CarlFKinstall oneiric, run app, compiz eats memory, doesn't release it when app is closed.  compiz trunk, no eat memory.00:11
CarlFKapp = dvswitch, it's from debian, http://alioth.debian.org/plugins/scmgit/cgi-bin/gitweb.cgi?p=dvswitch/dvswitch.git;a=blob;f=src/mixer_window.cpp;h=961f57a89a70f71ae2dc6e0d24e8086bf26df378;hb=HEAD00:12
=== eeejay_is_afk is now known as eeejay
CarlFKis ubuntu+1 installable, and have a different flavor of compiz?00:13
CarlFKwhat can I use to log compiz memory usage over a 12 hour period?01:17
bjsniderCarlFK, you could run it with valgrind01:24
CarlFKbjsnider: run compiz or dvswitch ?02:11
CarlFKguessing compiz - looking for how it is run... any tips?02:12
=== eeejay is now known as eeejay_is_afk
=== eeejay_is_afk is now known as eeejay
=== eeejay is now known as eeejay_is_afk
pittiGood morning03:48
pittijincreator: yes, so my question was why we need to put that snippet into language-selector when we can put it into ttf-nanum03:49
RAOFMorning pitti!03:51
pittihey RAOF03:52
RAOFxorg-server in oneiric-proposed is ready to be promoted to -updates, but launchpad times out when trying to copy it.  Just a heads up should you start your run through the sru queue :)03:53
bjsniderhyperair, is that libnotify fix getting into oneiric or am i the only one who gets to use it?03:54
hyperairbjsnider: it will be getting into oneiric.04:13
bjsnideroh, i thought i was special04:13
hyperairit's just that the fix i gave you was slightly sketchy, and i'd like to talk to the notify-sharp upstream before making an official upload04:13
hyperairand notify-sharp's upstream is kinda.. er.. missing.04:14
bjsniderthey're missing? hire indiana jones to find them04:16
hyperairwell i'm sure he'll be back04:16
bjsniderwho, indy?04:16
hyperairno i mean slomo04:20
pittiRAOF: ah, seems you are currently working on oneiric SRUs? doing natty/lucid queues then04:20
RAOFpitti: Yeah, making my way through the queue slowly!04:21
pittiRAOF: if you need a hand, I can start from the other end of the queue04:22
RAOFI suspect things are going to get less complicated than lightdm :)04:22
RAOFI'll be ok.04:23
pittiah, heh, yes04:23
pittiwell, I won't be bored today even without 2 hours of SRUs, still a lot to get done before UDS04:23
* RAOF heads out to collect Sam05:27
CarlFKI want to run compiz under valgrind - it it as simple as: pkill compbiz; valgrind compbiz?05:55
chrisccoulsongood morning everyone06:53
RAOFHey chrisccoulson07:01
RAOFHey ara ;)07:02
arahey RAOF07:03
chrisccoulsonhi RAOF, how are you?07:03
araI received your email, sorry I didn't reply yet. I will get mlegris to bring a hybrid system for you and one for bryceh to UDS07:03
RAOFchrisccoulson: I'm pretty well.  Strangely perking up after a slightly sleepy day's work ;)07:03
pittihey ara07:04
RAOFara: Woot!  Thanks!07:04
pittimorning chrisccoulson07:04
chrisccoulsonhi pitti07:04
brycehara, great thanks07:06
Sweetsharkpitti: does a blueprint starting with desktop-p but not tracking ubuntu, but libreoffice be tracked in status?08:29
pittiSweetshark: no, it only tracks ubuntu blueprints08:29
Sweetsharkpitti: thanks08:33
chrisccoulsonw00t, my 3g card just arrived09:12
chrisccoulsoni need to buy some screws though. why do people sell things like that without any screws?09:12
pittichrisccoulson: right in time for UDS :)09:52
chrisccoulsonpitti - yeah, but i'm currently desperately trying to find a M2x3 screw to fit it ;)09:55
chrisccoulsoni've already "borrowed" one from the keyboard to fit a bluetooth module09:55
chrisccoulsonah, it seems that my local maplin store has some in :)10:04
chrisccoulsonhmm, trying to use gdbus from jsctypes is a pain11:18
Sweetsharkpitti: what would be the right name for a rc-prerelease to ppa: libreoffice-3.4.4~rc1-0ubuntu1~ppa1?11:22
pittiSweetshark: looks ok11:25
Sweetsharkpitti: whats the orig-tarballname then? libreoffice-3.4.4_rc1.orig.tbz?11:45
