16:02:55 <slangasek> #startmeeting
16:02:55 * slangasek nudges the bot with his toe
16:02:55 <meetingology> Meeting started Wed Feb 20 16:02:55 2013 UTC.  The chair is slangasek. Information about MeetBot at http://wiki.ubuntu.com/meetingology.
16:02:55 <meetingology> 
16:02:55 <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
16:04:17 <slangasek> there we go
16:04:23 <slangasek> [TOPIC] Lightning round
16:04:39 <slangasek> $ echo $(shuf -e barry doko stgraber jodh ev bdmurray slangasek ogra cjwatson xnox stokachu)
16:04:42 <slangasek> slangasek stgraber jodh ogra cjwatson xnox doko barry bdmurray ev stokachu
16:04:45 <slangasek> hmm, I don't like these dice
16:04:50 <ev> I love these dice
16:04:52 * xnox *giggles*
16:06:15 <slangasek> - uploaded initial packaging of systemd services to raring last week, in preparation for the desktop team's sprint this week
16:06:46 <slangasek> - discussions around the UbuntuKylin flavor application at the TB
16:06:54 <ogra_> oops
16:07:02 <stokachu> gnome relying on systemd makes me soooo happy
16:07:10 <ogra_> everyone
16:07:12 <ogra_> ...
16:07:16 <ogra_> lets switch to KDE
16:07:29 * barry dusts off fvwm
16:07:33 <xnox> rumour has it, it will all be rewritten in QML soon =)
16:07:35 <stokachu> haha
16:07:36 <slangasek> - trying to figure out how to make ovmf use persistent storage for nvram variables
16:07:53 <xnox> that would be so awesome to have persistent nvram =)
16:08:23 <ev> GNOME? That thing still exists?
16:08:34 * xnox ponders if one would be able to test mac os x weird efi using a modified ovmf.
16:08:50 <slangasek> - working on bug #1130419, since bad timing of the steam client release and Ubuntu 12.04.2 makes a mess on 64-bit systems
16:08:53 <slangasek> (done)
16:08:54 <ubottu> bug 1130419 in apt (Ubuntu Precise) "apt resolver doesn't do sensible things when satisfying a cross-dependency on a virtual package (steam, wine)" [Critical,Triaged] https://launchpad.net/bugs/1130419
16:09:01 <stgraber> Feature work:
16:09:01 <stgraber> - Upstart (BLUEPRINT: foundations-r-upstart-user-session-enhancements)
16:09:01 <stgraber> - Helped fix and prepare the upstart-shutdown branch.
16:09:01 <stgraber> - Tested some chanegs to the way we deal with the environment.
16:09:01 <stgraber> - Published updated test packages and sent call for testing yesterday.
16:09:04 <stgraber> - Discussed event prefixing, decided not to actually change anything.
16:09:06 <stgraber> - Container (BLUEPRINT: servercloud-r-lxc)
16:09:09 <stgraber> - Release 0.9~alpha3.
16:09:11 <stgraber> - Updated the package in Ubuntu which brings us really close to upstream, dropping over 100 patches.
16:09:14 <stgraber> - Some more code reviews and forwarded most of our remaining delta upstream.
16:09:16 <cjwatson> slangasek: steam> oh, yes, sorry, way too many interrupts today and haven't got to that yet ...
16:09:17 <stgraber> - Networking (BLUEPRINT: foundations-r-networking)
16:09:20 <stgraber> - Still waiting on test results for the infiniband support, no other progress there.
16:09:23 <stgraber> Other work:
16:09:25 <stgraber> - Networking
16:09:28 <stgraber> - Didn't spend as much time as I'd have wanted on ofono support in NM but made all the property querying work. Now I need to figure out some random segfaults.
16:09:29 <slangasek> xnox: well, the darn thing is, ovmf has code already to read/write nvram state to disk... it's just not working, so I'm fighting with the thing to figure out where it's sending its debug output
16:09:31 <stgraber> - Release
16:09:33 <stgraber> - Edubuntu 12.04.2 LTS
16:09:36 <stgraber> - Discussed plans for Ubuntu Kylin
16:09:38 <stgraber> TODO:
16:09:41 <stgraber> - Continue the ofono/NM work.
16:09:43 <stgraber> - Try to finish any LXC feature work for this cycle (1 item left).
16:09:46 <stgraber> (DONE)
16:10:08 <slangasek> ogra_, stokachu: bah, don't worry, it's just a set of dbus interfaces for systemd :)
16:10:13 <cjwatson> slangasek: but I see David is way ahead of me anyway
16:10:40 <slangasek> cjwatson: right :-)  Can you follow up on his patch today, see if it works as advertised?
