15:02:17 #startmeeting 15:02:17 Meeting started Wed Oct 24 15:02:17 2012 UTC. The chair is slangasek. Information about MeetBot at http://wiki.ubuntu.com/meetingology. 15:02:17 15:02:17 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 15:02:26 [TOPIC] Lightning round 15:03:14 $ echo $(shuf -e barry doko stgraber jodh ev bdmurray slangasek ogra infinity cjwatson xnox stokachu) 15:03:17 bdmurray xnox slangasek doko jodh stgraber stokachu barry cjwatson ogra ev infinity 15:03:50 bug review of update-manager / ubuntu-release-upgrader bug reports 15:03:50 analysis and testing of bug 346386 / bug 24061 15:03:50 wrote a new test case for bug 346386 15:03:50 wrote a test case for ubuntu-release-upgarder bug 1065806 15:03:50 quantal-proposed upload for ubuntu-release-upgrader bugs 1070043, 1065806 15:03:54 Launchpad bug 346386 in apt (Ubuntu Precise) "[MASTER] Update fails with invalid package files with "Encountered a section with no Package: header"" [High,Fix committed] https://launchpad.net/bugs/346386 15:03:56 Launchpad bug 24061 in apt (Ubuntu Quantal) "GPG error with apt-get/aptitude/update-manager behind proxy (BADSIG 40976EAF437D05B5)" [High,Fix released] https://launchpad.net/bugs/24061 15:03:57 Launchpad bug 1065806 in ubuntu-release-upgrader (Ubuntu) "diff window is too small on upgrade" [High,In progress] https://launchpad.net/bugs/1065806 15:03:58 Launchpad bug 1070043 in ubuntu-release-upgrader (Ubuntu) "DistUpgradeApport.py should check to see if --tags is available" [High,In progress] https://launchpad.net/bugs/1070043 15:04:04 quantal-proposed upload for apport bug 1039220 15:04:04 precise-proposed upload for apport-symptoms bug 1070542 15:04:04 sru verification of apt bug 346386, 1050779, 1050791 15:04:04 sru verification of apport bug 1050853 15:04:04 sru verification of coreutils bug 1038468 15:04:05 Launchpad bug 1039220 in apport (Ubuntu) "don't report crashes for programs that don't match the file on disk (like for kernel crashes)" [Medium,In progress] https://launchpad.net/bugs/1039220 15:04:21 Launchpad bug 1070542 in apport-symptoms (Ubuntu Precise) "not possible to use ubuntu-bug ubuntu-release-upgrader-core on precise" [High,Fix released] https://launchpad.net/bugs/1070542 15:04:21 * mvo hugs bdmurray 15:04:22 Launchpad bug 1050779 in apt (Ubuntu Precise) "excessive backup pkgcache.bin files retained" [Medium,Fix committed] https://launchpad.net/bugs/1050779 15:04:23 sru verification of update-notifier bug 946718 15:04:23 Launchpad bug 1050791 in apt (Ubuntu Precise) "do not run into loop on new-pre-depends-breaks" [Medium,Fix committed] https://launchpad.net/bugs/1050791 15:04:23 sru verification of ubuntu-release-upgrader bug 1068389 15:04:23 test case creation for apt bug 1065429 15:04:23 reported ubuntu-release-upgrader bugs 1067542, 1067393, 1068874 15:04:24 tested vmware player 5 installer in quantal 15:04:26 release noted bug 993056 15:04:28 setup rls-r-tracking reports on cranberry 15:04:29 Launchpad bug 1050853 in Whoopsie "Not recorded whether an error is in a -proposed or -updates package" [Undecided,New] https://launchpad.net/bugs/1050853 15:04:30 Launchpad bug 1038468 in coreutils (Debian) "data loss on sort -u" [Unknown,New] https://launchpad.net/bugs/1038468 15:04:31 Launchpad bug 946718 in update-notifier (Ubuntu Precise) "backend_helper.py crashed with RuntimeError in add_signal_receiver(): To make asynchronous calls, receive signals or export objects, D-Bus connections must be attached to a main loop by passing mainloop=... to the constructor or calling dbus.set_default_main_loop(...)" [High,Fix released] https://launchpad.net/bugs/946718 15:04:32 Launchpad bug 1068389 in ubuntu-release-upgrader (Ubuntu) "P->Q - do-release-upgrade crashed with UnicodeEncodeError: 