[00:01] at the moment I can only see an error in timeout accounting causing this error [00:06] I doubt that is the case though [00:27] w00t, cleaning up the Incomplete list (marking as dupes, closing as Invalid, closing as fixed based on update, etc) has brought compiz back below 350 bugs [00:27] and now that's all bugs for all compiz source packages [00:29] https://bugs.edge.launchpad.net/gvfs/+bug/376145/comments/73 <- summary of my findings, now for bed [00:29] Launchpad bug 376145 in gvfs "gvfs-gdu-volume-monitor crashed with SIGSEGV in gdu_pool_get_presentables()" [High,Confirmed] [00:43] thanks james_w [00:44] there are no threads involved are there? [00:44] I can't find any [00:44] i don't think so [06:09] robert_ancell_: Just subscribed you to that bug I planned on filing against gdm. [06:09] TheMuso, thanks [06:38] Good morning! === asac_ is now known as asac [07:20] Good morning [07:26] hey pitti [08:37] hey pitti, there are a number of packages that depend on policykit-gnome, should these be updated to policykit-1-gnome? [08:37] robert_ancell: you mean apart from checkbox and screen-resolution-extras? (these two still need to be ported) [08:39] there is gnome-mount and gnome-volume-manager [08:40] ah [08:40] they weren't ported to pk-1, though, and are obsolete, too [08:40] ok, I wasn't sure if they're obsolete [08:40] apt-get removed! [08:41] robert_ancell: both still use hal, too, so they don't really use pk-1 [08:41] hello robert_ancell [08:41] robert_ancell: hal doesn't use any PK at the moment, so if you touch the packages, just drop the pk-gnome depends entirely, I'd say [08:42] seb128, hey [08:42] robert_ancell, good work on gdm and g-c-c ;-) [08:42] thanks! [08:43] I have to go, one last question - there are some UI changes to PolicyKit GNOME - are these out of scope for Karmic? [08:43] one is changing the "authentication failure" message to "incorrect password, please re-enter" [08:44] the other is selecting your user by default [08:44] from upstream? [08:44] seb128, pitti ^^^ [08:44] from me, as requested by design team and others [08:44] I wouldn't like to introduce an ubuntu specific string now, it breaks all translations again [08:45] same here [08:45] sure, I have no desire to push them through so I'll leave the patches in LP and let someone else fight for them if they think they're important [08:45] no objection to getting it upstream, of course [08:45] I'm pushing them upstream [08:45] selecting your user by default++ [08:45] selecting user by defualt sounds harmless [08:46] it's not really an UI change, more like a bug fix [08:46] ok, I'll try and get that one done (it is a subset of other changes). I think that will annoy people [08:46] I don't actually have that, though [08:46] I guess it only happens if you have more than one admin? [08:46] currently no user is selected by default [08:46] sorry, you must be right [08:46] ah, I see [08:47] I added my test user to admin, can reproduce now [08:47] that's annoying, yes [08:47] annoying huh? [08:48] gtg, see you all tomorrow [08:48] robert_ancell, see you [08:48] * pitti votes for changing the "auth fail" string to "This operation requires beta-three clearance" [08:49] or using sudo insults :) [08:49] pitti, do you know where the strings comes from there? [08:49] the auth failure one [08:49] hello everyone [08:49] it's not in pk-1-gnome? It really should [08:49] hey chrisccoulson [08:50] hey pitti [08:50] pitti, I was looking in polkit-1 [08:50] not -gnome [08:50] how are you today? [08:50] hey chrisccoulson [08:50] chrisccoulson: pretty good, although still a bit tired; hunting gdm/g-s-d bugs now [08:50] hey seb128 [08:51] whats wrong with g-s-d? [08:51] ah, keyboard issues [08:51] bug 421212 [08:51] Launchpad bug 421212 in gnome-settings-daemon "gdm ignores keyboard layout selection for variants" [High,Triaged] https://launchpad.net/bugs/421212 [08:51] yeah, i just saw that in my inbox [08:51] chrisccoulson, did you manage to solve this