16:10:40 * ogra_ notes down GNOME -> set of dbus interfaces for systemd
16:10:59 <cjwatson> slangasek: ok
16:11:09 <stokachu> i heard gnome is being included in the systemd package
16:12:04 <cjwatson> jodh is off, IIRC
16:12:06 <ogra_> i heard it becomes a kernel module
16:12:07 <doko> well, they wouldn't have any issue anymore with build deps
16:12:11 <stokachu> lol
16:12:26 <slangasek> jodh> correct
16:12:28 <ogra_> right after dbus moved into the kernel
16:12:29 <cjwatson> So ogra's turn
16:12:30 <slangasek> ogra_:
16:12:35 <ogra_> done:
16:12:48 <ogra_> * hunting installer breakage ...
16:13:06 <ogra_> (thanks xnox, cjwatson and seb for helping)
16:13:11 <ogra_> todo:
16:13:14 <xnox> doko: yeah it's very hard to do the whole Gnome module inclusion thing, so they decided that systemd is the last module they will include. If anything else is added, it must be in the systemd module.
16:13:20 <ogra_> * switch images to xz compression
16:13:27 <doko> that wsa short ;p
16:13:29 <ogra_> * make flavour images for nx7 work
16:13:33 <doko> that wsa short ;p
16:13:47 <xnox> doko: wait for my update.
16:13:49 <ogra_> * implement publisher code for tablet and phone dailies
16:13:52 <ogra_> ..
16:14:08 <cjwatson> ogra_: let me know if you need help with that last, since I gather it's urgent
16:14:20 <cjwatson> Finished and released 12.04.2.  (Was very little use to anyone on Friday.)
16:14:24 <cjwatson> foundations-r-phased-updates: Solicited and responded to another round of review.
16:14:27 <cjwatson> foundations-r-future-release-infrastructure: Made a bit more progress on rewriting cdimage in Python.  Not quite worked up the fortitude to tackle publish-release yet.
16:14:30 <cjwatson> Spent most of the rest of the week trying to make progress on rls-r-tracking-bug-tasks:
16:14:33 <cjwatson> * Fixed apt-clone autopkgtest failures (bug 1112472).  Did some general maintenance on apt-clone while I was at it.
16:14:35 <ubottu> bug 1112472 in apt-clone (Ubuntu Raring) "dep8 test failed because test data is too old" [High,Fix released] https://launchpad.net/bugs/1112472
16:14:36 <cjwatson> * Tweaked shim-signed to recommend secureboot-db (bug 1087843).
16:14:37 <ubottu> bug 1087843 in shim-signed (Ubuntu Quantal) "[MIR] secureboot-db" [Undecided,New] https://launchpad.net/bugs/1087843
16:14:38 <cjwatson> * Fixed syslinux to build with GCC 4.7 (bug 1127462).
16:14:39 <ubottu> bug 1127462 in syslinux (Ubuntu) "syslinux should be built using gcc-4.7 again" [High,Fix released] https://launchpad.net/bugs/1127462
16:14:41 <cjwatson> * Fixed grub-installer's transfer of user boot parameters to target system to work even with live filesystems (bug 664526).
16:14:42 <ubottu> bug 664526 in grub-installer (Ubuntu Precise) "setting nomodeset in grub, if live session was started with nomodeset" [High,Triaged] https://launchpad.net/bugs/664526
16:14:44 <cjwatson> * Fixed crash when double-clicking on partitioner controls (bug 1095684).
16:14:45 <ubottu> bug 1095684 in ubiquity (Ubuntu Raring) "Ubiquity crashes upon clicking 'New partition table' twice in the manual partitioning step during raring installation" [High,Fix committed] https://launchpad.net/bugs/1095684
16:14:46 <ogra_> cjwatson, not that urgent as i understood (for dailies that is) ... the demo image will be a simple cp on my side
16:14:47 <cjwatson> * Fixed ordering of encryption password strength indications (bug 1068391).
16:14:49 <ubottu> bug 1068391 in ubiquity (Ubuntu Raring) "Password strength bug" [High,Fix committed] https://launchpad.net/bugs/1068391
16:14:50 <cjwatson> * Made installer locale handling smarter to be able to cope with many mismatched language/location pairs (bug 1094872).
16:14:51 <ubottu> bug 1094872 in ubiquity (Ubuntu Raring) "Installation information cause letter format instead of DIN A4" [High,Fix committed] https://launchpad.net/bugs/1094872
16:14:53 <cjwatson> Cleared up image build failure due to libreoffice 4.0.0 packaging changes.
16:14:57 <cjwatson> ..
