[08:10] <didrocks> good morning
[08:11] <pitti> Good morning
[08:11] <dpm> morning didrocks
[08:11] <dpm> good morning all
[08:12] <didrocks> hey pitti, dpm
[08:31] <Sweetshark> G'Morning, all!
[08:32] <didrocks> hey Sweetshark :)
[08:32] <didrocks> how was your travel to return home?
[08:32] <pitti> hey Sweetshark, good morning! enjoyed fosdem?
[08:34] <Sweetshark> didrocks: better than expected: there had ben verz haevy rains in Hamburg this weekend and we expected trouble, but it went well ...
[08:36] <didrocks> Sweetshark: oh really? Julie told me that I missed a very sunny week-end in Lyon (with a little "ahahah, you should have been home, see!" :p)
[08:37] <pitti> yeah, it was awesome here as well, warm and sunny
[08:38] <Sweetshark> pitti: Absolutely! I think it was a great success for Libreoffice too. Some of the talks of the LO track had the room really packed. The booth was also very wellvisited (other than the OOo booth 3 Meters next to it) ....
[08:39] <pitti> lol - good to hear!
[08:42] <didrocks> and the booth was wayyyyy bigger. I found that fun :)
[08:56] <pitti> bonjour seb128
[08:57] <seb128> hello pitti
[08:57] <seb128> how are you?
[08:57] <chrisccoulson> hey seb128 and pitti, how are you?
[08:58] <didrocks> hey chrisccoulson, how are you? ;)
[08:58] <seb128> hey chrisccoulson, hey again didrocks
[08:58] <chrisccoulson> hi didrocks! i'm good thanks, how are you?
[08:58] <seb128> I'm fien thanks
[08:58] <seb128> I'm fine as well :-)
[08:59] <didrocks> chrisccoulson: little bit tired, but I'm fine, thanks :)
[09:02] <chrisccoulson> yeah, i'm quite tired too. i was rudely awoken at 6.30am by my daughter
[09:02] <chrisccoulson> i wish she would realise that waking up when it is still dark outside is not normal ;)
[09:02] <didrocks> urgh ;)
[09:15] <chrisccoulson> hey pitti - did you have a chance to try the latest version of my extension? i fixed a few more crashes on friday too
[09:15] <chrisccoulson> hopefully it stops crashing for you now ;)
[09:15] <pitti> chrisccoulson: hm, I did dist-upgrade, which version should I have now?
[09:15] <chrisccoulson> pitti - 0.4
[09:16] <pitti> chrisccoulson: I have that; so, let me switch back to unity and compiz and check
[09:16] <chrisccoulson> thanks :)
[09:16] <pitti> chrisccoulson: I noticed that under metacity/gnome (no global menu) I still have the "5 seconds of CPU until a click is processed" issue
[09:17] <chrisccoulson> pitti - try turning off the testpilot extension
[09:17] <chrisccoulson> i've seen a few people say that's eating CPU
[09:17] <pitti> testpilot?
[09:17] <chrisccoulson> pitti - yeah, it's called "Feedback" in the addons manager
[09:17] <pitti> oh, is that the "weekly report" kind of thing?
[09:18] <chrisccoulson> yeah, that's the one
[09:18] <pitti> chrisccoulson: ok, I'll try that, too, once it comes back
[09:18] <chrisccoulson> thanks
[09:18] <pitti> still early in the morning, it hasn't gone bad yet
[09:18] <chrisccoulson> testpilot will go away automatically when we get a RC build anyway
[09:19] <chrisccoulson> (which will be quite soon)
[09:25] <rodrigo_> morning
[09:25] <pitti> hey rodrigo_, had a nice weekend?
[09:26] <rodrigo_> hey pitti
[09:29] <seb128> hey rodrigo_
[09:29] <seb128> rodrigo_, how are you?
[09:30] <rodrigo_> hi seb128, I'm fine after a week of vacation, thanks :-)
[09:30] <rodrigo_> seb128, and you=
[09:30] <rodrigo_> ?
[09:30] <seb128> I'm great thanks!
[09:30] <seb128> rodrigo_, what do you plan to work on this week? ;-)
[09:31] <rodrigo_> seb128, unity a11y, gnome3 ppa, and anything other you tell me to do :-)
[09:31] <seb128> rodrigo_, the tomboy gconf cleaning? ;-)
[09:32] <rodrigo_> seb128, oh yes, been looking at that last week
[09:32] <Laney> did sandy say that was acceptable upstream?
[09:32] <seb128> oh, nice, do you think it's some not hard to do?
[09:32] <rodrigo_> well, it needs to be changed in gconf#, doing it in tomboy only is a lot of work
[09:33] <seb128> rodrigo_, what needs to the changed in gconf#?
[09:33] <seb128> rodrigo_, the goal was to use gconf# rather than the gconfpeditor thing no?
[09:33] <seb128> Laney, dunno, rodrigo_ said he would check
[09:33] <Laney> ok then
[09:34] <rodrigo_> seb128, gconfpeditor is in gconf#, and tomboy has a lot of classes hierarchy based on that, so that it works on windows only
[09:34] <rodrigo_> s/only/also
[09:35] <seb128> oh, k, sucks
[09:35] <seb128> well maybe not worth the work there
[09:35] <rodrigo_> well, I think we can remove the dep in gconf#, I'll let you know today
[09:35] <seb128> it's annoying we are like 2 rdepends away from dropping libbonoboui, etc
[09:35] <rodrigo_> yeah
[09:36] <seb128> not sure why the gconfpeditor bindings are in libgnomeui
[09:36] <seb128> that brings in gnomevfs etc
[09:36] <rodrigo_> yes
[09:36] <rodrigo_> need to move to my desktop pc, brb
[09:56] <dpm> hi pitti, morning. I saw that the full tarball export for the Lucid language packs worked - https://translations.launchpad.net/ubuntu/lucid/+language-packs, but I cannot see the packages in the PPA yet. Was there any build failure, or are they still building?
