15:02:05 <slangasek> #startmeeting
15:02:05 <meetingology> Meeting started Wed May 30 15:02:05 2012 UTC.  The chair is slangasek. Information about MeetBot at http://wiki.ubuntu.com/meetingology.
15:02:05 <meetingology> 
15:02:05 <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
15:02:28 <slangasek> [TOPIC] lightning round
15:02:31 <slangasek> echo $(shuf -e barry doko stgraber jodh ev bdmurray slangasek ogra infinity cjwatson)
15:02:36 <slangasek> cjwatson stgraber barry ev jodh doko bdmurray infinity ogra slangasek
15:02:45 <slangasek> oh
15:02:46 <slangasek> hmm!
15:02:49 <slangasek> did I pull the wrong list? :)
15:02:54 <doko> fired!
15:03:06 <barry> it always takes 2 attempts :)
15:03:06 <slangasek> stupid cut'n'paste
15:03:17 * xnox the lightning round is broken second meeting in a row.....
15:03:17 <jodh> there must be a charm to do this :)
15:03:19 <ogra_> doko, better than fried
15:03:22 <slangasek> $ echo $(shuf -e barry doko stgraber jodh ev bdmurray slangasek ogra infinity cjwatson xnox stokachu)
15:03:25 <slangasek> ogra doko cjwatson barry stokachu ev jodh infinity slangasek stgraber xnox bdmurray
15:03:28 <slangasek> is that better? :)
15:03:35 <stokachu> \o/
15:03:45 * ogra_ starts then
15:03:47 <ogra_> done:
15:03:48 <ogra_> * merges and FTBFS fixes (pending uploads)
15:03:48 <ogra_> * whitmonday on ... guess ... monday
15:03:48 <ogra_> * bugged infinity (who then bugged webops) enough to actually get our arm live builders back up
15:03:48 <ogra_> * synced the new flash-kernel from debian
15:03:48 <ogra_> todo:
15:03:52 <ogra_> * switch to livefs builds for arm
15:03:54 <ogra_> * test and fine bugs due to the flash-kernel switch
15:03:56 <ogra_> * prepare for A1 and try to get as many arm images working as possible
15:03:58 <ogra_> ..
15:05:46 <doko> - vacation last week
15:05:52 <doko> - Linaro connect this week
15:06:09 <doko> - fixing GCC 4.7 build failures (~80 and counting)
15:06:11 <doko> ..
15:06:17 <ogra_> .oO( why does he get all the candy and we dont )
15:06:24 <cjwatson> Packaging:
15:06:24 <cjwatson> * Various random syncs and merges.
15:06:24 <cjwatson> * Finished packaging OpenSSH 6.0p1, and merged to quantal.
15:06:24 <cjwatson> * Spent the best part of a day bisecting GRUB to track down what turned out to be a memory leak that disproportionately affected EFI systems due to the larger disk cache in use there.  Backported the fix.
15:06:26 <slangasek> doko: is that 80 fixed, or 80 to go?
15:06:28 <cjwatson> Launchpad:
15:06:31 <cjwatson> * Removed support for daily installer uploads, not used since the days of dak (bug 827965).
15:06:33 <ubottu> Launchpad bug 827965 in Launchpad itself "Ditch daily installers" [High,Fix released] https://launchpad.net/bugs/827965
15:06:34 <cjwatson> * Lots of work on the custom upload code (e.g. bug 827973); mostly extensive refactoring to support future changes.
15:06:35 <ubottu> Launchpad bug 827973 in Launchpad itself "Unify custom-upload filename parsing and handling" [High,In progress] https://launchpad.net/bugs/827973
15:06:37 <cjwatson> * Arranged for imports of Debian publication records to be a bit more responsive.
15:06:38 <slangasek> (are we counting up or down? :)
15:06:40 <cjwatson> * Landed the first part of my work on the queue API (bug 1006173), exporting the acceptFromQueue() and rejectFromQueue() methods on PackageUpload objects.  Should be deployed in a day or two.
15:06:42 <ubottu> Launchpad bug 1006173 in Launchpad itself "Queue tool requires direct DB access" [High,In progress] https://launchpad.net/bugs/1006173
15:06:44 <cjwatson> * Discussions of how to sanely lay out the rest of the queue API.  I have an initial version working from before these discussions, but it'll need to be upended somewhat now.
15:06:46 <doko> slangasek, 80+ fixed, see http://bugs.debian.org/cgi-bin/pkgreport.cgi?tag=ftbfs-gcc-4.7;users=debian-gcc@lists.debian.org
15:06:48 <cjwatson> * Fixed buggy Packages-arch-specific parsing (bug 917708).
15:06:49 <ubottu> Launchpad bug 917708 in Launchpad itself "Launchpad does not recognize Arch = any-arm" [Low,In progress] https://launchpad.net/bugs/917708
15:06:50 <cjwatson> To do: Have a quick look at bug 1003443, since we've had a number of reports of that.
15:06:51 <slangasek> ok
15:06:52 <ubottu> Launchpad bug 1003443 in ubiquity (Ubuntu Quantal) "Ubuntu Quantal Desktop 21120523 failed to install: ubi-timezone failed with exit code 2" [Critical,Confirmed] https://launchpad.net/bugs/1003443
15:06:53 <cjwatson> ..
