[00:58] <attente> robru: possibly, i'd test it, but my system's a bit borked atm
[00:59] <attente> robru: actually looks like it well is
[00:59] <attente> thanks for the ping
[00:59] <robru> attente, no hurry, I'm using libreoffice draw instead of dia anyway ;-)
[00:59] <robru> attente, but if you get time I can help you test it
[01:00] <attente> robru, thanks for the offer, don't worry though, i'll take care of it :)
[01:01] <robru> attente, great, thanks ;-)
[05:42] <pitti> Good morning
[06:16] <darkxst> pitti, which parts of systemd are missing with the proposed logind migration?
[06:16] <darkxst> i.e does journal work? and I am guessing the system unitdir does not exist though?
[06:19] <pitti> darkxst: the init bits are missing, as well as the journal
[06:19] <pitti> darkxst: the systemd unit dir already exists, but it won't be used
[06:21] <darkxst> right, I think gnome-shell mainly just uses the session tracking, but gdm might be interesting
[06:21] <pitti> darkxst: I went through that yesterday
[06:21] <pitti> gdm is also just using logind
[06:22] <darkxst> ok cool
[06:30] <pitti> darkxst: btw, do you know about http://codesearch.debian.net/ ?
[06:30] <pitti> darkxst: this is just plain awesome to answer questions like that
[06:31] <pitti> e. g. http://codesearch.debian.net/search?q=org.freedesktop.login1+package%3Agdm3
[06:32] <Quintasan> Good morning
[07:07] <darkxst> pitti, nope hadnt seen that before, thanks :)
[07:47] <jibel> good morning
[07:49] <didrocks> salut jibel!
[07:49] <jibel> bonjour didrocks
[07:59] <m4n1sh> mvo: can you please review the software sources merge proposal?
[08:03] <chrisccoulson> ooh, this is becoming more and more regular now: https://launchpadlibrarian.net/133901866/buildlog_ubuntu-precise-i386.firefox_20.0~b4%2Bbuild1-0ubuntu0.12.04.1~mfn1_FAILEDTOBUILD.txt.gz
[08:33] <mvo> m4n1sh: checking it out now, thanks for the reminder
[08:34] <seb128> hey desktopers
[08:34] <seb128> mvo \o/ wie gehts?
[08:34] <seb128> mvo, good to see you are still around ;-)
[08:34] <mvo> seb128: can't leave my friends alone :)
[08:35] <mvo> I guess its ok to upload software-properties ? freeze and all that?
[08:36] <seb128> mvo, upload away, it might end up being in the freeze queue for a day or two but that's ok
[08:36] <mvo> hrm, bzr out of sync but looks easy enough to fix
[08:37] <mvo> seb128: cool, will do, thanks
[08:39] <seb128> mvo, doko ... not surprised
[08:41] <mvo> lol
[08:44] <seb128> mvo, some things don't change :p
[08:44] <mvo> I was thinking exactly the same ;)
[09:04] <Laney> morning
[09:37] <Laney> Anyone else finding that shutdown/restart just brings them back to lightdm instead of actually doing the action requested?
[09:38] <seb128> wfm
[09:38] <seb128> maybe a logind/ck sideeffect
[09:38] <seb128> you played with that right?
[09:39] <Laney> indeed
[09:39] <Laney> pitti: have you seen that?
[09:40] <pitti> hey Laney
[09:40] <Laney> hey
[09:40] <pitti> Laney: right, I'm fixing that right now
[09:41] <Laney> ah, excellent
[09:41] <pitti> Laney: logind currently asks systemd to switch to shutdown.target, which doesn't work without systemd
[09:41] <pitti> Laney: can you confirm you are running gnome-session from the PPA?
[09:41] <pitti> Laney: raring's should talk to ConsoleKit
[09:41] <Laney> yes, 3pitti1
[09:41] <pitti> Laney: I'll add a fallback to "shutdown" and pm-utils
[09:41] <pitti> systemd drops pm-utils completely, but I think that ought to be discussed first, so I'll add it back
[09:49] <Laney> probably for the best
[10:00] <pitti> ah, that's working nicely
[10:42] <m4n1sh> mvo: this merge req https://code.launchpad.net/~manishsinha/software-properties/add-update-after-repo-add/+merge/150185
[11:09] <pitti> Laney: FYI, updated systemd is in the PPA now
[11:09] <pitti> Laney: that should work with shutdown, reboot, suspend, hibernate
[11:09] <Laney> cool, I'll try it when it gets published
[11:09] <pitti> it is
[11:09] <Laney> ;)
[11:12] <Laney> so it is! /me dist-upgrades
[11:14] <pitti> I'm not sure if dist-upgrading will restart logind properly; can you have a look while you are at it?
[11:14] <pitti> (i. e. at the output, whether you see some upstart message about restarting logind)
[11:14] <pitti> Laney: ^
[11:14] <Laney> I don't see anything like that
[11:15] <pitti> ah, indeed
[11:15] <pitti> Laney: so "sudo restart systemd-logind" it is for you
[11:15] <pitti>         dh_installinit --no-start --name systemd-logind --upstart-only
[11:15] <pitti> that might have something to do with it :)
[11:18] <Laney> seems to restart me properly \o/
[11:25] <pitti> Laney: fixed my local package to auto-restart logind on upgrades now
[11:25] <pitti> Laney: I also tested suspend, that works, too
[11:25] <pitti> Laney: btw, do you have consolekit installed still?
[11:26] <Laney> yes; I thought I'd run as it will be in the transition period
[11:26] <pitti> right
[11:26] <pitti> Laney: does user switchign work for you?
[11:26] <pitti> (from the indicator)
[11:26]  * Laney tries
[11:27] <pitti> I have consolekit purged to see the fallout, and the options went away
[11:27] <pitti> locking screen and clicking "switch user" there, works
[11:27] <Laney> yes, seems to
[11:27] <pitti> I suppose the indicator talks to CK directly; so that's on my list to port
[11:27] <Laney> to guest account and back
[11:27] <Laney> loginctl lists four sessions now
[11:27] <Laney> three with seats
[11:27] <pitti> Laney: thanks; so that makes another undeclared dependency :) but at least it's on the "to port list" in the spec
[11:28] <pitti> Laney: you, lightdm, and guest?
[11:28] <Laney> yes - guest twice (once with seat, once without)
[11:28] <pitti> the others should be in state "closing", if you run "loginctl show-session c2"
[11:28] <pitti> on the actual session numbber, of course
[11:28] <Laney> yes
[11:28] <pitti> ok
[11:29] <Laney> well, one closing one online
[11:29] <pitti> that doesn't directly hurt, but it's a bit of a wart (on my list, but I want to fix actual regressions first)
[11:29] <Laney> ooer
[11:29] <pitti> "ooer"?
[11:29] <Laney> if I switch back to guest again I'm asked for its password
[11:29] <pitti> oh, your guest session is running still?
[11:29] <Laney> yes, I user switched back to laney
[11:30] <Laney> perhaps that's a pre-existing bug
[11:30] <pitti> Laney: yeah, didn't transition gnome-screensaver yet (I think that's it)
[11:30] <seb128> Laney, oh, you uploaded indicator-weather, can I complain to you about how broken that package is? :p
[11:30] <Laney> no :P
[11:30] <pitti> but either way, I'll see how things communicate there; I guess gnome-session stops telling CK about "switchign user" as we already migrated that
[11:30] <seb128> :-(
[11:30] <Laney> I added a session job to make its breakage less onerous
[11:31] <seb128> I wonder how it made it to the archive
[11:31] <seb128> the postinst is doing stuff like
[11:31] <seb128> xdg-icon-resource install --theme hicolor --novendor --size 22 /usr/share/indicator-weather/media/icon.png weather-indicator
[11:31] <seb128> wth
[11:31] <Laney> O_O
[11:31] <seb128> yeah, wth!
