16:04:58 <slangasek> #startmeeting
16:04:58 <meetingology> Meeting started Wed Feb 15 16:04:58 2012 UTC.  The chair is slangasek. Information about MeetBot at http://wiki.ubuntu.com/meetingology.
16:04:58 <meetingology> 
16:04:58 <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
16:05:06 <slangasek> [TOPIC] Lightning round
16:05:12 <slangasek> ev: sorry ;)
16:05:15 <slangasek> $ echo $(shuf -e barry bdmurray cjwatson doko ev jhunt_ stgraber slangasek)
16:05:18 <slangasek> ev doko jhunt_ barry stgraber slangasek bdmurray cjwatson
16:05:22 <stgraber> ;)
16:05:37 <slangasek> doko: do you want to go ahead while ev is tea-ing?
16:07:13 <doko> ok
16:07:38 <doko> - LinaroConnect last week, cumbersome way back home
16:08:20 <doko> - forwarding GCC 4.7 upstream issues, another 4.6 test build (will hit precise this week)
16:08:38 <doko> - started openjdk-6 security updates
16:08:58 <doko> - eglibc fixes, thanks to cjwatson
16:09:01 <doko> done
16:09:40 <jhunt> Finished changes for bug 926468. Reworked early job logging code and
16:09:40 <jhunt> combined both into upstart-1.4-0ubuntu7~jh in my
16:09:40 <jhunt> https://launchpad.net/~jamesodhunt/+archive/upstart-job-logging PPA for
16:09:40 <jhunt> testing prior to FF (if you're not running it, please install and !)
16:09:43 <ubottu> Launchpad bug 926468 in upstart (Ubuntu Precise) "Stopping ssh with a logged in user causes init to spin at 100%" [High,In progress] https://launchpad.net/bugs/926468
16:09:43 <jhunt> Investigating test failures on PPC+ARM*. Spent a little time looking at
16:09:48 <jhunt> bug 931220/931584 with stgraber.
16:09:49 <ubottu> Launchpad bug 931220 in lxc (Ubuntu) "lxc-start should get a close-all-fds option to avoid weird issues" [High,Fix released] https://launchpad.net/bugs/931220
16:09:51 <jhunt>16:09:57 <jhunt> ... and provide feedback that is :)
16:10:16 <ev> back
16:10:23 <stgraber> ev: go ahead
16:10:24 <ev> apols for that
16:10:26 <ev> okay, cheers
16:10:44 <ev> - Lots and lots of work on the new Apport UI:
16:10:44 <ev> - More work on matching the KDE frontend to the GTK changes.
16:10:45 <ev> - Some bug fixes.
16:10:45 <ev> - Handle errors better in the information collection thread.
16:10:45 <ev> - Fix some issues around displaying multiple reports in one run.
16:10:46 <ev> - Unit test the GTK+ frontend, complete with a pair of bug fixes.
16:10:46 <ev> - Add KDE frontend unit tests, along with some more bug fixes.
16:10:46 <ev> - Remove old GTK+ dialogs.
16:10:47 <ev> - Put graphical tests under xvfb. Only use it if it's present.
16:10:47 <ev> - Don't hardcode Ubuntu.
16:10:55 <ev> - Move UI tests out of their respective modules, per pitti's request.
16:10:55 <ev> - Bring apport-cli into the new threaded way of collection problem
16:10:56 <ev> information. Update its calls to the UI backend.
16:10:57 <ev> - Merge review with Martin in:
16:10:57 <ev> https://code.launchpad.net/~ev/apport/whoopsie/+merge/92753
16:11:10 <ev> - Lots and lots of work around whoopsie's security review:
16:11:11 <ev> - Don't fail if there are no crash files (LP: #928735).
16:11:11 <ev> - Do not attempt to load the control center panel UI from the build
16:11:11 <ev> directory.
16:11:12 <ev> - Check the return value of the open call in get_system_uuid.
16:11:12 <ev> - Properly initialize libcrypt.
16:11:12 <ev> - Check that the call to gcry_md_open succeeds.
16:11:12 <ev> - Ensure that reading the SHA512 message digest succeeds.
16:11:13 <ev> - Protect against changes to the message digest length creating a security
16:11:13 <ev> vulnerability.
16:11:28 <ev> - Check the returncode of setenv.
16:11:28 <ev> - Use /var/lock/whoopsie instead of /tmp/.whoopsie-lock.
