[00:32] <TheMuso> yay for -1 day SRUs. :p
[04:24] <pitti> desrt: apport time delta> yes, we have that information at hand; I just wonder in what way this should be presented in the report; perhaps adding a new field if the delta between "crash happened" and "crash reported" is > 15 mins? or make better efforts to tell if it happened in the same session?
[04:24] <pitti> desrt: I don't think merely showing the exact time delta is very useful, as it's hard to search/filter for
[04:55] <RAOF> Wooo! Look at the 12.10 errors.ubuntu.com graph for colord. BOOYAH!
[05:06] <pitti> RAOF vs. D-Bus 1:1 ?
[05:07] <pitti> RAOF: wow, nice dip!
[05:07] <didrocks> guten morgen pitti
[05:07] <RAOF> pitti: I think I may have fixed ALL OF THE BUGS!
[05:07] <pitti> still twice as many as 12.04, but that might be due to people still needing to upgrade?
[05:08] <RAOF> Are the scales for 12.04 & 12.10 the same?
[05:08] <pitti> RAOF: indeed, I stopped getting annoying apport bubbles for this a while ago \o/
[05:08] <pitti> before that it seemed to be the way of quantal to say "good morning"
[05:08] <pitti> hey didrocks, wie gehts?
[05:08] <RAOF> Because I *know* that 12.04 is still terribly crashy and will remain so until colord moves into -proposed
[05:08] <didrocks> pitti: I'm ok, yourself?
[05:09] <pitti> quite fine
[05:10] <pitti> RAOF: hm, I'd expect that, otherwise the graph would be really confusing
[05:10] <pitti> RAOF: after all, the point of it is to allow comparisons?
[05:13] <RAOF> I have no idea what the values on the axes mean :)
[05:14] <RAOF> For example, on the 10th of October there were 0.11 $THINGS in 12.10.
[05:14] <pitti> I think it's "crashes per user in the selected time frame" (1 day, 1 month, etc.)
[05:14] <pitti> although, it's probably not that
[05:14] <RAOF> Heh
[05:14] <pitti> we don't know the #users
[05:15] <RAOF> This is making a fairly persuasive point for “does anyone know what the vertical axis means exactly” :)
[05:15] <RAOF> Presumably ev does.
[05:15] <pitti> so obviously it's "Kilojiffies by Milliblurbs"
[05:16] <maxb> The many ways of autostarting things in DMs are confusing :-/ .... indicator-cpufreq does it via /etc/xdg/autostart/, indicator-weather does it via X-GNOME-autostart-enabled=true in /usr/share/applications/ .... is one more correct ? Is there anywhere I can read about this stuff?
[05:16] <RAOF> maxb: Does X-GNOME-autostart-enabled=true actually do anything?
[05:17] <maxb> erm, good question
[05:17] <maxb> I assumed it did since I found it in an official quantal package, but...
[05:18]  * maxb logs out/in
[05:18] <RAOF> maxb: /etc/xdg/autostart or ~/.config/autostart should be the set of (non-default) things that get started. (Technically, that's $XDG_CONFIG_DIRS/autostart and $XDG_CONFIG_HOME/autostart)
[05:19] <maxb> Er, right, that'll teach me to trust stuff in the archive ;-)
[05:24] <BigWhale> Greetings all.
[07:32] <desrt> pitti: i was also thinking it could be useful if there was some global state flag set when logging out
[07:32] <desrt> or if we could find out if the machine rebooted since
[07:32] <desrt> or if something like the xdg session cookie or dbus bus address was different in the crashing session and the one that 'notices' the crash
[07:35] <desrt> these sort of things are more reliable than time passed (since that could happen for other reasons)
[07:35] <desrt> but still not as great as a flag set at logout time
[07:35] <desrt> because there could also be other reasons that a crash isn't noticed until logout/login even if the crash is not caused by the logout
[07:36] <larsu> desrt, good morning :)
[07:37]  * larsu is not used to seeing desrt around at this time
[07:46] <seb128> hey desktopers
[07:47] <didrocks> salut seb128
[07:47] <seb128> lut didrocks, ca va bien ?
[07:48] <didrocks> seb128: on fait rouler :)
[07:48] <didrocks> et toi?
[07:49] <seb128> tout va bien icic ;-)
[07:49] <pitti> desrt: re (sorry, was in a meeting)
[07:49] <pitti> bonjour seb128
[07:49] <seb128> pitti, salut, ca va ?
[07:49] <pitti> desrt: apport actually already makes some effort to determine whether the crash happens during session shutdown
[07:50] <pitti> desrt: and it's supposed to add a particular tag when you report a crash in a different session than the one it happened in, but that might be broken in some cases
[07:50] <desrt> larsu: hi :)
[07:50] <pitti> hey larsu, wie gehts?
[07:50] <pitti> seb128: bien, merci!
[07:50] <desrt> pitti: ah.  that's good news to know.  how do i look for that?
[07:50] <larsu> pitti, guten morgen! Gut, und selbst?
[07:51] <desrt> this channels speaks odd languages at this hour
[07:51] <larsu> desrt, you're starting to sound like tedg ;)
[07:52] <pitti> desrt: http://bazaar.launchpad.net/~apport-hackers/apport/trunk/revision/2058 is supposed to catch the crashes during shutdown, but apparenlty it's not sufficient
[07:53] <desrt> pitti: why would you think that this works?
[07:53] <desrt> DBUS_SESSION_BUS_ADDRESS doesn't magically disappear from the environment of a process when the bus goes away...
[07:54] <pitti> desrt: right, but gnome-session sohuld either be gone, or its state should be "shutting down"
[07:54] <desrt> oh
[07:54] <desrt> you just use the environment variable to contact gnome-session
[07:55] <desrt> right
[07:55] <pitti> desrt: hm, I recently added in_session_of_problem() which would tell you whether or not you report a crash that happened in the currently running session
[07:55] <pitti> desrt: but I just see that we don't seem to transform that into a tag
[07:55] <pitti> odd, I certainly discussed that with ev, but apparently it slipped through the cracks
[07:56] <desrt> pitti: glad to hear we're well on our way to a solution :)
[07:56] <pitti> desrt: you can kind of infer it from the presence or absence of an ~/.xsession-errors attachment
[07:56] <pitti> desrt: as we only attach it when the crash happened in the current session (otherwise xsession-errors is pointless)
[07:56] <desrt> ah.  neat trick.
[07:56] <pitti> but I guess a tag would be better
[07:56] <desrt> i'll check that for now
[07:57] <pitti> i. e. XsessionErrors
[07:57] <pitti> desrt: but it'd probably be best if you file a bug with the thing you are actually looking for (that wouldn't be a time delta for sure, but some more fundamental property), and then I'll see how to incorporate this?