[10:00] <pitti> hey dpm -- I'm currently building fresh lucid langpacks, with -base refresh
[10:02] <dpm> hey pitti, ah, is that not an automated process? I thought they'd be built in the PPA during the weekend
[10:02] <pitti> dpm: only the updates; they were actually, I got a few error messages from it
[10:03] <pitti> dpm: I disabled the lucid cron job now, and building from teh 0205 base export
[10:04] <dpm> pitti, ok. Shall I send the testing announcement to translators saying that packages will be built soon, or shall I wait on your confirmation?
[10:05] <pitti> dpm: I think you should wait until they are actually built
[10:05] <pitti> I'll ping you
[10:05] <pitti> I also need to test one first, for firefox
[10:06] <dpm> ok, will wait for your ping then. Thanks pitti!
[10:32] <seb128> rodrigo_, could you check with upstream about this g-s-d gdm race issue?
[10:32] <rodrigo_> seb128, yes
[10:33] <seb128> rodrigo_, I've assigned you the bug as well, that's something we should fix this cycle, we get extra users unhappy about it every day it seems
[10:33] <seb128> rodrigo_, thanks
[10:33] <rodrigo_> right
[10:34] <seb128> k
[10:35] <seb128> jibel: hey
[10:36] <seb128> jibel: bug #712630, the milestone seems wrong? you set a 10.04.3 one but the issue is 11.04 one
[10:36] <ubot2> Launchpad bug 712630 in gnome-session "Ubuntu Alternate: Empty desktop after installation with 2D fallback" [High,Confirmed] https://launchpad.net/bugs/712630
[10:38] <jibel_> seb128, sorry, monday morning :( and already 10.04.2 in mind. Fixing.
[10:38] <seb128> jibel_, no worry, thanks
[10:38] <seb128> didrocks, ^  seems a bug for you btw
[10:38] <didrocks> let me look
[10:39] <didrocks> ok, I'll have a look today, should install daily live on virtualbox
[10:40] <didrocks> hum, this was before alpha1
[10:40] <didrocks> alpha2
[10:40] <didrocks> I think it was the compiz crashing and shouldn't be valid starting from next Thursday with nux testing compiz as well
[10:41] <didrocks> jibel_: do you still have this virtualbox?
[10:42] <jibel_> didrocks, I can try to reproduce, but another user confirmed it this morning on #u-testing.
[10:42] <didrocks> jibel_: right, I just want to know if you can launch a shell and check that gnome-panel is running
[10:42] <didrocks> (even if not shown)
[10:43] <didrocks> if it's running, xwininfo -root -tree and put that in a log file
[10:43] <jibel_> didrocks, I trashed the VM, give me time to reinstall it and I'll tell you.
[10:44] <didrocks> jibel_: excellent, thanks :)
[10:44] <didrocks> jibel_: my bet is that it's a dup of https://bugs.edge.launchpad.net/ubuntu/+source/compiz/+bug/711378
[10:44] <ubot2> Launchpad bug 711378 in compiz "after compiz crashed, gnome-panel isn't mapped again" [High,Triaged]
[10:59] <dpm> pitti, also, could we have the Natty langpack uploads back now that we're post freeze?
[10:59] <pitti> dpm: already done
[10:59] <pitti> still haven't had time to look into the firefox translations
[10:59] <pitti> but I just fixed a bug in the po2xpi wrapper script
[10:59] <pitti> which broke firefox translations for lucid
[11:00] <pitti> (lucid now re-building)
[11:00] <pitti> this has affected the natty build as well
[11:00] <dpm> ah, bummer. Thanks pitti
[11:00] <pitti> but I want to produce a mini-tarball and integrate it into the test suite at some point
[11:01] <dpm> pitti, I'm not familiar with it, which test suite? What are these plans?
[11:01] <pitti> dpm: langpack-o-matic's
[11:03] <chrisccoulson> oh, we're going to see firefox translations this week?
[11:03] <chrisccoulson> nice :)
[11:03] <pitti> well, for lucid langpack updates
[11:03] <pitti> natty uses the new po2xpi branch
[11:04] <pitti> I don't know whether that one is actually working :)
[11:23] <jibel_> didrocks, gnome-panel is running, I attached the output of xwininfo to the report. I'll keep this VM for a few days if needed, let me know if you need other info.
[11:23] <didrocks> jibel_: ok, so it's bug #711378
[11:24] <ubot2> Launchpad bug 711378 in compiz "after compiz crashed, gnome-panel isn't mapped again" [High,Triaged] https://launchpad.net/bugs/711378
[11:24] <didrocks> jibel_: can you please attach it to this one and mark yours as dup?
[11:24] <jibel_> didrocks, sure
[11:24] <didrocks> it's not a gnome-session issue in any case
[11:24] <didrocks> jibel_: what I will do it with next nux release, the fallback will be handle diffenrently and we will avoid this bug
[11:25] <didrocks> but it will still be there, the restacking isn't working right
[11:25] <didrocks> thanks jibel_ :)
[11:31] <jibel_> didrocks, you're welcome, I've added a note about nautilus which suffers the same symptoms.
