted #action #agree #agreed #chair #commands #endmeeting #endvote #halp #help #idea #info #link #lurk #meetingname #meetingtopic #nick #progress #rejected #replay #restrictlogs #save #startmeeting #subtopic #topic #unchair #undo #unlurk #vote #voters #votesrequired 16:04:38 [TOPIC] Lightning round 16:04:47 $ echo $(shuf -e barry doko stgraber jodh ev bdmurray slangasek ogra infinity) 16:04:50 infinity barry stgraber ogra bdmurray slangasek doko jodh ev 16:04:58 infinity: looks like you get to go first for your first Foundations meeting :) 16:05:11 Fiiiine. 16:05:13 ohh man, this list is long ... 16:05:14 Worked on this week: - cdimage and debian-cd tweaks for ARM images and kubuntu-active - helped with general release management and queue management - filed tickets with IS for buildd precise upgrades - looked at fpc on armhf a bit more 16:05:18 Upcoming work: - need to chase down status of gnat on armhf with Linaro - more work on livefs-in-soyuz (rewriting from memory) - need to find time to debug the root cause of the qtwebkit-source FTBFS - finish the fpc on armhf port - dodge doko's attempts to dump the entire C toolchain on me - possibility of a (short) emergency VAC for personal reasons 16:05:23 Gah. 16:05:26 irrsi, damn you and your hatred of leading whitespace. 16:05:44 irssi, too. 16:05:48 Let's try that harder. 16:05:50 Worked on this week: 16:05:51 - cdimage and debian-cd tweaks for ARM images and kubuntu-active 16:05:51 - helped with general release management and queue management 16:05:51 - filed tickets with IS for buildd precise upgrades 16:05:53 - looked at fpc on armhf a bit more 16:05:56 Upcoming work: 16:05:58 - need to chase down status of gnat on armhf with Linaro 16:06:00 - more work on livefs-in-soyuz (rewriting from memory) 16:06:03 - need to find time to debug the root cause of the qtwebkit-source FTBFS 16:06:05 - finish the fpc on armhf port 16:06:08 - dodge doko's attempts to dump the entire C toolchain on me 16:06:11 - possibility of a (short) emergency VAC for personal reasons 16:06:13 There. 16:06:30 infinity, qtwebkit ... this is the linker needing too much memory, afaics 16:06:41 doko: Yes, but why? 16:07:05 because it's an insane amount of code? 16:07:15 doko: It's the same amount of code on other arches. :P 16:07:27 it was running out of memory on other arches too, before tweaking? 16:07:31 doko: Including armel, which doesn't have the issue. 16:07:47 infinity, yes, and webkit ftbfs on powerpc for the same reason 16:07:56 So, I dunno. We'll see. Maybe it can just be tweaked a bit, or maybe the linker on armhf is doing something dumb. 16:08:42 infinity: all done? 16:08:49 barry: Go nuts. 16:09:03 * barry can't go where he already is, but... 16:09:09 Go nutser. 16:09:11 bugs investigated, fixed, sponsored: bug 903126 (smc ftbfs), bug 903121 (crystalspace ftbfs), bug 27520 (cron multiple SRUs), bug 828731 (kexec-tools multiple SRUs), bug 848915 (update-manager without $DISPLAY), bug 556293 (apt global proxy), bug 942408 (python-support not cleaning up after upgrade). Sync'd wadllib 1.3.0-1. Also worked on Python issue 13703 (hash collision security issue) and Python 2.6.8rc1 released. To do: +1maint, 16:09:11 look at a few more bugs from my list, and prepping for Pycon next week. ∞ 16:09:16 Launchpad bug 903126 in smc (Debian) "smc 1.9-4ubuntu1 fails to build with cegui-mk2 0.7.5-7" [Unknown,New] https://launchpad.net/bugs/903126 16:09:17 Launchpad bug 903121 in crystalspace (Ubuntu) "crystalspace 1.4.0+dfsg-4ubuntu2 fails to build with cegui-mk2 0.7.5-7" [Medium,Triaged] https://launchpad.net/bugs/903121 16:09:18 Launchpad bug 27520 in cron (Ubuntu Lucid) "cron