15:07:19 <barry> short week due to usa holiday; working on the lazr.restfulclient stack (wsgi_intercept, lazr.authorization, oauth) - going slow but making progress; tracked down incompatibility in py3 version of httplib2 <http://tinyurl.com/7anbxcn>; reviewed slangasek's python-pam branch; updated transition tracker (now auto-updating, thanks xnox!); first dm dputs to debian for flufl.* packages; AM assigned, working through DD process; reviewed and
15:07:19 <barry> accepted pep 421 - need to review patches.  todo: continue w/lazr stack; done.
15:07:43 <stokachu> nothing to report. but i did work hard for it. done.
15:07:53 <stokachu> :D
15:07:59 <ev> - Continued work on the tastypie port of the web API. I've put this down for
15:08:00 <ev> now as it's taking longer than expected and I need to focus on client-side
15:08:01 <ev> work which has a FeatureFreeze deadline.
15:08:03 <ev> - Fixes to our handling of unicode data through the web API.
15:08:04 <ev> - Helped Marcus and Tom with some terminal screenshots.
15:08:05 <ev> - Reviewed and merged Dmitrijs' branch to remove migration-assistant.
15:08:06 <ev> - Dropped migration-assistant from quantal and precise-proposed.
15:08:12 <ev> - errors.ubuntu.com now greys out problems which it believes may be fixed (in
15:08:13 <ev> trunk).  This looks at whether the most recent published binary in Launchpad
15:08:15 <ev> is the most recent version daisy has seen and whether the linked bug is
15:08:15 <cjwatson> barry: Is the "py3 wsgi_intercept with half its brain removed" sketch plan working out?
15:08:16 <ev> marked as fixed. If all goes well here, we'll change this to hide problems
15:08:18 <ev> it believes are fixed.
15:08:19 <slangasek> cjwatson: queue api - woot!
15:08:19 <ev> - Started implementing Matthew's redesign of the diagnostic settings page.
15:08:20 <ev> This adds a checkbox for automatically sending reports and installing
15:08:21 <ev> updates when the computer is in an unusable state (can't login, etc).
15:08:24 <ev> - Split libwhoopsie out of whoopsie to handle generating the system
15:08:24 <ev> identifier. I'd appreciate any attention you can give the merge
15:08:26 <ev> (https://code.launchpad.net/~ev/whoopsie/libwhoopsie/+merge/107562) and I'll
15:08:27 <ev> return the favor. This is for:
15:08:29 <ev> - Whoopsie, obviously. (Not sure if it's best practice to link or build in).
15:08:30 <ev> - The preferences page in gnome-control-center, so we can have a button that
15:08:32 <ev> takes the user to a webpage of all the crashes that have occurred on their
15:08:33 <ev> computer.
15:08:34 <ev> - The metrics database.
15:08:35 <ev> - The hardware database.
15:08:38 <ev> - Call with Kate and others on the QA Dashboard project.
15:08:39 <ev> - Call with Ivanka on putting some structure to the various discussions going
15:08:41 <ev> on around metrics.
15:08:42 <ev> - Started researching the server side apport hooks. Some trickyness around
15:08:44 <ev> running them as the user who experienced the crash, given that they'll be
15:08:45 <ev> fed from the whoopsie process running as the whoopsie user. Suggestions
15:08:47 <ev> welcome.
15:08:48 <ev> - Started researching moving the Launchpad retracers onto the crash database
15:08:50 <ev> retracer system, teaching them to talk to Cassandra instead of the existing
15:08:51 <ev> sqlite database, and synthesizing a bug from the crash database retracers
15:08:53 <ev> when one does not already exist. I've asked pitti to review my plans.
15:08:54 <ev> - Started researching creating crash reports from unresponsive applications.
15:08:56 <ev> The plan at the moment is to hook into compiz via a new plugin (or patch the
15:08:57 <ev> Unity plugin) that checks the active property of the compiz window at the
15:08:59 <ev> glPaint interval, then hand off the $PID to a process that can ptrace so we
15:09:00 <ev> can gdb attach and bt, handing off to apport.
15:09:02 <ev> I've chatted with Neil about this, and he approves of the compiz side of the
15:09:03 <ev> plan. I've started a discussion with the security team about the ptrace
15:09:05 <ev> side (https://bugs.launchpad.net/ubuntu/+bug/1006398).
15:09:06 <ev> - Ongoing conversation with Francis, Robert Collins, and William Grant around
15:09:06 <ubottu> Launchpad bug 1006398 in Ubuntu "Bypassing ptrace restrictions for errors from hanging applications" [Undecided,New]
15:09:08 <ev> creating an API to map crash signatures to fixed versions of binary
15:09:09 <ev> packages. I need to write an LEP and largely do the work myself, but they'll
15:09:11 <ev> provide guidance.
15:09:12 <ev> - Merged Dmitrijs' usb-creator branch to clear partitions rather than entire
15:09:14 <ev> disks. Uploaded a new usb-creator.