[11:31] <didrocks> jibel_: yeah, basically everything which was launched by gnome-session before metacity (once compiz crashes…)
[11:32] <didrocks> jibel_: the "no effect" session is working well, isn't it?
[11:33] <jibel_> didrocks, apparently, yes.
[11:34] <didrocks> jibel_: ok, thanks for confirming, so it's really when compiz is crashing… ok, I asked smspillaz as I think he's the most knowledgeable on that
[11:38] <tkamppeter> pitti, hi
[11:45] <pitti> hey tkamppeter, how are you?
[12:01] <tkamppeter> pitti, fine
[12:01] <tkamppeter> pitti, I have added a Poppler-based pdftoraster filter to CUPS, to address bug 668800.
[12:01] <tkamppeter> pitti, see the BZR.
[12:01] <ubot2> Launchpad bug 668800 in cups "Printing speed problem" [High,In progress] https://launchpad.net/bugs/668800
[12:01] <tkamppeter> pitti, but it is not ready for upload yet, as now both cups and ghostscript provide pdftoraster, so the packages conflict.
[12:02] <pitti> tkamppeter: does that apply to Debian as well, or just our gs?
[12:06] <tkamppeter> pitti, what I want to do is an update-alternatived, but this needs a change in both cups and ghostscript, default should be the Poppler-based one.
[12:06] <pitti> I don't like abusing alternatives for that really
[12:07] <pitti> alternatives should be for user invoked programs
[12:07] <pitti> not for driver things like pdftoraster; cups should know which ones is better and use that, and if we have several, they should have a different name
[12:07] <pitti> tkamppeter: why not name the new cups one pdftoraster-poppler
[12:07] <pitti> ?
[12:07] <tkamppeter> pitti, this is no problem for UIbuntu, as I can quickly change ghostscript, too. In Debian we must wait for the Debian maintainer, so better disable it in Debian for the time being.
[12:08] <tkamppeter> pitti, yes, I could do so, and then by a .convs file control which one gets used.
[12:09] <tkamppeter> pitti, but now if I deliver a pdftoraster-poppler.convs from cups and a pdftoraster.convs from Ghostscript, which one will get used.
[12:10] <tkamppeter> pitti, probably I have to modify the cost factors in a way that the inactive gets a very high cost factor and if problems occur (bug reports) I can tell the user to change the cost factors to switch over.
[12:10] <pitti> tkamppeter: don't they have a priority?
[12:12] <tkamppeter> pitti, the priority is managed by the cost factors in the .convs files. I have worked out the cost factors so that a PDF workflow gets used by default. The only exception is to have a PS workflow if incoming data is PS and the printer is PS.
[12:12] <tkamppeter> pitti, so I will do the following:
[12:14] <tkamppeter> pdftoraster-poppler will get the cost factor with Ghostscript's pdftoraster had before and Ghostscript's pdftoraster will get commented out.
[12:14] <tkamppeter> pitti, these changes I will do via the debian/rules files to make them only affect the Debian/Ubuntu packaging. Upstream will not get changed.
[12:16] <pitti> tkamppeter: you mean a patch? sure
[12:29] <rodrigo_> bug #579300
[12:29] <ubot2> Launchpad bug 579300 in linux "Please disable CONFIG_SOUND_OSS* and CONFIG_SND_*OSS*" [Wishlist,Fix released] https://launchpad.net/bugs/579300
[12:41] <tkamppeter> pitti, now I have pushed the inclusion of the convs file in the PDF filter add-on. The "make install" of CUPS now installs /usr/lib/cups/filter/pdftoraster and /usr/share/cups/mime/pdftoraster.convs in addition.
[12:42] <pitti> good
[12:42] <pitti> thanks
[12:43] <tkamppeter> pitti, a Debian build should remove these two files, an Ubuntu build should rename them to /usr/lib/cups/filter/pdftoraster-poppler and /usr/share/cups/mime/pdftoraster-poppler.convs and do s/pdftoraster/pdftoraster-poppler/ in the .convs file.
[12:44] <pitti> tkamppeter: why does it need to be different on D and U?
[12:44] <pitti> can't we use the -poppler name in both?
[12:47] <tkamppeter> Theoretically we can do. Build and install will perfectly work in both distros. Problem is that then there are two filter workflows with exactly the same cost: The ones which use pdftoraster of Ghostscript and the ones which use pdftoraster-poppler. On Ubuntu I will fix this today by modifying ghostscript. On Debian I will inform the ghostscript maintainer but it can take some days until he switches over. Or can you also upload Debian
[12:47] <tkamppeter> 's ghostscript?
[12:49] <pitti> no, I can't
[12:49] <tkamppeter> pitti, if you do anything on cups to implement this, please do not change anything in debian/local/filters/pdf-filters/, as this directory is a copy of the upstream add-on package. The implementation of the soluttion of our problem should be done only in the distro packaging.
[12:50] <pitti> tkamppeter: can we wait with the cups upload until you discussed that with the gs maintainer?
[12:50] <pitti> tkamppeter: but shouldn't cups just put a heavier weight on the gs filter then? then you can have both installed, and switch over in the .conf if the gs one works better
[13:00] <mpt> mvo_, hi, I'd like to make lots of little tweaks to the USC reviews presentation. Which do you think would be faster for you: (1) I make a branch of the 3.0-reviews branch, which you then merge onto both 3.0-reviews and trunk; or (2) I give you a long boring list?