[07:57] <seb128> pitti, the issue is that XsessionErrors is only there when there are errors matching your regexp to catch
[07:58] <seb128> so it could be a new session or it could be that this issue didn't let any info in the log
[07:59] <pitti> that's right; but in practice we always seem to have some warnings
[07:59] <pitti> don't get me wrong, I don't think this is an ideal state, just saying where we currently use it
[07:59] <pitti> I'd love to make it more explicit
[08:00] <seb128> pitti, that's a nice trick, thanks for the hint ;-)
[08:00] <desrt> pitti: okay
[08:00] <desrt> https://bugs.launchpad.net/apport/+bug/1067646
[08:00] <ubot2> Launchpad bug 1067646 in apport "should report if crash happened at logout/shutdown" [Undecided,New]
[08:01] <pitti> desrt: thanks; following up there
[08:07] <pitti> seb128: ah, thanks for the other use cases of hw dependend service activation
[08:08] <pitti> seb128: I still don't think that this is the main solution for our performance problems, but it's interesting nevertheless
[08:08] <seb128> pitti, yw ;-) That's the thing I want upstart jobs the most for
[08:08] <Laney> hallo
[08:09] <seb128> pitti, well, it's not only performances, it's flexibility ... autostarts are pretty limited (only one condition) and happen only on session start
[08:09] <seb128> we have no way atm to bring up a service when a key is set
[08:09] <seb128> or bring it down when a key is unset
[08:09] <seb128> key = gsettings key
[08:09] <seb128> same for hardware events
[08:09] <seb128> Laney, hey, how are you?
[08:09] <pitti> *nod*; but would upstart be able to listen for gsettings changes?
[08:10] <pitti> seb128: that rather seems to be a job for g-s-d to me?
[08:10] <seb128> yes
[08:10] <seb128> that's one of the things we asked for at last uds
[08:10] <seb128> a gsettings bridge (or whatever they call the events generators)
[08:10] <pitti> just like g-s-d starts/stops the trackpad polling process depending on the gsettings key
[08:10] <rickspencer3> bonjour seb128, didrocks quel est la mots de la rue apropos le release?
[08:10] <rickspencer3> (sorry for changing the topic ;) )
[08:10] <rickspencer3> just humor me
[08:10] <pitti> seb128: this would be a lot more upstream friendly
[08:11] <pitti> bonjour rickspencer3
[08:11] <didrocks> rickspencer3: quantal? c'est déjà dépassé, on bosse sur les SRU :)
[08:11] <seb128> rickspencer3, salut, on dirait qu'il y a toujours quelque problème de secure boot sur certaines images
[08:11] <seb128> rickspencer3, sinon c'est ok
[08:11]  * rickspencer3 just sees "ok" and tunes out
[08:11] <bryce> ah, must be le french invasionne hourre  ;-)
[08:11] <rickspencer3> j/k
[08:11] <seb128> pitti, well, I don't want g-s-d to become more of a kitchen sink, it's not g-s-d's job to take down the gwibber lens when there is not internet connection
[08:12] <didrocks> bryce: "hourre"? "season" rather :p
[08:12] <rickspencer3> til "on bosse sur"
[08:12] <bryce> didrocks, hehe
[08:12] <didrocks> or seaaaazone :p
[08:12] <rodrigo_> hi
[08:12] <pitti> seb128: right, but for wacom it might be
[08:12] <rodrigo_> getting this error:
[08:12] <rodrigo_> dpkg-source: error: cannot represent change to mesa-9.0/bin/install-sh:
[08:12] <rodrigo_> dpkg-source: error:   new version is symlink to minstall
[08:12] <rodrigo_> dpkg-source: error:   old version is nonexistent
[08:12] <rodrigo_> no idea what it means :(
[08:12] <seb128> pitti, yes, my examples suck, I didn't stop to think about proper ones
[08:12] <didrocks> bryce: you shouldn't have showed up I guess ^ :p
[08:12] <rodrigo_> hey seb128, pitti, didrocks!
[08:12] <didrocks> hey rodrigo_ ;)
[08:13] <pitti> hey rodrigo_
[08:13] <seb128> rodrigo_, hey
[08:13] <seb128> rodrigo_, what version of dpkg do you use? maybe try asking on #ubuntu-devel, the dpkg people are not on #desktop
[08:14] <rodrigo_> 1.16.8 (amd64)
[08:14] <rodrigo_> seb128, ok, I'll ask there
[08:14] <Laney> urg
[08:14] <Laney> I'm getting loads of weird crashes
[08:14] <Laney> hope it's not faulty ram
[08:14] <bryce> didrocks, nah, I shoved mesa9 into x-updates, I deserve all the complaints
[08:14] <rodrigo_> bryce, oh, you already packaged it?
[08:14] <didrocks> bryce: heh ;)
[08:14] <rodrigo_> bryce, was trying to
[08:14]  * Laney is going to run a quick memtest
[08:15] <bryce> rodrigo_, no; tjaalton packaged mesa9 for quantal, I just copied it into x-updates ppa.  Is that where you're installing from?
[08:15] <rodrigo_> bryce, no, I was trying to package it for the PPA you created for us
[08:16] <rodrigo_> so, looking for the tjaalton's package, will just copy that one
[08:16] <bryce> rodrigo_, yeah that'd be best
[08:16] <rodrigo_> yes
[08:17] <bryce> rodrigo_, btw I didn't merge your PPA into x-updates yet, but would like to do so when you feel it to be ready
[08:18] <rodrigo_> bryce, yes, still waiting for more feedback, so not ready yet, will let you know as soon as it is
[08:18] <rodrigo_> it works for me perfectly, but that's not enough testing :)
[08:18] <bryce> :-)
[08:18] <rodrigo_> with the mesa 9 update we should get more testing
[08:18] <bryce> yeah agreed
[08:46] <mvo> I guess that is a known issue, but compiz/xorg are using all my cpu after some time and make the system sluggish - thisis a intel video card
[08:48] <chrisccoulson> good morning everyone
[08:50] <seb128> mvo, hey, not know afaik ...
[08:51] <seb128> mvo, did you see the GTK guys fixed your leak in git ;-)
[08:51] <seb128> chrisccoulson, hey, how are you?
[08:51] <chrisccoulson> seb128, yeah, good thanks. how are you?
[08:51] <seb128> chrisccoulson, I'm good thanks
[08:52] <seb128> chrisccoulson, shame that you didn't upload that tb messaging menu fix on monday, with the late respins it would have made it to the iso :-(
[08:52] <seb128> chrisccoulson, they moved things from proposed to the iso still yesterday
[08:52] <chrisccoulson> seb128, oh, i didn't realize that there were late respins :/
[08:52] <chrisccoulson> never mind :(
[08:52] <bryce> mvo, usually high xorg cpu is due to clients driving the X server beyond the norm. So try to identify which client app might be driving it.
[08:52] <seb128> chrisccoulson, can we get the SRU rolling anyway?