[11:31] <seb128> echo "Fixing incomplete weather icons..."
[11:31] <seb128> if [ ! -e "/usr/share/icons/gnome/22x22/status/weather-clouds.png" ] && \
[11:31] <seb128>     [ !  -L "/usr/share/icons/gnome/22x22/status/weather-clouds.png" ]; then
[11:31] <seb128>     ln -s /usr/share/icons/gnome/22x22/status/weather-few-clouds.png /usr/share/icons/gnome/22x22/status/weather-clouds.png
[11:31] <seb128>  
[11:31] <seb128> ...
[11:31] <Laney> quick, find out who sponsored it originally!
[11:32]  * Laney didn't look into the maintainer scripts
[11:32] <Laney> I just thought it would be an interesting case for a job as it tends to randomly crash
[11:32] <seb128> https://launchpad.net/~andrewsomething
[11:33] <seb128> seems to be responsible for those postinst hacks
[11:33] <seb128> Laney, yeah, I like the upstart job idea, I don't like the postinst in there :p
[11:33] <seb128> it was added in https://launchpad.net/ubuntu/+source/indicator-weather/11.04.10+repack-0ubuntu1
[11:34]  * seb128 goes to nuke that postinst and file some bugs
[11:38]  * Laney eyes the guest account's indicator-datetime and gsd gobbling a core each
[11:41] <seb128> Laney, cross ppas is going to be fun
[11:42] <seb128> Laney, ubuntu-desktop has versions of those made to use systemd-services rather than the old g-s-d's interface (which is dropped in the same ppa)
[11:42] <Laney> yeah ...
[11:42] <Laney> do I have ubuntu-desktop?
[11:42] <seb128> Laney, if you got g-s-d from logind's ppa it will probably break stuff
[11:42] <seb128> Laney, dpkg -l | grep gnome-settings-daemon?
[11:42] <Laney> I should probably disable some of this stuff :p
[11:42] <Laney> doing
[11:42] <seb128> Laney, dpkg -l | grep indicator-datetime?
[11:42] <seb128> shrug
[11:42] <Laney> it was a rhetorical question really
[11:43] <seb128> the indicator-weather do a "ln -s /usr/share/bla /usr/share/bla"
[11:43] <Laney> archive versinos of both of those
[11:43] <seb128> e.g symlink on itself
[11:43] <seb128> head -> desk
[11:43] <seb128> Laney, ok, dunno then, but it's likely to bite a few people, I hope we can get the MIR reviewed soon and stuff to land in the archive
[11:44] <Laney> both bts have g_main_context_find_source_by_id at the top
[11:45] <Laney> anyway I have some weird system so I won't worry too much until this stuff lands properly
[11:50] <Laney> ah
[11:50] <Laney> some apparmor denials in dmesg
[11:50] <Laney> [ 1884.171060] type=1400 audit(1363089002.042:1468943): apparmor="DENIED" operation="mkdir" parent=4087 profile="/usr/lib/lightdm/lightdm/lightdm-guest-session-wrapper" name="/run/user/123/dconf/" pid=4656 comm="indicator-datet" requested_mask="c" denied_mask="c" fsuid=123 ouid=123
[11:50] <pitti> Laney: yep, jsut fixed upstream
[11:50] <pitti> (in lightdm)
[11:50] <Laney> how timely
[11:51] <pitti> Laney: the BP has a work item for logind support for lightdm user switching now, so at least that's recorded, too
[11:52] <pitti> (it had before, but I clarified it)
[13:10] <pitti> Laney: FYI, if you change something in debian/control.in, you should run debclean
[13:10] <pitti> Laney: (just changing g-control-center deps, and your "Remove Recommends on gnome-online-accounts" is only becoming active now)
[13:36] <xnox> seb128: is there a new icon for the "Software & Updates" (aka software-sources ?!)
[13:36] <seb128> xnox, not that I know, you can try asking JohnLea
[13:44] <ogra_`> xnox, there was a discussion mpt triggered on the ubuntu-art ML
[13:45] <ogra_`> (not sure what the outcome was though)
[13:45] <ogra_`> something about parachutes :)
[13:45]  * xnox slowly backs away....
[13:45] <xnox> ... it's just an icon, it's sole purpose is to be pretty
[13:46] <ogra_`> parachutes can be pretty :)
[14:03] <xnox> pitti: gnome-power-manager the ver. string used is 3.6.0-1svn1 is that on purpose or accidental?
[14:05] <pitti> xnox: on purpose, as I committed the fix to debian svn and then uploaded that
[14:05] <pitti> xnox: i. e. that shouldn't stop autosyncs, and shoudl be an indication that the next debian upload doesn't need merging
[14:06] <xnox> pitti: right, well I'm about to spoil the fun with a -ubuntu1 https://code.launchpad.net/~dedunumax/ubuntu/raring/gnome-power-manager/bug-923292/+merge/148190
[14:07] <xnox> it's a trivial patch, not accepted upstream yet, but it's part of papercuts project.
[14:07] <xnox> trivial fixes to improve usability.
[14:07] <Laney> pitti: really? I don't see the dep here
[14:07] <Laney> I probably forgot to commit it though ...
[14:08] <pitti> xnox: well, it wasn't sent upstream :)
[14:08] <pitti> xnox: perhaps you could attach the patch https://bugzilla.gnome.org/show_bug.cgi?id=678115 while you're at it?
[14:08] <ubot2> Gnome bug 678115 in gnome-power-statistics "ctrl+w/ctrl+q shortcut for closing power statistics not working" [Minor,Unconfirmed]
[14:08] <xnox> yeah, i was about to do that.
[14:08] <pitti> Laney: ah, that could be it
[14:08] <Laney> like I guess I edited control.in; bzr bd -S (which runs clean); test; upload
[14:08] <Laney> so control never got updated in the vcs
[14:09] <Laney> that's one bit of the gnome team's workflow that I don't like
[14:10] <pitti> yeah, I trained myself to run debclean whenever I change control.in
[14:11]  * Laney nods
[14:15]  * xnox likes how update-maintainer learned to update *both* control.in and control
[14:39] <mterry> dobey, why are there two software center desktop files now?  They both are showing up in the dash
[14:39] <dobey> mterry: i have no idea. i didn't add any
[14:40] <mterry> dobey, OK, will bzr blame this :)
[14:41] <dobey> mterry: hrmm, looks like both the ubuntu branded one, and the unbranded one, are installed
[14:41] <mterry> mvo, ^
[14:41] <mterry> dobey, that would be fine, but they should have OnlyShowIn or NoDisplay or something so only one is shown at a time
[14:42] <dobey> hrmm
[14:43] <dobey> mterry: what version of software-center do you have installed?
[14:43] <mterry> dobey, 5.5.5-0ubuntu1
[14:43] <mvo> I'm on my way to the train, I can check in ~1h or so if the problem is still there
[14:44] <Laney> yeah I see that too
[14:45] <dobey> weird, ok
[14:47] <dobey> mterry: file a bug and assign to me (or assign to me if one is already filed)?
[14:48] <mterry> dobey, let me check if it's filed
[14:48] <psivaa> seb128: pitti: Is workspaces being disabled the default setting in raring? workspace switcher is not attached to the launcher by default.
