[08:27] morning everyone [08:33] hello crevette [08:33] salut huats === BugMaN1 is now known as bugman1 === bugman1 is now known as BugMaN [09:30] seb128: vino done [09:31] pochu: right, I've nothing it got uploaded, probably thanks to dholbach who did some sponsoring ;-) [09:31] yeah :) [09:39] mvo: [09:39] $ gnome-keyboard-properties [09:39] (gnome-keyboard-properties:14180): GLib-GObject-WARNING **: invalid (NULL) pointer instance [09:39] (gnome-keyboard-properties:14180): GLib-GObject-CRITICAL **: g_signal_connect_data: assertion `G_TYPE_CHECK_INSTANCE (instance)' failed [09:39] mvo: those warnings were not there before you update [09:40] mvo: I get them when starting gnome-keyboard-properties [09:40] seb128: hrm, I don't get those [09:41] seb128: could you please give me your gconf keyboard settings? I will try to reproduce [09:42] mvo: http://people.ubuntu.com/~seb128/gconf.log [09:43] mvo: the stacktrace [09:43] #5 0xb7ac1838 in IA__g_type_check_instance (type_instance=0x0) at /build/buildd/glib2.0-2.18.0/gobject/gtype.c:3814 [09:43] #6 0xb7abc559 in IA__g_signal_connect_data (instance=0x0, detailed_signal=0x805a9e1 "clicked", [09:43] c_handler=0x8051bd0 , data=0x88d1a00, destroy_data=0, connect_flags=0) [09:43] at /build/buildd/glib2.0-2.18.0/gobject/gsignal.c:2235 [09:43] #7 0x08051a1f in setup_xkb_tabs (dialog=0x88d1a00, changeset=0x0) at gnome-keyboard-properties-xkb.c:319 [09:50] mvo: frame 7 is [09:50] 319 g_signal_connect (G_OBJECT (WID ("button_make_default")), [09:51] seb128: thanks, I have it here now too [09:52] mvo: ok, my fault, I've fixed it :-p [09:52] oh? [09:52] mvo: I did apt-get install gnome-control-center to try the new version but that didn't upgrade capplets-data which has the required glade changes [09:53] right, that sounds like a bad dependency [09:53] * mvo fixes in bzr [09:54] seb128: hm, currently the dependency on capplet-data is gnome:version [09:55] seb128: is there a risk in making it binary:version ? [09:55] mvo: I guess it's really a corner case, don't bother [09:55] * mvo nods [09:55] seb128: can you make it crask after the update? [09:55] mvo: the issue is that strict versionning between arch all and any breaks installability on !i386 [09:55] right [09:56] mvo: no, but I think it's due to tjaalton changes, maybe partial upgrade there too, ie the new xbd-data not installed or something [10:24] seb128: I keep an eye on the issue, I just found and fixed one crash, but it does not really match the signature of the other bugreport. I will keep looking [10:24] mvo: ok [10:24] mvo: thanks [10:25] seb128: are you in charge of the rhythmbox package? [10:26] andreasn: yes [10:26] seb128: I just realized why the icon looks like crap in the Applications menu [10:27] it lacks a 24x24 size in icons/hicolor [10:29] ah, same thing upstream it seems [10:30] if I sent you a 24x24 icon, could you fix it? I want to submit it upstream as well, but maybe it's too late [10:30] seb128: bug 268470, I guess it needs a feature freeze exception [10:30] Launchpad bug 268470 in gtk-vnc "New upstream 0.3.7 release" [Wishlist,New] https://launchpad.net/bugs/268470 [10:30] soren: ^ [10:31] pochu: no, it's a GNOME requirement so it should be alright [10:32] great :) [10:32] andreasn: that should be fixed in the svn snapshot I uploaded to intrepid yesterday? [10:32] oh, cool [10:33] I tested the latest Alpha of Ubuntu, but if it's fixed already I can sleep calmly at nights again :( [10:33] :) I mean [10:59] brb [11:18] hmm, vinagre + scaling doesn't work here [11:19] (vinagre:31774): GdkGLExt-WARNING **: Window system doesn't support OpenGL. [11:19] err [11:19] glxgears doesn't either! [11:20] Xlib: extension "GLX" missing on display ":0.0". [11:20] heh, glxinfo segfaults [11:20] tjaalton: ^ this is an intel GMA 965, any idea? [11:21] pochu: do you have some nvidia-*177 installed? [11:21] pochu: sudo apt-get remove nvidia-glx-177 [11:22] seb128: yeah, that's installed. removing it, thanks [11:22] pochu: right, broken nvidia-kernel-common upload which pulled nvidia-glx.. [11:22] so