timeout not being at the correct value? [08:51] thanks for hunting the d-bus activation thing [08:52] seems there's no final understanding yet, but you are closing in? [08:52] seb128 - not yet. james_w did more work on that than me. all we know is it isn't timing out with the correct value [08:52] it times out much sooner than 25 seconds, which should be the default [08:52] about 3.5 seconds on my box [08:53] pitti, where is polkit-1-gnome being worked? fd.o? [08:54] chrisccoulson: is that a constant ratio? I. e. does it time out after 7 seconds when you request 50? [08:54] pitti, how are translators organized there, it seems poorly translated at least in french on karmic [08:55] seb128: http://git.gnome.org/cgit/PolicyKit-gnome/ [08:55] seb128: policykit-1 itself is on fdo [08:55] pitti - i think james_w discovered that it wasn't completely linear [08:56] seb128: there's a whole bunch of recent translation commits there [08:56] pitti, oh, lot of translations updates there, do you think you could snapshot git or just the .po after beta? [08:56] pitti, yes, GNOME translators are usually active ;-) [08:56] seb128: sure [09:06] pitti, you want me to do it? [09:07] seb128: oh, if you wish [09:07] pitti, ok, will do [09:07] merci [09:07] a-haaa [09:07] de-nada [09:08] so, I found the bug for one case of wrong keyboard variant handling, just to discover yet another one *sigh* [09:08] urg [09:08] the indicator message applet seems screwed there, brb [09:12] pitti - another bug in g-s-d? [09:13] more like a conceptual flaw [09:14] whats the issue now? [09:14] so, g-s-d tries to be clever [09:14] if /desktop/gnome/peripherals/keyboard/kbd/layouts is empty, it adds $GDM_KEYBOARD_LAYOUT to the gconf list [09:15] if it's nonempty, it tries to match $GDM_KEYBOARD_LAYOUT to one of the existing lists [09:15] neither actually works, since gdm and g-s-d have a different interpretation of the value, but that's trivial to fix (doing now) [09:15] ah, ok. thanks:) [09:15] but if you already have a gconf setting, like [us], and then select German keyboard, neither of those two cases matches, and the keyboard isn't set [09:16] so we need a third case to add the new layout to existing entries [09:16] and finally, gdm doesn't read the variant value from hal [09:18] * pitti can't believe that there isn't a g_string_replace() [09:19] or _subst() or whatever [09:40] when i install my jaunty on my pc it installs all the way then when i reboot to the new system it gives me the message "existing intel boot agent" how do i bypass that [09:42] when i install my jaunty on my pc it installs all the way then when i reboot to the new system it gives me the message "existing intel boot agent" how do i bypass that [09:43] DuckGod: that's a message from your network card when it's being initialised by the BIOS. Does it not say anything after that? [09:43] no it just stalls at that message [09:49] have any advice? [09:51] seb128 - i reverted the gnome-python-extras change which removed python-gnome2-extras last night, but i just realised that i forgot to push to bzr [09:51] and seb128 isn't here ;) [09:53] DuckGod: maybe check the boot settings in the BIOS, make sure it's actually going to try and boot from the hard disk [09:54] i took my ethernet card out and now it says reboot and select proper boot device [09:55] after that i went into bios and set it to bood from the hard drive and got the same message [09:55] so then i tried to boot from cd rom and got the same message [09:55] how is brainstorm.u.c used? that is, do the blueprints for discussion at UDS get created from the most popular ideas at brainstorm.u.c? [09:58] any advice? [10:00] DuckGod - you might have more luck in #ubuntu, which is for support [10:01] this is the only room that anyone awnsered me === dpm__ is now known as dpm [10:12] mat_t: can you tr [10:13] mat_t: can you try my new package, please? https://launchpad.net/~albertomilone/+archive/x-testing/+packages [10:14] tseliot: sure, will try