'ascii' codec can't encode character u'\xbb' in position 1: ordinal not in range(128) in DistUpgrade/DistUpgradeViewText.py", line 143, in showInPager" [High,In progress] https://launchpad.net/bugs/1068389 15:04:33 Launchpad bug 1065429 in apt (Ubuntu Precise) "Long passwords for authenticated repositories not handled well" [Medium,Fix committed] https://launchpad.net/bugs/1065429 15:04:34 Launchpad bug 1067542 in ubuntu-release-upgrader (Ubuntu) "possible for some do-release-upgrade crashes to be unreportable" [High,New] https://launchpad.net/bugs/1067542 15:04:36 Launchpad bug 1068874 in ubuntu-release-upgrader (Ubuntu) "does not need to enable apport" [Medium,New] https://launchpad.net/bugs/1068874 15:04:37 Launchpad bug 1067393 in ubuntu-release-upgrader (Ubuntu) "using a local mirror and an official mirror of -backports does not work" [Medium,In progress] https://launchpad.net/bugs/1067393 15:04:37 Launchpad bug 993056 in activity-log-manager (Ubuntu Quantal) ""Privacy" > "Diagnostics" > "Send error reports" can't be turned on or off" [High,Triaged] https://launchpad.net/bugs/993056 15:05:04 ⌁ done ⌁ 15:05:12 * Quantal Released. Yeah! 15:05:12 * Did merges (see lp:~xnox/*/r-proposed, mostly partman-* and related 15:05:12 packages). Still some todo. 15:05:12 * Staging ubiquity bugfixes in lp:~xnox/ubiquity/r-proposed 15:05:12 * Helping with python3.3 transition at the archive opening 15:05:12 * Added ubiquity slide decks https://wiki.ubuntu.com/Ubiquity/SlideDecks to 15:05:14 unblock design team on changes to installer. 15:05:23 .. 15:05:40 * xnox also started py3.3 by default ppa 15:06:31 xnox: mm? I thought that was already in hand (barry) 15:06:46 he's at a conf today 15:07:14 yes 15:07:15 me and doko finished up rebuilding public modules yesterday. and we'd rather get it done earlier than later. 15:07:19 ok 15:07:27 * xnox has nasty python bugs for barry =) 15:08:35 slangasek: You're up. :P 15:08:37 * Secure Boot follow-through: getting support for efi variable writing, efifb quirking 15:08:40 * UDS prep 15:08:43 * debugging heat issues on my laptop in quantal 15:08:45 * preliminary look at fixing cross-bootstrapping issues in raring (http://people.linaro.org/~wookey/buildd/quantal/sbuild-ma/status.html) 15:08:48 (done) 15:09:58 - quantal: two last java ftbfs for the release 15:09:58 - gcc multiarching: libgfortran-dev split & cross fixes 15:09:58 - gcc x32: gcc bits are ready, blocked on eglibc for further testing 15:09:58 - gcc aarch64: integrated fixes 15:09:58 - gcc upstream: armhf-triplet patches upstreamed, fixed libgo ftbfs on trunk 15:09:59 - binutils update to 2.23 just to find out that the upstream release is incomplete 15:10:00 - raring opening: binutils & gcc bits in place, waiting for kernel headers 3.7 (not anymore for the opening), and eglibc ... 15:10:03 - raring opening: python3.3 is supported, most extension modules are built for 3.3, bug fixes and no change uploads. test rebuild can start now 15:10:08 - finished openjdk security updates (not yet published) 15:10:10 (done) 15:10:32 * misc: on holiday yesterday. 15:10:32 * upstart: 15:10:33 - Fixed bug 1049820 and bug 888910. 15:10:33 - Added new test for stateful re-exec. 15:10:35 Launchpad bug 1049820 in upstart "Using kill signal SIGPWR results in system crash" [Medium,Fix committed] https://launchpad.net/bugs/1049820 15:10:36 - Currently chasing down a test race issue. 15:10:37 Launchpad bug 888910 in upstart (Ubuntu) ""make check" fails in sbuild chroot" [Undecided,Confirmed] https://launchpad.net/bugs/888910 15:10:39 * TODO: 15:10:42 - get stateful-reexec into raring. 15:10:46 - blueprints and UDS planning. 