no need to report it, I guess :) [11:23] brb [11:23] thanks folks [11:24] pochu: you're welcome [11:25] yay, scaling rocks! [11:25] but it doesn't work with compositing... [11:25] "Scaling does not work properly on composited windows. Disable the visual effects and try again." [11:31] seb128: bug 268484 [11:31] Launchpad bug 268484 in vinagre "New upstream 2.23.92 release" [Wishlist,Triaged] https://launchpad.net/bugs/268484 [11:31] pochu: cool thanks [11:33] seb128: no problem! [12:03] mvo, there is currently some discussion about a software online repo at #PackageKit [12:03] mvo, perhaps we could finally target something for 9.04 [12:03] mvo, hughsie will try to get some more information from the redhat guys about the amber project [12:21] <_8472> Hi, is there any chance to enable in Gnome this kind of METACITY setting "automatically give focus to newly created windows" ? I'm already searching for it for some time, but still no success. Always have found similar answer, it's not in Gnome ....... , but i don't like this answer. thx in advance [13:00] seb128: poppler 0.8.7-1 is in incoming or so [13:00] lool: right I noticed, will sync it thanks === reynaldo_ is now known as reynaldo [14:54] mpt, hello. [14:55] mpt, you did some work on specifing the maintenace level of a package. Could you please point me to the wiki page? [14:55] mpt, I would like to get this into packagekit too, but cannot find the page anymore [14:56] glatzor, sure, https://wiki.ubuntu.com/PackageMaintainednessPresentation [14:57] mpt, great! [14:57] mpt, have you found any time recently to work on your package management idea? [14:58] glatzor, it's not on my to-do list at the moment, but I'm quite hopeful it will be for the Jackalope [14:58] mpt, RedHat and Novell are both working on an online software catalog. [14:59] muahahaha [14:59] mpt, I would like to get this finally for Debian and Ubuntu [15:00] mpt, AFAIK they plan to make use of the packagekit browser plugin to allow to easily install, remove or run the software [15:00] (RedHat) [15:03] Hey can you people still mount USB keys? [15:03] I just tried a regular one with vfat and it wouldn't mount in nautilus [15:04] mpt, http://benjiweber.co.uk/blog/?p=10 [15:05] mpt, there are screenshots from the novell thing. but they use a java application :( [15:05] mpt, http://blogs.gnome.org/hughsie/2008/09/09/packagekit-web-plugin/ [15:06] mpt, here is a screenshot of the packagekit plugin. [15:07] a Java application powering a Web site, no less [15:08] oh, the packagekit one is a Web site too [15:13] mpt, the screenshot is just a demo of the plugin. the project of Redhat is called amber. [15:13] mpt, perhaps I can setup a local server soon [15:15] glatzor, plug-in for what? [15:16] Hey all. What sets XDG_SESSION_COOKIE? [15:17] some users are seeing problems when it is not available over an "ssh -X" connection [15:18] mpt, the packagekit web plugin is the small white box that you can see on the screenshots [15:18] mpt, it allows to interact with the local package manager. [15:19] I don't understand why it's a Web site in the first place [15:19] mpt, it represents the installation status and allow to install, run or remove an application [15:19] mpt, because you make use of non-static and large data [15:19] so? [15:19] mpt, only think of screenshots [15:19] Web browsers aren't the only programs that can do HTTP. [15:20] mpt, sure there is still wget :) [15:20] Putting it in a Web browser just makes it half as efficient and offers 100 more wrong things to click. [15:21] mpt, but how do you want to communicate the requirement to be "online" in a better way? [15:23] With a message, inside the package manager, asking you to either connect to the Internet or insert an appropriate disc [15:24] mpt, sorry, I have to leave. My break is over. and the patients are waiting for their medicine. see you! [15:24] ok :-) [15:24] (glatzor's a doctor?) [15:42] lool: Worked for me a couple weeks ago... [15:42] Yeah for me too [15:43] james_w: It's a cookie to keep track of