16:14:59 <cjwatson> ogra_: OK
16:15:02 <cjwatson> handy :)
16:15:16 <ogra_> as i understood dailies early next week should be fine
16:16:11 <xnox> * uploaded plymouth with support to ask questions (bug #509384 )
16:16:11 <xnox> .
16:16:11 <xnox> * uploaded mdadm merge with many changes, most importantly plymouth
16:16:11 <xnox> integration and initramfs support to assemble Intel Matrix Raid
16:16:11 <xnox> arrays (sometimes aka Intel Rapid Storage). (bug #1102499 )
16:16:11 <xnox> .
16:16:12 <ubottu> bug 509384 in plymouth (Ubuntu) "ubuntu-logo does not implement display_question callback" [High,Fix released] https://launchpad.net/bugs/509384
16:16:13 <xnox> * uploaded ntfs-3g package update (caused ABI break of testdisk, which
16:16:14 <ubottu> bug 1102499 in mdadm (Ubuntu) "Cannot see 'bootdegraded' question" [High,Fix released] https://launchpad.net/bugs/1102499
16:16:15 <xnox> got spotted and escalated all the way to debian ctte, now fixed in
16:16:17 <xnox> ubuntu and patch forwarded to debian). It solve a few compatibality
16:16:19 <xnox> issues with Windows8 ntfs updates.
16:16:21 <xnox> .
16:16:23 <xnox> * helped ogra to debug ubiquity nexus7 image (and as it turns out
16:16:27 <xnox> desktop images as well). In the end uploaded a workaround to not
16:16:29 <xnox> start whoopsie together with ubiquity-dm. The actual cause of the
16:16:31 <xnox> bug is believed to be deep in the network-manager generated bindings
16:16:33 <xnox> which Laney was diving into to fix. (Not sure what the update there
16:16:35 <xnox> is).
16:16:37 <xnox> .
16:16:39 <xnox> * tested unattended preseeding oem-config on the nexus7. It works \o/
16:16:41 <xnox> email send to utha-devel & ubuntu-installer. The wifi preseeding
16:16:43 <xnox> didn't work (by simply dropping network-manager config). I will try
16:16:45 <xnox> calling nmcli tool directly, and if that fails will pass this over
16:16:47 <xnox> to desktop/qa teams. Documented automatic-oem-config boot option on
16:16:49 <xnox> the https://wiki.ubuntu.com/DesktopCDOptions
16:16:51 <xnox> .
16:16:53 <xnox> * Implemented support to pass rebootcommand argument to the reboot
16:16:55 <xnox> syscall in the upstart's reboot utility. This allows to reboot into,
16:16:59 <xnox> e.g. fastboot mode on the nexus7. (Thus completing the unattended
16:17:01 <xnox> redeploy cycle for Nexus 7). Merge proposal on the way. I wonder if
16:17:03 <xnox> UEFI fastpath "reboot into firmware" is implemented the same
16:17:05 <xnox> way. Also interesting to know how one can query which rebootcommands
16:17:07 <xnox> are supported.
16:17:09 <xnox> .
16:17:11 <xnox> * bug #1080437 is essentially fixed, by using g-s-d on the images that
16:17:13 <ubottu> bug 1080437 in ubiquity (Ubuntu Raring) "no background during the 13.04 daily install" [High,In progress] https://launchpad.net/bugs/1080437
16:17:13 <xnox> are available. There is merge proposal to use some lxde tool for
16:17:15 <xnox> Lubuntu flavour and Xubuntu flavour is still needs to be
16:17:17 <xnox> fixed. Painting background with g-s-d is nice, since it support
16:17:19 <xnox> rotation & screen size changes out of the box (e.g. for nexus 7). At
16:17:21 <xnox> the same time noticed t
16:17:23 <xnox> At
16:17:25 <xnox> the same time noticed that bug #594233 has regressed as the
16:17:26 <ubottu> bug 594233 in ubiquity (Ubuntu) "Pressing ctrl-alt-T gets you a root terminal in oem-config" [Undecided,Fix released] https://launchpad.net/bugs/594233
16:17:29 <xnox> configuration option moved from gconf/metacity to
16:17:31 <xnox> gsettings/g-s-d. This is now fixed. At the same time cleaned up
16:17:33 <xnox> remaining usage of gconf in ubiquity, as actually all of those keys
16:17:35 <xnox> have migrated to gsettings now.
16:17:37 <xnox> .
16:17:39 <xnox> * Tested the upstart session login, fixed a race with g-s-d being
16:17:41 <xnox> spawned too many times. Still other fixes to do. Chatted with laney
16:17:43 <xnox> and forwarded instructions to him, he will be looking into / helping
16:17:45 <xnox> with upstart based sessions on desktop side. And since he uses
16:17:47 <xnox> xmonad session he wants a generic way to run existing xsessions.