15:09:15 <ev> - Webops got paged that the retracers are having ongoing connectivity problems
15:09:17 <ev> to Cassandra. It looks like we have a load problem somewhere in the system,
15:09:18 <ev> so I've dedicated a day to getting Cassandra state monitoring via JMX,
15:09:20 <ev> jmxtrans, and Graphite up and running:
15:09:21 <ev> https://rt.admin.canonical.com//Ticket/Display.html?id=53325
15:09:23 <ev> - Helped review Martin's merge to ubiquity that moves from jockey driver
15:09:23 <barry> cjwatson: more or less, yes.  the only parts we care about are the httplib2 injections (hence the py3 httplib2 bug discovered), and whatever else is necessary to make that work in the context of lazr.*  i have a branch, probably not pushed yet.
15:09:24 <ev> installation to ubuntu-drivers:
15:09:26 <ev> https://code.launchpad.net/~pitti/ubiquity/ubuntu-drivers-common/+merge/107744
15:09:27 <ev> - Started adding Graphite reporting to the API calls in
15:09:28 <ev> http://errors.ubuntu.com.
15:09:29 <ev> (done)
15:09:42 <jodh> * boot/upstart:
15:09:42 <jodh> - working on stateful re-exec blueprint.
15:09:42 <jodh> - updated stateful re-exec Design doc.
15:09:42 <jodh> - discussions with cjwaton and slangasek regarding upgrade scenarios.
15:09:45 <jodh> - added basic JSON serialisation/deserialisation for sessions.
15:09:48 <jodh> * archive/mountall: testing changes prior to first upload for bug 805509.
15:09:50 <ubottu> Launchpad bug 805509 in mountall (Ubuntu Precise) "mountall(8) man page does not list any of the 9 available options." [Wishlist,Triaged] https://launchpad.net/bugs/805509
15:09:52 <jodh> ┋
15:09:54 <barry> cjwatson: (still fiddling with the insanity of wsgi on py3 though, so it may not yet be completely right)
15:09:55 <jodh> 
15:09:59 <cjwatson> barry: *nod*
15:10:09 <slangasek> barry: +1 to your post to ubuntu-devel asking if we can prune jockey from the python3 dependency list ;)
15:10:20 <barry> slangasek: less is more :)
15:10:57 <slangasek> ev: if errors.ubuntu.com will hide problems it believes are fixed, how do we find out that it's wrong in this belief?
15:11:20 <ev> slangasek: indeed, I don't want to make it ever go away completely
15:11:36 <ev> but there may be value in splitting that into two views
15:11:39 <ev> or hiding it by default
15:11:50 <ev> we'll see how it goes though - I'm not definitely saying this will happen
15:12:07 <xnox> infinity is not here?!...
15:12:19 <cjwatson> he's connecting I think
15:12:37 <slangasek> ev: right - just want to know what that view is and make sure we're monitoring it in some fashion :)
15:12:44 * doko looks at the bar
15:12:51 <ev> slangasek: understood
15:12:53 <ogra_> xnox, he mailed iirc
15:13:28 <ogra_> being at connect it is a decision between bar and meeting ... ;)
15:14:11 <slangasek> jodh: oh cool, you're on top of those mountall merge requests then?  I'll clear those out of my queue
15:14:49 <slangasek> no infinity today, he's excused for it being 11pm in Hong Kong
15:15:00 <slangasek> as doko would've been, but apparently he chose IRC over beer :)
15:15:13 <slangasek> * short week, memorial day holiday in the US
15:15:13 <slangasek> * merge: autoconf, anacron (which now has upstart support in Debian)
15:15:13 <slangasek> * fix makedev (bug #1001460)
15:15:13 <slangasek> * SRU update-notifier for a regression caused by fixing the inclusion of translations, bug #1003100
15:15:15 <ogra_> or has beer with IRC
15:15:16 <slangasek> * SRU processing
15:15:17 <ubottu> Launchpad bug 1001460 in makedev (Ubuntu Quantal) "preinst fails in d-i" [Critical,Fix released] https://launchpad.net/bugs/1001460
15:15:18 <slangasek> * ported python-pam (not yet landed)
15:15:19 <ubottu> Launchpad bug 1003100 in update-notifier (Ubuntu Quantal) "package-data-downloader: KeyError: 'paquetes'" [High,Fix committed] https://launchpad.net/bugs/1003100
15:15:21 <slangasek> * reviewing merge proposals from xnox
15:15:23 <slangasek> ..
15:15:39 <jodh> slangasek: I'm testing your mountall changes now.
15:15:53 <stgraber> - Networking
15:15:53 <stgraber> - Spent some time testing candidates SRU for an IPv6 bug in Network Manager
15:15:56 <stgraber> - Pushed some bugfixes in isc-dhcp to quantal
15:15:59 <stgraber> - Prepared and uploaded SRU of isc-dhcp to 12.04
15:16:01 <stgraber> - SRU
15:16:04 <stgraber> - Pushed the next LXC SRU, containing quite a few bugfixes and reducing the delta with Quantal.
15:16:07 <stgraber> - Sponsored libgcrypt11 for stokachu
15:16:09 <stgraber> - Containers
15:16:09 <slangasek> jodh: hmm, the ones I staged in the branch?