[13:00] <mvo_> definetly (1)
[13:01] <mpt> ok
[13:01] <mvo_> thanks! looking forward to that merge
[13:01] <mvo_> mpt: let me know if you need anything (to find the right places etc)
[13:02] <pitti> chrisccoulson, dpm: argh, now the lucid langpacks built, but:
[13:03] <pitti> mozilla.tar.gz is only 233 bytes, and only has a folder
[13:03] <pitti> ./usr/lib/firefox-addons/extensions/langpack-de@firefox-3.6.ubuntu.com/
[13:03] <pitti> hm, I think I know why, hang on
[13:03] <pitti> it probably used the new branch
[13:03] <dpm> ah
[13:04] <pitti> if [ "$3" = lucid -o "$2" = maverick ]; then
[13:04] <pitti> it was called with lucid-proposed
[13:04] <tkamppeter> pitti, I have looked into it now, and we can simply let CUPS put a heavier weight on its own pdftoraster. This way we do not need any change on ghostscript and also no difference between Ubuntu and Debian. So I can complete the CUPS package.
[13:04] <dpm> phew, glad it was an easy fix
[13:07]  * pitti runs again, darn
[13:52] <kenvandine> good morning all
[13:52] <kenvandine> seb128, nudge about geoclue
[14:00] <mpt> mvo_, first silly question: The stars are currently red, but the only star graphics I see in the tree (in data/images/) are yellow and orange. Where are the red ones?
[14:00] <mpt> (orange one in data/emblems/)
[14:00] <seb128> re
[14:00] <seb128> seems I got disconnected
[14:02] <kenvandine> seb128, nudge about geoclue
[14:02] <kenvandine> :)
[14:02] <seb128> kenvandine, hey, oh right!
[14:02] <seb128> doing it now this time
[14:02] <kenvandine> :-D
[14:02] <kenvandine> thx
[14:02] <seb128> kenvandine, how are you?
[14:02] <kenvandine> great
[14:02] <kenvandine> and you?
[14:03] <seb128> I'm fine thanks
[14:03] <seb128> e gpg
[14:03] <seb128> ups
[14:06] <tkamppeter> pitti, cups package is tested and updated now so that it actually uses the new pdftoraster in both Debian and Ubuntu and without need of change on ghostscript. Can you please upload it? Thanks.
[14:06] <kenvandine> :)
[14:07] <pitti> tkamppeter: nice! yep, will do ASAP
[14:07] <seb128> kenvandine, ok, will do in a bit when indicator-appmenu will stop crashing my command lines when I try to ssh to the server
[14:07] <kenvandine> ugh
[14:07] <kenvandine> weird crash
[14:13] <mvo_> mpt: we draw them now manually, let me quickly check if you have the very latest version
[14:14] <mpt> mvo_, revno 1127
[14:14] <mvo_> mpt: please update to r1128 and see if that is better
[14:14] <mvo_> mpt: the stars should no longer be red, I fixed that a while ago
[14:14]  * bcurtiswx waves to room
[14:16] <mpt> mvo_, they're still red in r1128
[14:16] <mvo_> oh, hold on a sec
[14:17] <seb128> kenvandine, ok, done
[14:17] <kenvandine> seb128, thx!
[14:17] <seb128> kenvandine, it was just geoclue?
[14:17] <kenvandine> ubuntu-geoip too
[14:17] <seb128> both got mir acked right?
[14:17] <kenvandine> yup
[14:17] <seb128> kenvandine, ok, both done
[14:18] <kenvandine> seb128, thx!
[14:18] <seb128> yw
[14:18]  * kenvandine makes i-d build
[14:18] <mterry> seb128, what's this about indicator-appmenu crashing your ssh?
[14:18] <seb128> kenvandine, can you do some libdbusmenu backporting? the one commit from mterry after the release
[14:18] <seb128> mterry, hey
[14:18] <mvo_> mpt: please try r1129
[14:19] <seb128> jibel_, what was the bug number you pinged about before?
[14:19] <seb128> jibel_, the appmenu crasher
[14:20] <jibel_> seb128, tedg, bug 703988
[14:20] <ubot2> Launchpad bug 703988 in vim "(various) crashed with SIGSEGV in g_atomic_int_exchange_and_add()/g_variant_unref/?libappmenu.so/g_simple_async_result_complete" [Undecided,New] https://launchpad.net/bugs/703988
[14:20] <seb128> mterry, ^
[14:20] <tedg> jibel_, Yeah, I need to see if I can recreate that.
[14:21] <tedg> Doesn't look pleasant.
[14:21] <seb128> not sure why ssh-ing to the d-c triggers it but I couldn't keep it running enough to run a command
[14:21] <jibel_> tedg, just run vim-gnome in unity and wait
[14:21] <seb128> I had to turn off unity now
[14:21] <mpt> mvo_, ok, they're now orange instead of red, but where is that defined?
[14:22] <kenvandine> seb128, i'll backport it
[14:22] <seb128> kenvandine, btw nice that you got libindicate to build, that seems a workaround though?