16:17:49 <xnox> .
16:17:51 <xnox> * Working from the office this week, due to ongoing desktop
16:17:53 <xnox> sprint. You can tell I'm catching on evan's verbose changelog mode.
16:17:55 <xnox> .
16:17:59 <xnox> * Also uploaded adb from 4.2.2 android sdk into the nexus7 installer ppa.
16:18:01 <xnox> + blog. (for those not on raring).
16:18:03 <xnox> ..
16:18:05 <stgraber> xnox: you've been spending way too much time with ev ;)
16:18:06 <ogra_> you definitely work to close to ev now !
16:18:12 <ev> bwahahaha
16:18:14 <ogra_> *snap*
16:18:25 <doko> ev's company does do you any good
16:18:26 <stokachu> lp:1090512 - Need help in determing what rdepends I need to be concerned with and other pitfalls that may arise
16:18:27 <stokachu> lp:1068399 - Needs a ack/nack on getting this into precise
16:18:28 <stokachu> Most of my time has been utilized packaging Firefox and now Thunderbird ESR
16:18:30 <stokachu> (done)
16:18:38 <Laney> (I'm going to abuse cyphermox to help me on that)
16:18:43 <xnox> bug 1090512 bug 1068399
16:18:45 <ubottu> bug 1090512 in Quantal Backports "Please backport sudo 1.8.6p3-0ubuntu2 (main) from raring" [Undecided,New] https://launchpad.net/bugs/1090512
16:18:46 <ubottu> bug 1068399 in Precise Backports "Please backport parallel 20120422-1 (universe) from raring" [Undecided,New] https://launchpad.net/bugs/1068399
16:18:55 <stokachu> doh sorry i thought i had it switched to the bug
16:19:27 <xnox> np.
16:19:40 <Laney> ping micahg about parallel
16:19:42 <doko> - another GCC-4.8 test rebuild (Debian testing, done by lucas), submitted four GCC issues for ICEs
16:19:42 <doko> - oops, they did it again ... OpenJDK-6 security builds
16:19:42 <doko> - +1 maintenance work (ftbfs, -proposed migration, component mismatches, package removals). looks like I could do this for another few weeks full time
16:19:42 <doko> - tracked down upstart coding issue, develper trying to outsmart the compiler
16:19:42 <doko> - got rid of gcc-4.6 in main
16:19:43 <doko> - chase downs submitters of incomplete MIRs
16:19:45 <doko> - Linaro GCC 4.7-2013.02 release, including all cross builds
16:19:48 <doko> - Linaro GCC 4.6-2013.02 release
16:19:49 <doko> - binutils updates from the branch, including all cross builds
16:19:51 <stokachu> Laney: ok will do
16:19:51 <doko> - helped with the Canonical statement about the Python trademark issue
16:19:53 <doko> done
16:20:06 <cjwatson> ooh, is that statement published?
16:20:23 <doko> I don't know
16:20:23 <barry> cjwatson: it is
16:20:32 <stokachu> i think chinchilla would be an awesome replacement name
16:20:36 <barry> cjwatson: well, um, it was sent to the psf anyway
16:20:38 <barry> short week due to usa holiday.  bug 1097922.  friends package debugging.  u1 client oauth debugging (upstream issue 120)y.  click packages.  tablet py3 work.
16:20:40 <ubottu> bug 1097922 in newt (Ubuntu Raring) "byobu-config crashed with ImportError in /usr/lib/python2.7/dist-packages/snack.py: /usr/lib/python2.7/dist-packages/_snack.so: undefined symbol: _Py_RefTotal" [High,Fix released] https://launchpad.net/bugs/1097922
16:20:43 <barry> foundations-r-python3-oauth: done
16:20:44 <xnox> barry: public URL or it didn't happen =)
16:20:44 <cjwatson> gotcha, thanks, just noosey
16:20:46 <cjwatson> *nosey
16:20:48 <barry> foundations-r-python-versions: done (though porting work may be ongoing)
16:20:52 <barry> foundations-r-python33: done
16:20:56 <barry> ..
16:21:06 <barry> (good question as to whether the psf will publish them)
16:21:08 <stokachu> ah i went out of turn, sorry everyone
16:21:30 <xnox> stokachu: chin3 and chin2.7 don't sound like awesome shebang lines.