16:11:29 <ev> - Future proof by using getrlimit instead of explicitly closing STD*.
16:11:29 <ev> - Redirect stdin, stdout, and stderr to /dev/null.
16:11:29 <ev> - Ensure strings created in update_to_crash_file are NUL-terminated.
16:11:29 <ev> - Only process regular files in /var/crash.
16:11:30 <ev> - Replace calls to *alloc with g_*alloc, which calls abort on failure.
16:11:30 <ev> - Remove unused system_uuid pointer references.
16:11:43 <ev> - Recover from being unable to get the system UUID.
16:11:44 <ev> - Fix warnings in make check.
16:11:44 <ev> - Initialize all of curl.
16:11:44 <ev> - Redirect stderr to null in chgrp and chmod calls.
16:11:45 <ev> - Set home directory to /nonexistent.
16:11:45 <ev> - Enable libcrypt secure memory.
16:11:57 <ev> - Put the lock file in /var/lock/whoopsie/.
16:11:58 <ev> - Sanity check the CRASH_DB_URL environment variable.
16:11:58 <ev> - Add some more tests for parse_report, along with data from Jamie
16:11:58 <ev> Strandboge.
16:11:58 <ev> - Check handling of embedded NUL bytes.
16:11:59 <ev> - Verify that symlinks in /var/crash produce the correct error.
16:11:59 <ev> - Verify that keys without values in reports produce an error message.
16:11:59 <ev> - Ensure that the report does not start with a value.
16:12:07 <ev> - Correctly identify a report without spaces as malformed.
16:12:07 <ev> - Verify that directories in /var/crash produce the correct error.
16:12:07 <ev> - Ensure that blank lines in a report are treated as errors.
16:12:08 <ev> - Change the crash database URL to http://daisy.ubuntu.com :)
16:12:08 <ev> - Ensure that carriage returns are escaped.
16:12:08 <ev> - Do not start multiline values with a newline.
16:12:14 <ev> - Check that a valid report has the exact expected contents.
16:12:14 <ev> - MIR/security review with Jamie in:
16:12:15 <ev> https://bugs.launchpad.net/ubuntu/+source/whoopsie-daisy/+bug/913694
16:12:15 <ev> - Slideshow planning meeting with Christian and Dylan.
16:12:15 <ev> - Meeting to understand Canonical UK pension scheme.
16:12:16 <ubottu> Launchpad bug 913694 in whoopsie-daisy (Ubuntu) "[MIR] whoopsie-daisy" [Undecided,In progress]
16:12:16 <ev> (done)
16:12:56 <ev> I will need a freeze exception, but I'm buried in interview prep work, so I'll sort that later tonight
16:13:01 <stgraber> - Networking
16:13:01 <stgraber> - Waiting on SRU test results for ifenslave/vlan/bridge, I'm told I should have these later today
16:13:04 <stgraber> - Got daily IPv6 testing ported to regular bridge-utils so it can run within a container (yeah, container in containers)
16:13:07 <stgraber> - Arkose
16:13:10 <stgraber> - Updated the Arkose cli tool to show all the new features, did some minor bugfixing of the python module, updated translations and released 1.5
16:13:13 <stgraber> - Friendly-Recovery
16:13:15 <stgraber> - Updated menu landed, all options are always displayed now and switch from read-only to read/write is done when selecting an option requiring it
16:13:19 <stgraber> - Still working on a 'system information' option, should land later today or early tomorrow
16:13:21 <stgraber> - LTSP
16:13:24 <stgraber> - Improve ltsp-live for Edubuntu
16:13:26 <stgraber> - Lots of changes happening upstream to get the new shiny stuff ready by Feature Freeze. Testers are still busy and I expect an upload very shortly before FF.
16:13:30 <stgraber> - Merged nbd from unstable to get some fixes needed for LTSP.
16:13:32 <stgraber> - I've also started cleaning up the maintainer scripts and will spend some time to deal with our current installer bugs (ltsp-client-builder udeb)
16:13:35 <stgraber> - Upstart
16:13:38 <stgraber> - Tried to figure out why some jobs sometimes inherit an extra fd (a socket in this case), tracked down to something around libldap/libnss-ldap/libnss affecting
16:13:41 <stgraber> machines using libnss-ldap. This was breaking LXC's upstart job, now fixed in LXC by having lxc-start close the extra fds.