15:10:50 ⃕ 15:11:06 - On vacation Friday, Monday and Tuesday 15:11:06 - Release 15:11:07 - Released Ubuntu 12.10 15:11:07 - Networking 15:11:07 - Went through the bug reports, prepared some uploads for when raring opens 15:11:09 - Container 15:11:12 - Reviewed a few branches, fixed a bunch of upstream bugs, next upstream daily build should be good 15:11:15 - Installer 15:11:17 - Did some tests on the preseed generator script. Mostly working, need to check if I can do something 15:11:20 about the remaining keys easily. Will be uploaded as part of installation-report in raring. 15:11:23 - UDS 15:11:26 - Submitted usual networking session 15:11:28 - Planned session on making dist upgrade a bit more reliable (especially lts-to-lts) 15:11:31 - Going through notes to figure out if there's something else than should be discussed at UDS 15:11:34 - TODO 15:11:37 - Some more UDS preps 15:11:39 - The usual merges 15:11:42 - Push all lxc fixes to a bugfix branch so 0.8 can finally be released 15:11:44 (DONE) 15:12:21 Nothing on fire at the moment, requested a backport for pbuilder-scripts in precise, still working on appmenu-gtk multiarch solution. done. 15:13:08 cjwatson: I think we're barryless. 15:13:17 Spent all week working on raring-proposed -> raring auto-migration, both the migration code itself ("britney") and the necessary Launchpad patches. I'm currently waiting for one LP database patch to be deployed, and then for an LP code patch to be reviewed, landed, and deployed; after that we should be able to turn this all on, open raring, and start auto-syncs. 15:13:22 .. 15:13:49 ogra_: 15:13:51 (Oh, for clarity, those patches are the ones that forcibly redirect all uploads to -proposed.) 15:14:14 done: 15:14:14 * internal project work, to be presented at UDS 15:14:14 * registered some specs 15:14:14 todo: 15:14:14 * go to UDS, flesh out the spec content 15:14:15 * test multiarch cross compiling 15:14:16 .. 15:14:34 - Spending most of my time on my UDS plenary. 15:14:34 - We discovered that our current calculation for average errors per calendar 15:14:35 day (number of errors / unique users in the past 90 days) is incorrect. We 15:14:35 have good reason to believe that it is what caused the large spike around 15:14:35 release in the 12.10 graph: 15:14:35 https://bugs.launchpad.net/errors/+bug/1069827 15:14:38 Launchpad bug 1069827 in Errors "Error rate incorrectly spikes with any influx of machines" [Undecided,New] 15:14:41 We've been working on a fix for this. I implemented an early solution, but 15:14:41 we found out again that is was wrong. We'll be taking this to canonical-tech 15:14:41 and http://stats.stackexchange.com. 15:14:44 - Fixed a number of small issues in http://errors.ubuntu.com. We now identify 15:14:44 private bugs rather than treating them as regressions. 15:14:44 - Fixed some issues in the juju deployment code for the entire whoopsie daisy 15:14:45 project. Discussions with juju people on how to best represent the 15:14:45 schema creation dependency that the retracers have on the main daisy 15:14:45 instance. 15:14:46 (done) 15:15:01 Last week: 15:15:01 - Released a new version of some operating system or other 15:15:01 - Worked on the opening of raring 15:15:01 - Several merges, bug fixes, and SRUs relating to the opening 15:15:01 - Worked on eglibc 2.16 4-way merge between Debian, Ubuntu, schepler, and wookey 15:15:03 - Discussed glibc 2.17 roadmap with upstream 15:15:05 - Meetings, meetings, and more meetings 15:15:08 - Spent some time cursing at bug 997359 (jodh: Can we get this fixed?) 