16:21:35 <bdmurray> irc discussion with evan regarding the custom comparator for package versions and rolling it out
16:21:37 <bdmurray> worked on migration of bucketversion column family to one that uses the custom comparator
16:21:40 <bdmurray> pushed errors branch allowing one to use a first_appearence url parameter to the API
16:21:43 <bdmurray> submitted rt regarding updating errors to r270
16:21:47 <bdmurray> built and tested dpkgversiontype comparator on canonistack
16:21:47 <bdmurray> reviewed and tested evan's bucket page reworking
16:21:47 <bdmurray> modified bug bot to comment on sru bugs still needing verification
16:21:47 <bdmurray> wrote a tool to query the errors api regarding versions of packages from -proposed for errors
16:21:50 <bdmurray> modified errors regression checker to use date published and today as from and to dates
16:21:59 <bdmurray> subscribed desktop-packages and dx-packages to packages which they should be subscribed to
16:22:02 <bdmurray> reassignment of linaro packages in package to team mapping
16:22:05 <bdmurray> review of server team package to team mapping discrepancies
16:22:07 <bdmurray> updated meta release files for the 12.04.2 point release
16:22:10 <bdmurray> resolved an issue with cd_error being unset in the ubiquity source package hook
16:22:13 <bdmurray> worked on an upstart job to replace CD with packages functionality from update-notifier
16:22:16 <bdmurray> research into bug 1124330
16:22:17 <ubottu> bug 1124330 in whoopsie (Ubuntu) "[raring] Latest whoopsie 0.2.13 slows down boot process by 29 seconds!" [Undecided,Confirmed] https://launchpad.net/bugs/1124330
16:22:33 <bdmurray> regarding work items I'm working on the following:
16:22:39 * xnox ponders if it's the same network-manager bug....
16:22:43 <bdmurray> create and update a BucketSystems column family to keep track of unique systems in a bucket: INPROGRESS
16:22:55 <bdmurray> Import developer to package and developer to team mappings and filter the default view on errors.ubuntu.com to problems the developer and their team are responsible for: INPROGRESS
16:23:09 <bdmurray> and finished
16:23:19 <bdmurray> auto comment on bugs needing verification (indicate that it needs to happen in the next 15 days) at some interval (3 months)
16:23:36 <bdmurray> ⌁ done
16:23:38 <ev> - Slower week. Sick for much of it, but feeling infinitely better now.
16:23:38 <ev> - Rebuilt the errors.ubuntu.com bucket page to match the design:
16:23:39 <ev> https://errors.ubuntu.com/bucket/?id=%2Fusr%2Fbin%2Flsb_release%3AIOError%3A%3Cmodule%3E%3Amain%3Acheck_modules_installed%3Agetoutput%3Agetstatusoutput
16:23:39 <ev> https://wiki.ubuntu.com/ErrorTracker#Bucket_page
16:23:39 <ev> - The instances are now presented in an infinitely-scrolling table (sorting
16:23:39 <ev> by time coming soon). This was made possible by a rebuild of the Buckets CF
16:23:49 <ev> using a TimeUUID comparator. I have a ticket in process to drop the old CF.
16:23:49 <ev> - The graph is now based on average instances per calendar day (weighted
16:23:50 <ev> against the systems) rather than a raw instance count, and provides
16:23:50 <ev> individual lines for each release (the key is coming back soon).
16:23:50 <ev> - The package versions with this problem table will provide a
16:23:56 <ev> version-by-version breakdown of the problem in each Ubuntu version, but
16:23:56 <ev> right now it just displays the total count. This is waiting for the new
16:23:56 <ev> BucketVersions CF that uses a composite of my DpkgVersionType comparator
16:23:56 <ev> and a UTF8 string for the Ubuntu release.
16:23:56 <ev> - Wrote a script to rebuild the BucketVersions CF using different comparators
16:24:01 <ev> and used it to find out that there are a lot of binary crashes that we're not
16:24:01 <ev> retracing. This is because while we clear the Rabbit queues when they get too
16:24:01 <ev> big, we don't also clear the index in Cassandra for this. So while we remove
16:24:01 <ev> the core files we have to hand, we never ask users for a replacement. I'm
16:24:07 <ev> going to work with IS to get the process changed to account for this.
16:24:07 <ev> The rebuild script is running in a debugging mode on one of the retracer
16:24:08 <ev> machines for now, collecting some more background information. As soon as I
16:24:08 <ev> finish getting the new Cassandra version (1.2.1) ready for production, the
16:24:08 <ev> new column family set up in oops-repository, and everything deployed, I'll
16:24:15 <ev> ask webops to switch it over to write mode.
16:24:15 <ev> - Investigation work around Cassandra 1.2.1. They've finally built an ACL in
16:24:15 <ev> the core distribution again, but it's not complete in the released version.