it in a sec :) [10:14] thanks [10:15] i386 is https://launchpad.net/~albertomilone/+archive/x-testing/+build/1269831 [10:20] * tseliot can't see the cursor flicker any longer \o/ [10:21] tseliot, do you know about any upstream fix for dpms issues? [10:21] hey btw ;-) [10:21] seb128: for what card? [10:22] tseliot, intel there but I'm not sure it's specific to it [10:22] seb128: is there a bug report about it? [10:22] * tseliot needs more details [10:24] tseliot, see https://bugzilla.gnome.org/show_bug.cgi?id=596609 [10:24] Gnome bug 596609 in gnome-power-manager "screen wakes up right after it has been suspended" [Normal,Unconfirmed] [10:24] tseliot, upstream says there is 3 xserver changes [10:26] seb128: ok, let me try to reproduce the problem here. I might have a patch for that already [10:26] tseliot, ok, thanks [10:27] * tseliot can't lock the screen... [10:27] * seb128 hugs pitti for tackling xkb issues [10:28] tseliot, there is no need to lock anything [10:28] tseliot, set the dpms timeout to 1 minute and don't touch the box [10:28] the screen will flicker and come back [10:28] ok [10:28] you just have to wait one minute [10:29] I can do it ;) [10:29] I've confirmed the bug there without gnome-screensaver running [10:29] ;-) [10:29] * pitti hugs back seb128 [10:32] * tseliot can't reproduce the problem with the screensaver on and will try without it [10:33] tseliot, you have the screen cutting? [10:33] ie dpms kicks in and stay after 1 minute? [10:33] ok, I reproduced the problem [10:33] let me try a new patch [10:34] thanks [10:36] grrr, focus issues and closing wrong dialogs [10:52] tseliot: looks like this might have worked! :D [10:52] mat_t: great and thanks for testing :-) [10:54] tseliot: thank you. I'll keep you posted if anything pops up [10:54] mat_t: ok, good [10:55] tseliot: would you mind just marking the bug as fixed? [10:56] mat_t: sure, when I upload the new package [10:56] pitti, did you restart the retracers? [10:56] tseliot: of course. Thanks again [11:04] seb128: yesterday; are they broken again? I didn't get new mail [11:05] mat_t: please don't, it will be auto-closed by the upload [11:05] "fix committed" would be appropriate, though [11:06] pitti, the log didn't get updated for a day [11:06] pitti, but they are running [11:06] pitti, ie they seem stucked [11:06] I was not sure if you just restarted those [11:06] ie the log didn't get an update yet [11:06] or if they are stucked [11:06] hm; still in VTs debugging gdm, I can have a look later [11:06] do they have any lock files? [11:06] ok thanks [11:07] yes, lock file, apport processes running [11:07] yes, lock file, apport processes running and no log update [11:11] seb128: and it's not just updating chroots and the log is behind? (due to file buffering) [11:12] pitti, the log didn't change for 19 hours now [11:12] that sounds wrong then [11:12] 2009-09-30 15:42 log [11:13] ronne went down yesterday, and was rebooted [11:13] but I'm not sure how this could cause this [11:13] perhaps just kill them and remove locks? [11:13] ok, will try that [11:15] pitti, done, let's see what happens next [11:17] pitti: yes, that's exactly what I meant :) [11:30] aargh, stupid unreliable 3g connection! [11:38] seb128: ok, problem solved. We were missing on patch [11:39] tseliot, oh, good, thanks [11:39] seb128: it's really an xserver bug [11:39] can you reassign the launchpad bug and comment? [11:39] I will close the GNOME one [11:42] seb128: I've just added a comment in the gnome bug report (without closing it). I'll do the same for LP and assign it to xserver [11:42] tseliot, thanks [11:42] seb128: thanks for bringing the problem to my attention === MacSlow is now known as MacSlow|lunch [12:50] dpm, hye [12:50] hey [12:51] hey seb128 [12:51] dpm, there is quite some updated policykit-gnome-1 .po in GNOME git compared to karmic, do you think you could get those imported directly in rosetta by some way or is it easier if we do source changes in karmic? [12:58] is