15:15:09 Launchpad bug 997359 in libnih (Ubuntu) "nih uses eglibc private symbol __abort_msg" [High,Confirmed] https://launchpad.net/bugs/997359 15:15:10 This week: 15:15:13 - Finish up eglibc 2.16, including making future merges much less effort 15:15:16 - Revisit kernel cleanup, time permitting 15:15:19 - Do some pre-UDS prep and blueprinty things 15:15:21 - UDS 15:15:23 ☭ 15:15:45 finished? don't see it 15:15:56 infinity: I'm waiting for feedback on https://bugs.launchpad.net/libnih/+bug/997359/comments/2 :) 15:15:57 I think "this week" is a to-do 15:16:03 jodh: I realise there's a "rebootstrap nih/glibc every time" workaround for the above, but it would be lovely to just fix it. 15:16:22 infinity: do you expect 2.16 to land this week? 15:16:27 slangasek: Yeahp. 15:16:39 ok cool 15:16:48 You have until, uh, not quite sure if you want it before britney's all ready. 15:17:04 Probably Friday by this point (which is a bit unfortunate since I'm meant to be on holiday). 15:17:25 Might end up turning it on from the hotel :) 15:17:30 A pre-UDS one day rest? 15:17:52 Oh crap, it's almost the end of the year. I should take a holiday sometime too. 15:17:53 My family'll be with me so hacking time a bit limited. We'll see. 15:18:02 La la la. 15:18:31 infinity: APPROVED 15:18:34 * slangasek whistles and runs away 15:18:43 [TOPIC] Bugs 15:18:47 infinity, heh, me too 15:19:38 bdmurray: I guess we might be in a weird spot for this since raring has no new bugs and we're not putting as much effort into quantal SRUs... but any that you think we need to fix though? 15:19:52 jodh: Your maybe in comment #2 is a bit vague, but without digging into it further, I say "yes, that would be lovely". :P 15:20:37 Hmm.... we will have 12.04.2 on the New Years Eve. 15:20:43 slangasek: no, not yet. however I've been trying to verify the fix for bug 1066445 so that the apt sru which fixes bug 346386 can be released. 15:20:46 Launchpad bug 1066445 in apt (Debian) "apt-get crashed with SIGSEGV in pkgCacheGenerator::ListParser::NewProvides()" [Unknown,New] https://launchpad.net/bugs/1066445 15:20:47 Launchpad bug 346386 in apt (Ubuntu Precise) "[MASTER] Update fails with invalid package files with "Encountered a section with no Package: header"" [High,Fix committed] https://launchpad.net/bugs/346386 15:21:05 I've had no luck getting the failure in bug 1066445 though 15:21:13 xnox: 31 Jan, not 31 Dec :-) 15:21:21 that's the one that cjwatson suggested randomly permuting the cache size on, yes? 15:21:35 right, I've tried much smaller cache sizes 15:21:39 Yeah, it's probably a right pig to reproduce if you're unlucky 15:22:04 cjwatson: Oh good =) 15:22:07 Trying to reproduce on precise? 15:22:08 If it helps, you can use 'apt-cache gencaches' rather than 'apt-get update' (after the first time), which will be a lot faster since it doesn't download anything 15:22:11 would regression testing, dist-upgrades with the -proposed apt, be sufficient? 15:22:24 infinity: yes on precise 15:22:38 yeah, I'm kinda thinking we should just call it good with the regression test 15:23:00 I suspect regression testing may be all we can realistically get, in the absence of a deliberately malicious fuzzer tool 15:23:04 especially if we're pushing another apt SRU right behind it, the impact on users of having an extra SRU is minimal 15:23:12 Which arguably might be a good idea given the nature of this code 15:23:21 But it's not something we can knock together quickly 15:24:46 Unfortunately we don't have a /var/lib/apt/ tarball from that bug, otherwise reproduction would be a matter of dropping that in place and running 'apt-cache gencaches' 15:25:01 okay, I'll do one more test with apt from -proposed and then comment on the bug and tag it v-done 15:25:06 sounds good 15:25:22 Do make sure you're testing on amd64 15:26:03 slangasek: BTW, the reason it's more likely when upgrading to quantal than when upgrading within precise is bigger Packages files with more Multi-Arch entries 15:26:04 I am - is that mentioned in the bug at all? 15:26:06 * infinity just tried i386 and amd64 and failed to reproduce. 