[14:49] <dobey> ah there is a bug
[14:49] <pitti> psivaa: yes, it is; you can reenable them in the control cetner
[14:49] <psivaa> pitti: ack, thank you
[14:51] <dobey> meh, and package imports being broken is a PITA.
[14:52] <mterry> dobey, point me at bug
[14:52] <dobey> https://bugs.launchpad.net/ubuntu/+source/software-center/+bug/1153548
[14:52] <ubot2> Launchpad bug 1153548 in software-center "Two icons in Dash" [Undecided,Confirmed]
[14:52] <mterry> ah yes
[14:52] <dobey> i'll fix it
[14:52] <mterry> dobey, thanks!
[15:21] <doomlord> can anyone comment on how hard it would be to modify expo/scale plugins
[15:22] <doomlord> i have some customizations in mind , multi-level zoom for "expo" (eg show 2x2 or 3x3) and/or would like to be able to show desktop thumbnails alongside windows in "scale"
[15:22] <doomlord> does that sound difficult to do
[15:26] <robru> doomlord, are you talking about compiz plugins? Install ccsm and you can tweak compiz plugins to your hearts desire... not "officially" recommended though
[15:27] <doomlord> i'm talking modifying source
[15:28] <doomlord> these features aren't there
[15:28] <doomlord> i've got ccsm
[15:28] <robru> doomlord, then you're on your own ;-)
[15:28] <doomlord> ok i just figure someone may have dabled with the source and could comment
[15:29] <robru> doomlord, oh yeah, tons of developers on here... but I've never touched compiz source myself, just used ccsm
[15:29] <doomlord> last time i tried.. i found out that compiz & ubuntu repos are different :)
[15:29] <doomlord> couldn't get it to compile & run
[15:30] <qengho> Meeting in one hour yesno?
[15:30] <Laney> indeed
[15:30] <robru> qengho, yes I believe it's in an hour. I was in a panic to wake up just now only to discover that DST has shifted the meeting an hour forward...
[15:51] <Sweetshark> launchpad, deal with it: "finding revisions to fetch 1/598878" ...
[15:58] <czajkowski> Sweetshark: stop breaking LP :)
[16:01] <Sweetshark> czajkowski: Ill do it very tenderly!
[16:04]  * mlankhorst hands Sweetshark a tenderizer
[16:27] <doomlord> the fact i couldn't even compile compiz plugins before isn't encouraging
[16:28] <doomlord> but i was talking to someone who didn't know the differences in the ubuntu repositories
[16:28] <doomlord> There are many ways to acheive what i want
[16:30] <seb128> hey
[16:30] <seb128> it's meeting time!
[16:30] <robru> hey hey!
[16:30] <didrocks> hey!
[16:31] <seb128> Sweetshark, qengho, chrisccoulson, RAOF, didrocks, Laney, kenvandine, mlankhorst, cyphermox_, mterry, robru, tkamppeter, attente, desrt: hey
[16:31] <Laney> greetings
[16:31]  * Sweetshark reporting in.
[16:31] <seb128> https://wiki.ubuntu.com/DesktopTeam/Meeting/2013-03-12
[16:32]  * kenvandine waves
[16:33] <mlankhorst> g'day mate
[16:33] <robru> seb128, you are all an hour late, what is going on here? ;-)
[16:33] <qengho> Hah
[16:33] <seb128> lol
[16:33] <seb128> let's get started
[16:34] <seb128> Sweetshark, hey, anything interesting you are working on? any output for vUDS? ;-)
[16:34] <Sweetshark> seb128: 4.0.1 in raring breaking the installers. fix is in.
[16:35] <seb128> \o/ for 4.0.1 (not for breaking the installer)
[16:35] <seb128> thanks for the quick followup fix though ;-)
[16:35] <Sweetshark> output for vUDS? like the next one? I could bore you with some upstream progress ...
[16:35] <seb128> no, the one from last week
[16:35] <seb128> for those who had sessions
[16:36] <seb128> I don't think you had any on libreoffice so no worry ;-)
[16:36] <Sweetshark> other than that trying to corner the remaining unity integration issues and work on tinderboxes and more automated testing.
[16:36] <seb128> but it would be good to summarize the things decided for those who had sessions, took on new work
[16:37] <seb128> Sweetshark, thanks
[16:37] <qengho> Hello my Pretties!
[16:37] <qengho> * New stable released yesterday. Averaging about 1 every 5 days lately.
[16:37] <qengho> * chromium-browser autopkgtests. built-ins and maybe webapps!
[16:37] <qengho> * add search reporting for canonical's accountants and lawyers. Search -> $$$.
[16:37] <qengho> * armhf bugginess.
[16:37] <qengho> * will learn all I can about various webkit upstreams to prepare for packaging.
[16:37] <qengho> EOF
[16:37] <seb128> hey qengho ;-)
[16:38] <seb128> updated chromium \o/
[16:38] <seb128> good work there!
[16:38] <Sweetshark> seb128: no, vUDS had no libreoffice specifics, except for heart attack by the 15g-turns in a oiltanker that is libreoffice-packaging affected by 'lets get rolling' -- 'nah, lets keep stuff as is' ...
[16:38] <qengho> Evefry damn week.
[16:38] <seb128> they need to keep you busy... ;-)
[16:38] <seb128> qengho, thanks
[16:39] <seb128> chrisccoulson, hey
[16:39] <Laney> he got swallowed by a webkit debug build
[16:39] <chrisccoulson> hah
[16:39] <chrisccoulson> - Made a lot of progress with Chromium crashing on arm (lots of intense v8 debugging last week)
[16:39] <chrisccoulson> - Firefox / Thunderbird updates
[16:39] <chrisccoulson> - Fixed an issue where certain test harness failures in Firefox (eg, the testsuite hanging) weren't showing up as failures in Jenkins, despite skipping half of the test suite
[16:39] <chrisccoulson> - Running tests in an almost proper session now (with gnome-session, gnome-settings-daemon and a WM) - this resulted in additional failures!
[16:39] <chrisccoulson> - Cleaning up some of the monster Jenkins integration patches as well
[16:39] <chrisccoulson> EOF
[16:40] <seb128> chrisccoulson, still not at the bottom of the chromium issue?
[16:40] <seb128> that's a though one
[16:40]  * qengho weeps.
[16:41] <chrisccoulson> seb128, i've got a pretty good understanding of what's going on now, but i had to put it to one side yesterday to take care of other daily activities
[16:41] <seb128> k
[16:41] <seb128> chrisccoulson, thanks ;-)
[16:41] <seb128> good work!
[16:42] <seb128> didrocks, hey
[16:42] <didrocks> hey hey
[16:42] <didrocks> * growing tests on daily release landing and now classic maintenance of dailies with some power outage forcing to fix/workaround issues in the datacenter.
[16:42] <didrocks> * 100 scopes: strike 2, the return of the son. We are going to land it to raring. A PPA is under preparation with all this in (still moving) code. The current plan is to land those ~ 40 packages by end of next week. We'll need tester, I think kenvandine, robru and seb128 (+ popey's) help would be appreciated for installing the ppa once ready and report issues (only file and application lens will work
[16:42] <didrocks> at the beginning). Also note that we'll have to NEW them (seb128? :p) and MIR them (mterry? ;)). I'll give more info on people testing once the ppa is ready.
[16:42] <didrocks> * Adapated daily release to release to a ppa. Not as straightforward as it seems as we need to take into account that distro is a moving target as well. Preparing things to ensure that the testers are not screwed.
[16:42] <didrocks> * created WI for landing all components from where we are upstream for for kenvandine, cyphermox_, mterry, robru, and vorlon to look for. We need to have all that ready by the EOM in raring, under daily releases.