[08:53] <chrisccoulson> seb128, yeah, i'll upload it now
[08:53] <seb128> chrisccoulson, thanks
[08:56] <mvo> seb128: very nice
[08:56] <mvo> seb128: looks like its not one leak but multiple from looking at the git diff - but great to know that its fixed now
[08:57] <seb128> mvo, well, the specific one you pointed seems to be one commit to destroy a cairo surface
[08:57] <seb128> but cosimoc fixed other issues while he was at it
[08:58] <mvo> bryce: thanks, this is my usual workload, i.e. xchat, firefox, mutt, emacs - but it now stopped (or is smaller, more like 10% cpu when typing here)
[08:58] <mvo> bryce: is there anything I could do to debug when this happen next time? look at something like top to see what else is using high cpu?
[08:58] <mvo> (crude, I know :/
[08:59] <mvo> and here it is again
[09:03] <bryce> hmm, there is xrestop but it mostly reports memory usage
[09:04] <bryce> mvo, fwiw those four programs are also my four main programs.  (well, emacs -nox)
[09:04] <bryce> firefox plugins can sometimes drive X.
[09:04] <bryce> compiz plugins are another likely source, esp. if you have something non-default enabled
[09:05] <mvo> bryce: its a pretty stock install, let me try xrestop
[09:05] <bryce> mvo, if you kill a client app and the X.org CPU usage decreases, that's a good sign
[09:06] <mvo> bryce: well…
[09:06] <mvo> bryce: that caused a nice x crash
[09:06] <bryce> awesome
[09:06] <mvo> [200466.552] Caught signal 11 (Segmentation fault). Server aborting
[09:06] <mvo> bryce: was this a trick ;) ?
[09:06] <bryce> mvo, anything captured in /var/crash/ ?
[09:07] <tjaalton> oh there's a bug about xrestop crashing
[09:07] <bryce> tjaalton, oh?  just crashing itself or crashing xserver?
[09:07] <tjaalton> xserver
[09:07] <bryce> doh
[09:07] <bryce> tjaalton, bug #?  sounds bad
[09:08] <tjaalton> bug 1060059
[09:08] <ubot2> Launchpad bug 1060059 in xorg-server "Xorg crashed with SIGABRT in ResFindAllRes()" [Medium,New] https://launchpad.net/bugs/1060059
[09:09] <mvo> I reported it now using apport but it would not open a bugreport
[09:09] <mvo> (after asking me a bunch of question like if I am willing to help debug)
[09:09] <seb128> tjaalton, shrug, shouldn't "obvious xorg crasher" bugs be better triaged that non confirmed, medium, non assigned, non milestoned?
[09:09] <bryce> tjaalton, sheesh.  wtf xserver...
[09:10] <tjaalton> seb128: just remembered that it was reported
[09:10] <seb128> tjaalton, who do I assign it to? ;-)
[09:10] <seb128> it's milestoned for quantal-updates now fwiw :p
[09:10] <tjaalton> I've been assigned to other tasks for the next few weeks..
[09:11] <bryce> seb128, chill.  we have a lot of X crashes, but me and all the other X guys are tied up on various projects
[09:11] <bryce> seb128, assign to canonical-x or escalate with jason for critical issues
[09:12]  * mvo runs xrestop again just to see if its reproducable
[09:13] <seb128> bryce, sorry, it's always annoying when we gets bugs that lead users to have their session to go down and close,loose any open work
[09:13] <seb128> bryce, but understood, I know everyone is busy and stretched out :-(
[09:13] <bryce> mvo, there is also https://wiki.ubuntu.com/X/Troubleshooting/HighCPU however it doesn't say much more than I've already mentioned.
[09:14] <seb128> bryce, I've set it to triaged/high for now
[09:14] <bryce> seb128, actually I set it triaged/high, but thanks :-)
[09:14] <seb128> bryce, https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/1065113/+activity :p
[09:14] <ubot2> Launchpad bug 1065113 in xorg-server "Xorg crashed with SIGABRT in free()" [High,New]
[09:14] <seb128> bryce, yw ;-)
[09:15] <seb128> shrugh, wrong bug
[09:15] <seb128> anyway moving on ;-)
[09:15]  * seb128 is trying to clean up the "quantal milestoned bugs" list
[09:15] <mvo> pretty reliable
[09:16] <bryce> seb128, but yes for the future either assign to me or to canonical-x, and I'll ensure they get assigned out to the right folks
[09:16] <seb128> bryce, thanks
[09:16] <bryce> seb128, sadly I guess *most* X.org bugs are of the "session goes down; lose your work" variety
[09:16] <mvo> maybe we should stop recommending using it for now
[09:17] <bryce> mvo, yeah now that I know, for sure.  Actually getting the crash fixed should be a top priority now.
[09:17] <seb128> bryce, yeah, I figured so...
[09:17] <mvo> lol - xvfb-run xrestop  is fine
[09:18] <mlankhorst> bryce: well would help if we could start testing the quantal stack too but I kind of want the multiarch bug fixed first before I announce it for wider testing
[09:18] <tjaalton> seb128, bryce: I should have a few spare cycles today, so I'll at least make sure it's sent upstream
[09:19] <mlankhorst> this particular bug looks like it should be catchable with valgrind
[09:19] <tjaalton> might even get a 0-day update for it, we'll see
[09:19] <tjaalton> mlankhorst: don't mind you looking at it ;)
[09:19] <tjaalton> maybe it was during the weekend when I saw duflu's bug about it go by
[09:20] <tjaalton> so didn't react or test
[09:20] <seb128> tjaalton, thanks
[09:20] <mlankhorst> "Xorg crashed with SIGABRT in free()"
[09:20] <bryce> mvo, typically the next step I suggest after identify possible client causes is to set up xtrace between server and (suspected) client, and see if what X calls it's making.  Most (but not all) high X.org CPU issues are due to an out of control client making excess X calls.
[09:20] <mlankhorst> that's not really something we'd instantly notice though :x
[09:21] <didrocks> desrt: hey, I think we'll need your help :p
[09:21] <bryce> tjaalton, thanks
[09:21] <tjaalton> mvo: it's also possible that the gpu is (partially) hung
[09:22] <tjaalton> check dmesg
[09:22]  * mlankhorst updating quantal
[09:22] <tjaalton> i've seen that on precise a couple of times
[09:23] <mvo> mlankhorst: that is blocking on me right ? that multiarch bug? sorry for that
[09:24] <mlankhorst> np :)
[09:24] <mvo> tjaalton: nothing suspicious in dmesg afaict
[09:25] <tjaalton> mvo: and the system is not swapping?
[09:26] <mvo> tjaalton: I doubt it, haven't checked, but it has 8gb of ram (can't check now as the session crashed so all memory hungry stuff is gone)
[09:26] <tjaalton> ah, ok :)
[09:26] <mvo> tjaalton: but I will remember to check next time :)
[09:26] <mvo> could be firefox eating all my ram of course
[09:27] <tjaalton> in theory there could be a huge pixmap leak or such causing it to swap
[09:27]  * mvo nods
[09:27] <mvo> good to know
[09:27] <mlankhorst> and here I was hoping we didn't have such problems :p
[09:27] <tjaalton> but those tend to stay, not vanish by itself
[09:27] <tjaalton> oh well, lunch ->
[09:33] <mhr3> seb128, gvfsd-http still leaking pretty bad, i think we can rule out that it was just outdated pkg on my system
[09:34] <mhr3> seb128, can you rep? just open the dash and do random searches in shopping/music
[09:34] <seb128> mhr3, you know how to valgrind and track leaks right, patches are welcome ;-)
[09:35] <mhr3> seb128, i just can't take the joy from you :)
[09:37] <seb128> mhr3, but yeah, confirmed
[09:39] <mlankhorst> seb128: what screen reader does ubuntu use?