15:26:25 cjwatson: fun 15:26:30 bdmurray: No, but it only bit amd64 originally 15:26:50 I wonder if bug 1016040 might have indirectly fixed bug 1066445? 15:26:53 Launchpad bug 1016040 in apt (Ubuntu Quantal) "apt_check.py crashed with SIGSEGV in FileName()" [Medium,Confirmed] https://launchpad.net/bugs/1016040 15:26:54 Launchpad bug 1066445 in apt (Debian) "apt-get crashed with SIGSEGV in pkgCacheGenerator::ListParser::NewProvides()" [Unknown,New] https://launchpad.net/bugs/1066445 15:27:12 jodh: Not that way round 15:28:19 bdmurray: is that everything? 15:28:34 slangasek: that is all I have for today 15:28:39 bdmurray: ok, thanks 15:28:41 [TOPIC] UDS 15:28:50 blueprints 15:29:18 anything that you expect to have a session on at UDS, get your blueprint registered and proposed for uds-r by *tomorrow* 15:29:44 and *before* Monday, I want everyone to have gone through the schedule and subscribed themselves to the sessions they plan to attend 15:30:11 for any session that you would ask to have moved to work around a conflict, you should be marking yourself "participation essential" 15:30:53 I do not want you coming to me during the week and asking to move sessions because of a conflict - *tell the system* that it's a conflict and let the auto-scheduler do its job, please :) 15:31:11 jodh: So - I suppose it's possible, but I think it's relatively unlikely as my observations while debugging 1066445 were that it tended to crash immediately 15:31:36 (ok, if you only find out during the week that you have to be at two sessions at once, come to me all the same - but don't let this happen for things you should know about in advance) 15:31:48 clear? 15:32:04 slangasek, what if we bribe you with lots of beer for it ? 15:32:09 * infinity gets the impression that slangasek feels strongly about this topic. 15:32:10 dont work ? 15:32:23 :) 15:32:38 jodh: It seems unlikely since the crash was due to dereferencing a pointer into memory that had only *just* been deallocated - I don't think there's any way it could end up overwriting a stat buffer or similar 15:33:06 ogra_: my fee for moving sessions during the week is that I will look at you with an evil eye 15:33:21 this is a flat rate and not subject to negotiation 15:33:21 hahaha 15:33:26 k 15:33:35 also of note: 15:33:37 slangasek: Your evil eye tends to be accompanied with a friendly smile, it's rather confusing. 15:33:49 this time around we're trying something different in terms of session types 15:33:51 infinity, thats what makes it so scary 15:34:02 in addition to blueprint sessions, we can also have workshop sessions 15:34:13 (workshop/hacking sessions) 15:34:17 sweet ! 15:34:26 and also, presentation sessions 15:34:28 SO 15:34:30 * xnox thinks the laser strength shooting from the evil eye will vary in kW 15:34:36 * ogra_ fells like rolled back 6 years 15:34:51 *before* we started scheduling everything madly 15:35:00 for the first one, if you're being asked for blueprints to be added to UDS for the umpteenth time, and are thinking "we don't need another session on that, we just need to sit down and DO it" 15:35:29 - then there's a possibility of doing just that. Let me know and we'll see if scheduling a workshop session instead of a blueprint session makes sense. 15:35:46 Is there a way to represent that in the LP UI? 15:35:59 Like, a blueprintless blueprint? :P 15:36:09 Or do we need to ask someone (like you) to schedule a workshoppy thing? 15:36:12 informational ? 15:36:25 ogra_: Informational would seem to map better to the presentation sessions. 