15:16:12 <stgraber> - Pushed some improvements upstream, did some more work on the API design
15:16:15 <stgraber> - Kernel SRU validation for apparmor bugfix required by LXC in 12.04
15:16:17 <stgraber> - Implemented workaround for conflict with dnsmasq by shipping a dnsmasq.d config file
15:16:20 <stgraber> - ISO tracker
15:16:23 <stgraber> - Finished user interface for the testcase management changes
15:16:25 <stgraber> - Fixed the API and updated python-qatracker, unit test says it's all good :)
15:16:28 <stgraber> - Did some other minor improvements to the DB schema
15:16:30 <slangasek> jodh: perhaps I should push the rest of my changes here - I was planning 2.37 to go straight to Debian unstable
15:16:31 <stgraber> - Fixed remaining bits of the admin UI
15:16:33 <stgraber> - Weekly meeting with QA, got some more changes to do
15:16:36 <stgraber> - Other
15:16:38 <stgraber> - Setup fortnightly 12.04.1 team meeting, starting tomorrow at 14:00 UTC in #ubuntu-meeting
15:16:41 <stgraber> - Discussed API key policy with pastebin.com, ported to new API and SRUed to everything
15:16:44 <slangasek> jodh: and it's just been blocked because we need plymouth updated in Debian first
15:16:44 <stgraber> - Updated a bunch of bugs to go have the required SRU fields
15:16:47 <stgraber> - TODO this week
15:16:49 <stgraber> - Continue with ISO tracker changes based on the list from QA and work items
15:16:52 <stgraber> - Monitor current SRUs and get them all tested as soon as they hit -proposed
15:16:55 <stgraber> - Prepare for alpha1
15:16:56 <jodh> slangasek: ok - i'll hold fire.
15:16:58 <stgraber> (DONE)
15:17:35 <xnox> == ubiquity-lvm-luks ==
15:17:35 <xnox> * Working with mpt on the design
15:17:36 <xnox> * LINK https://wiki.ubuntu.com/Ubiquity/AdvancedPartitioningSchemes
15:17:36 <xnox> * Prototyping / testing changes in ubiquity
15:17:37 <xnox> == foundations-q-btrfs-requirements ==
15:17:37 <xnox> * btrfs-tools sponsored!
15:17:38 <xnox> * needs discussion about backporting to precise (not sure about
15:17:38 <xnox> kernel version dependencies, as precise has 3.2)
15:17:39 <xnox> == other ==
15:17:39 <xnox> * e2fsprogs sponsored!
15:17:40 <xnox> * going to debconf, organised travel
15:17:40 <xnox> * shepherding boost1.49 transition
15:17:41 <xnox> * uploading gcc 4.7 patches to debian (~7)
15:17:41 <xnox> * replying to merge request reviews...
15:17:47 <slangasek> stgraber: did you get my note about Debian bug #673490?
15:17:49 <ubottu> Debian bug 673490 in bridge-utils "serious doubts about /lib/udev/bridge-network-interface" [Important,Open] http://bugs.debian.org/673490
15:17:56 <doko> slangasek, I'll catch up later =)
15:18:02 <slangasek> doko: heh
15:18:10 <xnox> short...
15:18:29 <stgraber> slangasek: yes, will try to find some time to look at it/reply later this week
15:18:36 <doko> xnox, thanks for the gcc work
15:18:37 <slangasek> stgraber: ok - just
15:18:39 <ogra_> xnox, nah, stokachu was short :)
15:18:46 <stokachu> i win
15:18:47 <slangasek> stgraber: ok, just making sure you got it and I didn't mis-send
15:19:09 <ogra_> xnox, and he worked harder than you ! (he said so !!!)
15:19:19 <stokachu> hah
15:19:29 <bdmurray> bug triage of initramfs-tools package install failures
15:19:29 <bdmurray> updated bug-bot to handle live media package install failures with a full /cdrom
15:19:31 * xnox is shunned again ^_^
15:19:32 <bdmurray> ran bug-bot against old apport-package install failures on old releases
15:19:34 <bdmurray> bug-bot expand logging messages to be more verbose about all the actions taken
15:19:42 <bdmurray> modifications to bug-bot to deal with ubiquity bug reports differently
15:19:42 <bdmurray> fixed https://wiki.ubuntu.com/X/Bugs/UpdateManagerWarningForI8xx url in the ubuntu wiki which update-manager links to
15:19:45 <bdmurray> wrote a bug pattern for update-notifier bug 1003100 regarding paquetes
15:19:47 <ubottu> Launchpad bug 1003100 in update-notifier (Ubuntu Quantal) "package-data-downloader: KeyError: 'paquetes'" [High,Fix committed] https://launchpad.net/bugs/1003100
15:19:47 <bdmurray> merge proposal for grub2 with modification to apport package hook
15:20:09 <bdmurray> consolidation of man-db driver installation on Live Media bug reports into bug 985636
15:20:11 <ubottu> Launchpad bug 985636 in man-db (Ubuntu) "package man-db 2.5.7-4 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,Incomplete] https://launchpad.net/bugs/985636
15:20:12 <bdmurray> investigation into apport-collect failure on precise - LP: #1004029
15:20:15 <bdmurray> email to ubuntu-devel regarding migration-assistant
15:20:17 <bdmurray> patch piloting
15:20:20 <bdmurray> Stable Release Updates team training
15:20:26 <bdmurray> ⌁ done ⌁
15:20:51 * xnox is the 'paquetes' bug due to switching default locale to french?