individual sessions. It's used by stuff like policykit and consolekit. I would imagine that it wouldn't work over an ssh connection anyway as DBus is not being routed across that session. [15:44] tedg: yeah, this is consolekit [15:44] james_w: Probably an app bug that it doesn't work gracefully without having it. [15:45] bug 231246 [15:45] Launchpad bug 231246 in policykit-gnome "'Unlock' button in admin utilities greyed out" [Undecided,Invalid] https://launchpad.net/bugs/231246 [15:59] tedg, seb128 call [15:59] lool: There [16:00] MacSlow: too [16:00] lool: yeah, I was just going to call ;-) [16:00] lool: I didn't know you were joining, good ;-) [16:00] * MacSlow is in [16:01] james_w: I guess perhaps the admin tools shouldn't require the local console? Seems, at least in the LTSP case, that shouldn't be required. [16:01] tedg: yeah, I'm not sure if this is the actual problem, or just a red-herring [16:04] lool, tedg, seb128: also maybe join #advisoryboard on irc.gimp.org [16:41] seb128: gvfs-trashd is seriously FUBAR [16:41] gicmo: indeed [16:41] gicmo: we have over 60 crash duplicates now in launchpad [16:42] seb128: ok, I really need to look at this [16:43] seb128: can you point me at one? [16:43] I mean bug report in lp [16:43] gicmo: https://bugs.edge.launchpad.net/ubuntu/+source/gvfs/+bug/252174 [16:43] Launchpad bug 252174 in gvfs "gvfsd-trash crashed with SIGSEGV in g_main_context_dispatch()" [High,Triaged] [16:43] wtf, in g_main_context_dispatch?! [16:44] gicmo: the top of the stacktrace is missing and I didn't manage to know why, we already talked about that [16:44] gicmo: see http://bugzilla.gnome.org/show_bug.cgi?id=547568 which could be the same issue [16:44] Gnome bug 547568 in trash backend "gvfsd-trash crashed with SIGSEGV in g_main_context_dispatch()" [Critical,Unconfirmed] [16:45] seb128: Ill see what I can do [16:45] gicmo: http://bugzilla.gnome.org/show_bug.cgi?id=547726 too [16:45] Gnome bug 547726 in hal volume monitor "gvfsd-trash crashed with SIGSEGV in dbus_error_is_set()" [Critical,Unconfirmed] [16:45] gicmo: valgrind has a "==28416== Jump to the invalid address stated on the next line" error [16:47] wtf is trash doing that others dont [16:48] I mean its only trash, crashing all the fucking time [16:48] and hadess branched gvfs without asking me [16:48] to commit his stupid cd text xattr patch [16:48] super duper [16:50] seb128: you remember when did this start? [16:50] gicmo: the version where the monitors have been splitted [16:52] oh, who'd guessed that now ;-) [17:05] seb128: yo [17:06] seb128: soo, walters told me, and he is right, it looks like its doing dbus stuff from different threads [17:07] gicmo: oh [17:07] in do_mount [17:07] gicmo: read http://bugzilla.gnome.org/show_bug.cgi?id=551337 on that topic [17:07] Gnome bug 551337 in general "gedit crashes browsing fileselector ssh locations" [Critical,Unconfirmed] [17:13] seb128: so, we call enumerate_root_trashdir () in the trash backend on a second thread [17:13] which causes g_daemon_vfs_init () to be called [17:13] http://launchpadlibrarian.net/16686194/stacktrace-gvfsd-trash.log [17:13] see #11 0x00007f398b088f50 in g_daemon_vfs_init (vfs=0x227cf20) [17:13] which then loads the proxy monitors [17:14] and dbus stuff etc [17:14] seb128: gicmo: does this crash happen with GIO_USE_VFS=local ? [17:14] seb128: maybe we can start the gvfsd-thrash daemon with GIO_USE_VFS=local and see if it goes away [17:15] can we reproduce that reliable? [17:15] gicmo: ugly workaround? [17:15] seb128: just trying to find out if that is the cause at all [17:15] gicmo: no, it happens randomly once a day or something [17:15] teh suck [17:15] its prolly a race [17:15] gicmo: but your "don't use dbus in different threads" seems a good explanation [17:16] gicmo: did you read the bug I just pointed? [17:16] gicmo: in fact that's an another issue [17:17] seb128: its another one, but quite interesting as well [17:21] trashdir = 0x227f930 "/home/username/.local/share/Trash" [17:21] topdir = 0x2283620 "/home/username/.local/share" [17:21] wtf [17:21] who calls himself username? [17:22] i wonder if launchpad automatically replaces /home// with username [17:22] ;-) [17:22] walters: no [17:23] walters: that be nice though [17:23] right [17:23] yeah, would make sense for privacy [17:23] walters: so, I guess its crashing at vfs->async_bus = dbus_bus_get_private (DBUS_BUS_SESSION, NULL); [17:24] which we call on thread 2 [17:24] I am pretty sure that we already called dbus_bus_get on thread 1 before because we already have talked to the daemon over dbus [17:24] walters: any idea if that is illegal? [17:25] what i would say is that all code should have a good idea of what thread it is going to run on [17:25] if code was intended to be run from the main thread it's going to be a problem [17:26] just calling dbus_bus_get_private from a thread should be OK; it will hand you back a new connection and that should be safe [17:27] hmm [17:28] then maybe the dbus code that gets pulled in froom the volume proxy monitors the the clurprt [17:28] curlprit [17:36] gicmo: any informations which would be useful to provide? [17:42] 18:22 < walters> i wonder if launchpad automatically replaces /home// with username [17:42] that's done by apport, bug 192786 [17:42] Launchpad bug 192786 in apport "Anonymize user and host name" [High,Fix released] https://launchpad.net/bugs/192786 [17:44] makes sense [17:44] the bug report is funny, btw :) [17:45] yeah, totem is definitely a source of amusing bug reports [17:46] one thought i had on this is to simply make bug reports non-public by default; i.e. you have to have an account on the system to access them [17:46] that at least somewhat mitigates the slashdot effect [17:46] s/bug report/stack traces/ i mean [17:48] crashes are private until someone with access to them marks them as public [17:48] ah, ok [17:48] is it just my keyboard or keyboard shortcuts are case-sensitive? [17:58] seb128:well trying to run the daemon with GIO_USE_VOLUME_MONITOR=unix and see if that helps could be of use [17:58] but oh well [18:01] gicmo: is that going to provide useful informations? [18:01] gicmo: I can do that, but if that turns to workaround the bug what do we do next? ;-) [18:01] walters: [18:02] hmm might it be that its _g_dbus_connection_integrate_with_main () being called from the thread oder something [18:02] I am just blindly guessing here [18:03] gicmo: that sounds very likely to break [18:04] also, using mainloop integration causes stuff to be dispached from the mainloop (thread 1), right? [18:10] the proxy monitors also do a dbus_connection_send_with_reply_and_block () [18:10] which will happen on thread 2 [18:10] we call IsSupported () on org.gtk.Private.RemoteVolumeMonitor [18:13] gicmo: maybe take that discussion on #gnome-hackers? [18:13] gicmo: higher change to have people jumping in there ;-) [18:14] well, I am pretty sure doing dbus_connection_send_with_reply_and_block () on a second thread is kinda bad [18:16] easiest fix would be to force gvfs/gio being initilized from the main thread [18:31] gicmo: could be worth to discussing in with davidz anyway, since he did some of the changes he might be wanting to fix it too ;-) [19:04] nice job on that pidgin/empathy usability report [20:11] hello [20:14] I have an error with packageKit ppa, should I report it upstream or on lp ? [20:17] hi crevette! [20:17] crevette: if it's with the ppa, report it to glatzor [20:18] hey pochu [20:19] crevette: hey, if you jump in #packagekit then I can have a look for you [20:19] let's jump === crd1b is now known as crdlb [21:47] james_w: seb128 had this idea about bzr-buildpackage --upload that would automatically add a tag to the upload and do a dput, what do you think? is that something worth adding a whishlist bug for bzr-buildpackage? [21:47] * mvo likes the idea [21:48] mvo: yeah, it's something that I would like to do [21:48] is lack of time the problem? [21:48] however, I haven't thought through how to handle dput/dupload, multiple hosts, extra options, etc. [21:48] not so much [21:49] it's not high