[14:22] <seb128> kenvandine, thanks
[14:22] <jibel_> tedg, I also now get the same trace with gucharmap running on classic desktop without effects
[14:23] <kenvandine> seb128, yeah... no idea why i had to do that
[14:23] <kenvandine> i so wanted to figure out wtf happened
[14:23] <mvo_> mpt: check softwarecenter/view/widgets/reviews.py (self.bg_color, self.fg_color)
[14:23] <kenvandine> but i spent too much time on it... :)
[14:24] <tedg> jibel_, Hmm, yeah, I don't :(
[14:25] <micahg> hi seb128, I prepared the pidgin update as discussed on Friday, it's a merge proposal w/ubuntu-desktop and ubuntu-dev as reviewers, so hopefully on both sponsorship lists
[14:25] <mterry> seb128, bookmarked, i can look at it eventually if no one else does
[14:25] <tedg> jibel_, Can you run dbusmenu trunk and see if it happens for you?  There are a couple of fixes there that may have gotten this one as well.
[14:26] <jibel_> tedg, ok
[14:26] <seb128> tedg, natty has your most recent commit backported
[14:26] <seb128> tedg, the only other commit was mterry's update fix, would that make a difference for that crash?
[14:27] <mterry> seb128, I don't think it would
[14:27] <tedg> seb128, No, it shouldn't.  Okay, so that's not a useful exercise :)
[14:27] <tedg> Wonder why it doesn't happen here then.
[14:27] <seb128> tedg, how did you try?
[14:28] <tedg> gdb vim.gnome -ex run and gdb gucharman -ex run
[14:28] <seb128> not sure what makes it crash, got several softwares crashing this way today
[14:28] <seb128> but it seems the easiest way there to crash something is to ssh from a command line
[14:28] <seb128> not sure if it has to do with title updates
[14:29] <tedg> seb128, Hmm, odd.  Why is there no retrace?  Are the retracers down?
[14:29]  * tedg hopes not
[14:29] <seb128> let me check
[14:29] <kenvandine> seb128, mterry: was there a specific bug for that dbusmenu fix i am backporting?
[14:30]  * tedg challenges seb128 to do it without using SSH ;)
[14:30] <mterry> kenvandine, bug 709839
[14:30] <ubot2> Launchpad bug 709839 in libdbusmenu "some apps are missing menu items" [Undecided,Confirmed] https://launchpad.net/bugs/709839
[14:30] <kenvandine> thx
[14:31] <seb128> tedg, ok, retracers are down on a stupid upgrade error, I'm cleaning that and restarting
[14:31] <tedg> seb128, Great, thanks!
[14:31] <kenvandine> seb128, does that crash make you not be able to type in any window?
[14:31] <seb128> kenvandine, no
[14:32] <kenvandine> humm... i just tried to ssh and suddenly lost all keyboard input
[14:32] <kenvandine> well, in windows
[14:32] <kenvandine> could get to a VT
[14:32] <kenvandine> mouse worked though, just couldn't type in anything
[14:32] <kenvandine> appmenu crashed too
[14:33] <seb128> kenvandine, you might have hit this one
[14:33] <seb128> bug #714531
[14:33] <ubot2> Launchpad bug 714531 in indicator-appmenu "unity-panel-service consuming very high cpu time" [Undecided,New] https://launchpad.net/bugs/714531
[14:33] <seb128> kenvandine, chrisccoulson said it was blocking applications for him sometimes
[14:33] <kenvandine> ah
[14:33] <kenvandine> that would make sense
[14:34] <seb128> not a great start of week for appmenu-gtk ;-)
[14:34] <kenvandine> didn't look at the load
[14:34] <kenvandine> hehe :)
[14:34] <kenvandine> yeah
[14:35] <seb128> tedg, ok, retracers should be running again, let's wait a bit
[14:36] <didrocks> pitti: can you accept the invit for ~ubuntu-core-dev to be part of ~utouch-packaging (seeing the Technical board is the administrator)
[14:37] <pitti> didrocks: ~utouch-packaging should have a contact email set first
[14:37] <didrocks> pitti: oh right, for the spamming issue, let me try to fix it
[14:37] <pitti> otherwise every ubuntu-core-dev will get bug mail
[14:37] <didrocks> pitti: yeah, I keep forgetting about that one, sorry :)
[14:38] <mterry> So...  looking for some help here in debugging SWT Java apps and appmenu: GTK+ in SWT sees an XID that is reliably 48 higher than the XID that libwnck sees for the window.  This mismatch means appmenu doesn't show anything.  Where should I be looking to explain that 48?  There don't seem to be any offscreen rendering or parenting issues that I can see
[14:38] <tedg> smspillaz, ^ any ideas?
[14:40] <smspillaz> mterry: why are you using libwnck for menus ?
[14:40] <smspillaz> libwnck only tracks windows in _NET_CLIENT_LIST
[14:40] <mterry> smspillaz, its being used through bamf on the appmenu side
[14:40] <smspillaz> so you're using libwnck to get the XID of the actual application then?
[14:40] <tedg> smspillaz, He's looking for the XID of the window.
[14:41] <smspillaz> and by "window" you mean, the actual window and not the X window of the menu right?
[14:41] <mterry> smspillaz, right
[14:41] <mterry> smspillaz, the gdk window of the toplevel
[14:41] <smspillaz> mterry: right
[14:42] <smspillaz> mterry: so libwnck will get the one out of _NET_CLIENT_LIST (which works fine, otherwise there'd by no window decorations) - how are you getting the xid from GTK+ ?
[14:43] <mterry> smspillaz, GDK_WINDOW_XID of the toplevel that contains the menu.  I'm thinking I have the wrong toplevel, but that would be odd, as it does contain the menu.
