16:02:23 <slangasek> #startmeeting
16:02:23 <meetingology> Meeting started Wed Nov  9 16:02:23 2011 UTC.  The chair is slangasek. Information about MeetBot at http://wiki.ubuntu.com/AlanBell/mootbot.
16:02:23 <meetingology> 
16:02:23 <meetingology> 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
16:02:51 <slangasek> $ echo $(shuf -e cjwatson barry doko stgraber jhunt ev bdmurray slangasek)
16:02:54 <slangasek> bdmurray stgraber barry jhunt ev cjwatson doko slangasek
16:02:58 <slangasek> ah, topic first :P
16:03:04 <slangasek> [TOPIC] lightning round
16:03:06 <slangasek> and go :)
16:03:22 <ev> whoops, missed that.  Who am I after?
16:03:30 <cjwatson> ev: jhunt
16:03:34 <ev> cool, thanks
16:03:42 <bdmurray> Attended UDS P in Orlando Florida
16:03:43 <bdmurray> updating blueprints with notes from UDS Precise
16:03:43 <bdmurray> investigation into bug 882255 regarding passwordless login and changing of password
16:03:45 <ubottu> Launchpad bug 882255 in gnome-control-center (Ubuntu) "No administrative actions possible (password refused) after enabling passwordless login" [High,Triaged] https://launchpad.net/bugs/882255
16:03:45 <bdmurray> added lp_work_item_editor to firefox-lp-improvements
16:03:54 <bdmurray> uploaded new version of firefox-lp-improvements
16:03:54 <bdmurray> review of Oneiric nominations for foundations packages
16:03:54 <bdmurray> update bug bot to deal with bugs with gzipped attachments better
16:03:54 <bdmurray> wrote bug bot function to comment on no package bug reports
16:04:04 <bdmurray> done
16:04:19 <stgraber> - UDS last week
16:04:19 <stgraber> - Looked a bit at bug 876829 (few patches around, none really fixing the issue) and bug 823366 (need to do some testing)
16:04:23 <stgraber> - Testing tracker work
16:04:23 <ubottu> Launchpad bug 876829 in ifupdown (Ubuntu Precise) "Oneiric's ifupdown breaks ip aliases" [High,Triaged] https://launchpad.net/bugs/876829
16:04:24 <ubottu> Launchpad bug 823366 in ifenslave-2.6 (Ubuntu) "bond_primary is ignored in /etc/network/interfaces" [Undecided,Incomplete] https://launchpad.net/bugs/823366
16:04:25 <stgraber> - http://91.189.93.73 is usually up to date with what's in the branch
16:04:28 <stgraber> - Secured the XML-RPC API (authentication using a login on the website + a key in the user profile)
16:04:31 <stgraber> - Initial work on notifications
16:04:33 <stgraber> - Changed theme to gain more screen space
16:04:36 <stgraber> - Merged initial report from Jean-Baptiste
16:04:38 <stgraber> - Update schema and API to support multiple result per user with link to hardware profile and a note with each build (rebuild reason)
16:04:41 <stgraber> - Improved migration/upgrade scripts
16:04:44 <stgraber> - Played a bit trying to get upstart:i386 to install on my laptop, uploaded most of its dependencies to my PPA but still won't work. Will poke slangasek :)
16:04:47 <stgraber> - TODO
16:04:49 <stgraber> - Get a test environment for bug 823366
16:04:52 <stgraber> - Get a test environment for iscsi, then look at the open-iscsi merge (hopefully fixing a few bugs in the process)
16:04:55 <stgraber> - Continue testing tracker work, now working on finishing the user UI and notifications
16:04:58 <stgraber> - Make sure all my blueprints are up to date on LP (should be) and cleanup the work items list a bit
16:05:01 <stgraber> (done)
16:07:25 <barry> since uds-p: merged claws, claws-plugins patches from official PPA; bug 884890; working on blueprints; start investigating python 3 for update manager and python-dbus; done.