16:13:44 <stgraber> - Containers
16:13:46 <stgraber> - Uploaded a new LXC with reworked template and Serge's patch to have our consoles and ttys in /dev/lxc, fixing upgrade/dist-upgrade in LXC for good. (bug 927519))
16:13:47 <ubottu> Launchpad bug 927519 in lxc (Ubuntu) "makedev fails to upgrade in lxc container with error 'Device or resource busy'" [Medium,Fix released] https://launchpad.net/bugs/927519
16:13:49 <stgraber> - Updated shadow to add the new LXC consoles and ttys to securetty
16:13:52 <stgraber> - Added a network-interface-container upstart job to ifupdown to simulate net-device-added for 'lo' in containers
16:13:55 <stgraber> - Disable udevadm trigger --action=add in containers (bug 925122) - commited to the packaging branch, waiting on upstart to build on armel/armhf/powerpc.
16:13:56 <ubottu> Launchpad bug 925122 in udev (Ubuntu) "container's udevadm trigger --add affects the host" [Medium,Fix committed] https://launchpad.net/bugs/925122
16:13:58 <stgraber> - Other
16:14:00 <stgraber> - Helped some friends of the LTTng project, sponsored the new LTTng2 stack (now in NEW) and filed the needed package removal bugs for the old broken stack
16:14:03 <stgraber> - Will release pastebinit 1.3 later today, waiting for some extra translations at the moment
16:14:06 <stgraber> (done)
16:15:21 * slangasek runs out of breath reading ev's report
16:15:50 <ev> hah
16:16:15 <jhunt> ENOSPC over here! :)
16:17:35 <slangasek> stgraber: huh, why does lxc need different names for its ttys?
16:18:34 <stgraber> slangasek: lxc was bind mounting its ttys from outside the container to /dev/tty[1234] which is all fine, except when you deal with anything like udev/makedev/... which tries to move them, remove them, ...
16:19:04 <stgraber> slangasek: so the easiest solution to fix it until we get a proper device namespace and can stop with the bind mounts was to use another location
16:19:14 <slangasek> hmm, ok
16:19:29 <slangasek> right, my turn
16:19:33 <slangasek> * short week, was off Monday
16:19:34 <slangasek> * interviewing
16:19:34 <slangasek> * helping test/merge upstart fixes to be landed in precise today
16:19:34 <slangasek> * resolvconf finagling
16:19:34 <slangasek> * discussions about how to minimize future breakage from eglibc uploads
16:19:36 <slangasek> * libnih1 uploaded for essential-ish pre-dependency issues
16:19:38 <slangasek> * working with IS to get qemu-linaro situated to where it can be used as an emulated armel buildd running on top of a hardy VM
16:19:41 <slangasek> * minor eglibc work
16:19:44 <slangasek> * virtual sprint planning
16:19:47 <slangasek> ..
16:19:49 <slangasek> questions?
16:20:22 <bdmurray> review of bug pattern merge proposals from Vadim Rutkovsky
16:20:22 <bdmurray> fixing of SRU for bug 873424 after discussion with verifier
16:20:23 <ubottu> Launchpad bug 873424 in update-manager (Ubuntu Oneiric) "ask me later fails" [Medium,Fix committed] https://launchpad.net/bugs/873424
16:20:24 <bdmurray> test and fix for bug 901381 regarding ubiquity's apport behavior
16:20:26 <ubottu> Launchpad bug 901381 in ubiquity (Ubuntu) "excepthook in gtk_ui.py should write a crash report instead of calling ubuntu-bug" [High,Fix released] https://launchpad.net/bugs/901381
16:20:36 <bdmurray> triage of bug 922330 regarding wubi installing precise
16:20:36 <bdmurray> fixed lp_buttontags to make tagging bugs easier
16:20:36 <bdmurray> modifications to lp-grab-attachments to download bugs about a project
16:20:37 <ubottu> Launchpad bug 922330 in Wubi "Wubi.exe from 11.10 official CD installs 12.04 snapshot if CD not available" [Medium,Won't fix] https://launchpad.net/bugs/922330
16:20:54 <bdmurray> updated bugsquad tags to include new ubiquity bug tags
16:20:55 <bdmurray> research into and fixing of universe packages assigned to the foundations team in the package to team mapping
16:20:57 <bdmurray> upload of update-manager to lucid-proposed which adds in apport package hook
16:21:00 <bdmurray> fixed bug 878335 regarding moving bugs to grub-installer from ubiquity
16:21:01 <bdmurray> creation of packageset recent package bug graphs
16:21:02 <ubottu> Launchpad bug 878335 in apport (Ubuntu) "ubiquity source package hook should move more bugs to grub-installer" [Medium,In progress] https://launchpad.net/bugs/878335
16:21:07 <bdmurray>16:22:00 <cjwatson> Interviewing.