daemon caches user-non-existent lookup results, causing "ORPHAN" message and skipping jobs for all LDAP/NIS-defined users" [Medium,Fix committed] https://launchpad.net/bugs/27520 16:09:19 Launchpad bug 828731 in kexec-tools (Ubuntu Natty) "kdump functionality not working as expected when /boot is a separate partition" [Medium,In progress] https://launchpad.net/bugs/828731 16:09:20 Launchpad bug 848915 in update-manager (Ubuntu Precise) "cryptic traceback when running update-manager w/no X $DISPLAY" [High,Triaged] https://launchpad.net/bugs/848915 16:09:58 - Networking 16:09:58 - Blogged about the changes to DNS in 12.04: http://www.stgraber.org/2012/02/24/dns-in-ubuntu-12-04/ 16:10:01 - Tested and applied patch to fix netcfg's dns-nameservers handling (bug 932275) 16:10:02 Launchpad bug 932275 in ifupdown (Ubuntu) "Bogus or no nameserver information written to /etc/network/interfaces" [High,Triaged] https://launchpad.net/bugs/932275 16:10:04 - TPM 16:10:06 - Fixed bug 934799 so that trousers now installs properly 16:10:07 Launchpad bug 934799 in trousers (Ubuntu) "trousers start script has exit status 137" [Undecided,Fix released] https://launchpad.net/bugs/934799 16:10:09 - Installer 16:10:11 - Last minute fixes of the installer slideshows (all flavours) and uploaded for beta1 16:10:14 - Fixed bug 898278 and bug 936115 in ubiquity. 16:10:16 Launchpad bug 898278 in ubiquity (Ubuntu Precise) "Upgrade menu option should not appear for old releases" [High,Fix released] https://launchpad.net/bugs/898278 16:10:17 - Tracked down bug 939450 to a gtk regression 16:10:18 Launchpad bug 936115 in ubiquity (Ubuntu Precise) "ubiquity crashed with TypeError in partman_popup(): popup() takes exactly 7 arguments (6 given)" [High,Fix released] https://launchpad.net/bugs/936115 16:10:19 Launchpad bug 939450 in ubiquity (Ubuntu Precise) "ubiquity crashed with TypeError: argument of type 'NoneType' is not iterable in ubi-partman.py" [High,Triaged] https://launchpad.net/bugs/939450 16:10:19 - Uploaded a workaround for bug 645449 that in turn caused bug 942560, now woking on the right way of fixing this ... 16:10:23 - Also started debugging bug 940908 but it's lower priority than the other one 16:10:25 Launchpad bug 645449 in ubiquity (Ubuntu Precise) "Ubiquity hangs at Keyboard layout if you use keyboard to navigate / select" [High,Fix released] https://launchpad.net/bugs/645449 16:10:26 - Noticed during ISO testing that Wubi in Edubuntu doesn't allow me to install it (only offer the "reboot to live environment" option), apparently caused by broken free space calculation 16:10:26 Launchpad bug 942560 in ubiquity (Ubuntu) "keyboard layout screen - Keyboard navigation broken" [High,Triaged] https://launchpad.net/bugs/942560 16:10:27 Launchpad bug 940908 in ubiquity (Ubuntu Precise) "Keyboard layout not set on persistent USB image" [High,Confirmed] https://launchpad.net/bugs/940908 16:10:30 - Containers 16:10:32 - Some more Ubuntu server guide review for the LXC section 16:10:35 - Fixed some more bugs in the Ubuntu LXC template to do with sudo group in older releases. 16:10:38 - Other 16:10:40 - Wrote a new queuebot, currently running in #ubuntu-release 16:10:43 (done) 16:10:49 infinity, ftbfs on armel too 16:11:02 barry: I assigned you a couple of triaged bugs on update-manager that looked like easy wins; do you think you could get to those this week? 16:11:17 doko: Only due to a broken .symbols file. 16:11:18 slangasek: yep, their on my list for this week 16:11:20 doko: The link succeeds. 