[16:42] <didrocks> adapted*
[16:43]  * Laney fears the 100 scopes landing
[16:43] <Laney> is that 100 source packages?!
[16:43] <seb128> Laney, going to be fun indeed
[16:44] <seb128> Laney, 40 packages from what didrocks just wrote
[16:44] <didrocks> Laney: no, only 40 :)
[16:44] <Laney> oh yes
[16:44] <didrocks> see, just saved 60!
[16:44] <robru> Laney, trust me, 100 new source packages is better than dozens of things packaged in a single source package. we have a situation like that with webapps and any time you want to fix one little bug in one little thing, it's a whole new giant source package mess.
[16:44] <seb128> didrocks, nice summary, lots going on ... I'm surprised the IRC server didn't kick you out for spamming ;-)
[16:44] <didrocks> seb128: yeah, I wondered at some point to write a line and then another :p
[16:45] <didrocks> Laney: btw, you just volonteered for reviewing the FFe
[16:45] <didrocks> kthxbye
[16:45] <seb128> didrocks, thanks
[16:45] <seb128> Laney, hey
[16:45] <Laney> hur hur
[16:45] <Laney> A probably incomplete list of things I worked on lately - GNOME updates (various syncs, merges and plain upstream updates) - Release management work now that Feature Freeze has started (reviewing FFes; if you think you still want to land a feature then talk to me / the release team) - vUDS & associated discussions around rolling releases
[16:45] <Laney>  - Looking at uptart in the user session and trying to write soeme sample jobs. I don't see a clear list of targets in the BP(s); will try to come up with one myself. - Trying out pitti's logind packages and finding / being confused by bugs that pitti inevitably already knew about when I got around to IRC-reporting them. :-)
[16:45] <Laney> \0
[16:45] <Laney> OH
[16:45] <Laney> irssi broke my paste, sorry
[16:46] <attente> !v
[16:46] <ubot2> Factoid 'v' not found
[16:46]  * popey quickly books holidays far away from scopes
[16:46] <didrocks> popey: toooo late, I talked to your manager ;)
[16:46] <popey> bah!
[16:46] <popey> looking forward to it.. ☺
[16:46] <popey> \o/ new stuff
[16:47] <seb128> Laney, @upstart session: yeah, we didn't really do the work to list target, I can help looking at that this week if you want
[16:47] <seb128> Laney, thanks
[16:47] <Laney> would be good
[16:47] <Laney> np
[16:47] <seb128> kenvandine, hey
[16:47] <kenvandine>  * signon-ui for Qt5 that doesn't break the desktop
[16:47] <kenvandine>    - mardy is working on webkit1 bindings for Qt5
[16:47] <kenvandine>    - I have a backported patch for qt5 that adds xembed support, I'll land that after I get some testing done
[16:47] <kenvandine>  * Gwibber/Friends
[16:47] <kenvandine>    - We really want to replace gwibber on the desktop with Friends, requires a few features to land to get the new Gwibber client fully functional (this week) then we'll request a FFe
[16:48] <kenvandine> and i'll be out tomorrow
[16:48] <kenvandine>  /done
[16:48] <seb128> enjoy the day off!
[16:48] <seb128> kenvandine, thanks
[16:48] <seb128> mlankhorst, hey
[16:48] <mlankhorst> touch bug https://bugs.freedesktop.org/show_bug.cgi?id=56578 - xorg 1.14 input testing, plymouth/kernel/lightdm/xorg-server race nailing with help of tjaalton
[16:48] <ubot2> Freedesktop bug 56578 in Server/Input/Core "race condition with active/passive grabs when opening menus with touch" [Normal,Assigned]
[16:48] <mlankhorst> and trying to make unbind for drm work correctly
[16:49] <mlankhorst> ^D
[16:50] <seb128> mlankhorst, thanks
[16:50] <seb128> cyphermox_, hey
[16:51] <cyphermox_> hey!
[16:51] <cyphermox_> so just now been fixing up two MIRs for a merge that hasn't landed yet for indicator-session (bugfix)
[16:51] <cyphermox_> spent part of last week fixing up patches for upstream inclusion, they all got landed, for nm-applet
[16:52] <cyphermox_> updated nm/nm-applet on raring, left with the vpn plugins to updates but will need FFE
[16:52] <cyphermox_> other than that, blocked on a hug libcolumbus transition for which code isn't quite ready.
[16:52] <cyphermox_> \x00
[16:52] <cyphermox_> hug=hud
[16:52] <mterry> boo
[16:52] <seb128> ok
[16:52] <mterry> hoped for a pkg called hug
[16:52] <seb128> lol
[16:52] <seb128> cyphermox_, thanks
[16:52] <cyphermox_> there's hugin
[16:52] <Laney> Maintainer: Daniel Holbach
[16:53] <mterry> heh
[16:53] <seb128> ;-)
[16:53] <cyphermox_> oh, nice :)
[16:53] <seb128> mterry, hey
[16:53] <mterry> - Light week between vUDS and a day off
[16:53] <mterry> - Some MIR work
[16:53] <mterry> - Filed branch for start of LightDM support in phablet greeter, waiting a review and me adding tests
[16:53] <mterry> EOF
[16:53] <seb128> mterry, thanks
[16:53] <seb128> robru, hey
[16:53] <robru> hey hey!
[16:54] <robru> * implemented since_id= for twitter, vastly reducing bandwidth usage by avoiding redownloading duplicate tweets, also fixed a bug where friends was re-notifying heinously stale tweets
[16:54] <robru> * bring up dbus interface earlier during init, solving some dbus timeout issues
[16:54] <robru> * vast reduction in complexity of db schema, making it way easier for consumers of friends to filter the db based on common criteria (this will make facebook and twitter phone apps much simpler)
[16:54] <robru> * added GPS location info to db schema, allowing us to leverage location data provided to us by social networks. currently only implemented for foursquare but I intend to add facebook soon.
[16:54] <robru> * put together a sweet flowchart of the friends stack architecture: http://ubuntuone.com/2W6HSyoN9tWUcAZL0k7Z1f (btw Sweetshark, used libreoffice draw for this, pretty slick tool! I used to use Dia but Draw is way nicer).
[16:54] <robru> ^D
[16:54] <seb128> nice
[16:54] <seb128> robru, thanks
[16:54] <seb128> tkamppeter, hey
[16:55] <tkamppeter> UDS session about printing: Mobile with reduced, driverless (support for IPP Everywhere/CUPS Raster, PDF, PS, PCL) printing stack, Poppler as only renderer, no GS, if possible daemons on-demand, print dialog in Touch SDK
[16:55] <tkamppeter> https://blueprints.launchpad.net/ubuntu/+spec/client-1303-printing-stack-with-mobile-in-mind
[16:55] <tkamppeter> system-config-printer: Printer driver auto-download working correctly from both GNOME's own printer setup tool and the system-config-printer GUI
[16:55] <tkamppeter> Updates for Feature Freeze: HPLIP 3.13.3, foo2zjs 20130306, foomatic-db 20130308, cups-filters 1.0.30, QPDF 4.0.1. Planned is also Ghostscript 9.07, waiting for update of LCMS, see LP: #1126427, also CUPS 1.6.2 to fix LP: #1069671
[16:55] <ubot2> Launchpad bug 1126427 in lcms2 (Ubuntu) "LCMS2 needs multi-threading fixes to work with the new Ghostscript 9.07" [High,New] https://launchpad.net/bugs/1126427
[16:55] <ubot2> Launchpad bug 1069671 in cups (Ubuntu Quantal) "no print queues displayed in pure client mode" [Medium,Triaged] https://launchpad.net/bugs/1069671
[16:56] <tkamppeter> Had to rush in all new versions due to real existance of FF instead of RR came on short notice.