there a reason that something would behave differently when you run the libtool wrapper script vs the executable in .libs? [12:58] using libs from the same build or system ones [12:59] ie if your build has a binary and libraries the wrapper will use the libraries from the local build [12:59] where just running the binary in .libs will not [12:59] ah [13:00] thanks [13:00] you're welcome [13:00] do you know of some magic to get gdb to "run" the wrapper script? [13:00] no [13:00] but you can LD_LIBRARY_PATH=.libs gdb ... [13:00] libtool --mode=execute gdb hell [13:00] documentation to the rescue :-) [13:01] hehe [13:01] I usually use the LD_LIBRARY_PATH trick which works fine too [13:05] seb128, for source packages, I think it's best to do the changes in them and import them through the normal process. If you provide me a tarball with the POT template and all the PO files I think I can upload that as well, but they should match the template in the package [13:05] hey james_w [13:05] you had any more thoughts on this timeout issue?@ [13:05] hey chrisccoulson [13:05] I'm looking at it now [13:06] cool:) [13:06] dpm, well the package doesn't change any po [13:06] trying to track down which code path causes the timeout error to actually be returned [13:06] it's proving rather tricky :-) [13:06] yeah, i'm not too sure either [13:06] I might need a DBus built with -O0 [13:06] i tried following the code in libdbus last night [13:07] there are two functions that cause a timeout error to be sent, and breaking on either of them doesn't trigger [13:07] but i got tired and went to bed ;) [13:07] dpm, so it's basically importing the po from the git source, I can also copy those in the package and upload but that will be after beta [13:08] chrisccoulson, so you do sleep sometimes? ;-) [13:08] seb128 - i slept for 4 hours last night ;) [13:08] urg [13:08] that's more than usual! [13:08] you don't need much sleep or you are just counting on weekends to catch up on the extra hours? [13:09] i tend to just catch up at the weekend [13:09] i sleep a lot at the weekend! [13:11] hmm [13:12] could it be the timeout error isn't the client-side one? [13:12] the server is sending a timeout error to the client? [13:12] that would explain the symptoms a bit better [13:12] not why the call timeout changes it though [13:15] james_w - i'm not sure where the timrout message delivery is dispatched from. i know the message is a client-side one, but i don't know if it is triggered off the client's main loop, or whether it is triggered from the daemon [13:15] that's the bit i didn't understand yet [13:16] the message is pre-allocated client side [13:16] (I think) [13:16] I can find only two functions that access it [13:16] neither seems to be called here [13:16] building with -O0 now to help [13:17] I'm pretty sure the timeout counting is purely client-side [13:17] hmm, maybe not [13:17] yeah, the actual message itself is purely client-side [13:17] but i'm not so sure about the timing [13:17] i couldn't find any main loop interaction when i followed the code last night [13:18] seb128, where's the upstream git repo? http://git.gnome.org/cgit/PolicyKit-gnome/ is still the old policykit, isn't it? [13:19] dpm - 0.94 is polkit-1 [13:19] ok, -O0 helps a lot [13:22] I can now break on the error return [13:22] chrisccoulson, ah, thanks. I'm guessing that the master branch is also polkit-1. Should I fetch the po files from 'master', seb128, or which one is used in the src package? [13:23] james_w - you'll have the issue solved by the time i finish work at this rate ;) [13:23] chrisccoulson: you loosened the lid :-) [13:23] not sure I'll fix it, but hopefully we can get beyond "wtf?" [13:24] dpm, http://git.gnome.org/cgit/PolicyKit-gnome/ is still the old policykit, isn't it? <- no [13:24] dpm, it's the new one [13:24] ok, so it really does think it has timed out [13:25] dpm, yes master is the one to use [13:25] seb128, ok, I'll have a look into it, see if I