16:24:15 <ev> We can continue using SimpleAuthenticator (text files) for now and switch
16:24:21 <slangasek> stokachu: sudo reverse-depends> I'm not all that familiar with the backports process, but it seems to me like the point is there to test *all* of them to spot any issues?  If the question is which revdeps are likely to break, I think the answer is "none"
16:24:21 <ev> over to the Cassandra keyspace backed authentication backend when 1.2.2/3 is
16:24:22 <ev> out.
16:24:22 <ev> - Bundling Cassandra 1.2.1 with simple authentication and my DpkgVersionType
16:24:22 <ev> comparator. Testing on Canonistack.
16:24:27 <ev> - Code review for Brian.
16:24:27 <ev> - Discussion with Bryce about Xorg needs around the error tracker. They'd like
16:24:27 <ev> to retain some core files as they need to see locals in the backtrace output.
16:24:27 <ev> We agreed to build a small fixed size queue of Xorg core dumps.
16:24:30 <ev> They also want to be able to sort by the Last seen field in the most common
16:24:30 <ev> problems table, ellipsize on the First/Last seen fields instead of the
16:24:30 <ev> signature column, and search by packageset.
16:24:31 <ev> (done)
16:24:47 <doko> that was short
16:24:52 <stgraber> ev: you loose, xnox's was longer ;)
16:25:08 <stgraber> ev: but you were sick, I guess that's a valid excuse
16:25:12 <slangasek> bdmurray, ev: is that custom comparator going to be released, btw?  Seems like the sort of thing that would find occasional reuse elsewhere
16:25:25 <stokachu> slangasek: yea i was afraid i'd have to test them all
16:25:31 <ev> slangasek: https://github.com/evandandrea/cassandra-dpkgversiontype
16:25:47 <slangasek> stokachu: or alternately, none ;)  But that's a question of the backports process
16:27:59 <slangasek> ev: yay - you should send an RFP to Debian, see if someone on the java team wants to maintain a package for it
16:28:19 <slangasek> any other questions re: statuses?
16:29:09 <ev> slangasek: tricky - cassandra moves quicker than Debian's glacial pace and thus the packages live in the Apache debian repository
16:29:24 <ev> so if it were packaged, the dependencies wouldn't be solvable in Debian
16:29:37 <slangasek> ah, it has cassandra-specific deps?
16:29:44 <slangasek> maybe less interesting then, indeed
16:29:50 <ev> it's built against the Cassandra API, yeah
16:30:07 <slangasek> right, I assumed it was something generalizable that cassandra would call out to
16:30:13 <slangasek> [TOPIC] Bugs
16:30:26 <slangasek> bdmurray: what's new with six legs and wings?
16:30:41 <bdmurray> well there is the previously mentioned bug 1124330
16:30:42 <ubottu> bug 1124330 in whoopsie (Ubuntu) "[raring] Latest whoopsie 0.2.13 slows down boot process by 29 seconds!" [Undecided,Confirmed] https://launchpad.net/bugs/1124330
16:31:20 <slangasek> has anyone else seen that bug?
16:31:25 <slangasek> I rebooted yesterday, saw nothing of the sort
16:31:37 <ev> given that it's installed on everyone's system and the delta from 0.2.9 to 0.2.13 didn't touch anything upstart related, I call shenanigans
16:31:49 * barry neither
16:31:58 <bdmurray> I've not seen it
16:32:04 <ev> installed on everyone's system> I'd have quite a few people beating down my door
16:32:32 <slangasek> so probably unrelated to the whoopsie update
16:32:57 <slangasek> though it may still be triggered by whoopsie in some fashion
16:33:36 <ev> nothing depends on it
16:34:14 <xnox> well from 0.2.9 -> 0.2.13 we have a switch from direct dbus to nm-bindings-to-dbus and that broke nexus7 and desktop images when no desktop-session is running.
16:34:38 <xnox> so if someone happens to boot into something like ubiquity-dm.....
16:34:38 <ev> yeah, that was strange
16:36:45 <slangasek> so, on the boot charts it looks like things go to sleep exactly at the point of NM activity
16:37:43 <xnox> and whoopsie can start before network-manager (as there is no start on as far as I can see)
16:37:56 <xnox> thus doing dbus activation of network-manager?!
16:38:40 <ev> hm, I would say it should start on network-manager, but it doesn't explicitly need it
16:38:41 <slangasek> I don't see ubiquity in the chart, fwiw; though there is 'preload', which points to this user tinkering to try to speed up their system
16:38:53 <ev> if it cannot talk to NM, it's supposed to carry on like it has an active internet connection
16:39:21 <xnox> and you are using async NM calls? cause otherwise it will block on getting the dbus proxy for NM objects.