[09:42] <seb128> mlankhorst, orca
[09:42] <seb128> mlankhorst, https://launchpad.net/ubuntu/+source/gnome-orca
[09:42] <mlankhorst> ah k
[11:35] <didrocks> pitti: would you mind bumping https://launchpad.net/~ubuntu-desktop/+archive/ppa/+build/3910394 please? :)
[11:36] <didrocks> pitti: it contains a worrying upgrade issue fix for SRU0 and I want it to have wider testing
[11:36] <didrocks> (for compiz)
[11:38] <chrisccoulson> didrocks, you're stealing my PPA builders? ;)
[11:39] <didrocks> chrisccoulson: do you have an urgent upgrade issue? :p
[11:39] <chrisccoulson> heh, not today ;)
[11:39] <seb128> chrisccoulson, just upload that tb fix to proposed :p
[11:40] <chrisccoulson> seb128, i uploaded it this morning ;)
[11:40] <didrocks> chrisccoulson: see, I can steal them gladely then! :)
[11:40] <chrisccoulson> it's not been approved yet though
[11:40] <seb128> chrisccoulson, \o/
[11:57] <pitti> didrocks: done
[11:57]  * didrocks hugs pitti
[11:58] <pitti> didrocks: je suis désole, j'étais au déjeuner
[11:58] <didrocks> pitti: tu as de la chance, j'ai pas pu encore en avoir! à cause de ce *ù$*$ù**mmfs de bug :-)
[11:59]  * pitti hugs didrocks
[12:06] <FlyingElvis> hello everyone...im new to ubuntu...just having ditched windows in the last week and put ubuntu on every computer in the house, and while everything is going fine...is there a FULL list of keyboard commands?
[12:06] <mlankhorst> hold the key between control and alt?
[12:07] <FlyingElvis> yes i know of that...but thats all the commands?
[12:08] <FlyingElvis> all and all im really impressed with ubuntu...and the rest of the family are finding it easy to use also
[12:15] <mlankhorst> :-)
[13:03] <cyphermox> morning :)
[13:13] <seb128> cyphermox, hey, how are you?
[13:30] <cyphermox> seb128: not bad, you?
[13:34] <seb128> cyphermox, I'm good thanks
[13:50] <chrisccoulson> maaaaan, trying to get the background of a tooltip to render on to the surface of another window is a PITA
[14:00] <sil2100> desrt: ok, so currently the decision is - no per-relocatable-schema overrides in gsettings?
[14:01] <desrt> yes
[14:01] <desrt> the unity package should override all of the plugins that it wants to
[14:01] <desrt> not just the plugin list
[14:02] <didrocks> hum the discussion started on another channel?
[14:02] <sil2100> desrt: currently it's a bit hard with the current way things are
[14:02] <didrocks> desrt: the issue is that we have 2 profiles
[14:02] <didrocks> desrt: and the profiles are different depending on the session you start
[14:03] <didrocks> (and it's using different locations)
[14:03] <sil2100> Since in this case we won't be able to use overrides for this purpose really
[14:04] <didrocks> hey desrt btw :p
[14:04] <didrocks> (sorry, jumping into the discussion as it seems you are speaking about the issue we had)
[14:04] <sil2100> Since the case is we need like 2 different overrides, each for every profile, as didrocks says - so not sure if we can use overrides in this way
[14:05] <sil2100> didrocks: good that you joined in, I was expecting that even! :)
[14:05] <didrocks> sil2100: not sure what the beginning of the discussion was, but well, I think that desrt needs and deserves to have the full story :p
[14:06] <didrocks> otherwise, this means a dconf *writes* at first login :)
[14:06] <didrocks> and I know desrt doesn't like that!
[14:06] <desrt> i don't want to add a complicated feature to gsettings that will really only support a single (very weird) usecase
[14:06] <didrocks> desrt: what do you suggest that we do in that case?
[14:06] <desrt> and hopefully this one user will be gone soon enough...
[14:07] <desrt> didrocks: install overrides
[14:07] <didrocks> desrt: but we already do that
[14:07] <desrt> and don't expect a gnome session to work on the same machine as a unity session
[14:07] <sil2100> didrocks: I know you'll be mad at me, but I gave desrt the whole story on priv... then I got enlighted that I'm talking in private, so I moved the discussion here
[14:07] <sil2100> ;(
[14:07] <didrocks> desrt: ok, so waiting for the fallback session to die
[14:07] <didrocks> sil2100: yeah, please stop this habit :(
[14:08] <sil2100> didrocks: it's getting better though, since I understood this myself and corrected my behavior this time ;)!
[14:09] <desrt> so here's one possibility
[14:09] <chrisccoulson> heh, https://twitter.com/opensourcerer/status/258569723481055233 ;)
[14:09] <desrt> the compizconfig abstraction is pretty... impressive
[14:09] <desrt> why not just have a second set of schemas in there?
[14:09] <desrt> like, instead of installing overrides, the unity session could install new copies of the schemas
[14:09] <desrt> with some prefix or something....
[14:10] <jcastro_> chrisccoulson: do not listen to him, get to work.
[14:10] <chrisccoulson> lol
[14:10] <desrt> chrisccoulson: you be scrollin', they be hatin'
[14:10] <sil2100> I think that's what was being done with GConf
[14:10] <didrocks> desrt: I think that relocatable schemas were exactly for those purposes, isn't it?
[14:10] <desrt> didrocks: no
[14:11] <sil2100> desrt: so I think that's why we're using relocatable schemas
[14:11] <desrt> didrocks: they were made for the case where you have multiple accounts, for example
[14:11] <desrt> and each account has user/password/host/port/etc type settings
[14:11] <sil2100> Ok, in this use-case, you only need one pair of overrides usually
[14:11] <didrocks> desrt: don't you think we can have "types of accounts" with exactly the same parameter, by different default values?
[14:11] <desrt> didrocks: i guess it's unusual to have some account parameters set by default
[14:12] <desrt> for only one account....