[14:43] <mterry> smspillaz, XID is just an xlib Window pointer, right?
[14:44] <smspillaz> mterry: yeah, that's right
[14:44] <mterry> smspillaz, so what would result in a 48 byte difference between the Real window and my window?
[14:44] <smspillaz> mterry: let me think
[14:45] <smspillaz> mterry: the only thing that would explain that is that the one in _NET_CLIENT_LIST and the gdk one are different windows
[14:45] <smspillaz> mterry: sudo apt-get install wmctrl and then wmctrl -l and see if your gdk xid is in there
[14:46] <mterry> smspillaz, it's not
[14:46] <mterry> smspillaz, or at least, it's not in the list of windows wnck knowss about
[14:46] <mterry> I haven't done wmctrl, but I assume it's the same list?
[14:46] <smspillaz> it should be
[14:46] <smspillaz> I'm thinking there could be a bug somewhere though
[14:46] <smspillaz> so wmctrl -l will just tell you whats in _NET_CLIENT_LIST
[14:46] <smspillaz> and then we'll know if wnck or bamf is buggy
[14:51] <mterry> smspillaz, wmctrl agrees with wnck
[14:51] <smspillaz> uh-huh
[14:51] <smspillaz> mterry: and your window gets decorated?
[14:53] <pitti> WAH! third attempt of langpack building failed again *sigh*
[14:53] <seb128> jibel_, could you submit one of the crashes you get with the crashdump as well?
[14:54] <jibel_> seb128, okay, uploading a 42MB report for simple-scan now.
[14:56] <seb128> jibel_, or install the dbgsym locally if you prefer
[14:56] <seb128> brb
[14:58] <mterry> smspillaz, whoops, I disconnected.  Let me recap what I may have sent into the void
[14:58] <mterry> smspillaz, wmctrl agrees with wnck
[14:58] <mterry>  There's no secret transient-for window...
[14:58] <mterry>  smspillaz, but remember that this is GTK inside of Java's SWT library.  Could it be doing some funky proxying or some such?  The SWT library code doesn't appear to be, but maybe there's something there
[15:00] <jibel_> seb128, bug 714630
[15:00] <ubot2> jibel_: Error: Could not parse data returned by Launchpad: list index out of range (https://launchpad.net/bugs/714630)
[15:00] <seb128> jibel_, thanks, let's wait for retracing now
[15:01] <smspillaz> mterry: swt is just funky in general
[15:02] <smspillaz> mterry: maybe let me know the gdk xid of the window and then post your xwininfo -root -tree ?
[15:02] <smspillaz> so we can see where it is sitting
[15:02] <mterry> good idea, trying
[15:05] <smspillaz> mterry: just be weary of swt in general
[15:05] <smspillaz> mterry: they make all these stupid assumptions
[15:05] <smspillaz> for example, they'll check the window manager string for "kwin" and then assume that you're double reparenting only in that case
[15:05] <mterry> smspillaz, http://pastebin.ubuntu.com/563902/ : wnck sees 0x06a0004a and gtk sees 0x06a0007a  (they are both in the wininfo tree)
[15:05] <seb128> tedg, the unity-panel-service crashes this way, http://paste.ubuntu.com/563903/
[15:05] <seb128> tedg, known issue?
[15:06] <seb128> tedg, ignore that, libdbusmenu is not uptodate and it lacks debug symbols
[15:06] <mterry> smspillaz, oh sorry, chop leading 0 for both of those
[15:06] <smspillaz> yeah, saw
[15:07] <smspillaz> mterry: is 0x6a0004a a menu or a toplevel ?
[15:07] <mterry> smspillaz, toplevel
[15:07] <smspillaz> 140x39 seems like a weird size for a toplevel
[15:07] <smspillaz> ok
[15:07] <mterry> smspillaz, it's a test window
[15:07] <smspillaz> ok
[15:07] <smspillaz> so 0x6a0004a is in fact the managed window (reparented)
[15:08] <smspillaz> mterry: is your menu 200x200 ?
[15:08] <mterry> smspillaz, not likely
[15:08] <smspillaz> mterry: can you post xwininfo -id 0x6a0007a -all ?
[15:09] <mterry> smspillaz, http://pastebin.ubuntu.com/563904/
[15:11] <smspillaz> right, so it is one of those stupid unmapped "lets store properties on these" windows
[15:11] <smspillaz> also override-redirect
[15:11] <smspillaz> so it might be the menu
[15:12] <smspillaz> or just a property storage window
[15:12] <didrocks> pitti: https://launchpad.net/~utouch-packaging has a contact now
[15:12] <mterry> smspillaz, it is definitely a toplevel GtkWindow
[15:13] <mterry> smspillaz, that is holding the menu anyway (may be some fake container window just for the menu...)
[15:13] <smspillaz> that sounds messy ....
[15:13] <smspillaz> thought his is sun-awt we're talking about
[15:13] <smspillaz> *this is
[15:14] <mterry> smspillaz, and I agree it seems to be unmapped until I realize it
[15:14] <mterry> smspillaz, it's eclipse swt, whatever that is.  openjdk
[15:15] <mterry> smspillaz, ok, so gtk is seeing the wrong window.  there is probably something I can do here...  I'll dig some more, thanks smspillaz
[15:16] <smspillaz> np
[15:26] <seb128> tedg, do you need details on this bug then?
[15:28] <tedg> seb128, Which one?  Did it get retraced?