16:39:26 <ev> I guess there's no way to express, "ideally wait for network-manager, but if it doesn't show up, still start whoopsie"
16:39:29 * xnox should poke that code.
16:39:35 <slangasek> ev: but if the way it checks for NM is over dbus, and this auto-launches NM, that could be a problem
16:39:39 <ev> or is that effectively dbus service activation
16:39:40 <slangasek> (maybe)
16:40:11 <xnox> ev: well we can start on started network-manager or (something that is much later but present on all systems)
16:40:18 <slangasek> the only dbus-activated service I see in NM is /usr/share/dbus-1/system-services/org.freedesktop.nm_dispatcher.service
16:40:42 <ev> xnox: is it safe to assume everyone has NM?
16:40:46 <slangasek> no
16:40:48 <xnox> no.
16:40:56 <ev> xnox: there are no async libnm calls
16:41:05 <ev> at least not for what we need: http://projects.gnome.org/NetworkManager/developers/libnm-glib/09/NMClient.html
16:41:18 <slangasek> (not if you want whoopsie to work on servers in any capacity)
16:41:28 <ev> yeah, and I do
16:41:33 <slangasek> delayed whoopsie startup can be done as 'start on stopped rc RUNLEVEL=[2345]'
16:41:55 <slangasek> though that's not /guaranteed/ to take longer than starting up NM itself
16:41:58 <xnox> ev: that is no much different from the udisks2 bindings.
16:42:03 <xnox> s/no/so/
16:42:06 <slangasek> however, as mentioned the only dbus-activated NM serice is the dispatcher
16:42:11 <slangasek> which I don't think whoopsie would be using?
16:42:18 <ev> nope
16:42:34 <slangasek> ok
16:42:41 <slangasek> who can follow up on this bug?  xnox?
16:43:00 <ev> http://bazaar.launchpad.net/~daisy-pluckers/whoopsie/trunk/view/head:/src/connectivity.c is pretty simple
16:43:04 <xnox> slangasek: ok. i'll poke it a bit more.
16:43:19 <xnox> unless ev is in love with network-manager? =)
16:43:20 <slangasek> xnox: thanks
16:43:26 <slangasek> bdmurray: other bugs?
16:43:31 <ev> xnox: stab stab
16:43:34 <ev> thanks :)
16:43:38 <bdmurray> the oem team pinged me about bug 1097570
16:43:41 <ubottu> bug 1097570 in grub2 (Ubuntu) "grub2-signed can not find the right device when there are two filesystems containing the file '.disk/info'." [High,Triaged] https://launchpad.net/bugs/1097570
16:44:12 * xnox .o0( How to tell apart identical twins?! )
16:44:25 <bdmurray> cjwatson: you'd said at one point in time that wouldn't make 12.04.2 because it was rather hard
16:44:31 <cjwatson> Indeed
16:44:51 <cjwatson> My best idea is to put a UUID in that file and match on that instead
16:44:54 <cjwatson> Or something similar
16:44:59 <cjwatson> But I haven't yet checked whether that's feasible
16:45:08 <cjwatson> (Probably a different file, actually)
16:45:38 <cjwatson> bdmurray: I can take assignee on that - probably more sensible since I wrote the code
16:45:45 <cjwatson> And I'll put it on rls-r-tracking so I remember
16:46:10 <bdmurray> cjwatson: okay, sounds good
16:46:25 <bdmurray> and they also mentioned bug 967229
16:46:28 <ubottu> bug 967229 in plymouth (Ubuntu) "Text mode shown briefly with various "cryptic" texts when logging out or shutting down" [High,Triaged] https://launchpad.net/bugs/967229
16:46:39 <cjwatson> Heh, now *that's* a long-running problem
16:46:51 <cjwatson> (Sorry, that's my only useful input)
16:47:25 <slangasek> I saw an update on that one from jmleddy yesterday
16:47:43 <bdmurray> right
16:48:03 <slangasek> the printf/clear seems to be what we discussed with apw several udses ago but never implemented
16:48:51 <slangasek> for the record, there's no way to prevent this from going into text mode briefly on shutdown - it's not an upstart ordering bug
16:49:29 <slangasek> the architecture just doesn't allow for it to be done sanely - the X server won't leave the VT in graphics mode
16:49:44 <slangasek> (and probably shouldn't, for purposes of maintainability/debuggability)
16:49:50 <slangasek> so we do need to clear the text
16:50:49 <slangasek> bdmurray: IIRC we said we wanted to clear the text there on VT7 after starting lightdm?