16:07:25 <barry> 
16:07:26 <ubottu> Launchpad bug 884890 in python-tz (Ubuntu) "Failure to instantiate timezones with well-known names in 12.04" [High,Fix released] https://launchpad.net/bugs/884890
16:08:48 <slangasek> jhunt: you're up
16:08:50 <jhunt> UDS last week (still recovering!). This week have worked on updating the
16:08:50 <jhunt> cookbook and going the bugs (had some significant contributions from the
16:08:50 <jhunt> community which is great!!) Worked on a fix for bug 829980 - currently
16:08:50 <jhunt> awaiting feedback on my ppa build from users with "dodgy" battery
16:08:50 <jhunt> hardware. Plan then to merge pad data -> blueprints, and finish rework
16:08:52 <jhunt> on job logging.
16:08:53 <ubottu> Launchpad bug 829980 in upstart (Ubuntu) "upstart-udev-bridge eats 100% cpu calling dbus_message_iter_append_basic()" [Medium,In progress] https://launchpad.net/bugs/829980
16:08:55 <jhunt> ⟌
16:09:00 <jhunt> soz - didn't see the end marker :)
16:09:45 <ev> - Getting back up to speed after UDS.  Filed expenses, filled out the survey.
16:09:45 <ev> - Discussion with Robert around the replacement Launchpad OOPS system he's
16:09:46 <ev> building, using Cassandra as a backend.  Set up a call with him to brain
16:09:46 <ev> dump the results of our UDS discussions on the crash DB for Thursday night.
16:09:46 <ev> - Helping someone from the XMBC team get ubiquity working on their stripped
16:09:47 <ev> down live CD.
16:09:47 <ev> - Blueprints and workitems drafting.
16:09:47 <ev> - Trying to get an apport-retracer instance up and running in the Canonicloud.
16:09:48 <ev> - Looking at AMQP vs Cassandra's asynchronous triggers as a way of triggering
16:09:48 <ev> retraces of submitted crashes.
16:09:48 <ev> (done)
16:10:01 <slangasek> jhunt: understandable, since he cheesed out and didn't use unicode :)
16:10:13 <jhunt> tsk tsk!
16:10:48 <slangasek> ev: what's XMBC expand to?
16:10:56 <ev> xbox media center
16:11:00 <ev> XMBC*
16:11:07 <ev> err XBMC
16:11:11 <slangasek> ah, ok
16:11:15 <ev> apparently I just cannot type that acronym
16:11:43 <cjwatson> Getting the +1 maintenance team going: this month it's me, mterry, and cyphermox once he gets back from vacation.
16:11:46 <cjwatson> Lots of build fixes and transitions and the like.  Broke cmake/armel (not my fault!), analysed and reported the compiler bug that broke it, and worked around it.  Broke yelp and fixed it.
16:11:51 <cjwatson> Patch pilot on Tuesday morning.
16:11:53 <cjwatson> Fixed casper bug 882308 by request of OEM^WPES.
16:11:56 <ubottu> Launchpad bug 882308 in casper (Ubuntu Oneiric) "Choosing simplified chinese in isolinux menu causes corrupted characters and untranslated installer" [High,In progress] https://launchpad.net/bugs/882308
16:11:56 <cjwatson> Experimenting with multiarch cross-building.  I at least seem to have a usable chroot now and have fixed a few packages as a result.
16:11:59 <cjwatson> We almost have the new stonking cdimage master machine!  Working with IS on getting it set up properly, and doing some preparatory work on antimony.
16:12:03 <cjwatson> Drafted foundations-p-image-build-pipeline.  Three or so more to do.