priority for this cycle, but I will have time to do it for the next one [21:49] it's more a lack of experience, as I hardly ever upload, and when I do it's very basic [21:50] then seb128 is the right person, he uploads a lot :) [21:50] ah ah [21:50] *A LOT* [21:50] :-) [21:50] mvo: I upload often but don't use bzr when you don't make me do it ;-) [21:50] you will like it someday :P [21:51] the thing just hate me [21:51] every time an upload should take 10 seconds it takes 10 minutes ;-) [21:53] mvo: joke aside this one was alright but I think I'll not like the extra steps to edit patches [21:53] which extra steps? [21:53] james_w: there is only the debian directory in bzr, how do I use cdbs-edit-patches? [21:54] yeah, that is something that is currently not ideal [21:54] I would love to have bzr-edit-patch or somesuch that exports the tarball, runs cdbs-edit-aptch (or dpatch-edit-patch or quilt) in there and on exit adds the patch to bzr [21:55] and auto detectin of --merge and --native would rock! [21:56] (well, bzr-buildpackage rocks already, but that would make it rock even more) [21:56] "bzr bd-do 'dpatch-edit-patch 01-foo'" [21:57] "bzr bd-do" to get a shell to do whatever you want [21:57] you need to "bzr add" afterwards, I haven't decided if that should be automatic yet [21:58] the name is completely non-obvious as well [21:58] oh, nice [21:58] * mvo hugs james_w [21:58] it's just cribbed from svn-buildpackage [21:59] I'm not a fan of debian-only though, the branches I create will be full source [21:59] bzr: ERROR: unknown command "bd-do" [21:59] * mvo needs to go to bed [21:59] seb128, I'm testing gtkmm now [22:00] * seb128 installs bzr-buildpackage [22:00] NCommander: 2.13.8? ;-) [22:00] ah :-) [22:00] seb128, ARGH [22:00] * NCommander swears [22:00] NCommander: new pangomm too [22:00] * NCommander twitches [22:00] You just like inflicting pain on me [22:01] james_w: cool, that works, thanks for the hint ;-) [22:01] seb128, rolling them now [22:01] seb128, please commit them for me onto pkg-gnome once I roll the changeset [22:02] NCommander: oh, you like to complain apparently, http://download.gnome.org/sources/libgweather/2.23/libgweather-2.23.92.tar.gz for you [22:02] hey [22:02] seb128, :-P! [22:02] NCommander: they added python binding and it's using cdbs *g* [22:02] I'm rolling gtkmm now [22:02] .... [22:02] You'll get gtkmm and pangomm from me for now :-P [22:02] ;-) [22:02] lut crevette [22:02] seb128: no problem, glad it didn't hate you :-) [22:03] hey seb128 [22:03] seb128: is it too late to upgrade nemiver to latest version, we're still in 0.5.2 [22:03] NCommander: those updates are no good for debian yet, they don't have cairo 1.7 and gtk 2.14 [22:04] Ok [22:04] I thought gtk at the very least was in experimental [22:04] crevette: version are frozen [22:04] NCommander: no [22:04] seb128, building gtkmm now [22:04] I seen bluez-gnome is old too [22:04] crevette: you need to ask for a freeze exception, should be easy to get [22:05] okay [22:05] don't have much time now, we'll see this week end [22:05] seb128, will you commit pangomm for me? [22:05] NCommander: where? [22:05] seb128, 127.0.0.1, it will go to my PPA if it builds [22:05] NCommander: debian can't update, the new version requires a new pango which requires a new cairo [22:05] seb128, Oh! [22:06] * NCommander hasn't even looked at the build-deps for pangomm yet [22:06] seb128: don't you handle freeze exceptions for GNOME packages in universe? ;) [22:06] pochu: that's not a GNOME packages, that's a random GNOMish application, can't handle every of those [22:07] oh, ok [22:07] pochu: you are welcome to ask for the freeze exception though ;-) [22:07] seb128, gtkmm still needs to be updated in Debian once all its build-deps are in place [22:07] you could just ack it ;) [22:08] pochu: I'm not in the corresponding motu team to do that [22:08] pochu: and GNOME packages don't need freeze exception btw since GNOME has a standing exception ;-) [22:11] seb128, rolled pangomm, testing now [22:13] seb128: I'm referring