16:11:22 cheers 16:11:41 Worked on fix for bug 940290 (updating tests atm...) Investigating bug 16:11:41 883618 - looks like "something" is calling an xdg-* command or possibly 16:11:41 /usr/share/acpi-support/screenblank, but tricky to pin this down due to 16:11:41 the time taken to perform the upgrade. Wishing we had opengrok to help 16:11:42 Launchpad bug 940290 in upstart "console log consumes fds upon job restart/respawn" [Medium,Confirmed] https://launchpad.net/bugs/940290 16:11:44 track this down. :). Did some investigation on bug 935296 and 16:11:45 Launchpad bug 935296 in java-atk-wrapper (Ubuntu Precise) "java-atk-wrapper prevents java applications from starting" [High,Confirmed] https://launchpad.net/bugs/935296 16:11:47 established bug is with AT-SPI/ATK. 16:11:51 ⌛ 16:12:03 slangasek: if you have a preference for priority, pvtmsg me your list 16:13:36 barry: ack 16:13:38 ogra_: you're up 16:13:47 jodh, is serial console support still on your list ? 16:13:57 done: 16:13:57 * started researching why ubiquity kde bits end up on ac100 images 16:13:57 * look into bug #924018 (Preseeding doesn't work with oem-config) ... made sure preseed values get applied at all, now on to why oem-config doesnt pick up 16:13:57 * tried to spend a little time with bug #747229 (screen color change when oem-config-remove runs in serial consoles) 16:13:58 * closed a handfull of bugs in the "rotting ubuntu-arm bugs" list, ongoing process 16:14:00 * started researching why the keyboard selection gets completely skipped in oem-config in some preinstalled images (bug #925482) 16:14:00 Launchpad bug 924018 in ubiquity (Ubuntu) "Preseeding doesn't work with oem-config" [Medium,Confirmed] https://launchpad.net/bugs/924018 16:14:02 Launchpad bug 747229 in ubiquity (Ubuntu Precise) "weird color change during oem-config debconf package removal step in serial installs" [Medium,Confirmed] https://launchpad.net/bugs/747229 16:14:03 Launchpad bug 925482 in ubiquity (Ubuntu Precise) "no keboard selection offered in oem-config on precise ac100 images" [High,New] https://launchpad.net/bugs/925482 16:14:03 * did general image testing for omap4 and ac100 as well as typical milestone paperwork (tech preview wiki, manifest etc) 16:14:06 todo: 16:14:08 * fix whatever causes the kde bits to show up 16:14:10 * more oem-config bugwork 16:14:12 * stopwatch initrdless boot 16:14:14 * look across the armel ftbfs list 16:14:16 .. 16:14:36 ogra_: Which kde bits? 16:14:59 infinity, well, ubiquity-frontend-kde and all its deps end up on the ac100 images 16:15:34 not sure why ... but its pretty visible since akonadi is the first app in the unity app lens now :P 16:15:49 they also dont get removed 16:15:51 .. 16:15:57 bug triage of iso-testing bug reports 16:15:57 bug triage of update-manager bug reports 16:15:59 discussion with mvo regarding NvidiaDetection and datadir and distribution upgrades (fixed) 16:16:02 ogra_: I'll poke at it, it's my code breaking. :P 16:16:09 you sure ? 16:16:13 *nod* 16:16:17 modifications to the bug pattern for bug 541595 to avoid incorrect dupication 16:16:18 ogra_: it's in there somewhere. I'll see if I can sift it slightly 16:16:18 closer to the surface. 16:16:18 16:16:18 16:16:19 Launchpad bug 541595 in dpkg (Ubuntu) "[Master] package failed to install/upgrade: package is already installed and configured" [High,Confirmed] https://launchpad.net/bugs/541595 16:16:21 bug bot fix for no package bugs without a SourcePackage line in the bug description 16:16:24 wrote launchpadlib code to show the 20 hottest bugs about packages to which the foundations team is subscribed 16:16:27 blog post regarding recent package bug tasks charts 16:16:33 infinity, i was suspecting ac100-tarball-installer vs jasper 16:16:42 experimentation with recent