16:12:05 <cjwatson> --
16:13:09 <slangasek> ah, and doko's off this week, so it's just me
16:14:41 <slangasek> halfway through my own blueprint drafting
16:14:41 <slangasek> upgraded to precise this week, have started reporting bugs
16:14:41 <slangasek> filed bugs for the various unity2d inconveniences I hadn't gotten around to reporting before
16:14:41 <slangasek> merged console-setup, will try to forward some changes upstream later to reduce our delta
16:14:41 <slangasek> merged user-setup
16:14:42 <slangasek> pushed a patch to sbuild so that it understands multiarch build-dep syntax
16:14:42 <slangasek> worked on checking the multiarch cross-buildability of a few packages in the base system
16:14:43 <slangasek> got bored with that and decided to go straight for qt4-x11 instead; pushed a bunch of M-A patches
16:14:43 <slangasek> (done)
16:15:20 <slangasek> anyone have questions?
16:16:28 <slangasek> [TOPIC] Bugs
16:16:44 <slangasek> bdmurray: anything of interest from your side?
16:17:11 <bdmurray> Only one bug 834516 which you've touched recently
16:17:13 <ubottu> Launchpad bug 834516 in user-setup (Ubuntu) "Encryped home directory selected but not encrypted" [High,New] https://launchpad.net/bugs/834516
16:18:05 <slangasek> cjwatson, ev, stgraber: have you seen any other reports of that issue?
16:19:09 <stgraber> not that I can remember.
16:20:21 <cjwatson> nor I
16:20:35 <bdmurray> I don't see "Cannot get ecryptfs version" in any ubiquity bug report log files
16:20:51 <slangasek> huh
16:21:03 <cjwatson> although it rings a bell, I think I remember fixing something similar
16:21:22 <bdmurray> InstallationMedia: Ubuntu 11.10 "Oneiric Ocelot" - Alpha amd64 (20110817)
16:21:40 <slangasek> oh, it was installed from an alpha?  should we close this out as fixed?
16:22:18 <bdmurray> Well should ubiquity have stopped at that point?
16:22:22 <cjwatson> hm, I was thinking of bug 820460 which was different
16:22:25 <ubottu> Launchpad bug 820460 in user-setup (Ubuntu Oneiric) "User not created with encrypted home partition: user-setup exit code 32" [Critical,Fix released] https://launchpad.net/bugs/820460
16:22:37 <jrgifford> there was this bug, which is the exact opposite. - bug #840570
16:22:38 <ubottu> Launchpad bug 840570 in ubiquity (Ubuntu) "Forced to encrypt my home directory during Oneric install" [Undecided,New] https://launchpad.net/bugs/840570
16:23:43 <slangasek> heh
16:24:23 <slangasek> stgraber: do you think you could follow up on this and see what the source of that error message is?  I'll bet it was a problem with the actual kernel during the alpha and should be closed
16:24:49 <bdmurray> I mean it seems wrong to me for the installation to have proceeded even though ecryptfs wasn't available.
16:24:56 <stgraber> slangasek: yep, I can do that
16:24:57 <cjwatson> hmm
16:25:13 <cjwatson> we || true the adduser call I think because it's possible that the user might already exist in some contenxts
16:25:16 <cjwatson> *contexts
16:25:23 <slangasek> bdmurray: ah, true
16:25:40 <cjwatson> I agree that we shouldn't proceed if it's clearly broken; we just need to be careful not to go too far the other way
16:25:46 <slangasek> so we shouldn't just close the bug anyway, we ought to communicate the missing encryption to the user
16:25:51 <slangasek> stgraber: still yours if you want it :)
16:26:47 <cjwatson> (that || true is very old; I think it may date back to base-config)
16:27:02 <cjwatson> it's in r1 of user-setup
16:27:49 <stgraber> slangasek: I can certainly add that to my low priority list :) If we skip the user creation, we should at least show a warning I'd think
16:29:09 <slangasek> stgraber: I think it's a high urgency to confirm that the underlying issue is fixed in the final release, and low urgency (but not low priority) to not silently skip encryption
16:30:07 <stgraber> slangasek: sounds good
16:30:25 <slangasek> [LINK] http://reports.qa.ubuntu.com/reports/kernel-bugs/reports/rls-p-tracking-bugs.html
16:30:41 <slangasek> I think this link has been passed around; hopefully everyone has seen it
16:31:04 <ev> fancy pant
16:31:04 <ev> s
16:31:28 <slangasek> this is the list of bugs that I've committed you to fix for the release so far :P
16:32:15 <slangasek> there are 33 tasks on there for foundations... which is high, but I don't see any bugs on that list that I think are ok for us to ship with
16:32:32 <slangasek> if you disagree you can argue it :), but I'd much rather we focus on getting them fixed
16:33:04 <slangasek> since this is an LTS and we're emphasizing quality, the list reflects that
16:33:10 <barry> slangasek: is there a difference between "Canonical Foundations Team" and "Ubuntu Foundations Team"?