15:20:59 <bdmurray> heh
15:21:01 <slangasek> no
15:21:05 <ogra_> only for the desktop though
15:21:06 <slangasek> that would be paquets
15:21:34 <slangasek> this was because the translation teams for ALL the languages on the Iberian subcontinent were ambitious and translated a variable name that wasn't expected to be translated ;)
15:21:39 * xnox should check upon / verify the grub2 sru
15:22:06 <ogra_> heh
15:22:27 <slangasek> (and python's Template routine bails with eparse on a variable it can't substitute)
15:23:22 <slangasek> any other questions about any of the above?  concerns about the switch to French locale and keyboard? :)
15:23:24 * barry hides
15:23:59 * slangasek ponders shutting down bug reporting against the resolvconf package entirely
15:24:21 <slangasek> can we have a bug pattern that matches the package name and just redirects to stgraber's blog? :)
15:24:23 * xnox uses US layout on the UK keyboards. I think I will survive fr locale, as I already fight all the 'helpful' geo-location keyboard suggestions.
15:24:38 <slangasek> us dvorak alternate international ftw
15:24:55 <barry> slangasek: on an mbp keyboard
15:25:00 <slangasek> nooooo
15:25:00 <xnox> slangasek: lucky me, I used resolvconf for like ~2 years now =)))) I was happy about the switch.
15:25:17 <slangasek> xnox: I'm not sure how you managed while the package in Ubuntu was in such a state, but ok :)
15:25:22 <slangasek> [TOPIC] Bugs
15:25:23 * ogra_ never got the hang of dvorak ... its so much work to scrape off all the letters off the keys and add them in a different order
15:26:09 * xnox ogra_ you know that you can pull out the key with the letter sticker and rearrange ? =)
15:26:22 <bdmurray> bug 985636
15:26:23 <ogra_> nah, thats cheating !
15:26:26 * xnox not so much on my Microsoft Ergonomic keyboard
15:26:27 <ubottu> Launchpad bug 985636 in man-db (Ubuntu) "package man-db 2.5.7-4 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,Incomplete] https://launchpad.net/bugs/985636
15:26:50 <bdmurray> looks to me like people were trying to install nvidia in the live environment after installing ubuntu
15:27:04 <stokachu> Needs sponsor: * http://launchpad.net/bugs/974054 (oneiric is affected as well)
15:27:05 <cjwatson> is there any way we can make these bugs stop landing on man-db?
15:27:06 <ubottu> Launchpad bug 974054 in isc-dhcp (Ubuntu Oneiric) "dhcpd attempts to use /var/run/dhcpd.pid, AppArmor errors" [Undecided,Confirmed]
15:27:08 <stokachu> Needs guidance for getting bug more attention: * http://launchpad.net/bugs/520386 (mir: https://bugs.launchpad.net/ubuntu/+source/netcf/+bug/904014)
15:27:11 <ubottu> Launchpad bug 520386 in libvirt (Ubuntu Lucid) "libvirt-bin hypervisor does not support virConnectNumOfInterfaces / unable to create domain with virt-manager using network bridge" [Low,Confirmed]
15:27:11 <cjwatson> they're basically never its fault
15:27:12 <ubottu> Launchpad bug 904014 in netcf (Ubuntu) "[MIR] netcf" [Undecided,In progress]
15:27:29 <bdmurray> these what?
15:27:46 <cjwatson> all the gazillion bugs of the general shape of 985636
15:28:03 <cjwatson> a bunch of them were bugs in some package management programs, which I recall fixing a cycle or two ago
15:28:03 <bdmurray> right so any package install failure shouldn't be about man-db?
15:28:18 <slangasek> man-db has a trigger, which is why it bears the brunt
15:28:35 <cjwatson> right, the same goes for other packages with triggers and I don't know that a simple blacklist is right
15:28:40 <ogra_> yeah, we should probably make that silent or dont say the package name in the output
15:28:43 <stgraber> stokachu: I'll look at the oneiric part of bug 974054
15:28:45 <ubottu> Launchpad bug 974054 in isc-dhcp (Ubuntu Oneiric) "dhcpd attempts to use /var/run/dhcpd.pid, AppArmor errors" [Undecided,Confirmed] https://launchpad.net/bugs/974054
15:28:49 <cjwatson> we want to somehow figure out how to identify the actual package at fault
15:28:58 <stokachu> stgraber: cool thanks
15:29:00 <cjwatson> the problem is that due to aforementioned package manager bugs the dpkg terminal log is empty
15:29:05 <slangasek> heh, yes
15:29:21 <slangasek> can we just sunset those buggy versions of the package manager?  or SRU?
15:29:30 <cjwatson> I'm kind of at the point where I'd say that if the terminal log is that useless (only started/ended entries) and there's nothing else of much use, we should just unreportable the bug
15:29:30 <slangasek> that particular bug was reported against Ubuntu 10.10
15:29:53 <xnox> stokachu: the netcf / libvirt bugs affect me. Anything needed there in particular?