package bug tasks and showing duplicate bugs 16:16:43 experimented with recent bug tasks graph changes (periods show 1-7 and 8-14) 16:16:46 created a bugs receiving duplicates chart 16:16:52 troubleshooting an mdadm recovery issue - bug 925280 16:16:52 push ubiquity apport package hook modifications to fix bug 874727 16:16:53 call with evan regarding crash database 16:16:53 Launchpad bug 925280 in mdadm (Ubuntu) "Software RAID fails to rebuild after testing degraded cold boot" [Medium,Confirmed] https://launchpad.net/bugs/925280 16:16:55 Launchpad bug 874727 in apport (Ubuntu Precise) "media error bug report blocking is incomplete" [High,Fix released] https://launchpad.net/bugs/874727 16:17:02 ␗ 16:17:07 bdmurray: hmm, is there more than one "already installed and configured" bug, or was the bug pattern duping things that were altogether unrelated? 16:17:36 infinity, both the bfd and the gold linker fail to link this 16:17:36 slangasek: no, not altogether unrelated - just a bit overzealous 16:18:01 slangasek: it was a small number of incorrect dupes and I fixed the most recent ones 16:18:09 ok 16:18:15 doko: Ahh, I hadn't tested with gold yet (though I note that the build is attempting to use gold and failing to do so... Do we not carry the --with-ld GCC patch?) 16:19:02 infinity, use -B/usr/lib/gold-ld to use gold 16:19:13 /usr/bin/ld: fatal error: libQtWebKit.so.4.9.0: mmap: failed to allocate 4275704736 bytes for output file: Cannot allocate memory 16:19:13 collect2: ld returned 1 exit status 16:20:01 bdmurray: I don't think I've seen this hottest-bugs list yet - anything interesting there? 16:20:57 * worked on: 16:20:57 * ARM team hand-off 16:20:57 * driving discussion with Debian about upstart 16:20:57 * beta-1 tending 16:20:57 * NFS fixes (bug #939232) 16:20:59 Launchpad bug 939232 in libnfsidmap (Ubuntu) "nfs mount fails and rpc.idmapd fails" [Medium,Fix released] https://launchpad.net/bugs/939232 16:21:00 * resolvconf: cope with historically broken netcfg. (bug #932275) 16:21:00 Launchpad bug 932275 in ifupdown (Ubuntu) "Bogus or no nameserver information written to /etc/network/interfaces" [High,Triaged] https://launchpad.net/bugs/932275 16:21:02 * upload activity-log-manager to fix an upgrade issue (bug #939842) 16:21:03 Launchpad bug 939842 in activity-log-manager (Ubuntu) "package activity-log-manager-control-center 0.9.1-0ubuntu2 failed to install/upgrade: trying to overwrite '/etc/dbus-1/system.d/com.ubuntu.WhoopsiePreferences.conf', which is also in package whoopsie 0.1.10" [Undecided,Fix released] https://launchpad.net/bugs/939842 16:21:04 * upcoming: 16:21:07 * planning finalized for virtual sprint (Mar 19-21) 16:21:09 slangasek: that's for later! the hottest bug is twice as hot as the 2nd hottest and we should sort it out 16:21:10 (done) 16:21:32 ogra_: Definitely my bug. Fixing. 16:22:06 bdmurray: ok :) 16:22:10 - integrated icedtea-web pages upstream, update package to current 1.2 branch 16:22:10 - update and integrate cjk font patches into openjdk-6 and -7 16:22:10 - investigate test suite hang in openjdk-6, ix86 only, for now just disable the tests 16:22:10 - gcc-4.6.3~rc1 packaging 16:22:10 - prepare gcc-linaro updates based on the existing 2012.02 release 16:22:11 - cherry pick gcc-linaro ARM patches from Linaro 16:22:15 - help janimo to investigate the chromium build failure on armhf 16:22:17 - investigate and forward some general gcc-4.6 ICEs (ppl, webkit) 16:22:19 - investigate qtwebkit-source build failures on ARM 16:22:21 - prepare eglibc debian merge (libc6-dbg multiarch installability for non-ARM) 16:22:23 - dump eglibc on adconrad, do it right! 