16:33:44 <slangasek> barry: two different teams in launchpad, the latter includes some community folks
16:33:55 * barry nods
16:34:45 <slangasek> so the challenge I'd like to set for the team is to get 20 of those bugs fixed by the end of the year
16:35:00 <slangasek> no reason to wait until after FF to start fixing them :)
16:35:09 <slangasek> what do you think?
16:36:08 <cjwatson> it can't hurt to try; do you have the current set recorded somewhere so that tracking won't be thrown off by additions?
16:36:23 <slangasek> not yet, but I'll snapshot it today
16:36:56 <slangasek> there are a lot of installer bugs on there to go around, fwiw
16:37:17 <slangasek> so anyone who isn't already familiar with the installer and would like to learn more, now's your chance :)
16:37:36 <jhunt> those two plymouth bugs are the same issue. Would it make sense to make one a dup of another?
16:38:24 <jhunt> only reason I say that being some folks have expressed the view that they don't get notified of changes to the bug their bug is a dup of.
16:38:49 <slangasek> I don't remember why they wound up split
16:39:17 <slangasek> I think it was because the issue was rereported (849414) in a later release than when we thought we'd fixed it
16:39:30 <slangasek> jhunt: if you merge them, please make sure the bug description of the master says something sensible :)
16:39:42 <jhunt> slangasek: righto.
16:39:47 <slangasek> (e.g., reproduce instructions)
16:40:14 <slangasek> cjwatson: that reminds me... is there anything other than ^C that would generate SIGINT on console?
16:40:31 <cjwatson> some random keycode or other with the console in raw mode
16:40:41 <slangasek> heh
16:41:18 <slangasek> just trying to figure out if the bodgy SIGINT handler in the plymouth theme really does explain all the problems we've seen, or if as upstream says there's some other issue
16:41:20 <cjwatson> much like, er is it Enter?, turns into SIGQUIT
16:41:21 <jhunt> slangasek: I've got a systemtap script that logs all SIGINTs system wide if anyone has a system that actually exhibits the plymouth failure on those bugs.
16:41:37 <slangasek> jhunt: \o/
16:42:04 <slangasek> jhunt: we should probably take upstream's advice and nix the SIGINT handler anyway, and then see if the problem goes away
16:42:10 <jhunt> Taking a closer look at the sigint handler, it *does* look safe which is even more mysterious. The only way I've managed to recreate the issue is to fire SIGINTs at plymouthd semi-randomly.
16:42:41 <jhunt> right. All the plugins have the same code seemingly (copy'n'paste.. ? :)
16:42:50 <slangasek> :)
16:43:36 <jhunt> I'm just not sure what the implications of removing that handler are for the plugins - it's not as if the handlers aren't doing anything.
16:43:38 <cjwatson> I don't agree that all of them are safe
16:44:02 <cjwatson> looking at the script handler as an arbitrary example, it frees memory in a signal handler, which is unsafe
16:44:05 <slangasek> the implication is that ^C does nothing, which seems reasonable anyway
16:44:15 <slangasek> and consistent with the other themes
16:44:51 <jhunt> are you sure? I was thrown off track initially as the handlers themselves do dangerous things like modifying linked lists. *But* the handlers are not called when the signal arrives - they are called on the next main loop iteration.