[14:12] <didrocks> desrt: ok, I'm fine with the hack for quantal, as I think the profile thingy will die soon
[14:12] <desrt> i'll look the other way when you write to dconf :)
[14:14] <didrocks> desrt: heh, and worse, I have to use dconf itself because of the revert :p
[14:14] <didrocks> desrt: but we'll discuss this at UDS I guess
[14:14] <didrocks> (revert bug)
[14:14] <sil2100> ;)
[14:15] <desrt> didrocks: it's always your option to install a dconf override :)
[14:15] <desrt> it's really the same issue as we've had 1000 times before
[14:15] <didrocks> desrt: heh, indeed :)
[14:15] <desrt> we want one set of settings in one session and another in another
[14:17] <didrocks> desrt: rings me a bell :p
[14:20]  * sil2100 remembers setting once both gsettings and dconf override for the same thing just to be sure it's overridden
[14:21] <sil2100> Not really proud of that one
[14:35] <chrisccoulson> oh crap
[14:35] <chrisccoulson> firefox is loading 2 different copies of sqlite in to memory
[14:35] <chrisccoulson> it's own copy, and the system copy :(
[14:36] <chrisccoulson> i wonder how this has not completely broken already
[15:06] <tkamppeter> pitti, hi
[15:35] <seb128> Laney, do you run your glib update?
[15:35] <Laney> yeah
[15:35] <Laney> well, without the cherry-pick
[15:35] <seb128> Laney, does "gresource details /usr/bin/nautilus" works?
[15:36] <Laney> seb128: nope, "gresource is built without elf support"
[15:36] <seb128> Laney, ok, your "drop libelf-dev build-depends" is buggy :p
[15:36] <Laney> huh, weird
[15:36] <seb128> $ gresource details /usr/bin/nautilus
[15:36] <seb128> nautilus    206 u /org/gnome/nautilus/icons/knob.png
[15:36] <seb128> nautilus    482 u /org/gnome/nautilus/icons/thumbnail_frame.png
[15:36] <seb128> nautilus  18004 c /org/gnome/nautilus/nautilus-bookmarks-window.ui
[15:36] <seb128> ...
[15:37] <seb128> on quantal
[15:37] <seb128> well
[15:37] <seb128> https://launchpadlibrarian.net/117502434/buildlog_ubuntu-quantal-i386.glib2.0_2.34.0-1ubuntu1_BUILDING.txt.gz
[15:37] <seb128> checking for LIBELF... no
[15:37] <seb128> checking for elf_begin in -lelf... yes
[15:37] <seb128> checking for elf_getshdrstrndx in -lelf... yes
[15:37] <seb128> ...
[15:37] <Laney> yeah i was going to compare that
[15:37] <Laney> still testsuiting atm
[15:37] <Laney> so it did do something before
[15:38] <seb128> Laney, well anyway gresource doing "gresource is built without elf support" is the reason why I added that build-depends in quantal :p
[15:38] <seb128> not sure why Josselin though it was the wrong lib...
[15:38] <Laney> no, it's fine, I just believed the changelog from joss
[15:38] <seb128> there are 2 elf libs though, one used to be in universe
[15:39] <seb128> so we might prefer the other one compared to debian (I know that was the case for bud-buddy by then)
[15:39] <Laney> it's true that libelf-dev doesn't ship a pcfile
[15:41] <rickspencer3> seb128, does turning on accessibility slow down the desktop a little? and is accessibility required to be on to use onboard?
[15:42] <seb128> Laney, right, but the configure doesn't check for one, it checks for .h also
[15:42] <Laney> correct
[15:42] <Laney> putting it back
[15:42] <bcurtiswx> kenvandine, using chromium-browser how does webapps let me (for example) work with gmail? i installed unity-webapps-gmail but i dont' know how it's supposed to work
[15:43] <seb128> Laney, thanks
[15:43] <seb128> rickspencer3, my gut feeling is that it does slow down thing yes, and it's not required for onboard afaik
[15:43] <pitti> hello tkamppeter
[15:45] <kenvandine> bcurtiswx, you should get prompted to integrate
[15:45] <kenvandine> and you shouldn't need to install the package yourself
[15:45] <kenvandine> that should happen automatically
[15:45] <bcurtiswx> hmm
[15:45] <bcurtiswx> kenvandine, even if I already had a gmail account setup and logged in before webapps was integrated ?
[15:45] <kenvandine> yes
[15:45] <kenvandine> well
[15:46] <kenvandine> installing the package isn't integrating
[15:46] <kenvandine> you need to click yes in the infobar
[15:46] <kenvandine> when it prompts you
[15:46] <kenvandine> does it work in firefox?
[15:47] <bcurtiswx> kenvandine, yes it does (noticed i haven't used firefox in a while)
[15:48] <kenvandine> so you got prompted, and after accepting it you see the launcher for it?
[15:48] <rickspencer3> thanks seb128
[15:48] <bcurtiswx> i get prompted yes, haven't gone through to see the launcher.  I wanted to use chromium-browser
[15:48] <kenvandine> bcurtiswx, you should get prompted in chromium as well
[15:48] <kenvandine> oh
[15:49] <kenvandine> do you have unity-chromium-extension installed?
[15:49] <bcurtiswx> i did not
[15:49] <bcurtiswx> installing now
[15:49] <bcurtiswx> is it not a dep on chromium-browser?
[15:49] <kenvandine> nope
[15:53] <bcurtiswx> kenvandine, in order for the gmail webapp to work with chromium it seems to need an "untitled" chromium icon even with the gmail icon in unity
[15:56] <tkamppeter> pitti, hi
[15:57] <pitti> hello tkamppeter
[15:57] <kenvandine> bcurtiswx, yes
[15:57] <kenvandine> that activates the spread
[15:57] <kenvandine> for current webapp browser windows
[15:58] <tkamppeter> pitti, it is about the missing /etc/apparmor.d/local/usr.sbin.cupsd file in our cups package. Is this preventing Quantal from being freshly installed or running as live CD?
[15:58] <bcurtiswx> kenvandine, that just doesn't seem right to me. i will never need to click on that "untitled" chromium icon it's a waste of space in my launcher bar
[15:58] <pitti> tkamppeter: it shouldn't even start if it that were the case, so the image build should fail
[15:59] <kenvandine> bcurtiswx, i didn't design it :)
[15:59] <tkamppeter> pitti, bug 1026921
[15:59] <ubot2> Launchpad bug 1026921 in cups "package cups 1.5.3-0ubuntu1 failed to instal/upgrade: subprocess installed post-installation script returned error exit status 1" [High,Triaged] https://launchpad.net/bugs/1026921
[15:59] <bcurtiswx> who did ?
[16:00] <tkamppeter> pitti, is this file really required (explicitly referenced by another file in standard config)?
[16:00] <pitti> tkamppeter: I don't know, I just saw the report today the first time; jdstrand should know
[16:01] <kenvandine> bcurtiswx, nuthinking
[16:01] <kenvandine> bcurtiswx, file a bug explaining what you don' tlike
[16:01] <bcurtiswx> kenvandine, of course :)
[16:02] <jdstrand> tkamppeter, pitti: an #include file in an apparmor profile must exist, yes
[16:02] <kenvandine> i've expressed my opinion on the topic already :)
[16:02] <bcurtiswx> kenvandine, do you have a bug about it ?