15:30:07 <slangasek> cjwatson: maybe it would be better to intercept and report it against the package manager itself, so we know if there are problems with log breakage?
15:30:15 <cjwatson> quite possibly
15:30:17 <stokachu> xnox: someone recently created a ppa for lucid so probably needs some testing
15:30:20 <cjwatson> bug 680328 is the one I remember
15:30:22 <ubottu> Launchpad bug 680328 in qapt (Ubuntu Maverick) "Many postinst scripts fail using either PackageKit, or QApt" [Undecided,New] https://launchpad.net/bugs/680328
15:30:29 <cjwatson> we never did get round to SRUing that to maverick, and can't now
15:30:30 <stokachu> xnox: i know serge is swamped but this bug is a high visibility one
15:30:48 <slangasek> bdmurray: ^^ do you want to take the action to get package install reports with broken logs reported against the package manager?
15:31:02 <bdmurray> and we might be able to find out which package manager from the history log file - if it has data
15:31:09 <cjwatson> so we could just bin any such bugs from 10.10; >=11.04 would be something different
15:31:10 <bdmurray> slangasek: yes
15:31:29 <slangasek> [ACTION] bdmurray to get package install reports with broken logs reported against the package manager instead of a random package with a trigger
15:31:29 * meetingology bdmurray to get package install reports with broken logs reported against the package manager instead of a random package with a trigger
15:31:33 <slangasek> bdmurray: thanks
15:32:21 <bdmurray> regarding the bug in question - is it possible to try and install nvidia on live media anymore?
15:33:17 <slangasek> stgraber: 974054> thanks for taking
15:33:22 <slangasek> bdmurray: is that meant to be blocked?
15:34:06 <slangasek> I guess you'd get inconsistent results with nvidia installed in the live env, since you can't update the initramfs
15:34:20 <slangasek> and would therefore get the drm driver loaded at next boot for plymouth
15:34:21 <stgraber> slangasek: it's roughly the same fix I uploaded to quantal and precise, just didn't think oneiric was also affected.
15:34:23 <bdmurray> slangasek: I believe so, I'll go back and look again
15:34:56 <slangasek> cjwatson, ev, stgraber: any of you recall if we're blocking nvidia driver installation in the live env?
15:35:19 <ev> not offhand
15:35:19 <cjwatson> well, there's no way to stop you *trying* to apt-get install it; I don't recall what happens though ...
15:35:20 <bdmurray> That's what I remember and email thread on ubuntu-devel with pitti and bryce talking about it
15:35:48 <slangasek> well, I guess we don't want the GUI to facilitate this bit of foot shooting :)
15:36:14 * xnox as a user, jockey didn't run & didn't offer to install those in the livecd. Only on installed system. But yeah people follow 'guides' and 'tutorials' and install it byhand.
15:36:29 <slangasek> bdmurray: I guess maybe it's worth you following up with them then, since no one here knows for sure
15:36:45 <slangasek> bdmurray: did you have more bugs?  Let's look at the two stokachu mentioned and then we can circle around to any more you have for us
15:36:45 <bdmurray> http://ubuntu.5.n6.nabble.com/Installing-drivers-on-USB-sticks-td727548.html
15:36:50 <stokachu> who is the goto person for nfs?
15:36:53 <bdmurray> slangasek: okay
15:37:24 * xnox slowly hides from stokachu
15:37:38 <slangasek> stokachu: I generally tend to the client side; bryce happens to look after the server side with a community hat on; but xnox can be your point of contact for nfs going forward ;)
15:37:55 <ogra_> stokachu, a hybrid between xnox (filesystems) and stgraber (networking) .... "xngraber"
15:38:02 <stokachu> sweet :D
15:38:19 <stokachu> nothing atm but im working an autofs case i may speak to one of you about
15:38:20 <xnox> what is the question about nfs?
15:38:35 <slangasek> whimper, autofs
15:38:48 <slangasek> we still have shutdown ordering bugs in that package, unless something's changed recently
15:38:54 <stokachu> yes... 1500+ mounts all hanging on autofs reload
15:39:03 <slangasek> lovely
15:39:07 <ogra_> ugh
15:39:16 <slangasek> yeah, that's one to discuss out of band :)
15:39:18 <stokachu> exactlyyy
15:39:28 <cjwatson> speaking of autofs, does anyone fancy merging our autofs5 changes into the new autofs source package in Debian?
15:39:46 <cjwatson> (well, "new", it used to be in lucid before the rename to autofs5, but it looks like now it's been renamed back
15:39:49 <cjwatson> )
15:40:22 <xnox> [ACTION] xnox to merge autofs
15:40:22 * meetingology xnox to merge autofs
15:40:27 <slangasek> xnox: thanks :)
15:40:30 <ogra_> wow, brave !