pittiSweetshark: everything up to the first dash, i. e. libreoffice-3.4.4~rc1.orig.*11:48
pittierr, libreoffice_3.4.4~rc1.orig.*11:48
=== eeejay_is_afk is now known as eeejay
m4n1shcan someone review my merge request https://code.launchpad.net/~manishsinha/gedit/enable-zeitgeist-datasource-plugin/+merge/8056113:35
dobeymterry: i made the avahi change use breaks/replaces, and added a couple comments re: package splitting to the proposal13:42
mterrydobey, reading13:43
=== m_conley_away is now known as m_conley
mterrydobey, I still think it's technically incorrect.  I understand that it is just a port of the old system though.  What's more important than any of this is what the Debian maintainer does.  I'm assuming you haven't opened a bug in Debian about this yet?  Until that is settled, it's probably not a good idea to further split the gir packages.  So I'm fine with the branch as is, but please send the patch to Debian too, so we can eventually merge13:46
mterryI can sponsor today13:47
dobeyi haven't opened a bug on debian, no. really, i have no idea how to deal with getting changes into debian directly.13:49
mterrydobey, oh, there's a neat program called submittodebian.  Run it in the source directory after you make a change like this and it will guide you through.  You basically get a chance to edit the diff and then write an email13:50
pittigood bye everyone, see you on Monday! (I'm off tomorrow)14:28
dobeypitti: see you sunday :)14:30
=== eeejay is now known as eeejay_is_afk
mterrydobey, ok, now when I build your avahi package, the .typelib files get installed to /usr/lib/i386-linux-gnu/girepository-1.0/ instead of /usr/lib/girepository-1.0/.  Which means they don't get found by Python.  Do you know why that would be?14:44
dobeymterry: no idea. that happenened here too. presumably because debuild is passing --libdir=/usr/lib/i386-linux-gnu?14:47
mterrydobey, maybe change where they are installed in the .install file then, assuming this is an unexpected change (I could believe it's a good thing for multiarch though).  But at least Oneiric's python doesn't pick up that location14:48
mterrypitti, do you know if gobject-introspection typelibs are supposed to be installed in a multiarch location?14:49
dobeyi think pitti's gone until UDS :)14:49
mterrycjwatson, ^ ?14:50
dobeymterry: anyway, there's no rush on the avahi package. it's for precise. we can discuss in orlando if that makes it easier14:50
mterrydobey, fair14:50
dobeymterry: i am more concerned with the banshee SRU :)14:50
dobeybut i'm not sure what to do about the arbitrarily weird race condition inside banshee for that :-/14:51
m4n1shis powerpc still a supported architecture?14:53
m4n1shwhat is a main package fails to build on powerpc?14:53
mterrydobey, yar.  But it's a regression that's worse than slow startup  :-/14:53
m4n1shmain -> main repo14:53
dobeymterry: any ideas on how to debug it?14:55
mterrydobey, I'm not very familiar with mono.  But using actual debuggers probably isn't very good for a race anyway.  Can always fall back to printf debugging14:56
dobeymterry: right. problem is that when i tried to add that printf, the issue went away; and even removing the printf, i can't get it to happen again :(14:57
mterrydobey, oh.  :-/14:57
mterrydobey, did you reinstall from repos or from a build you made of no-printf?14:57
dobeymterry: that was building it from straight git master14:58
mterrydobey, if you reinstall oneiric's, do you get the bug?  That might be an interesting starting point14:58
mterrydobey, maybe git fixed something14:59
dobeyno, it was happening from git master. i added the logging and it went away, i removed the logging and rebuilt, and it was still not coming back up15:00
dobeyso basically, it was magic :(15:00
dobeymterry: and i'm testing this on 11.04, so it's not that. it's definitely an issue in banshee, i just have no idea how to isolate where it's happening :(15:01