[16:56] <tkamppeter> One day after UDS.
[16:56] <seb128> tkamppeter, lot of good work there
[16:56] <seb128> yeah, it has been an issue for several of us
[16:57] <tkamppeter> And still not ready, GS and CUPS still to come, but CUPS is a bug fix release, so not related to FF.
[16:57] <seb128> tkamppeter, we should discuss, at some point this week, on whether we want to keep the GNOME printing panel or go back to system-config-printer for raring
[16:57] <seb128> I will ping you about it tomorrow
[16:57] <seb128> (need to run into another meeting that this one)
[16:58] <seb128> tkamppeter, thanks
[16:58] <seb128> attente, hey
[16:58] <attente> seb128, hey
[16:58] <attente> not much from me, spent the last week fighting with autopilot, getting integration tests into unity-gtk-module, and doing a bit more rhythmbox profiling
[16:58] <attente> just going to fix a unity-gtk-module bug robru pointed out to me last night, but afterwards planning to switch focus to the ibus indicator
[16:58] <attente> EOF
[16:58] <seb128> \o/
[16:58] <tkamppeter> seb128, for me seems that s-c-p gets another cycle, as there is no GUI to share the local printers for getting accessed byu remote machines.
[16:58] <seb128> attente, nice coincidence, I planned to talk to you about the keyboard indicators today, to see if you would be interested working on those
[16:59] <seb128> attente, let's chat about it a bit later
[16:59] <tkamppeter> Receiving shares (at least Bonjour from CUPS 1.6.1 servers) is always on currently.
[16:59] <attente> seb128, sure
[16:59] <seb128> tkamppeter, yeah, I think we should go back to s-c-p as well
[16:59] <seb128> attente, thanks
[16:59] <seb128> desrt, hey
[17:00] <seb128> not desrt it seems
[17:00] <seb128> so me
[17:00] <seb128> - hosted a client track for vUDS
[17:00] <seb128> which took some 2 days
[17:00] <seb128> - updated some icons for design (which involved some patching and migrations)
[17:01] <seb128> - some desktop updates
[17:01] <seb128> - tested the systemd-services transition, filed a FFe for it
[17:01] <seb128> - looked at some bugs, SRU
[17:01] <seb128> - tried to update our blueprints, postponed some work, still way behind the trendline
[17:02] <ivanka> hey didrocks
[17:02] <seb128> especially after vUDS and getting some new specs

[17:02] <ivanka> didrocks, how do I see blueprint bits that are assigned to me?
[17:02] <desrt> seb128: hi
[17:02] <desrt> sorry.  was mid-rebase :)
[17:02] <seb128> ivanka, launchpad.net/~/+upcomingwork
[17:02] <seb128> ivanka, https://launchpad.net/~seb128/+upcomingwork
[17:02] <seb128> ups
[17:02] <seb128> ivanka, https://launchpad.net/~/+upcomingwork
[17:03] <didrocks> ivanka: you probably want to see the WI assigned to you as well: http://status.ubuntu.com/ubuntu-raring/u/ivanka.html (well 1 WI it seems)
[17:03] <seb128> desrt, no worry ... we do team round and summary of what we did in the week ;-)
[17:03] <ivanka> cheers seb128 and didrocks
[17:03] <desrt> seb128: let me know if you want to come back to me :)
[17:03] <seb128> desrt, please go ahead, we were just done
[17:03] <seb128> I was last and just finished my update
[17:04] <desrt> did a lot of small items related to the datetime stuff
[17:04] <desrt> did a systemd upstream patch to add the CanNTP property, did one for our (old) version of systemd too
[17:04] <desrt> also patched upstream g-c-c to use this property
[17:04] <desrt> did some MIR paperwork
[17:04] <desrt> and started on writing the service for the dconf security proxy stuff (for app confinement)
[17:05] <desrt> just working on some glib stuff now (fixing test failures noted by pitti's jenkins stuff)
[17:05] <desrt> (eof)
[17:05] <seb128> desrt, thanks
[17:06] <seb128> desrt, speaking about glib, you missed the recent release rounds it seems, is that wanted? ;-)
[17:06] <seb128> current tarball is almost a month old
[17:06] <desrt> seb128: mclasen and i fight over who does the releases sometimes :p
[17:06] <desrt> but thanks for noticing.  i'll do one soon.
[17:06] <seb128> thanks
[17:06] <seb128> ok, any other topic from anyone?
[17:06] <seb128> comments/questions/...?
[17:06] <desrt> seb128: gtk 3.8 for 13.4?
[17:07] <seb128> desrt, nop, missed that shot
[17:07] <desrt> fair enough
[17:07] <desrt> got the new harfbuzz/pango stuff though, right?
[17:07] <seb128> we are past FFe and I can't see a reason to push for it
[17:07] <seb128> yes
[17:07] <desrt> good
[17:07] <seb128> new glib as well
[17:07] <desrt> that will be a big memory win
[17:07] <seb128> new GTK would have been nice, but it has still some known regressions not fixed
[17:07] <desrt> theme stuff mostly, right?
[17:08] <seb128> no, dnd (due to the clock work), screensaver turning your screen white
[17:08] <seb128> the current tarball makes gedit segfault when you scroll (that's fixed in git though)
[17:08] <desrt> neat.
[17:08] <desrt> ya... enough reasons to wait :)
[17:09] <seb128> right
[17:09]  * desrt has been tracking rawhide... some things are kinda sketchy there
[17:09] <seb128> other comments? ;-)
[17:09] <cyphermox_> about rawhide? j/k
[17:09] <didrocks> robru: mterry: cyphermox_: think about starting on the touch apps packaging for raring
[17:10] <didrocks> ask me anything if you see the WI don't make sense :)
[17:10] <cyphermox_> didrocks: is the hud stuff ready?
[17:10] <robru> didrocks, yes, I have some questions
[17:10] <didrocks> cyphermox_: not yet, but please, feel free to check directly with tedg
[17:10] <didrocks> robru: sure?
[17:10] <robru> didrocks, but I'm not awake yet. how much longer will you be online for?
[17:11] <didrocks> robru: I give you 50 minutes for a good coffee! :)
[17:11] <robru> didrocks, hehe, yeah. let me just think up some questions then I will ask
[17:11] <didrocks> sure ;)
[17:13] <seb128> ok, meeting wrap I guess, thanks everyone
[17:13] <didrocks> thanks
[17:14] <Laney> thanks \o/
[17:14] <desrt> seb128: so FFe is approved
[17:14] <desrt> does this mean you're going to start landing the changeset?
[17:15] <desrt> or are we waiting for jbicha to do testing?
[17:16] <seb128> desrt, we are waiting for
[17:16] <seb128> 1- the MIR to be acked
[17:16] <seb128> 2- the beta1 freeze to end (that's this week)
[17:16] <seb128> 3- then we can upload
[17:17] <Laney> you can still upload
[17:17] <Laney> everything critical ought to be blocked from migration
[17:17] <desrt> seb128: MIR was acked, no?
[17:26] <robru> didrocks, ok, I sent you an email with a dozen questions. ;-)
[17:27] <didrocks> robru: got it
[17:28] <kenvandine> robru,  the schema changes and the qml-friends changes, lets propose them for merging into "raring" branches so we can do reviews there
[17:28] <kenvandine> robru, and block on the FFe for landing in trunk
[17:28] <robru> kenvandine, I'm not really familiar with that process, can you elaborate a bit?
