=== crd1b is now known as crdlb [00:06] tap tap tap is thing on? === hggdh is now known as hggdh|away === crd1b is now known as crdlb === NCommander is now known as NC|Elmor_Fudd === NC|Elmor_Fudd is now known as NCommander === crd1b is now known as crdlb === crd1b is now known as crdlb [09:07] morning all [09:54] lut huats [09:54] mvo: when you have a moment could you look at the new comments on http://bugzilla.gnome.org/show_bug.cgi?id=545123 and tell me if that makes sense? [09:54] Gnome bug 545123 in general "Please support the SmInteractStyleNone again" [Minor,New] [09:55] hello seb128 and mvo === MacSlow_ is now known as MacSlow [10:04] seb128: yes, will do [10:04] hey huats [11:24] seb128: i think swfdec needs to be newed [11:24] asac: right, will do that in a minute [11:24] cool [11:41] seb128: what gtk will intrepid final ship? [11:41] 2.14? [11:41] asac: 2.14.n [11:42] why? [11:42] seb128: swfdec has a problem which appears to be fixed on gtk-trunk [11:42] http://bugzilla.gnome.org/show_bug.cgi?id=548993 [11:42] Gnome bug 548993 in gdk "regression: gdk 2.13 leaves stray windows in certain cases" [Major,New] [11:43] swfdec author said that it will be fixed in next 2.13.x release. so we probably dont need to do anyhting here and verifying when that gets into intrepid is enough [11:44] right [11:44] the next gtk tarball should come soon [12:21] asac: Oh that's actually a gtk bug; I'm glad it's goind to be fixed, it has annoyed me big time [12:21] lool: right. i think even some of the windows we see with flash get probably fixed by that [13:41] hi [13:46] heya seb [13:47] seb128: http://xoomer.alice.it/bat/tmp/application-x-debian-package.tar.bz2 [13:47] hello lapo [13:47] cool, thanks === asac_ is now known as asac === seb128_ is now known as seb128 [14:21] tedg: btw you mailed me about a change which was required for the new gdm, the way to start it I think some time ago [14:21] tedg: any reason you didn't commit that to the bzr so people actually get the fix? [14:21] seb128: I made my own branch and committed it there so that you could look at it before committing it to the main branch. [14:22] seb128: lp:~ted-gould/+junk/gdm-snap [14:22] tedg: it has been week ago, it didn't turn to work as you expected? [14:22] tedg: would be nice to merge such changes quickly [14:23] The change was in the .debs that I made. [14:23] In general, the new GDM didn't work as expected :) [14:23] tedg: the .deb that you made but that nobody knows about, I point users to the team bzr to build a snapshot [14:25] seb128: Okay, so do you think the change is good? I'd be happy to merge it. [14:30] tedg: looking at the diff, the email change is not required, not sure about all the usplash things you didn't document those [14:30] seb128, Hmm, I only remember changing on line... [14:31] tedg: http://bazaar.launchpad.net/~ted-gould/+junk/gdm-snap/revision/4986 [14:31] seb128: Oh, and the e-mail, yeah, that's not required. [14:32] tedg: what was the rational again about the background thing? I don't find your mail now [14:33] seb128: It wasn't starting in the background for me. It would just block on package upgrade and init. [14:33] what do you mean? [14:33] seb128: I'm not sure about the usplash stuff, is that from a patch that got left applied? [14:33] no idea about that either [14:34] I though you fixed "login doesn't work after boot" [14:34] So when the running "/etc/init.d/gdm start" that would block, forever. [14:34] not package upgrade issues [14:34] oh [14:34] weird, it didn't for the previous gdm [14:34] So I first noticed it on the upgrade, but then I found that it also happened on boot. [14:40] what happened on boot? [14:41] GDM would start, but I think it would block init. So if you killed it the rest of init would run, and then it would restart. [14:41] seb128: http://brainstorm.ubuntu.com/idea/11844/ you can kill this one once done with patching g-i-t [14:42] tedg: ah ok, so just adding the --background seems to be a good idea, I'll do that later [14:42] lapo: ok thanks [14:46] pitti, how much can I bother you with more dbus/ck-related questions? [14:46] MacSlow: on that box where you try gdm: if you boot it normaly and use the standard gdm, does your GNOME session have a CK seat? (ck-list-sessions) and does it complain about errors? [14:46] pitti, I still need to get intrepid on the laptop (iwl3945 still refuses to work on it :/ ) [14:47] MacSlow: as much as needed to get it working for you to unblock your work :) [14:47] MacSlow: shouldn't be different in hardy [14:47] MacSlow: I'm trying to find out what's broken in your dbus setup [14:47] pitti, no normal gdm/gnome-session works on hardy ... e.g. calling echo $XDG_SESSION_COOKIE in my current session in a gnome-terminal yields a proper value [14:48] MacSlow: ck-list-sessions works, too? and doesn't complain about "cannot lookup session yadayada"? [14:48] pitti, yes [14:48] that works too [14:48] MacSlow: so if you stop that session and gdm, and start your custom gdm, does ck-list-sessions still work afterwards and gives you a session for the new gdm? [14:49] pitti, no ... only when I try my custom installed gdm I've all these problems [14:49] MacSlow: you don't restart dbus or consolekit in between or anything? [14:49] no ... but just to be sure I can try exactely that right now once sec [14:56] So, I added some icons to the FUSA applet. And I got it working with uuencode/decode so that they'd go in the diff. But since it's my first time doing something like this, could someone review this patch? http://people.ubuntu.com/~ted/85_5_status_icons.patch [14:58] pitti, so I just exited my gnome-session, as