16:22:00 <cjwatson> Another try at fixing ports installs (bug 919356).
16:22:00 <cjwatson> Investigation of fixes for glibc 2.15 regressions.
16:22:02 <ubottu> Launchpad bug 919356 in choose-mirror (Ubuntu) "Default hostname needed for PowerPC" [Medium,Fix released] https://launchpad.net/bugs/919356
16:22:12 <cjwatson> Added a mock GResolver to the ubiquity test suite to make it more reliable in various contexts (the powerpc build, stgraber's LXC setup).  Would have been much quicker except that https://bugzilla.gnome.org/show_bug.cgi?id=669847 meant I had to do this in C.
16:22:14 <ubottu> Gnome bug 669847 in gio "GSimpleAsyncResult introspection annotates methods as functions; unusable for creating results for errors" [Normal,Unconfirmed]
16:22:16 <cjwatson> Clearing out http://lintian.ubuntuwire.org/tags/preinst-uses-dpkg-maintscript-helper-without-predepends.html.  Very slow and tedious, but everything is uploaded or has patches in a BTS somewhere now.
16:22:20 <cjwatson> Made a pass through some outstanding team work items with Steve and postponed a few.
16:22:23 <cjwatson> Tested transfer speed between cdimage and ftpmaster for foundations-p-image-build-pipeline.
16:22:26 <cjwatson> Reorganised the ubiquity test suite to be able to run in a DEP-8-compliant package tester (i.e. testing the installed version of the package).
16:22:29 <cjwatson> Working on a refactoring of how ubiquity removes about-to-be-copied files from the target system that fixes a slew of bugs and is more testable.
16:22:32 <cjwatson> ..
16:23:04 <jhunt> cjwatson: have we got the infrastructure to run DEP-8 tests yet?
16:23:36 <cjwatson> autopkgtest exists; I don't know its current status, and we don't yet have it set up in a datacentre machine or anything
16:23:57 <cjwatson> in theory should be a simple matter of sysadmin
16:24:15 <jhunt> great as I'd like to be able to run the upstart tests in this way too.
16:24:36 <cjwatson> You should be able to give autopkgtest a spin locally if you have time :)
16:25:03 <slangasek> stgraber: this udev change would more than slow the boot, it would prevent events from being triggered entirely... this requires a versioned dep on upstart.
16:25:31 <cjwatson> I admit I didn't actually try it out, merely ran the test in the appropriate environment under strace to ensure that it was testing the installed system and read the spec for the control file very carefully
16:25:52 <jhunt> cjwatson: ok, ta.
16:26:06 <slangasek> is the QA team involved with the autopkgtest discussions?
16:26:11 <stgraber> slangasek: indeed, I'll add the versioned dep.
16:26:51 <cjwatson> slangasek: there were QA people present at the appropriate UDS sessions, and jibel has a work item that came out of that; can't speak for after that
16:27:01 <slangasek> stgraber: please check that such a versioned dep doesn't cause upgrade problems, too...
16:27:29 <slangasek> cjwatson: ok
16:28:11 <slangasek> cjwatson: just wondering if this is something that should be set up to run as part of the QA lab
16:30:39 <cjwatson> eventually, yes; the blueprint calls for initial metadata to be added to a few key packages, and for something to be written that's able to feed it into a putative "should this be allowed to migrate from -proposed" tool
16:30:48 * slangasek nods
16:30:50 <cjwatson> https://blueprints.launchpad.net/ubuntu/+spec/foundations-p-upload-intermediary
16:31:45 <slangasek> any other questions?
16:32:30 <slangasek> [TOPIC] Bugs
16:33:29 <bdmurray> I've run across a couple this week worth discussing
16:33:46 <bdmurray> Is bug 930304 something our team should fix?