15:36:32 indeed 15:36:41 and for the second one, if there are topics that you think it would be useful to share with the larger Ubuntu community - like something you'd really like to give as a plenary but there wasn't a slot for - feel free to propose a presentation session 15:36:41 but could be abused for it 15:36:48 slangasek: is summit actually still reading the participation essential bit from LP? the UI suggests that it only uses LP to check if you want to attend and that you need to poke the track lead to be marked as essential (the UI says "If you feel that your participation is required for this topic, please contact Steve Langasek..."). Not that I have any conflict at the moment, just found it a bit confusing... 15:36:49 infinity: you ask a track lead to add it 15:36:55 slangasek: Check. 15:37:30 stgraber: I'm not sure; the summit UI is showing ALL subscribers to ALL sessions with a star. I'll look into that today and get back to you 15:37:51 We could nick higlight cjohnston intentionally for once, and ask him. 15:38:13 stgraber: http://summit.ubuntu.com/uds-r/attendee_review/21047/foundations-r-prior-release-feedback/ appears to reflect the blueprint P.E. status 15:38:22 Though he's probably so fed up with people actually wanting cjwatson that he no longer responds. 15:38:36 It happens both ways round ... 15:39:12 slangasek: ok. I also got one case where I had to manually subscribe on summit as even though I was subscribed on LP it wouldn't show it on summit. But as I said, no conflict so far, so I'll start pinging people if a conflict shows up... 15:39:16 stgraber: so there appears to be an additional flippy bit that I have access to, to make people "required" as opposed to "very interested in attending" 15:39:41 but I'm hoping they aren't making me set that for every single person who's required to be in a session, because that would be horrid and non-scalable 15:40:29 [TOPIC] AOB 15:40:42 any other topics burning a hole in the pockets of your minds? 15:40:44 slangasek: Ugh. So, essential isn't actually essential? 15:41:02 "Please mark me super-duper-essential"? 15:41:15 infinity: I don't know. There's a perception on the part of other track leads that the launchpad field has been abused by people who aren't essential, which has never been my experience 15:41:33 so it may be that the scheduler has two different levels now 15:41:46 "try to schedule around this person" vs "require this person" 15:41:53 so do we open the archive in any case tomorrow? 15:42:08 regardless of any missing stuff? 15:42:13 I'll try to get clarification. But needless to say, I'm not going to go through by hand and edit status fields one by one for every one of you on summit. :P 15:42:31 doko: I don't think the "force everything into -proposed" stuff is going to have quite landed by then 15:42:36 slangasek: But that sounds like oodles of fun. OODLES. 15:42:52 And we're so close by now, and that will make so much difference, that I'm not willing to open without that at this point 15:43:01 ahh, ok. 15:43:28 My database patch is second in the queue, so that won't land until at least tomorrow morning (if I'm lucky) 15:43:42 cjwatson, do you expect to happen his during UDS? 15:43:44 My code patch hasn't yet been completely reviewed, although I've had some comments 15:44:25 Then it needs to go through EC2 (~4 hours, strictly after the DB patch lands), buildbot (~40 mins), QA (not too arduous), and be deployed 15:44:56 Subject to review, it should be possible to get it all done before UDS 15:44:58 cjwatson: ndt or fdt? 15:45:06 infinity: fdt for the db patch, ndt for code 15:45:33 [fastdowntime / nodowntime] 15:45:35 Yeah, Friday sounds reasonable then. 15:45:36 ok, I think, I'll send an email about when to expect the opening. will send you the draft 15:46:37 I thought I kind of covered that with my -devel mail earlier 15:46:43 But as you like 15:48:50 ok - anything else? 15:50:20 #endmeeting