15:41:02 * xnox is that a suicide or something. I'm not core-dev so one of you will have to review it =))))
15:41:05 <cjwatson> cool, thanks
15:41:11 * xnox laughs
15:42:06 <stokachu> would firefox be considered a desktop or foundations area
15:42:16 <slangasek> ogra_: what's the worst that can happen if he breaks every filesystem package in the archive during his probation period? ;)
15:42:20 <slangasek> stokachu: desktop
15:42:21 <bdmurray> what? desktop!
15:42:22 <xnox> we used to have a separate team for it
15:42:52 <slangasek> bug #520386
15:42:52 <stokachu> ok cool b/c this is the ugliest FF bug
15:42:54 <ubottu> Launchpad bug 520386 in libvirt (Ubuntu Lucid) "libvirt-bin hypervisor does not support virConnectNumOfInterfaces / unable to create domain with virt-manager using network bridge" [Low,Confirmed] https://launchpad.net/bugs/520386
15:43:12 <slangasek> so that one needs a core-dev for an SRU, by the look of it
15:43:19 <cjwatson> yeah, over to desktop with an extra hard kick
15:43:37 <slangasek> it looks like it needs fixed in quantal first though, according to the bug state?
15:43:40 <ogra_> slangasek, we hire him and assing all these packages to him forever ?
15:43:46 <ogra_> *assign
15:43:53 <slangasek> stokachu: is 520386 fixed in quantal?
15:44:12 <stokachu> slangasek: i don't believe so ill need to double check; progress has been moving very slow on it
15:44:14 <slangasek> ogra_: ah, the sisyphus approach to performance management
15:44:21 <ogra_> :)
15:44:34 <stokachu> is serge in here?
15:44:46 <ogra_> doesnt look like
15:44:52 <slangasek> stokachu: nope - he's server team; you'll probably want to touch base with him directly then
15:44:56 <ogra_> (if you mean hallyn)
15:45:04 <slangasek> stokachu: once there's a fix in quantal we can certainly help shepherd it through for SRUing
15:45:09 <stokachu> ah, yea serge hallyn, ok ill ping him
15:45:12 <stokachu> slangasek: ok good deal
15:45:34 <stokachu> ogra_: whats his nick on irc?
15:45:38 <slangasek> stokachu: 'hallyn'
15:45:41 <ogra_> hallyn
15:45:43 <slangasek> he's on #ubuntu-devel
15:45:48 <ogra_> and -server
15:45:54 <stokachu> ok ill pop over there after meeting
15:45:55 <ogra_> and sometimes in -kernel
15:46:10 <slangasek> stokachu: any other bugs that want discussing?
15:46:25 <slangasek> stokachu: congrats on the progress on the NSS+ldap+gcrypt nightmare bug, btw :)
15:46:44 <stokachu> slangasek: thanks -- could not have done it without Mr Chu :)
15:47:07 <stokachu> slangasek: oh i did get an email from someone who i think is upstream maintainer about it, ill forward it to you
15:47:53 <slangasek> stokachu: ok cool
15:48:05 <stokachu> slangasek: sent
15:48:05 <slangasek> bdmurray: how about from your side... any other bugs?
15:48:33 <bdmurray> bug 349469 has some recent debugging data in it
15:48:35 <ubottu> Launchpad bug 349469 in debconf (Ubuntu) "debconf: DbDriver "config": /var/cache/debconf/config.dat is locked by another process: Resource temporarily unavailable" [Medium,Triaged] https://launchpad.net/bugs/349469
15:50:32 <slangasek> yes
15:50:59 <slangasek> the common denominator seems to be aptd
15:51:39 <slangasek> but it doesn't make sense why the dpkg lock doesn't get in the way first
15:52:02 <slangasek> anyone have any theories how this could happen?
15:52:20 <slangasek> if not, I can follow up with bdmurray to try to formulate a plan of attack for debugging
15:53:00 <cjwatson> Is the dpkg lock applied before preconfiguration?
15:53:15 <slangasek> ah, of course it wouldn't be
15:53:29 <cjwatson> haha
15:53:31 <cjwatson> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=469354
15:53:32 <slangasek> and apt only checks the dpkg lock, it doesn't set one of its own?
15:53:33 <ubottu> Debian bug 469354 in debconf "debconf: dpkg-reconfigure should acquire dpkg lock" [Normal,Open]
15:53:59 <slangasek> bdmurray: woo, triaged ;)
15:54:21 <cjwatson> but when I looked at the recent process tree, I didn't see any other apt/dpkg/debconf-like processes
15:54:22 <xnox> in apt there is a window of oportunity when no apt/dpkg lock exist, as a user found out, between the two resolution loops you can e.g. start synaptic and create a deadlock.
15:54:29 <xnox> not sure if it's the same.
15:54:37 <slangasek> cjwatson: which ones were you looking at?
15:54:37 <cjwatson> one package management process does not a lock collision make?
15:54:38 * xnox this maybe old info
15:54:56 <slangasek> cjwatson: the process trees are being provided by the users after hitting the bug
15:54:59 <cjwatson> xnox: simply starting synaptic wouldn't take the debconf lock
15:55:11 <slangasek> so the collided process has exited non-zero by that point
15:55:14 <cjwatson> ah
15:55:20 <xnox> cjwatson: true, you need to launch a synaptic action for example.