16:33:47 <ubottu> Launchpad bug 930304 in eglibc (Ubuntu) "gdm is attempted to be reloaded even when not runnning" [Medium,Triaged] https://launchpad.net/bugs/930304
16:34:01 <slangasek> yes
16:34:22 <bdmurray> okay
16:34:28 <slangasek> I'll probably wind up doing that one with my Debian eglibc committer hat
16:34:39 <slangasek> (feel free to assign to me)
16:34:51 <doko> ohh, nice =)
16:35:17 <bdmurray> regarding bug 818760 should the timeout be increased or might it be something with the server?
16:35:18 <ubottu> Launchpad bug 818760 in update-manager (Ubuntu) "update-manager crashed with timeout in readline(): timed out" [High,Confirmed] https://launchpad.net/bugs/818760
16:35:33 <bdmurray> uri=urllib2.urlopen(req, timeout=20)
16:35:34 <cjwatson> in principle it's always possible for network communication to fail for all kinds of reasons
16:35:41 <bdmurray> from the traceback
16:36:04 <cjwatson> u-m has to cope somehow without crashing
16:36:26 <cjwatson> and preferably without stalling the UI forever while it tries to talk through a firewall that's dropping all its outbound packets, say
16:37:17 <slangasek> right, so it needs to be fixed to not crash
16:37:58 <bdmurray> got it
16:39:37 <bdmurray> bug 926581 - that should have tasks for the meta package or the package manager?
16:39:39 <ubottu> Launchpad bug 926581 in msttcorefonts (Ubuntu) "eula not shown when installed via restricted-extras" [High,Triaged] https://launchpad.net/bugs/926581
16:39:46 <mvo> (fwiw, this stuff should be running in a thread already, but indeed, it should not crash)
16:40:46 <ev> ah, debconf. Can we set fire to it yet?
16:40:54 <cjwatson> :-P
16:41:12 <ev> :D
16:41:15 <slangasek> no, but feel free to fix its protocol ;)
16:41:26 <cjwatson> what's the protocol-level problem here?
16:41:31 <slangasek> here? none
16:41:40 <ev> sure, I'll just make it NOP every instruction
16:41:43 <slangasek> but I was offering ev productive outlets for his anger ;)
16:42:35 <ev> give me the source code to dpkg and hook me to a turbine. I'll power London with my rage. ;)
16:42:40 <slangasek> bdmurray: seems like a synaptic problem
16:42:59 <bdmurray> and software-center (that's what I tested)
16:43:17 <bdmurray> er, not software-center
16:43:20 <bdmurray> okay, thanks!
16:43:23 <slangasek> oh
16:43:27 <slangasek> sorry, read to fast
16:43:35 <slangasek> I read comment 3 but didn't digest 4
16:44:02 <bdmurray> well my grammar didn't help I'm sure
16:44:45 <slangasek> that's a puzzling one
16:44:52 <slangasek> anyone want to take a closer look at it?
16:46:24 <slangasek> well, I'll have a peek then
16:46:40 <slangasek> bdmurray: any other bugs?
16:47:22 <bdmurray> just wubi bug 862003
16:47:23 <ubottu> Launchpad bug 862003 in Wubi "wubi installs successfully but ends in 'permission denied' error due to the presence of a virtual or readonly drive; this only happens running wubi.exe standalone i.e. when installing using the tar.xz preinstalled image" [High,Triaged] https://launchpad.net/bugs/862003
16:48:48 <cjwatson> ... how's that Windows hire going? ;-)
16:49:11 <slangasek> ... workin' on it
16:49:16 <slangasek> add that bug to the queue :P
16:49:58 <bdmurray> I'll have to make a queue but okay
16:50:12 <slangasek> [TOPIC] AOB
16:50:15 <slangasek> anything else?
16:51:20 <ev> filing a freeze exception for the crash database stuff as mentioned above
16:51:30 <ev> not a lot more work, just more than I think I'll manage tonight
16:52:01 <cjwatson> the freeze is actually 2100 tomorrow, I believe
16:52:22 <ev> yay, that buys me a bit more time
16:52:26 <slangasek> :)
16:52:42 <ev> should I request an exception, just in case, or is mentioning it here enough to make the release team aware that I'm cutting it close?
16:53:02 <slangasek> it's already on Kate's list
16:53:11 <slangasek> as a "late-landing change"
16:53:25 <cjwatson> https://wiki.ubuntu.com/PrecisePangolin/FeatureLanding
16:53:29 <slangasek> yep
16:53:50 <ev> wow, hyper-organized
16:53:58 <ev> cheers guys
16:54:10 <slangasek> #endmeeting