[16:02] <jdstrand> otherwise the parser will fail
[16:02] <kenvandine> no
[16:02] <pitti> jdstrand: curious how this ever worked for a default  install then
[16:02] <kenvandine> i just complained to him
[16:02] <jdstrand> does it use dh_apparmor? it will create it for you
[16:03] <pitti> ooh
[16:03] <pitti> jdstrand: yes, that's it
[16:03] <pitti> jdstrand: it's created in postinst, I had assumed it was a conffile
[16:03] <pitti> jdstrand: thanks!
[16:03] <pitti> tkamppeter: so I guess the reporter just deleted it
[16:04] <jdstrand> pitti: right-- the local/* is intentionally *not* a conffile. this is for site local additions so people don't have to deal with conffile prompts on upgrades
[16:04] <pitti> tkamppeter: merely creating an empty file should be okay, or sudo dpkg-reconfigure cups should also do
[16:15] <tkamppeter> pitti, so there is no bug which have to be fixed?
[16:15] <pitti> tkamppeter: apparenlty not, it can go back to invalid
[16:21] <tkamppeter> pitti, can you upload Cairo for me, bug 1063618? Thanks.
[16:21] <ubot2> Launchpad bug 1063618 in cairo "PDF printing output of evince broken with PDF file from "Der Spiegel"" [High,In progress] https://launchpad.net/bugs/1063618
[16:34] <seb128> tkamppeter, pitti: I can upload cairo if pitti didn't start on it
[16:35] <BigWhale> seb128, great thread about GNOME plans on ubuntu-desktop.
[16:35] <pitti> I didn't yet, sorry (still in #u-server debugging juju)
[16:35] <tkamppeter> seb128, OK, thanks in advance
[16:35] <pitti> seb128, tkamppeter ^
[16:35] <seb128> BigWhale, thanks ;-)
[16:35] <seb128> pitti, ok, I'm looking at it you can consider yourself unpinged
[16:36] <seb128> tkamppeter, thank you for backporting those patches ... did you confirm they fix your issue?
[16:51] <tkamppeter> seb128, yes, they fix the problem for me.
[16:52] <seb128> tkamppeter, can you update the bug with the SRU informations? (Impact, Test Case, Regression Potential)?
[17:12] <tkamppeter> seb128, will do
[17:13] <seb128> tkamppeter, thanks, I'll uploaded the package for you
[17:26] <robru> mhr3: ping
[17:29] <mhr3> robru, pong
[17:29] <robru> mhr3: hey. so kenvandine tells me that you are the guy to talk to about Dee indexing and performance
[17:31] <mhr3> robru, try me :)
[17:31] <robru> mhr3: we have a Dee.SharedModel and the problem is that we're trying to populate it with rows that come from a highly-duplicated source, so we need quite a bit of duplicate-checking logic before adding rows...
[17:31] <robru> mhr3: so our original solution was to index the row data in a python dict, and then consult the dict before publishing new rows to the model
[17:32] <robru> mhr3: that worked fine at first, but now we're persisting the sharedmodel inside a resourcemanager, and what's happening is that when we restore the sharedmodel, it has a bunch of rows that are no longer in our dicts (because the python script starts up fresh with empty dicts...)
[17:32] <robru> mhr3: so I wrote a really naive piece of code that just scans the sharedmodel on startup, and populates the dicts with the index data we need
[17:32] <robru> mhr3: but we were hoping that Dee might have a better solution for this.
[17:33] <robru> mhr3: https://code.launchpad.net/~robru/friends/persistence/+merge/130009 here is some code to consider
[17:34]  * mhr3 looking
[17:35] <robru> mhr3: it gets complicated because we can't just index on a single column; we have two columns that must be unique when concatenated, and then we have this other column that's a list, and the list is going to have a lot of elements, but those elements also need to be uniquely indexable.
[17:35] <robru> mhr3: (that sounds a bit insane when I say it like that... maybe we need to consider a different model schema ;-)
[17:35] <mhr3> sounds likely :)
[17:36] <mhr3> but, as long as there's a way to keep the rows sorted, not even multiple columns is problem
[17:36] <mhr3> then you get fast lookups
[17:36] <robru> mhr3: ok, so we can introduce a new column that contains a unique key, that won't be hard. but that other column that's a list, we still need to somehow index the rows based on the elements of that list.
[17:38] <mhr3> robru, did you see insert_row_sorted?
[17:38] <robru> mhr3: no ;-)
[17:38] <robru> mhr3: I'm quite new to Dee stuff.
[17:40] <mhr3> if you can come up with a cmp_func that will keep the model uniquely sorted, then you can find whether a particular row is already in the model in log(n), ie fast :)
[17:40] <mhr3> (find_row_sorted)
[17:40] <robru> mhr3: ok, I think I follow, but that's still only half the problem.
[17:41] <micahg> Laney: I'm planning on doing a webkit update for quantal for 1.10.1, is there anything that I need to include (or wait for you to figure out about it, I see you have a PPA build going)
[17:41] <robru> mhr3: that takes care of the two columns that we need to concatenate for uniqueness (I'll make a new column for that post-concatenation, and then we can insert and find on that, sorted)
[17:42] <Laney> micahg: do you have something special over that build?
[17:42] <robru> mhr3: but also there's this other column that's a list. we need a way to query the model for "does this value exist in any of the lists in any of the rows in this particular column"
[17:42] <robru> in a fast way
[17:42] <Laney> it's going in that ppa so I can test arm
[17:42] <mhr3> robru, so you need two indices?
[17:42] <robru> mhr3: yes
[17:43] <robru> mhr3: the concatenated column to index on will just be utf8 text, so a simple lexical cmp should work fine.
[17:43] <mhr3> it's doable with filtermodel, but you'll have to rebuild the filtermodel everytime
[17:44] <mhr3> ie. the filtermodel will be sorted in a different way than the original model (but it still points to the same data), so you'll just use the list to create your new sorting
[17:44] <robru> mhr3: but the other problem is that because we've de-duplicated messages, we have a many-to-one relationship between message ids and rows in the model. so we're storing these ids in a list. so each row has multiple ids, but we occaisionally will just have one of the ids, and then need to find that row. currently we're doing it with a python dict
[17:44] <mhr3> (oh by everytime i mean only when the process starts)
[17:45] <mhr3> robru, so one row has multiple ids? and those are stored in the list?
[17:47] <robru> mhr3: sorry, I was simplifying a bit ;-) we have a column that's of type 'aas', that is, an array of an array of strings. the inner arrays are all of length 3, and the outer array will be of varying lengths from row to row. some rows will only have one inner array, some rows will have many. but occaisionally we're going to only know one of the inner arrays, and we need to be able to find the row using just the inner array.
[17:48] <mhr3> well you need a way to create an index from that, otherwise you can't do fast lookup
[17:48] <robru> mhr3: so basically one column in a row might look like [['website', 'account_id', 'message_id'], ['other_site', 'other_acct', 'other_msg']] and we'll need to find that row using only ['website', 'account_id', 'message_id']
[17:49] <robru> mhr3: that's what I'm asking you! ;-) how do we index this properly?