15:55:29 <cjwatson> xnox: which would take the apt lock
15:55:33 <cjwatson> modulo that debconf bug
15:55:47 <cjwatson> Still, all that would do is move the lock collision somewhere else
15:55:51 <cjwatson> It would probably still show up
15:55:55 * xnox will try to reproduce my old deadlock, didn't see it happen in a while
15:56:07 <cjwatson> This isn't a deadlock, it's just a failure to acquire the lock
15:57:40 <xnox> ok.
15:59:06 <slangasek> followed up to the bug
15:59:29 <slangasek> cjwatson, bdmurray: thanks, nice to have a prospect finally of fixing that one
15:59:41 <cjwatson> heh, snap
15:59:47 <cjwatson> except you linked to the Ubuntu bug with the same number
15:59:53 <slangasek> heh, sorry
16:00:13 <slangasek> I typed 'Debian' in my head
16:00:23 <slangasek> bdmurray: anything else?
16:00:34 <bdmurray> no that's good for me
16:00:38 <slangasek> ok, thanks
16:00:45 <slangasek> [TOPIC] Python porting virtual sprint
16:01:12 <slangasek> as I mentioned earlier, we'll be doing a 3-day virtual sprint next month for python porting
16:01:26 <slangasek> June 11-13
16:01:40 <slangasek> so now you have it mentioned somewhere logged ;)
16:02:48 <slangasek> I'll send email along soon-ish with the exact plans for the hours we'll be working... but it should be less of a shift for everyone compared with the last sprint we did, because we'll run two groups with 4-hour overlap instead of moving everyone to the dead center
16:03:00 <slangasek> [TOPIC] AOB
16:03:02 <slangasek> anything else?
16:03:09 <ev> https://code.launchpad.net/~ev/whoopsie/libwhoopsie/+merge/107562 - would love some extra eyes on that one
16:03:16 <slangasek> I was just going to mention that actually :)
16:03:20 <ev> it's just C, it won't bite
16:03:35 <slangasek> ev: why use c99 instead of $(CC)?
16:03:58 <ev> slangasek: $(CC) with the c99 option?
16:04:04 <slangasek> if you need to
16:04:06 <cjwatson> Yeah, you should always use $(CC) for x-compiling support
16:04:07 <ev> I do
16:04:19 <slangasek> then yeah, especially now that there's a lib, $(CC) would be a good idea
16:04:20 <ev> need to use c99 features, that is
16:04:26 <cjwatson> (not that whoopsie wouldn't need more work for that)
16:04:44 <slangasek> do you actually have to invoke gcc with -std=c99 to make use of c99 features?
16:04:46 <cjwatson> no multiarch support?
16:04:57 <xnox> [ACTION] slangasek to send an email about Python sprint
16:04:57 * meetingology slangasek to send an email about Python sprint
16:05:01 <slangasek> I thought you only have to do that to *disallow* code that's *not* c99-compliant
16:05:02 <ev> ah yes - I hadn't considered that. I was more thinking to hell with people trying to run it under clang
16:05:14 <cjwatson> we don't care about that for the rest of the distro ;-)
16:05:19 <slangasek> right :)
16:05:42 <ev> slangasek: if I run it without, it cries about variables defined in loop constructs
16:05:44 <ev> if memory serves
16:06:10 <slangasek> anyway, I'll try to send some more feedback on that merge request... would be good if others could as well
16:06:15 <ev> thanks!
16:06:18 <slangasek> it's just the addition of a shared library in C, it won't bite ;)
16:06:21 <slangasek> anything else?
16:06:21 <jodh> ev: I'm up for reviewing that too.
16:06:24 <ev> and I will gladly return the favor
16:06:43 <xnox> [ACTION] more reviews for ev's merge proposal
16:06:43 * meetingology more reviews for ev's merge proposal
16:06:50 <xnox> [LINK] https://code.launchpad.net/~ev/whoopsie/libwhoopsie/+merge/107562
16:06:53 <cjwatson> if you haven't already, you should nm -D the shared library to make sure only the symbols you absolutely need are exported
16:06:57 <ev> so do subscribe the team to merges. I'll try to get to yours quickly, so you're not left waiting to land branches.
16:06:59 * xnox meetingology fail
16:07:21 <cjwatson> and in answer to your earlier question, you should link whoopsie dynamically against libwhoopsie rather than (effectively) shipping two copies
16:07:24 <slangasek> [LINK] https://code.launchpad.net/~ev/whoopsie/libwhoopsie/+merge/107562
16:07:30 <slangasek> hmmph
16:07:43 <xnox> LINK https://code.launchpad.net/~ev/whoopsie/libwhoopsie/+merge/107562
16:07:52 <cjwatson> try #link
16:07:53 <ev> cjwatson: ahhh, noted, thanks! dynamically link> thanks bunches for clarifying that. I couldn't find a good answer elsewhere.
16:07:59 <slangasek> #link https://code.launchpad.net/~ev/whoopsie/libwhoopsie/+merge/107562
16:08:09 <cjwatson> there's basically hardly ever a good reason to link statically
16:08:29 * xnox so we test audio & irc bots during team meetings
16:08:36 <ev> *cough* Go
16:08:42 <slangasek> #endmeeting