root /etc/init.d/gdm stop and started just the upstream gdm [14:59] tedg: doesn't look correct [14:59] pitti, starting that upstream gdm failed with the error-message ** (gdm-binary:9272): WARNING** : Couldn't connect to system bus: Failed to connect to socket /opt/gdm-new/var/run/dbus/system_bus_socket: Connection refused [15:00] tedg: either you do upstream change and roll a new tarball and doesn't bother using uuencode [15:00] tedg: or you add those as packaging change and install them in the debian/rules [15:00] MacSlow: right, it's looking for the wrong socket; it needs to use the system one in /var/run/dbus/ [15:01] pitti, but there is this file/socket [15:01] pitti, does that matter? [15:01] MacSlow: you mean /opt/gdm-new/ socket exists? [15:01] MacSlow: that won't work [15:01] you can delete that [15:02] seb128: I guess I'm a little confused. I thought it was bad to do a new upstream tarball in general as it isn't "upstream" but it seems like if it's done in debian/rules it would be nearly impossible for upstream to accept the patch if they wanted to in the future. [15:03] That's why I did it all inline as a patch, including the uudecode stuff. [15:06] pitti, ok I'll delete it and try again? BTW, I have the line "/opt/gdm-new/share/dbus-1/system-services" in /etc/dbus-1/system.conf, which I just commented out. That was a hint from Jon earlier this week. [15:06] MacSlow: you shouldn't attempt to run a second dbus under /opt; everything shuold use the normal system dbus [15:06] otherwise you need a second consolekit, hal, and everything as well; that's totally unnecessary [15:07] pitti, I didn't restart dbus at all or tried to start the on in /opt/gdm-new [15:07] MacSlow: but what is weird is that the dbus socket location shuold only be known to libdbus, and you certainly didn't rebuild that in /opt? [15:07] that line is just a left-over from earlier attempts [15:07] MacSlow: shuoldn't be necessary [15:08] pitti, well I did install dbus, hald, ck with upstream gdm initially because I was told newer version of those were needed [15:08] MacSlow: we have the latest dbus and hal in intrepid [15:08] oh, wait, hardy [15:08] pitti, do you hint that I try to rebuild gdm with the hardy-supplied ones? [15:09] I don't have intrepid in a working state yet [15:09] MacSlow: that would be my first shot; if that doesn't work, you could take the intrepid dbus source package and build/install on hardy (that shouldn't cause any problem) [15:09] hardy's hal should be alright [15:10] so what should i try next? intrepid dbus on hardy? [15:13] MacSlow: yes, and revert all the dbus config changes you made [15:13] changes reverted [15:15] MacSlow: just weird that gdm needs such a new dbus... [15:15] MacSlow: maybe you can just try building against the normal hardy dbus [15:15] that'd certainly be easiest [15:17] pitti, hm... gdm's configure didn't complain using hardy's dbus now [15:17] odd [15:17] so much the better [15:18] pitti, ok "ldd gdm-binary | grep dbus" reports the system-wide dbus library used now [15:19] pitti, I'll now repeat the inital test with stopping hardy's gdm, firing up upstream gdm and checking ck-list-sessions [15:19] MacSlow: good luck! [15:19] * MacSlow would sell his soul to the devel if it helped [15:19] devil rather :) [15:24] pitti, that first test yielded a small progress ... this time ck-list-sessions reported the session from the upstream gdm [15:25] yay, so dbus works with the gdm [15:25] pitti, but gdm itself was "stuck" and complained about missing gnome-session in the install-prefix where I put it [15:26] pitti, so dependency hell is moved a little further ... gnome-session [15:27] MacSlow: ugh; you can't configure it for standard prefix and run gdm out of the built tree? [15:27] MacSlow: well, if not, /opt/gdm-new/usr/bin -> /usr/bin symlink should do (or similar) :) [15:28] MacSlow: back in some 45 minutes [15:28] pitti, ok [16:10] pitti, just a quick update ... the issue with $XDG_SESSION_COOKIE is now solved ... thanks a lot for the help! [16:13] MacSlow: what was it? [16:15] seb128, wrong dbus [16:16] seb128, and misleading info reagrding the really required dbus-version for upstream gdm [16:16] ah [16:16] you should really use intrepid for devel work [16:17] seb128, initially I wanted something stable ... and not introduce more moving targets than necessary [16:18] well it means that you have to backport all the things you need where you could just dist-upgrade [16:18] and things don't break that much around you [16:31] yay [16:33] seb128 is this the right thing to do when libgnomekbd have symbols removed and soname changed to .3 http://paste.ubuntu.com/43372/ ? [16:33] ember: I already packed this update [16:33] and evince [16:34] cool, but is correct the diff? [16:34] they are blocked due to CD builds, not the right time to change sonames [16:35] ember: yes, no need to add a shlibs though [16:35] I was just about to go [16:35] hmm ok thanks for the info [16:35] I'll set up a new system to claim updates soon, I've discuss that with dholbach, we start duplicating work too much there [17:01] ember: thanks for your updates! I will sponsor them as soon as intrepid is open again :) [17:01] np, thanks === mcasadevall_ is now known as NCommander === hggdh|away is now known as hggdh [19:21] * mpt wonders why Nautilus offers an "Open in Text Editor" menu item when his iPod is selected [19:26] mpt: To change the lyrics of the songs! Duh! [19:29] oh yeah [19:29] In that case it should offer an "Open in Text Editor Backwards" item, so I can read all the subliminal messages [19:34] I think you should submit a feature request. :)