can upload those and I'll let you know about the outcome [13:26] dpm, there has been a few string additions since our version but not string change to existent strings [13:26] dpm, thanks [13:26] ok [13:27] dpm, otherwise maybe you can drop an email to translator to say they might want to review and upload their locale one? [13:27] ie that git has lot of updates and could be a good idea to review and upload the updated upstream po to rosetta for karmic [13:30] seb128: pitti: I'm still getting a system beep on battery low on my Mini 9 - is this a known issue? [13:31] mat_t: hm, no; do you have "pcspkr" in /etc/modprobe.d/blacklist.conf? [13:31] * mat_t checks [13:33] james_w - can you work out where the timeout message is dispatched from in GDB then? [13:33] (i can't do much from here at the moment) [13:34] pitti: I've got blacklist pcspkr [13:34] chrisccoulson: dbus/dbus-connection.c [13:34] calls to complete_pending_call_and_unlock [13:35] in _dbus_connection_block_pending_call [13:35] the failure is the one at the end [13:37] I'm not quite sure what the success hits though [13:37] it's either something to do with the goto [13:37] or /* check to see if we already got the data off the socket */ [13:37] /* from another blocked pending call */ [13:38] or DBUS_DISPATCH_DATA_REMAINS [13:39] pitti, retracing are happily retracing now [13:39] mat_t: does it actually sound like the pc speaker, or is it more like a real system sound? [13:39] seb128: \o/ [13:39] pitti, thanks for the suggestion of just restart everything ;-) [13:39] pitti: sys speaker [13:40] hmm, I'm thinking maybe compiz should just not ship the gconf plugin at all [13:40] james_w - how do you get _dbus_verbose() to print the debug messages? [13:40] might be useful [13:40] or kconfig [13:40] I just did that :-) [13:40] too easy to screw things up [13:40] pitti: really loud and awful :) [13:40] s/_dbus_verbose/_dbus_warn/ for the easy way [13:40] Compiz doesn't support that fancy new sound because it doesn't use libcanberra [13:40] so you get system beep [13:42] arrrgh [13:42] I wasted two hours to track down a single-character fix [13:42] hrmpf gdm [13:42] but keyboard variants are finally working [13:42] pitti, it's better than two hours to not track down the issue ;-) [13:43] * seb128 hugs pitti [13:43] you rock dude [13:43] * pitti hugs back seb128 [13:43] do you upload g-s-d and gdm now? [13:43] seb128: I'll commit it [13:43] I can still upload after beta freeze [13:43] ok, feel free to upload [13:43] just in case we get another fix in between [13:43] there's already a gdm in the queeu [13:43] right [13:43] * pitti leaves the VTs and starts a real X session again [13:44] pitti, there is a gdm sponsoring request pending too if you want to look at that while you are at it === korn_ is now known as c_korn [13:52] seb128: I'll have a look [13:52] thanks [13:55] dbus_connection_send_with_reply_and_block(): Waited 5057 milliseconds and got no reply [13:56] that's a lot less than the timeout, so why do you think it is a timeout? [13:58] mat_t: can you please check "lsmod | grep pc" whether you have pcspkr or snd_pcsp loaded? [13:59] pitti: ok [14:01] got it [14:01] how did this ever work? [14:06] mac_v: I was just pestering asac about the new karmic artwork for network manager's notification area applet not showing a little sub-icon for VPNs and he redirected me to you :) [14:07] mac_v: please shout if you want a bug filing or somethign testing :) [14:08] Ng: are you using the distro version of humanity or from bzr? [14:08] mac_v: distro [14:08] Ng: yeah , it had problems... it is fixed in lp:humanity [14:08] ok cool :) [14:09] Ng: it will be immediately updated after beta [14:09] \o/ === MacSlow|lunch is now known as MacSlow [14:09] Ng: could you test the bzr and see it you still have problems? if still having a problem , a screenshot would be nice [14:13] james_w, nice, what was it? [14:13] chrisccoulson, vuntz managed to track the keyring timeout issue or