[17:49] <mhr3> i'm trying to remember if you can put the same backing row multiple times into the filtermodel, but i'm not sure
[17:49] <robru> mhr3: currently we keep those triples as dict keys, and the dict values are RowIters. it works ok but we have to rebuild the dict every launch.
[17:52] <mhr3> well, you'd have to rebuild it even if you used dee, don't think you can avoid that
[17:53] <robru> mhr3: we were hoping that whatever indexing options Dee gives us, we could persist with ResourceManager ;-)
[17:53] <mhr3> you could try denormalizing the model, so you'd end up with two (or more)
[17:54] <robru> mhr3: I don't know what that means ;-)
[17:54] <mhr3> it means you have too much data in that single cell :)
[17:54] <micahg> Laney: no, I just wanted to push it as a security update :)
[17:55] <robru> mhr3: is there any way that I could have a second Model that has an one column for the ids, and a second column for rowiters that point back to the first Model? eg, any way to store a DeeRowIter inside a SharedModel in a meaningful way?
[17:56] <micahg> Laney: I probably won't upload at least until tomorrow sometime (Monday would be fine as well)
[17:57] <mhr3> robru, no, you'd need to basically copy a column with the unique id into the second model (where it wouldn't be unique)
[17:57] <Laney> micahg: feel free to sponsor my upload into -security if you like ;)
[17:57] <Laney> save me the SRU pain
[17:57] <robru> mhr3: oooooohhhhhhhh! that's actually not half bad!
[17:57] <micahg> Laney: yeah, that could work, could you let me know how the tests go
[17:57] <mhr3> robru, we're basically talking sql tables here, you know that? :)
[17:58] <Laney> ok
[17:58] <robru> mhr3: yeah, I know ;-)
[17:58] <Laney> should pop out towards the end of tomorrow
[17:58] <robru> mhr3: not my fault, it's the data's fault for being structured this way. I'm just trying to figure out the most efficient DeeModel schema to represent it...
[17:58] <mhr3> robru, and since you need quite complex indexing i'm starting to wonder why your data isn't in sql? :)
[17:59] <robru> mhr3: because Dee.SharedModels are just so sexy, and we needed to be able to share this data over dbus *anyway*
[18:00] <mhr3> +1 on the sexiness :)
[18:01] <robru> mhr3: I actually like your idea of using the unique key non-uniquely in a second model, I think I can make that work. and if you say that the *_sorted methods are log(n) then this should be fast! also being able to persist that second model will stop us having to re-scan at launch. brilliant!
[18:01] <robru> mhr3: thanks!
[18:02] <mhr3> robru, the only problem is that you then need to make sure the two models are always in sync, otherwise dragons will come
[18:03] <mhr3> if for example you manage to write to disc one but not the other... uh oh
[18:04] <mhr3> plus it increases required storage space :/
[18:05] <micahg> Laney: actually, nevermind, I thought they would've included CVE fixes, but it's bug fix only, so I won't be pushing it after all
[18:05] <micahg> ogra_: ^^ laney is testing webkit on arm, I won't be touching it until the next point release
[18:08] <robru> mhr3: what is the storage space like for ResourceManager?
[18:09] <robru> mhr3: also, I was curious if you knew of any upper limits regarding how many rows a SharedModel can reasonably contain before performance becomes an issue.
[18:11] <mhr3> robru, well, everytime you init a shared model it's broadcast on the bus, so you'll probably end up being limited by dbus msg size
[18:11] <robru> mhr3: the whole model is transmitted in a single message?
[18:12] <robru> mhr3: does it have any ability to split that up over messages?
[18:12] <mhr3> currently no
[18:12] <mhr3> if you need to transfer that much data over dbus, it's bad
[18:12] <robru> mhr3: I ask because as it currently stands, we have written an algorithm that endlessly gathers data and dumps it into a SharedModel, forever. And the model persists, and there's not yet any logic for expiring old data. so it's just going to grow and grow and grow, and I'm wondering what kind of limits we can expect to hit.
[18:13] <mhr3> you should add that logic :)
[18:13] <robru> I will, but that's why I'm asking. should I discard the oldest data after there's 1,000 rows in the model? 10,000? a million? I have no idea.
[18:14] <mhr3> i'd make sure the model isn't larger than 8mb
[18:15] <robru> mhr3: and how would i measure that? can the resourcemanager tell me how much space it's taking?
[18:15] <robru> mhr3: can I find the file where it's persisting it to?
[18:15] <robru> ;-)
[18:16] <mhr3> figure out how big is a row on average on go by that
[18:16] <mhr3> but yea, it's saved somewhere, just not sure where exactly :)
[18:16] <robru> mhr3: hmmmm, that's highly variable ;-) but ok, I'll work on some averages.
[18:18] <mhr3> robru, you probably want to look at FileResourceManager.primary_path prop
[18:18] <robru> mhr3: ah, thanks
[18:19] <mhr3> hm, it even has getter method
[18:33] <tkamppeter> seb128, pitti, bug 1063618 updated
[18:33] <ubot2> Launchpad bug 1063618 in cairo "PDF printing output of evince broken with PDF file from "Der Spiegel"" [High,In progress] https://launchpad.net/bugs/1063618
[18:40] <seb128> tkamppeter, thanks
[19:04] <desrt> m_conley: ping
[19:04] <m_conley> desrt: pong
[19:04] <desrt> m_conley: i'm right next door :)
[19:04] <desrt> with will (new hire on the desktop team)
[19:04] <m_conley> desrt: you're in the community space? Cool. :)
[19:05] <chrisccoulson> desrt, oh, nice. lucky you ;)
[19:05] <m_conley> desrt: haven't seen you in a while - been travelling I guess?
[19:05] <desrt> yup
[19:05] <desrt> get to leave again on saturday for UDS
[19:05] <m_conley> classic desrt
[19:05] <desrt> your security situation here is getting pretty intense, eh?
[19:05] <m_conley> desrt: in what way?
[19:05] <m_conley> (I don't remember what it was like when you were here last...)
[19:05] <desrt> someone guarding the door, no longer open to non-mozilla hackers, etc.
[19:06] <m_conley> ah
[19:06] <m_conley> well
[19:06] <m_conley> they need to be vouched
[19:06] <desrt> you going to CPH, btw?
[19:07] <m_conley> desrt: hm, not sure what CPH is, so probably not. I'm dialing back the travel for a little while. I want to stay put for a little while. :)
[19:07] <desrt> copenhagen
[19:07] <Nafallo> m_conley: Copenhagen
[19:07] <m_conley> ah
[19:07] <m_conley> that's where the next UDS is, I guess? Yeah, I remember andreasn mentioning it
[19:08] <desrt> he's not on the formal invite list but i think he's going to pop in for a couple of days
[19:08] <desrt> his brother lives just across the river
[19:08] <m_conley> right, right
[19:08] <m_conley> I remember now
[19:08] <m_conley> cool. :)
[19:08] <m_conley> desrt: I'll probably join you over there in a bit
[19:09] <desrt> cool
[19:14] <chrisccoulson> desrt, m_conley, we need a UDS near the mozilla office ;)
[19:14] <m_conley> We do! Come to Toronto!