16:22:24 (done, but still dumping ;) 16:22:27 * slangasek eyes the snow 16:23:35 * infinity doesn't see anyone named adconrad. 16:23:56 heh 16:24:24 and jodh went and ev is on the list but is at conference today 16:24:26 that's a fake dump so that you are not prepared for the real one ;-P 16:24:48 [TOPIC] Bugz 16:25:38 so our hottest bug is bug 659438 16:25:39 Launchpad bug 659438 in apt (Ubuntu) "Installation/Removal fails because of package which could not be located (failure in apt.Cache.required_download)" [High,Triaged] https://launchpad.net/bugs/659438 16:26:19 I believe there is some bug status confusion on it 16:26:21 bdmurray: And if it didn't OOPS LP, we could look at it. 16:26:34 it worked for me 16:26:49 Three timeouts in a row for me. 16:26:52 Oh well. 16:26:56 use +text 16:27:02 if you want the gist of it 16:27:11 Ahh, it just rendered on the fourth go. 16:27:32 hmm, some dupes there 16:28:09 only dupe 903269 has apport data and is from precise 16:28:56 I wonder if we should just set that as the master bug 16:30:52 It might also be that its the same traceback but a different bug 16:31:04 could be 16:31:33 ogra_: (Fix committed and uploaded, BTW) 16:31:47 merci ! 16:31:58 * ogra_ strikes it off his todo 16:32:10 bdmurray: followed up on the bug and dropped the "fix committed" 16:32:17 er no 16:32:22 attempted to, but now I've hit an lp timeout ;) 16:32:43 bdmurray: I'll try to follow up on that asynchronously then. what else do you have for bugs? 16:33:25 ok that's way too much white stuff falling from the sky 16:33:26 bug 941922 - I couldn't decide how important that should be 16:33:30 Launchpad bug 941922 in update-manager (Ubuntu) "do-release-upgrade races puppet for file contents" [Undecided,New] https://launchpad.net/bugs/941922 16:34:23 bdmurray: hmm, I'd say medium at most 16:35:02 seems like people using puppet should be disabling it during a dist-upgrade, no? 16:35:17 I'm inclined to say that automagic configuration management and package managers just plain can't run concurrently. 16:35:35 the definitely should turn off puppet or any other configuration management tool during upgrade 16:35:44 *they 16:35:47 Not sure if that's something that should be fixed in apt/dpkg/rpm/etc/etc, or just in puppet (or in the minds of admins) 16:35:54 bdmurray: about #659438> do the reproduce steps actually work? 16:36:31 mvo: the test case in that? I haven't tried that yet but can today. 16:36:45 stgraber, well, as ugly as it is, then there should probably a debconf note telling them 16:36:45 we *could* turn off pupett automatically before the upgrade if that is something that people want, I'm just not sure that this is uncontroversial 16:37:24 bdmurray: that would be great, TBH the instruction does not look correct to me but I may be wrong about it 16:37:29 (like we do on libc upgrades) 16:37:34 mvo: Sure, you could do it in a high level tool like update-manager, but that still feels like it's broken for low-level package management. 16:37:49 * mvo nods at infinity 16:38:03 mvo: Maybe what this really wants is puppet shipping an apt-utils hook that turns itself off and on. 16:38:15 that sounds sane to me 16:38:15 mvo: Still doesn't cover the dpkg case, but it covers most everything else. 16:38:57 infinity: good point, that sounds like the best way forward, dpkg::{pre,post}-invoke 16:39:05 * infinity nods. 16:39:27 so we can reassign it to puppet and let the server team decide the severity? :) 16:39:30 And a dependency on apt-utils, so it works. (or is apt-utils no longer required for dpkg::-invoke hooks?) 16:39:56 the invoke hooks should work without apt-utils 16:40:00 Oh, shiny. 