[17:29] <kenvandine> robru, branch trunk to a raring branch
[17:29] <kenvandine> then propose your branches to that instead of trunk
[17:29] <kenvandine> so we have a place to do reviews without the merger landing them in trunk
[17:29] <robru> kenvandine, so you want the branch to be lp:~super-friends/friends/raring ?
[17:29] <kenvandine> yeah
[17:29] <kenvandine> things we are queuing to land in raring, pending a FFe approval
[17:30] <kenvandine> if that gets approved, we'll merge into trunk
[17:30] <robru> kenvandine, funny, a lot of what you are telling me I was just asking didrocks about in the email I sent him ;-)
[17:30] <kenvandine> haha
[17:30] <kenvandine> i was just discussing that with him too :)
[17:30] <didrocks> yeah, you are spamming me! :)
[17:30] <robru> kenvandine, this situation has been on my mind for a few days :-P
[17:31] <kenvandine> mine too
[17:32] <robru> kenvandine, ok, raring branch created, will make new MP now
[17:32] <kenvandine> thanks
[17:33] <robru> kenvandine, ok, so if I understood you correctly then this should be what you asked for: https://code.launchpad.net/~robru/friends/simple-schema/+merge/152971
[17:34] <robru> didrocks, there are other questions in that email though ;-)
[17:35] <kenvandine> i love this:
[17:35] <kenvandine> While we're at it, why are likes being stored as a double?
[17:35] <kenvandine> That doesn't even make sense. Go home Gwibber, you're drunk.
[17:35] <robru> ;-)
[17:35] <kenvandine> haha
[17:35] <didrocks> robru: as I'm giving more info on the email, let me CC kenvandine :p
[17:36] <didrocks> answered
[17:37] <robru> didrocks, thanks
[17:37] <didrocks> yw!
[17:37] <didrocks> seeing my current state of tireness, I hope I was clear :)
[17:37] <didrocks> if not, do not hesitate to harass me :)
[17:37] <robru> harass harass!
[17:38] <didrocks> :)
[17:40] <robru> kenvandine, so that branch I just made at lp:~super-friends/friends/raring, is that equivalent to lp:friends/raring ?
[17:41] <kenvandine> no
[17:41] <robru> kenvandine, is there a way to link those? to save on typing ;-)
[18:02] <robru> kenvandine, are you reviewing my branch still?
[18:03] <kenvandine> robru, i reviewed it, just about to test it out
[18:04] <robru> kenvandine, it's good. keep an eye out for your foursquare messages, which will now be geotagged in the model ;-)
[18:05] <robru> kenvandine, when you get around to implementing that in the qml frontend, you can take the location name and linkify it to point at google.com?q=lat,long and then people can click on it to get a map. ;-)
[18:05] <kenvandine> nah... we'll just display a map in the UI :)
[18:05] <robru> kenvandine, that too ;-)
[18:06] <kenvandine> robru, you should update the changelog in your branch
[18:07] <robru> kenvandine, oh, what? I thought the autolanding did that for me.
[18:08] <kenvandine> now i am confused... i think it kind of does
[18:08] <robru> didrocks, help ^^ ;-)
[18:08] <kenvandine> didrocks has yelled at folks for not changing it
[18:09] <kenvandine> i've been updating the changelog :)
[18:09] <didrocks> hum
[18:10] <didrocks> sorry? what branch?
[18:10] <didrocks> you bumped something?
[18:10] <robru> didrocks, so I've got a massive diff I want to merge into our special raring branch we just created
[18:10] <robru> didrocks, but I didn't touch the changelog
[18:10] <robru> didrocks, am I supposed to be writing my own changelog entries?
[18:11] <robru> I mean debian/changelog
[18:13] <didrocks> robru: you are not forced to if you didn't bump the version
[18:13] <didrocks> robru: but you should link to a bug :)
[18:13] <robru> didrocks, but these are new features, they don't have bugs
[18:13] <didrocks> robru: well, if it's a new interesting feature, you should put it in the changelog then if you don't want a bug
[18:13] <robru> didrocks, ok, will do
[18:13] <didrocks> robru: it's as you wish, for important things to mention, either bug or direct writing
[18:13] <didrocks> bug will populate for you
[18:13] <didrocks> otherwise, put it yourself :)
[18:16] <robru> kenvandine, ok, pushed updated changelog with new features listed.
[18:17] <kenvandine> thanks
[18:17] <seb128> desrt, sorry, was in a meeting, no, systemd's MIR has not been approved yet, it's being reviewed by the security team
[18:18] <desrt> seb128: i meant the FFe was approved
[18:18] <desrt> i think i misunderstood what you said was pending approval
[18:18] <seb128> desrt, right, and I said we need the MIR approved before we can upload ;-)
[18:18] <seb128> e.g we need systemd-services in main
[18:19] <desrt> right
[18:19] <seb128> ok, all good ;-)
[18:35] <robru> kenvandine, ugh, just noticed my debian/changelog changes conflict with trunk, do you want me to merge it or did you do it alreayd?
[18:39] <kenvandine> i didn't
[18:39] <kenvandine> merge it from trunk
[18:39] <kenvandine> your branch is great btw!
[18:41] <robru> kenvandine, thanks. just merged and pushed
[18:41] <robru> kenvandine, I mean pushed to the same feature branch for the mp. please merge the mp ;-)
[18:54] <robru> kenvandine, I'm gonna add location support to facebook, do you want it in the same MP or should I start a new MP for that?
[18:54] <kenvandine> new one
[18:55] <kenvandine> this will get merged in a few minutes
[18:55] <robru> kenvandine, ok, great
[19:03] <achiang> cyphermox_: just kicked off an nm-applet run using your dbg packages and collecting logs
[19:03] <achiang> cyphermox_: i'll come back tomorrow after i reproduce :)
[19:04] <cyphermox_> awesome
[19:04] <cyphermox_> here today it'll never work, so I'm on something else
[19:04] <cyphermox_> will get to reproducing that bug on thursday
[19:22] <kenvandine> robru, i sent you a MP for qml-friends... not sure how you feel about reviewing Qt stuff :)
[19:44] <robru> kenvandine, great, I do want to get more into Qt ;-)
[20:09] <doomlord> do apple ptents mean 'scale' in linux has to wobble (the way it interpolates position & velocity)
[20:18] <mterry> robru, didn't you make a wiki page to list common inlining steps?  I want to check it over again
[20:19] <robru> mterry, yeah, lemme find it again
[20:20] <robru> mterry, here it is: https://wiki.ubuntu.com/DailyRelease/InlinePackaging didrocks had moved it ;-)
[20:23] <mterry> robert_ancell, heyo!  I wanted to poke you about some pending reviews: https://code.launchpad.net/~mterry/unity-greeter
[20:23] <robert_ancell> mterry, consider me poked
[20:23] <mterry> robert_ancell, jpds wanted us to SRU the gsettings and dbus interface ones if we could
[20:28] <robert_ancell> mterry, the ones with jenkins failure I didn't change the main status - I'll let you fix those parts
[20:29] <mterry> robert_ancell, oh, hm.  I assume they're bogus..  will check
[20:40] <mterry> fginther, for unity-greeter autolanding, I assume I need a jenkins login to restart a jenkins job?  I don't recall if I have one, but there doesn't seem to be a signup option
[20:43] <fginther> mterry, let me check
[20:43] <mterry> fginther, this is on jenkins.qa.ubuntu.com
[20:44] <fginther> mterry, jenkins.qa.ubuntu.com doesn't execute any jobs, it's basically just a viewer
[20:45] <fginther> mterry, i don't have a login there either
[20:46] <fginther> mterry, is something not merging? The usual process to retrigger autolanding is to approve the MP.