16:51:13 <slangasek> in which case putting this in the lightdm job might be most sensible
16:51:19 <mdeslaur> that doesn't work
16:51:24 <cjwatson> mm, I would recommend clearing it just before exiting X as well - it's not uncommon for stuff to be left on vt7 later on
16:51:24 <bdmurray> slangasek: yes, that sounds right
16:51:31 <slangasek> mdeslaur: why not?
16:51:32 <cjwatson> shouldn't hurt to do both
16:51:54 <slangasek> cjwatson: I think at the time, we argued that anything writing to VT7 *while* lightdm is running is itself buggy and we should know about it and fix it at the source
16:51:59 <mdeslaur> you can clear the screen when you shutdown lightdm, but the processes that get shutdown in parallel by upstart still print junk before plymouth comes up
16:52:35 <cjwatson> I'm not sure I buy that argument - writing to the console under X isn't intrinsically buggy
16:52:49 <slangasek> cjwatson: the argument is that they shouldn't be outputting to the console
16:52:55 <slangasek> they should be outputting somewhere more useful
16:53:00 <cjwatson> Maybe
16:53:38 <slangasek> mdeslaur: likewise, jobs shutting down in parallel to lightdm shouldn't be writing anything here; the output should be sent somewhere more appropriate (possibly plymouth, possibly upstart log)
16:53:39 <cjwatson> However, if I boot without quiet (or even without splash), I still don't want startup junk on my screen when I shut down
16:54:10 <cjwatson> sending to plymouth is hard when plymouthd hasn't restarted yet :)
16:54:21 <slangasek> yeah
16:54:26 <mdeslaur> slangasek: yes, that makes sense
16:54:34 <cjwatson> and sending to the upstart log requires knowing what you're doing in advance
16:54:38 <slangasek> but we could make sure plymouthd starts sooner on shutdown, while leaving the splash deferred until lightdm shuts down
16:54:50 <cjwatson> I think we have to do that, yes
16:55:00 <slangasek> anyway, I think having the lightdm job clear the VT on startup is per-se correct
16:55:02 <cjwatson> similar to the startup architecture
16:55:21 <slangasek> so we should probably do that as a first step that covers the 90% case, I think?
16:57:21 <slangasek> bdmurray: are you happy moving ahead with that part of it?
16:58:43 <bdmurray> slangasek: sure
16:58:49 <slangasek> ok, thanks
16:58:50 <slangasek> any other bugs?
16:59:11 <bdmurray> bug 1126107 - I haven't had a chance to test it yet though
16:59:12 <ubottu> bug 1126107 in ubiquity (Ubuntu) "Ubiquity does not start during raring desktop installations using libvirt and KVM " [High,Confirmed] https://launchpad.net/bugs/1126107
17:00:04 <xnox> it should work now that whoopsie got a work around.
17:00:17 <slangasek> -> fix released then?
17:00:22 <xnox> wait, this is different.
17:00:54 <xnox> but needs a retest before proceeding with images from 18th or better.
17:01:18 <cjwatson> we didn't get a daily build this morning due to libreoffice-related breakage, but I'm respinning for that at the moment
17:01:31 <cjwatson> however the last daily build booted fine
17:01:36 <cjwatson> (just tested)
17:02:02 <xnox> I'll test more, for my current merge-proposal in a moment. And I'll mark it incomplete if I cannot reproduce this any more.
17:02:30 <cjwatson> I'm just leaving a comment too
17:02:40 <slangasek> ok, seems like that bug's in hand then
17:02:45 <slangasek> bdmurray: next? :)
17:02:54 <bdmurray> bug 1127867 regarding precise
17:02:57 <ubottu> bug 1127867 in mdadm (Ubuntu Precise) "/etc/udev/rules.d/65-mdadm.vol_id.rules uses non-existent 'vol_id'" [High,New] https://launchpad.net/bugs/1127867
17:03:25 <xnox> mine. Thanks for pointing out, didn't see it.
17:04:07 <slangasek> is that a file generated in the current (precise) package, or an upgrade leftover?
17:04:44 <slangasek> anyway, left to xnox (who is very popular today with the bugs)
17:04:45 <xnox> it's a regression between -release & -updates (where we switched to upstream udev rules)
17:04:51 <slangasek> ok
17:04:54 <xnox> as far as I can tell by quickly scanning.
17:05:40 <slangasek> bdmurray: any more?
17:05:48 <bdmurray> slangasek: nope, that's it!
17:05:51 <slangasek> great
17:05:53 <slangasek> [TOPIC] AOB
17:05:55 <slangasek> anything else?
17:07:01 <slangasek> #endmeeting