mterrydobey, if you can make guesses about where in the code this might happen, you can add delays to try to exacerbate the race15:02
dobeywell, adding delay "fixes" it, as evidenced by the 10s it takes to load the u1ms link15:04
mterrydobey, right, I'm just saying that it might help debugging wise, if you can ensure that you always hit it15:16
dobeywell, more so, it ensures that i never hit it, it seems15:17
dobeyeven just adding the one call to log which source was being selected, caused it to not happen15:17
sroeckerhi, does anyone else have problems with their theme and icons?15:20
sroeckergnome-menu or language-pack update today seems to be the culprit15:24
mterrysroecker, in 11.10? no15:26
sroeckermterry: yes, 11.10. a few hours before everything seemed right, now the theme and icons are the default ones15:26
mterrysroecker, I'm not sure myself, but you might have more success in #ubuntu, the support channel15:27
sroeckeron two computers15:27
sroeckerok, just wanted to inform you before lots of users are starting to complain ;)15:27
mterrysroecker, fair  :)15:28
mterrysroecker, if #ubuntu doesn't help, try filing a bug15:28
sroeckeryep, will do15:28
sroeckermterry: ubuntuforums.org/showthread.php?p=11397977 seems to be another user with my problem15:39
hyperairokay, nautilus is driving me nuts. can someone help me test (in oneiric) whether the lookahead works more than once?16:03
hyperairi.e. use nautilus's look-ahead search thing to locate a folder, press enter, and then type something again into the look-ahead search (to match something in the folder you just entered)16:04
hyperairdoes it append to the previous lookahead query, or does it start from a blank slate again (as it should)?16:04
Laneyworks properly16:06
hyperairoh ffs.16:14
hyperairit works in the guest session16:14
hyperairwtf happened?!16:14
bjsniderhyperair, busted userland config file i'd say17:08
hyperairbjsnider: yes, that much was obvious.17:08
hyperairbjsnider: but what got busted?17:09
hyperaircertainly not nautilus's setinsg17:09
bjsniderwhy not?17:09
hyperairbjsnider: because i purged everything under /org/gnome/nautilus in dconf and it's still there17:18
bjsnideri c17:19
hyperairthis is annoying17:19
dobeyhyperair: oh, the type-ahead search in nautilus?17:58
hyperairdobey: yeah. you notice that issue too?17:59
dobeyhyperair: it's all kinds of messed up in nautilus 3.2; there's a bug report on it, which is also filed upstream17:59
ubot2Gnome bug 662591 in File and Folder Operations "can't use backspace in typeahead (goes back one folder)" [Normal,Unconfirmed]17:59
dobeyhyperair: lots of people have.17:59
hyperairdobey: i think i can work around this..17:59
dobeyhyperair: best workaround i think is to hit C-l and use the completion feature on that18:00
hyperairheh lol18:00
hyperairdobey: no i mean i'm digging around inside nautilus-icon-container.c now..18:01
dobeyreset it to the 2.32 version? ;)18:01
hyperairdobey: read my comment there. there seems to be some weird issue where both the icon-container (the widget that you can click and select icons in) and the search entry both appear to be focused at the same time18:02
hyperairwhich obviously should not happen18:02
dobeyhyperair: i don't think the search entry is actually focused, hence the problem. there is some weird issue where if you try to use typeahead, after you used typeahead to find a folder, and pressed RET to open the folder, it goes nuts and stops working right18:03
hyperairthis looks like where things are screwing up.18:03
hyperairdobey: yeah, i figured out why.18:03
hyperairdobey: when the search entry is focused, enter causes the thing to get cleared18:03
hyperairwhen not, it just activates teh selected item18:04
hyperairso it looks the same, but doesn't clear the search entry18:04