16:40:05 Even better, then. 16:40:17 Feels like the cleanest solution to me, short of doing it at the dpkg level. 16:40:21 * mvo tries to think of a reason why it wouldn't but fails and gets more tea to keep the engine running 16:40:28 mvo, infinity: one of you want to reassign to puppet with that explanation? 16:40:35 What about the case where an admin runs do-release-upgrade by 16:40:36 mistake? Wouldn't it be better to have the command print a warning an 16:40:36 exit until admins intervene to manually switch off puppet/cfengine/bcfg2/etc? 16:40:36 16:40:45 s/an/and/ 16:41:08 jodh: if you run "do-release-upgrade" by mistake, you get to keep both pieces 16:41:28 jodh: Well, with the apt hook solution, there's no real need for warning. puppet just gets disabled (and re-enabled) every time apt runs. 16:41:28 we're not responsible for our users hiring minimally competent sysadmins ;) 16:41:31 jodh: if we go with the suggestion from infinity that would be fine because puttet would shutdown when dpkg is invoked by apt which it should. and if at this point it was still a mistake there are more problems :) 16:42:34 slangasek: I can follow up on the bug post-meeting, since mvo seems to have remained silent. :P 16:42:41 infinity: thanks :-) 16:42:52 infinity: I was about to say that I can do it but you deserve credit for the plan so … ;) 16:43:00 bdmurray: any more? 16:43:13 If a system is being managed by puppet, is it ever valid to run 16:43:13 do-release-upgrade though? 16:43:13 16:43:25 bug 942030 16:43:26 Launchpad bug 942030 in ubiquity (Ubuntu) "'Take a photo' preview pane not showing live preview" [High,Confirmed] https://launchpad.net/bugs/942030 16:43:55 jodh: Sure. 16:44:33 jodh: d-r-u is just a flimsy wrapper around apt. In the best case (ie: if we don't have horrible bugs we're hiding and working around in update-manager), it literally does nothing but install and remove packages. :P 16:44:38 bdmurray: I guess I'll take that one once I'm done with the keyboard stuff then, will need to figure out a way to expose a webcam to kvm ;) 16:44:52 jodh: And if "running apt" is an invalid thing for puppet users to do, there's a problem. 16:44:52 heh 16:45:38 that's all then 16:47:55 [TOPIC] Patch piloting 16:48:01 just as dholbach leaves ;) 16:48:13 :) 16:48:27 Well timed. 16:48:43 I know I missed my last pilot session due to connectivity issues while I was at Connect. :/ 16:49:00 just to reiterate... we need to make sure patch piloting is happening, to keep contributions flowing in 16:49:09 our track record on this has not been great 16:49:43 if circumstances prevent you from piloting on your assigned day (because it overlaps with a conference for example), by all means swap days 16:50:04 doko: looks like you're scheduled to patch pilot tomorrow - is that going to work ok? 16:50:21 urgh 16:50:32 yeah, will try to start early 16:51:02 regarding patches I also wanted to mention https://bugs.launchpad.net/~foundations-bugs/+patches 16:53:01 bdmurray: do you know if that list is going up/down? 16:54:18 slangasek: I would guess up but don't have numbers right in front of me 16:54:31 that's a good list of bugs for folks to draw on as well when patch piloting, anyway - probably a lot of these (particularly the older ones) should have the patch tag dropped because they don't really fix the bug 16:54:49 maybe stgraber wants to review the patch in bug #20420 :) 16:54:51 Launchpad bug 20420 in netcfg (Ubuntu) "default to wired network interface if connected" [Wishlist,Confirmed] https://launchpad.net/bugs/20420 16:55:34 any other concerns on patch piloting? 16:55:49 [TOPIC] AOB 16:55:57 or other business? 16:56:24 #endmeeting