[20:46] <mterry> fginther, ah...  How is are the links that jenkins leaves in comments like "to restart this job..." supposed to work?
[20:46] <fginther> mterry, ah that.
[20:46] <mterry> fginther, oh fair.  I was just thinking I'd actually do what jenkins suggested  :)
[20:47] <mterry> robert_ancell, I no longer have permissions to set the merges Approved...  :(  But doing so will retry the jenkins runs
[20:47] <fginther> mterry, that's a bug
[20:47] <fginther> it might even have a number ...
[20:48] <robert_ancell> mterry, oh really?
[20:48] <fginther> mterry, https://bugs.launchpad.net/jenkins-launchpad-plugin/+bug/1149232
[20:48] <ubot2> Launchpad bug 1149232 in jenkins-launchpad-plugin "link to retrigger a job is wrong when publishing jobs to jenkins.qa.ubuntu.com" [Undecided,New]
[20:48] <robert_ancell> mterry, oh, it's probably because you're not in the PS team right?
[20:49] <robert_ancell> mterry, I just invited the canonical desktop team, but it looks like someone has to ack that
[20:50] <seb128> robert_ancell, ok, accepted it
[20:50] <robert_ancell> mterry, try now
[20:50] <seb128> did I just sign up for extra work for our team? :p
[20:53] <robert_ancell> seb128, totally. I just unsubscribed ps :)
[21:17] <marga> desrt, you are not around, are you?
[21:17]  * marga in the wrong tz.
[21:18] <marga> Running 'sudo dconf update' is causing cinnamon to die with sigabort.  I'm trying to find out what is causing that, but with not much luck up to now.
[21:20] <marga> I get this funny message: GLib:ERROR:/build/buildd/glib2.0-2.32.3/./glib/gvariant-serialiser.c:1320:g_variant_serialised_n_children: code should not be reached
[21:20] <marga> It's always funny when something prints stuff like "this should never happen"
[21:54] <desrt> marga: hi.
[21:55] <desrt> marga: almost definitely a refcounting problem somewhere in cinnamon
[21:55] <desrt> marga: a backtrace would be very highly useful
[21:55] <desrt> marga: valgrind would probably also find the issue
[21:57] <marga> desrt, I've been struggling with the backtrace
[21:58] <marga> I have it, it's big, but gdb doesn't show anything useful when bt
[21:58] <desrt> probably you need to install some dbgsyms packages
[21:59] <marga> Maybe, but i'm used to at least seeing _something_
[21:59] <marga> In this case it's such a level of nothingness that I think something else is wrong
[21:59] <desrt> paste?
[22:00] <marga> http://paste.ubuntu.com/5609068/
[22:01] <marga> desrt, ^
[22:01] <desrt> cool
[22:01] <desrt> pretty useless
[22:01] <marga> indeed
[22:02] <marga> I think it's because it's an abort instead of segfault
[22:02] <desrt> nah... gdb is perfectly capable of getting reasonable data out of aborts
[22:02] <marga> it's aborting because glib's g_assert_not_reached ();
[22:02] <desrt> it's more likely because memory is getting trashed
[22:02] <desrt> does valgrind say anything useful?
[22:03] <marga> uhm... let's try that.
[22:07] <marga> well, valgrind complained about so much stuff that it's hard to find the real one
[22:07] <desrt> ya... so start with the first one
[22:07] <desrt> this is C
[22:07] <desrt> once things start going off the rails you can't really trust anything else after that
[22:07] <desrt> ...including gdb backtraces, as it were
[22:07] <marga> yeah... But this is gtk
[22:07] <marga> valgrind always complains about gtk stuff
[22:08] <desrt> are you looking at leaks or memory errors?
[22:08] <desrt> leaks ya... but it ought not to be complaining about invalid reads/writes
[22:08] <marga> I just did a valgrind run without params
[22:08] <marga> It's complaining a lot about Conditional jump or move depends on uninitialised value(s)
[22:08] <desrt> can you paste the whole log?
[22:08] <marga> And also about invalid reads
[22:09] <marga> I don't have it all... It went over my scroll limit.  I'm going to re-run this outputting to a file
[22:09] <desrt> thanks
[22:13] <marga> http://paste.ubuntu.com/5609093/
[22:13] <marga> desrt, that's a big paste for you :)
[22:13] <desrt> the zlib ones are probably harmless
[22:13] <desrt> ==7250== Conditional jump or move depends on uninitialised value(s)
[22:13] <desrt> ==7250==    at 0x51222CF: ??? (in /usr/lib/libmuffin.so.0.0.0)
[22:13] <desrt> ==7250==    by 0xAB99258: g_settings_get_mapped (in /usr/lib/x86_64-linux-gnu/libgio-2.0.so.0.3200.3)
[22:13] <desrt> ==7250==    by 0x5121CFA: meta_prefs_init (in /usr/lib/libmuffin.so.0.0.0)
[22:13] <desrt> ==7250==    by 0x511F500: meta_run (in /usr/lib/libmuffin.so.0.0.0)
[22:13] <desrt> this is the interesting one for sure
[22:14] <desrt> see if you can get dbgsyms for libmuffin and have another go at it
[22:14] <marga> Uhm, I think I'll need to rebuild it.
[22:14] <desrt> why?
[22:15] <desrt> is this something you custom-built locally?
[22:15] <marga> sort of.  It's the ppa, downloaded, built and signed
[22:15] <marga> So, it's the same code, but re-built internally, for security reasons
[22:16] <desrt> neat.
[22:16] <desrt> on one hand that seems excessively paranoid
[22:16] <marga> heh :)
[22:16] <desrt> but on the other hand it's like... if you have the resources to pull that off, why not?
[22:16] <desrt> and clearly you do
[22:16] <marga> Well, I'm going to rebuild this and everything... I hope it doesn't take long
[22:17] <mlankhorst> desrt: kind of useless if you can sneak your evil thing into the source code though ;)
[22:18] <marga> well, we are supposed to review the changes before rebuilding
[22:18] <marga> but yeah, it can go wrong.
[22:18] <mlankhorst> *cough* openssh-server
[22:19] <marga> indeed
[22:19] <marga> I was thinking the same
[22:40] <marga> Oh, how I hate cdbs...
[22:40] <marga> So, I'm rebuilding muffin, but it's getting stripped, even though I'm using the DEB_BUILD_OPTIONS=nostrip,noopt, that I've always used for this.
[22:41] <marga> And finding out why means getting into this cdbs mud :-/
[22:41] <Laney> aren't they whitespace separated?
[22:41] <marga> Laney, not according to policy
[22:41] <marga> and cdbs shouldn't require them to. But I'll try that anyway.
[22:41] <Laney> you're reading a different policy to me then ;-)
[22:43] <marga> uhm, they added this whitespace thing recently, I guess.  I'm old :)
[22:43] <marga> let's try again with whitespace...
[22:44] <Laney> not that I'm saying cdbs can't be independently buggy :p
[22:44] <marga> it might be that it's not buggy, it's just a big mess
[22:44] <marga> man dh_strip says that if "nostrip" is present, it won't strip
[22:44] <marga> Why should cdbs meddle with my env options?  It really shouldn't but...
[22:45] <marga> In any case, I do hope this is it
[22:47] <marga> yay, not stripped.  Thanks Laney.