to this mail, in which the MOTU council delegates the power of ack'ing feature freeze requests for GNOME packages to you: https://lists.ubuntu.com/archives/ubuntu-motu/2008-September/004648.html [22:13] seb128, /usr/share/gnome-pkg-tools/1/rules/check-dist.mk:19: Unknown distribution: intrepid [22:13] Is that a gnome bug? [22:13] er, known [22:13] pochu: ah, they did that again this cycle ;-) [22:13] :) [22:13] NCommander: that's a debian thing to not upload experimental packages to unstable, you can ignore it on ubuntu [22:14] seb128, I was just not sure if I just never noticed it before, or if it was new since my reinstall [22:14] seb128, pangomm builds and installs fine, on its way to my PPA [22:14] pochu: ok, if you try a debian update and it builds fine I can do syncs for you [22:14] NCommander: good [22:14] NCommander: you didn't notice before [22:16] seb128, pangomm installs in a chroot, but since the updated gtkmm isn't available just yet, it won't install on a production system since gtkmm has rdepends [22:16] (attempting to deconfigure gtkmm, unable because gnome-*) [22:16] right [22:16] I'm just waiting for gtkmm's new version to finish building [22:16] I doubt the pangomm will break things and there is no such much using gtkmm anyway [22:25] seb128, gtkmm builds fine [22:27] NCommander: good, so it's sponsoring time ;-) [22:27] seb128, they're both in my PPA [22:28] Launchpad is crawling for me [22:29] seb128, https://edge.launchpad.net/~sonicmctails/+archive [22:30] NCommander: btw gtkmm also has a new version [22:31] seb128, I uploaded the new version [22:31] seb128, it just hasn't popped up yet [22:31] (I rolled both pangomm and gtkmm) [22:31] excellent ;-) [22:31] See, I do amazing work sometimes [22:31] seb128, pangomm built on all PPA archs, gtkmm hasn't shown up yet -_-; [22:32] uh oh [22:32] Rejected: dm:manphiz@gmail.com may not upload/NMU source package pangomm [22:32] WTF O_o; [22:33] NCommander: that's a debian error right? [22:33] yeah [22:33] I've never gotten that one before [22:33] NCommander: well, that's probably the first time a dm try to upload one of the packages you maintain [22:33] whoops [22:34] I don't think I set Uploaders: >.>; [22:34] Probably why it was rejected [22:34] could be yes [22:35] * NCommander needs people to do backport uploading for me [22:36] NCommander: trying to ask on #ubuntu-motu? [22:36] Only core-dev can upload for backports [22:36] And really only backports should [22:37] NCommander: I think technically you should have bumped the pangomm shlibs [22:38] I'll upload anyway, nothing is using it yet [22:38] seb128, which shlibs, the soname didn't change [22:38] NCommander: soname and shlibs are different things [22:38] NCommander: shlibs should be updated every time something is added to the api [22:39] when the soname change the package name change [22:39] the shlibs is what is used to build depends [22:39] hrm [22:39] wait [22:39] if the current shlib version is 2.13.7 and a package build using pangomm [22:39] did you already upload? [22:39] it'll get a depends on (>= 2.13.7) [22:40] Oh [22:40] not if 2.13.8 has an new symbol [22:40] I thought you meant the .shlibs file [22:40] the application might not run on 2.13.7 and require this symbol [22:40] Sorry, I thought I lost my mine [22:40] *mind [22:40] NCommander: no, I mean shver in debian/rules [22:40] d'oh [22:40] NCommander: and no I didn't upload [22:40] I completely forgot about that [22:40] do you want to fix it? [22:40] On my non-CDBS packages, I have a rule that autodetermines that [22:40] I can't reupload to my PPA, it will reject due to the same version [22:41] I'll just edit the debian/rules for you [22:41] THank you [22:41] ok? [22:41] Yeah [22:41] cool [22:41] Sorry, that completely slipped my mind [22:41] Don't do gtkmm then, it probably needs that too [22:41] that's alright ;-) [22:41] I'll do the same change if required [22:41] SHVER := 1:2.13.5 - gtkmm really needs it [22:41] I completely forgot that existed [22:42] * NCommander holds head in shame [22:43]