[19:14] <m_conley> we...have....some cool stuff sometimes
[19:14] <chrisccoulson> heh, i'm sure you do :)
[19:15] <m_conley> we have a pretty cool lake
[19:15] <m_conley> the tower's all crazy-lit at night
[19:15] <m_conley> and we have crazy varieties of food
[19:15] <seb128> but you don't let insiders in anymore it seems? :p
[19:15] <m_conley> I'm talkin' Toronto
[19:15] <Nafallo> that might be a tad much of the crazies ;-)
[19:16] <chrisccoulson> i'd enjoy it. it would make a nice change from the US and europe ;)
[19:16] <Nafallo> how about crazy Internet speeds? :-P
[19:16] <seb128> you start sounding like desrt
[19:16] <chrisccoulson> heh
[19:16] <m_conley> for MoTo, you just need to be vouched, I believe. I think it also helps if you're a Mozillian.
[19:16] <seb128> he keeps advertising the variety of food you can get in Toronto and how cheap it is
[19:17] <Nafallo> hrm. isn't Toronto Canada or something?
[19:17] <m_conley> Toronto is indeed Canada.
[19:17] <m_conley> WE ARE CANADA
[19:17] <Nafallo> that might have been the reason it's not on the list then ;-)
[19:17] <Nafallo> I can't remember the reason, but I think there might be one.
[19:18] <m_conley> polar-bear aversion?
[19:19] <Nafallo> ehrm, Canada != the North Pole, no :-)
[19:19] <Nafallo> and I don't think the North Pole have sufficient Internets.
[19:19] <Nafallo> the polar bears ate them all!
[19:19] <m_conley> this is true
[19:21] <Nafallo> that said, I don't think bears would be a problem.
[19:22] <Nafallo> concidering they went to Australia back in the days, with 90% of the world's dangerous animals... I don't think bears would come close :-)
[19:24] <desrt> m_conley: toronto is not "real canada", i hear :)
[19:25] <desrt> m_conley: but seriously... back me up on the insane variety of awesome food at cheap prices
[19:25] <m_conley> this is true
[19:25] <m_conley> we have a helluva lot of what desrt said
[19:25] <chrisccoulson> i like the sounds of this a lot
[19:27]  * desrt note sthat he has not been to khao san road in a while...
[19:27] <m_conley> it's delicious
[19:48] <Nafallo> London for UDS... just rent the Olympic Stadium or something...
[19:53] <xnox> Nafallo: the Olympic Park is getting disassembled and sold in parts. One building sold to germany, the other elsewhere. Some buildings will stay, but the stadium will be smaller....
[19:53] <Nafallo> yeah, I heard something about that.
[20:17] <chrisccoulson> yay, no menubar in panel with the latest firefox nightly
[20:17] <chrisccoulson> FML
[20:18] <seb128> chrisccoulson, I told you that last week!
[20:18] <chrisccoulson> seb128, oh, the nightly only broke for me today :/
[20:19] <Nafallo> ♥ vimperator
[20:19] <Nafallo> :emenu FTW
[20:23] <chrisccoulson> i don't use vimperator. the only addons i install are development aids or addons that i've written ;)
[20:23] <Nafallo> :-P
[20:23] <Nafallo> be without menu then :-D
[20:24] <Nafallo> it also makes people think twice before they try to borrow your browser ;-)
[20:25] <chrisccoulson> it's ok. i've found the problem and fixed it already ;)
[20:25] <Nafallo> *sigh* developers... ;-)
[20:26] <attente> seb128: i'm trying to use quilt to apply the patches on gtk+ minus the menu proxy patch
[20:26] <attente> but getting an error due to the later patches depending on that one
[20:27] <attente> is there a way to resolve it without resorting to hand-modifying the patches?
[20:28] <seb128> attente, you can use quilt -f to force it anyway, fix the conflicts and then quilt refresh to update it
[20:31] <seb128> bbiab
[20:56] <attente> seb128: how do you turn off the symbol checks and variants when running debuild?
[21:01] <seb128> attente, edit debian/rules
[21:01] <seb128> DEB_DH_MAKESHLIBS_ARGS_$(SHARED_PKG) += -V --add-udeb=$(UDEB_PKG) -- -c4
[21:01] <seb128> drop the "-- -c4"
[21:01] <seb128> same for the gail line
[21:01] <attente> thanks
[21:02] <desrt> not dropping the --add-udeb?
[21:02] <seb128> attente, you can man dpkg-gensymbols to see the -c level details
[21:03] <desrt> we want to only build the 'normal' variant...
[21:03] <seb128> desrt, well you can hack the build to not generate an udeb if you want, that was not the question
[21:03] <desrt> seb128: 'checks and variants'
[21:03] <seb128> desrt, that's not an available feature, I've a local diff hack I apply for that though ...we should turn it into an official build option
[21:04] <seb128> desrt, good exercice for attente? ;-)
[21:04] <desrt> heh
[21:04] <desrt> getting rid of 'make check' would also be nice
[21:04] <desrt> seb128: you know that he has not officially started work yet, right? :)
[21:06] <seb128> desrt, yeah, you are the ones coming with questions though ;-)
[21:06] <seb128> desrt, so I'm just saying 'patches are welcome' :p
[21:48] <FlyingElvis> https://lh6.googleusercontent.com/-ATG9fJYoB-k/UH7GCH-atxI/AAAAAAAAeiQ/btFIneKjAsg/s582/MRsays.png
[22:54] <xnox> should I be making merge proposals for lp:ubuntu/indicator-session or lp:indicator-session . It's a bugfix for 13.04.
[23:49] <bjsnider> jbicha, bug 1067933
[23:49] <ubot2> Launchpad bug 1067933 in mutter "Reloading shell with Alt-F2 r loses all workspaces but first 2" [Undecided,New] https://launchpad.net/bugs/1067933
[23:49] <bjsnider> did everything there i know how to. i have seen gnome bugs linked, and i don't know how to go about that.
[23:51] <jbicha> bjsnider: do you know how to be a debdiff?
[23:51] <jbicha> *make a debdiff?
[23:51] <RAOF> bjsnider: You're after the “Also affects project” link, and then you just past in your bugzilla url.
[23:52] <sarnold> hehe, *be* the debdiff. *feel* the debdiff. *become* the debdiff. *be one* with the debdiff.
[23:53] <bryce> http://www.markshuttleworth.com/archives/1195
[23:53] <bryce> Raring Ringtail :-)
[23:53] <jbicha> because here's the deal: it's not important enough of a bug to me for me to cherry-pick that patch...but I'm happy to see other people get involved in (gnome) packaging so I'll sponsor it for you if you do the work