figure what is wrong rather [14:13] should be fixed for karmic too ;-) [14:14] mac_v: yep that puts a little padlock on fine :) [14:14] :) [14:15] mac_v: bluetooth? [14:15] Amaranth: done ;) [14:15] seb128: there was just one gdm sponsoring bug left, I cleaned up the other ones; committed now [14:15] mac_v: after beta? :) [14:15] yup [14:15] Amaranth: its been updated in the bzr since nearly a week, but we are in freeze :( [14:16] seb128: http://paste.ubuntu.com/282982/ [14:16] pitti: nope, seems like I don't [14:16] seb128: I didn't figure it out [14:16] seb128: I actually don't get it at all :-) [14:16] mat_t: hm, then I'm really puzzled what produces it on your machine.. [14:17] * mat_t is puzzled, too [14:17] mac_v: hey yeah, the bluetooth icon looks great :) [14:17] pitti: I'm going to run a fresh install and see what happens [14:17] mac_v: what's the little /!\ on the network manager icons? [14:17] mat_t: or just try a live system [14:18] Ng: screenshot pls ? [14:18] mac_v: yes, that should be gone [14:18] mac_v: a little warning you added [14:18] pitti, thanks! [14:18] pitti: hey, the notify-osd icons are not being used (nor are they intalled by default) [14:19] mat_t: what should be gone... ? i dont know what to remove without a screenshot ;) [14:19] kwwii: hm, works here.. [14:19] pitti: you see the icons with the gradients in the notify bubbles? [14:19] mac_v: a little red /!\ emblem you added to the ad-hoc icon [14:19] or do you see simple grey icons? [14:19] mac_v: http://mairukipa.tenshu.net/screenshots/2009-10-01-icons.png (I have wired and wireless and VPN active atm) [14:19] kwwii: notify-osd-icons is also in http://cdimage.ubuntu.com/daily-live/20090929.2/karmic-desktop-i386.manifest [14:19] mat_t: that is need for the ad-hoc [14:20] mac_v: can you make that yellow? [14:20] Ng: argh! nm-applet is totally mesed up [14:20] mac_v: could that be because I didn't restart it maybe? [14:20] Ng: blame asac ;0 [14:20] kwwii: it's some light gray; not the gray/blue loudspeaker from hicolor, but the gray-only abstract loudspeaker symbol [14:20] mac_v: as discussed, security is not the only property this is supposed to communicate ;) [14:20] asac: so should i remove it? [14:20] i thought you wanted it [14:21] mac_v: the exclamation mark? [14:21] yeah [14:21] mac_v: i didnt push for any change [14:21] james_w, oh, it's a dbus bug then [14:21] oh ... ok.. removing [14:21] yeah [14:21] mac_v: i thought what you did initially was ok [14:21] james_w, good work! [14:21] pitti: right, those are the wrong ones as well :( [14:21] a [14:21] asac: me too!... mat_t confused me ;p [14:21] mac_v: though not optimal ... but we discussed that the optimal way would be to make a separate ad-hoc sectrion [14:21] pitti: those are the panel icons from Humanity [14:21] we will keep this in mind for the applet rewrite next (nm) cycle [14:21] asac: yes, that we can do for Lucid [14:21] pitti: I'm confused about your comment in bug 596897 [14:21] Error: Launchpad bug 596897 could not be found [14:22] asac: mpt will be leading that [14:22] asac: wwan and ad-hoc are also using device icons in the panel! :/ [14:22] pitti: you say 'gdm separates layout and [14:22] variant with a space' [14:22] pitti: I think part of the problem is that they are being installed into /usr/share/notify-ods/icons/gnome/ [14:22] kwwii: I have this one: http://people.canonical.com/~pitti/tmp/notify.png [14:22] mat_t: i think whatever mpt is leading isnt upstream ;) [14:22] pitti: but the gdm patch in the other bug actually uses a tab as separator [14:22] gnome bug #596897 [14:22] Gnome bug 596897 in plugins "Ignores variants in $GDM_KEYBOARD_LAYOUT" [Major,Unconfirmed] http://bugzilla.gnome.org/show_bug.cgi?id=596897 [14:22] mat_t: but better discuss that somewhere else [14:22] ;) [14:22] pitti: hrm, that is the right icon [14:22] mclasen: after the last two hours of reading gdm code I actually found out that gdm seems to use a wild mix of tabs and spaces :-( [14:23]