[15:29] <seb128> tedg, not yet the retracing, I was just wondering if you managed to get it
[15:30] <tedg> seb128, No, but I can't run Unity anymore with the xserver transition :-/
[15:30] <tedg> seb128, So that's probably why I can't recreate it.
[15:30] <tedg> seb128, Hoping the stack trace makes it clear.
[15:30] <seb128> tedg, k
[15:34] <mterry> smspillaz, think I got it.  both windows were being passed to appmenu-gtk's gtk module, but we were ignoring the Real one due to a separate bug
[15:36] <seb128> tedg, ok, I think it's a sideeffect of bug #714531
[15:36] <ubot2> Launchpad bug 714531 in indicator-appmenu "unity-panel-service consuming very high cpu time" [Undecided,New] https://launchpad.net/bugs/714531
[15:37] <seb128> tedg, http://paste.ubuntu.com/563916/
[15:37] <seb128> mterry, ^
[15:38] <seb128> tedg, it starts happening to lot of applications once appmenu get stucked eating cpu
[15:38] <seb128> which is the bug I just pointed before
[15:38] <seb128> the pasteb has a stacktrace of a nautilus crash
[15:38] <tedg> seb128, Yeah, I think there's some race there.  Cool, I'll look at the trace.
[15:39] <mterry> seb128, tedg: seems like a clear bug that appmenu-gtk tries to unref the variant even if it's null there.  The unref should go after the if (error != NULL) block in bridge.c
[15:40] <tedg> mterry, Yup, I'll put a branch up for review.
[15:47] <tedg> Not sure about the CPU time one.  I got that once, and all it was doing was looking up in a hashtable... I wonder if the other bug was causing a DoS on it.
[15:47] <seb128> tedg, the cpu one seems to come from the fix for bug #706941
[15:47] <ubot2> Launchpad bug 706941 in indicator-appmenu "unity-panel-service crashed with SIGSEGV in g_type_check_instance_cast()" [High,Fix released] https://launchpad.net/bugs/706941
[15:48] <seb128> tedg, see the comment from jdstrand there
[15:53] <chrisccoulson> tedg - the "looking up in a hash table" is the bug i mentioned last week where it spins in that while loop
[15:54] <chrisccoulson> the hash table lookup happens inside that loop
[15:54] <tedg> chrisccoulson, Ah, okay.
[15:54] <tedg> Let me try the Evo thing.
[15:54] <chrisccoulson> tedg - i was going to take a look at it once i've got the latest ffox uploaded
[15:55] <chrisccoulson> i ended up getting distracted and fixing a bug in the thunderbird menus though ;)
[15:55] <chrisccoulson> just doing ffox now
[15:59] <tedg> Uhg, of course nothing happens under GDB :-(
[16:01] <seb128> it doesn't happen everytime for sure, seems a race or something
[16:01] <chrisccoulson> it happens quite frequently here (every 10 minutes or so with my normal workflow)
[16:03] <kenvandine> seb128, how long should it take for geoclue to show up in main?
[16:04] <seb128> kenvandine, one publisher run, i.e one hour
[16:04] <seb128> kenvandine, https://launchpad.net/ubuntu/+source/geoclue/+publishinghistory
[16:04] <kenvandine> 2 hours now... not there
[16:04] <kenvandine> oh, good tip
[16:04] <seb128> kenvandine, it has been publisher 59 minutes ago
[16:04] <kenvandine> humm
[16:04] <kenvandine> the LP page still says univ
[16:04]  * kenvandine retries build
[16:05] <kenvandine> oh yay... built :)
[16:06] <kenvandine> now for the eds integration :)
[16:06] <kenvandine> tedg, did you review that yet?
[16:12] <tedg> kenvandine, Yeah, and I've got a patch that I sent to klattimer, just pinged him about that.
[16:12] <tedg> kenvandine, Works here locally for me though :)
[16:12] <kenvandine> great
[16:13]  * tedg is reminded he's going to a hockey game this weekend ;)
[16:18] <pitti> chrisccoulson, dpm: there's somethign seriously broken with the 10.04 firefox exports apparently
[16:18] <chrisccoulson> :/
[16:18] <chrisccoulson> doesn't sound good :(
[16:18] <pitti> they keep being just 230 bytes long, and this time I hardcoded teh usage of the old po2xpi checkout
[16:18] <pitti> dpm: do you know if the export structure on LP changed at all?
[16:19] <pitti> it now doesn't seem to work with either
[16:19] <pitti> (the old or new po2xpi)
[16:21] <dpm> pitti, bummer. The only changes I know were: new upstream imports from bzr feature/message sharing with upstreams (that came on the last LP rollout last month); translation exports were broken a couple of weeks ago
[16:21] <pitti> seems I need to dissect the exported tarball tomorrow, and make this easier to test
[16:21] <pitti> turnaround times of 2 hours suck for that
[16:22] <pitti> dpm: I'll try to make some time for that tomorrow morning
[16:22] <pitti> dpm: anyway, if they miss 10.04.2, do you see any harm done?
[16:22] <pitti> dpm: the current ones should work well enough, and we can update them after 10.04.2 again?
[16:22] <tedg> chrisccoulson, I actually think it's something with mterry's patch to indicator-appmenu.  It seems to be when the window is destroyed.  I'll look some more.