[22:47] <Laney> woot. happy debugging!
[22:47]  * Laney goes to bed
[23:00] <marga> Backtrace: http://paste.ubuntu.com/5609198/
[23:01] <marga> Valgrind: http://paste.ubuntu.com/5609202/
[23:05] <bryce> marga, install some -dbg / -dbgsym packages ;-0
[23:07] <marga> heh
[23:07] <marga> I think I found the bug
[23:07] <marga> I mean the valgrind bug
[23:07] <marga> ... I mean, I found the bug that causes valgrind to complain.
[23:08] <marga> I'm unsure if that bug is really what's causing cinnamon to misbehave, but I'm going to do some monkey patching
[23:10] <desrt> marga: pretty obvious problem here...
[23:10] <desrt>   MetaVirtualModifier mods;
[23:10] <desrt>   if (!string_value || !meta_ui_parse_modifier (string_value, &mods))
[23:10] <desrt> ...
[23:10] <desrt>   if (mods != mouse_button_mods)
[23:10] <desrt> if string_value is NULL then 'mods' will be used uninitialised
[23:11] <desrt> which it may well be because of   string_value = g_variant_get_string (value, NULL);
[23:11] <marga> I'm looking at the code and I'm disliking the whole thing
[23:11] <marga> many things done wrong.
[23:11] <desrt> the reason you see it on 'dconf update' is simple: it broadcasts that _all_ settings may have changed
[23:11] <desrt> including ones that aren't set at all
[23:11] <desrt> they have an implicit assumption here that if a key changed they should be able to read it
[23:11] <marga> desrt, didn't we fix that?
[23:11] <desrt> marga: different pieces of software can be differently affected?
[23:12] <marga> Sorry, sorry, I'm misremembering.  We fixed gconf plugin's behavior on receiving that broadcast
[23:12] <desrt> yup
[23:12] <desrt> other software can misbehave in the same way
[23:12] <marga> indeed
[23:12] <desrt> and i guess (unfortunately for you) not too many people are using 'dconf update'
[23:12] <desrt> so these types of issues are rarely found
[23:12] <marga> right
[23:12] <desrt> anyway... i don't think this is the cause of your crash
[23:12] <desrt> but probably it's a similar issue in the same code
[23:13] <marga> We will find aaaaaall the dconf update bugs.
[23:13] <desrt> ==23835==    at 0x5138AFF: button_layout_equal (/usr/local/google/home/marga/muffin/src/core/prefs.c:1247)
[23:13] <desrt> this one is way more interesting...
[23:14] <marga> that's the one that I've just "patched" (i.e. forced it to return true)
[23:14] <marga> To see if it was the cause
[23:14] <desrt> man
[23:14] <desrt> button_layout_handler is a mess of a function
[23:15] <marga> yes
[23:15] <marga> that's the one I was cringing at
[23:18] <desrt> hmm
[23:18] <desrt> this is not the problem either
[23:19] <marga> From the backtrace it seems that the problem is in the  settings_changed function, that calls: g_settings_get_value(settings, key), where settings is a GSettings object and key is "keybindings/"
[23:19] <desrt> although clearly there is yet another unrelated issue there
[23:19] <desrt> ya.  exactly.
[23:19] <marga> But that looks like a valid call to me...
[23:20] <desrt> huh.
[23:20] <desrt> you maybe found a GSettings bug here
[23:20] <desrt> it depends what the docs say
[23:20] <marga> omg
[23:20] <desrt> ...if they say anything at all
[23:21] <desrt> huh.
[23:22] <desrt> can you install the glib dbgsyms and go one more time?
[23:23] <desrt> they're part of the standard glib packaging, so hopefully you don't need to rebuild....
[23:23] <desrt> valgrind is not being helpful here.... you want the gdb backtrace
[23:23] <desrt> lolwut
[23:23] <desrt> #15 0x00007ffff1e4a541 in ?? () from /usr/lib/x86_64-linux-gnu/libgio-2.0.so.0
[23:24] <desrt> #16 0x00007ffff60c21e4 in ffi_call_unix64 () from /usr/lib/libmozjs185.so.1.0
[23:24] <desrt> #17 0x00007ffff60c1c21 in ffi_call () from /usr/lib/libmozjs185.so.1.0
[23:24] <desrt> the invalid call is coming from javascript code?!
[23:24] <marga> cinnamon has some js plugins
[23:24] <marga> like the panel applets or something
[23:24] <marga> I'll look for glib debug
[23:27] <desrt> please do
[23:27] <desrt> i don't trust this backtrace identifying these as belonging to libmozjs
[23:27] <desrt> i have an extremely difficult time imagining anyone silly enough to write something to do this kind of patching of signals from JS
[23:28] <marga> heh...
[23:29] <marga> New backtrace: http://paste.ubuntu.com/5609269/
[23:29] <desrt> lovely
[23:31] <desrt> so there is at least one gsettings bug here
[23:31] <desrt> g_settings_get() should be crashing sooner
[23:31] <desrt> on a key with '/' in the name....
[23:31] <desrt> with a better assert message
[23:32] <marga> is it wrong that they key is /keybindings?
[23:32] <desrt> keybindings/
[23:32] <desrt> and yes...
[23:32] <marga> yes, that... It's wrong?
[23:32] <desrt> i'm reading the code over and over and i'm unsure how that could leak into the code
[23:32] <desrt> are you using a particularly old glib?
[23:32] <marga> precise's
[23:32] <desrt> that'll do it
[23:32] <desrt> cherry-pick this into your glib: ed492a5de2a217c08bccd4032b1122fba2f25c6f
[23:32] <desrt> that'll fix it
[23:33] <marga> 2.32.3-0ubuntu1
[23:33] <marga> Ok... Let's see if I can decipher that :)
[23:33] <desrt> ya... this fix was made in the 2.33 series
[23:33] <desrt> so this is indeed your problem
[23:34] <desrt> this actually fixes both problems: the problem of g_settings_get() rejecting sooner and the problem of the '/' getting there in the first place
[23:35] <desrt> Laney: around?
[23:35] <marga> Probably not... said going to bed almost an hour ago
[23:35] <desrt> i'm going to cherry-pick this over into the 2.32 branch upstream
[23:36] <desrt> and i'll try to find someone to SRU it tomorrow
[23:36] <marga> You know you rock, right?
[23:36] <desrt> :)
[23:36] <desrt> if you could test the patch in the meantime it would help with the SRU
[23:36] <marga> Yes, yes, I found the commit already, now trying to find how to download it
[23:37] <desrt> https://git.gnome.org/browse/glib/patch/?id=ed492a5de2a217c08bccd4032b1122fba2f25c6f
[23:39] <desrt> so it seems i did the cherry-pick at the time i made the change... and upstream release 2.32.4 already has the fix...
[23:39] <desrt> so someone just needs to package that
[23:41] <desrt> marga: did you already have a bug open about this?
[23:41] <marga> Not in launchpad
[23:41] <desrt> k.  i'll file that.
[23:41] <marga> This just hit us today, I've been debugging for about 4 hours
[23:46] <marga> rebuilding glib sure takes a long time :)
[23:46] <desrt> lots of tests
[23:46] <desrt> and i think the debian package builds several variants
[23:47] <marga> Yeah, that's likely
[23:47] <marga> Please give me the bug number, so I can chime in once this is done.
[23:49] <desrt> https://bugs.launchpad.net/ubuntu/+bug/1154370
[23:49] <ubot2> Launchpad bug 1154370 in ubuntu "[precise SRU] crash due to improper handling of "/" in GSettings" [Undecided,New]