16:44:59 <cjwatson> oh, except that the signal handler doesn't work that way
16:45:06 <cjwatson> you're right, I missed the indirection through the event loop
16:45:07 <jhunt> so, in principle it *should* be perfectly safe.
16:45:20 <jhunt> yeah - it's a mare of code layers.
16:45:47 <cjwatson> well
16:45:51 <jhunt> all the handlers do is write the signal number through a pipe and return.
16:45:52 <cjwatson> ply_write calls assert
16:46:09 <cjwatson> assert is not async-signal-safe
16:46:16 <jhunt> oooh!
16:46:16 <cjwatson> although that should only actually matter if the assert fails, I think
16:46:29 <cjwatson> and if that happened then it would fall over anyway
16:46:29 <slangasek> well, we do have failing asserts in some of the reports
16:46:52 <jhunt> frankly, I wish there were more asserts ther. No alloc calls are ever checked if I recall.
16:47:17 <jhunt> however, they would segv fast I'd hope as the address is immediately dereferenced
16:47:30 * slangasek nods
16:47:34 <jhunt> hope not being a word I like to use in combo with programming... :)
16:47:39 <slangasek> ok, we can chew on that more offline :)
16:47:42 <slangasek> [TOPIC] AOB
16:47:44 <jhunt> right
16:47:45 <slangasek> anything else?
16:47:54 * cjwatson tries to find the list of async-signal-safe functions; I keep losing it
16:48:08 <jhunt> can anyone offer any thoughts on bug 861268?
16:48:09 <ubottu> Launchpad bug 861268 in nvidia-graphics-drivers (Ubuntu) "text corruption in terminals (xterm, urxvt) and emacs" [Undecided,Incomplete] https://launchpad.net/bugs/861268
16:48:36 <slangasek> jhunt: did you not manage to pin down the X guys at UDS?
16:48:45 <slangasek> or did they tell you "oh yeah, binary driver, not our problem"? :)
16:49:02 <cjwatson> oh, there we go, it's buried in http://pubs.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html#tag_15_04
16:49:06 <bdmurray> Looking at recent ubiquity bugs some seem to be about Natty.  I wonder if people are installing it thinking it is the latest.  Or if commenting on them and pointing them to 11.10 makes sense.
16:49:17 <jhunt> slangasek: semi - I spoke to Chris who thought it could relate to the nvidia driver, but the nouveau driver behaves in the same way.
16:49:22 <cjwatson> send() and write() are async-signal-safe, so I've got nothing else that would break in that way
16:49:47 <ev> bdmurray: please point them at 11.10
16:49:53 <slangasek> +1
16:49:54 <ev> we don't fix bugs in non-LTS releases post-release
16:49:58 <jhunt> cjwatson: I've got some tools and scripts that we can use to try and pin this down. Might be worth a bit of pair-debugging?
16:50:03 <ev> for ubiquity, that is
16:50:11 <cjwatson> jhunt: yep, not that I can reproduce it myself
16:50:14 <bdmurray> well so if the same bug happens in 11.10 they'll be at the same place though
16:50:17 <cjwatson> ev: well, except for when we do, but yeah :)
16:50:32 <cjwatson> I think we'll want to do more post-release fixing given the existence and advertisement of live-build
16:50:54 <cjwatson> but sure, it's good to make sure that the bugs are current ...
16:50:57 <ev> I think testing the latest release is a reasonable first pass at triaging
16:51:00 <ev> indeed
16:51:49 <jhunt> cjwatson: signal(7) also lists async safe calls.
16:52:07 <ev> tmux is made of awesome
16:52:30 <cjwatson> jhunt: ah yes, thanks
16:52:42 <cjwatson> I always look for it in the base-definitions section of POSIX and it's not there
16:52:59 <jhunt> ev: +100
16:53:39 <slangasek> #endmeeting