[16:23] <dpm> pitti, I think translators would only be unhappy if we do not update them at all, so I think if we miss the 10.04.2 window but we update them afterwards, that'd be a good compromise
[16:23] <seb128> grrrrrrrrr at update-manager
[16:23] <seb128> stupid thing
[16:24] <pitti> dpm: in the ideal case I could fix it tomorrow morning and get them into -proposed by tomorrow afternoon, then we could at least still verify a handful and get them into -updates
[16:24] <pitti> dpm: 10.04.2 final images will be built on Friday
[16:24] <pitti> seb128: ?
[16:25] <seb128> pitti, I keep getting that gnome-policykit-daemon crash
[16:25] <dpm> pitti, ok. Thanks a lot for working on this
[16:25] <seb128> which makes update-manager hang because it doesn't handle the daemon crash
[16:25] <seb128> which trigger apport which somewhat screw my compiz with this invisible dialogs
[16:28] <seb128> which forces me to restart unity which leads to that indicator-appmenu crash
[16:28] <seb128> pitti, in summary just messed my running session ;-)
[16:29] <chrisccoulson> i've just gone back to metacity + gnome-panel for now ;)
[16:29] <seb128> grrrr, come on
[16:30] <pitti> same here; this morning all my terminals crashed after just seconds, repeatedly
[16:31] <chrisccoulson> the invisible window bug was driving me crazy, i can't get any work done with it
[16:31] <seb128> kill unity-panel-service if that happens
[16:34] <mvo_> seb128: didn't I submit a patch so that the session restars policykit-daemon?
[16:35] <seb128> mvo_, it does, but update-manager is frozen for 5 minutes then seems to timeout, display an error dialog about aptdaemon and resume working
[16:35] <seb128> mvo_, seems also gnome-policykit became crash land since the session patch, there is something wrong with it, we try to debug a bit at the rally with pitti but it collects duplicates
[16:40] <mvo_> seb128: ok (sorry, on the phone)
[16:40] <seb128> mvo_, (no hurry and nothing to be sorry about either ;-)
[17:17] <mpt> mvo_, actually, it wasn't the staging server, it was me removing rr.layout()
[17:21] <mpt> ... or maybe it wasn't me after all
[17:21]  * mpt shakes fist
[17:27]  * mpt -> home
[17:36] <seb128> kenvandine, https://bugs.launchpad.net/bugs/714531
[17:36] <ubot2> Launchpad bug 714531 in indicator-appmenu "unity-panel-service consuming very high cpu time" [Undecided,New]
[17:37] <seb128> kenvandine, seems another candidate for patch backporting ;-)
[17:37] <kenvandine> yaya
[17:37] <kenvandine> will do
[17:38] <seb128> thanks
[17:38] <seb128> mterry, I'm assigning you bug #708041
[17:38] <ubot2> Launchpad bug 708041 in indicator-appmenu "unity-panel-service crashed with signal 5 in g_closure_invoke()" [Low,Confirmed] https://launchpad.net/bugs/708041
[17:39] <seb128> it's a bit weird, the crash is because it can't get a dbus connection
[17:39] <seb128> well in any case it's collecting some duplicates so would be nice to make it be quiet
[17:41] <seb128> rodrigo_, I've assigned bug #641120 to you, would be nice to if you could try to figure what's going on there, it's a frequent crasher
[17:41] <ubot2> Launchpad bug 641120 in gnome-settings-daemon "gnome-settings-daemon crashed with SIGSEGV" [Medium,New] https://launchpad.net/bugs/641120
[17:42] <rodrigo_> seb128, ok
[17:42] <seb128> rodrigo_, thanks
[17:47] <pitti> good night everyone!
[17:47] <highvoltage> good night pitti
[17:48] <rodrigo_> bye pitti
[17:48] <didrocks> have a good night pitti
[17:54] <mterry> seb128, k
[18:07]  * rodrigo_ goes out for a bit also
[18:07] <rodrigo_> later all
[18:24] <Amaranth> hrm, everyone on my system is crashing in glib
[18:24] <Amaranth> appears to be an assertion in g_atomic_int_exchange_and_add
[18:31] <bcurtiswx> kenvandine, geoclue in ubuntu \o/ , any plans to enable geolocation in empathy now?
[18:32] <kenvandine> not yet
[18:33] <bcurtiswx> kenvandine, OK
[18:33] <kenvandine> i can
[18:33] <kenvandine> i can't even test that
[18:33] <kenvandine> since gtalk doesn't let you do that
[18:34] <bcurtiswx> kenvandine, isn't it an empathy thing where if who people are using geoclue in empathy you can see where they are?
[18:34] <bcurtiswx> s/who/two
[18:35] <kenvandine> no
[18:35] <kenvandine> the jabber server has to support an extension to handle the location data
[18:35] <kenvandine> google doesn't
[18:35] <bcurtiswx> kenvandine, what does? i have an account in a lot of protocols
[18:36] <bcurtiswx> GSOC project ;)
[18:37] <kenvandine> not sure, it is a PEP extension
[18:37] <kenvandine> jabber.org supports it
[18:40] <didrocks> good night everyone!
[18:42] <bcurtiswx> yeah it's only jabber
[18:42] <bcurtiswx> with the lone exception of gtalk.. lol
[18:42] <bcurtiswx> fyi http://coccinella.im/servers/servers_by_pubsub_pep.html
[19:05] <Amaranth> wow I don't even remember filing bug 432536, probably because it was 1.5 years ago and was a simple rename :/
[19:05] <ubot2> Launchpad bug 432536 in libindicate-qt "incorrect binary package name" [High,Fix released] https://launchpad.net/bugs/432536