dobeyyeah that code doesn't look kosher ;)18:04
hyperairadditionally, the search entry only works at all because events are being forwarded (wrongly) into the search entry from the icon-container.18:04
hyperairspecifically this part18:05
hyperairnotice how backspace is filtered out, which is why your backspaces don't work on the search entry18:06
hyperairand /, and delete18:06
hyperairin fact, they're handled all the way up there, so things work out really strangely.18:06
dobeyhrmm, i recall backspace working *before* i hit enter to open the searched-for directory18:11
desrtdbarth: hi :)18:11
dobeyanywya, i agree nautilus is broken :)18:11
dobeyand sadly, so is banshee, which i have to debug :-/18:12
hyperairdobey: does backspace still work for you before you hit enter?18:15
dobeyhyperair: let me boot up oneiric and see18:16
dobeyhrmm. this banshee race seems to have 2 issues as a result18:18
dobeyhyperair: yeah, backspace works fine before i hit enter18:19
hyperairdobey: ugh. this makes things really weird.18:19
dobeyreally weird seems to explain what i have been dealing with lately, fairly well :)18:20
dobeyi wish i could understand what banshee is doing exactly here18:21
hyperairdobey: which bug?18:29
dobeyhyperair: there's a race at start-up, so when we try to switch the source, it doesn't happen.18:32
hyperairadd printfs?18:32
hyperairi mean Console.WriteLines18:32
dobeyhyperair: but when i added a logging call, it started working, except now it switches, but the selection in the treeview is still wrong18:32
hyperairurgh, that really sucks18:33
dobeyyeah. it's blocking my SRU :(18:33
hyperairyay, my fix works!18:37
hyperairdobey: build nautilus with that patch in place and you should get back your typeahead18:38
dobeyhyperair: shouldn't lines 25 and 26 be swapped in that patch?18:39
dobeyugh, banshee package behaves the same as from git master :-/18:41
dobeyand i didn't even add/remove logging with the packaged version18:41
hyperairdobey: apparently not. i copy-pasted the code from a little further down the same functino.18:43
=== zyga is now known as zyga-afk
hyperairdobey: just before gdk_event_free, the value of new_event->window was restored to what it was originally.18:43
* hyperair shrugs18:43
hyperairah whoops, but i should have set ->window = gtk_widget_get_window of search_entry or something18:44
dobeyoh i guess line 26 is just not necessary18:44
dobeyas also is line 24 probably not necessary18:44
dobeyah, what an awesome fix18:58
dobeyGLib.Idle.Add ftw18:58
mterrydobey, for banshee?19:07
dobeymterry: yes19:08
dobeymterry: also; ICP? o_O19:08
mterrydobey, :)  Fixes it to wait for idle loop?19:08
mterrydobey, heh.  Not my usual music choice19:08
dobeymterry: it appears to a problem with calling a callback handler19:09
=== zyga-afk is now known as zyga
m4n1shjbicha: do you have a bit of time for another merge request?19:48
m4n1shjbicha: this merge request https://code.launchpad.net/~manishsinha/gedit/enable-zeitgeist-datasource-plugin/+merge/8056119:49
dobeymterry: ping; can you test the updates i just pushed to my banshee SRU branch? should fix the race for you as well20:00
mterrydobey, ok20:00
mterrydobey, can you link me to the branch?  I'm on a different machine right now20:03
mterrydobey, nm, got it20:03
dobeybbiab even20:23
dobeytotally lost track of time. and i had an appointment at 4 :-/20:23
dobeyerr, or will have that a bit later i guess :-/20:26
dobeymterry: please tell me it works for you now20:30
mterrydobey, will do.  I have to install a dev environment on this machine, building banshee now20:30
dobeyah ok20:31
mterrydobey, looks good to me!20:39
mterrydobey, will sponsor into oneiric-proposed20:39
dobeymterry: yay!20:40
=== m_conley is now known as m_conley_away
=== SoulShadow is now known as Shadows`class
=== tkamppeter_ is now known as tkamppeter

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!