[09:46] <pitti> hey seb128
[09:46] <seb128> hello pitti
[09:47] <pitti> seb128: can we attack the libgphoto automounting today? what needs to be done for this? (I can take the f-spot changes in any case, but I'm unsure about the gvfs change)
[09:48] <seb128> pitti: sure, let me do the gvfs change
[09:48] <pitti> seb128: ok, merci; once that's in, I'll look after f-spot
[09:55] <seb128> pitti: uploaded
[09:56] <pitti> wow, that was quick *hug*
[09:56] <seb128> ;-)
[09:56]  * seb128 hugs pitti
[09:56] <pitti> I'll let it hit the archive, and until then continue to attack the sponsoring queue
[09:56] <pitti> I managed to cut quite a dent into it
[09:56] <seb128> I noticed the uploads, good work there ;-)
[09:58] <mvo> seb128, eh, that was cheating, right? you had the change prepared already ;) ?
[09:59]  * seb128 slaps mvo
[09:59] <seb128> mvo: no but I looked at it yesterday and knew what to change ;-)
[10:07] <pitti> seb128, mvo: did either of you already look into the patch in bug 159996 ?
[10:07] <seb128> pitti: mvo did, he commented on the bug and the new patch has only been added yesterday
[10:08] <pitti> right
[10:08] <pitti> just wondering who should look into it now
[10:08] <seb128> the new patch looks correct
[10:08] <seb128> mvo: do you have any gnome-control-center pending upload?
[10:10] <seb128> pitti: mvo or I will do it
[10:10] <pitti> ok, thanks; I'll grab other stuff then
[10:11] <mvo> seb128, not right now, but I had the patch on my list of things to look at for today (that list is growing at a alarming rate)
[10:12] <seb128> mvo: ok, it's probably a quick one so I'll not steal if from you ;-)
[10:12] <mvo> not sure if "stealing" is the right word here :)
[10:12] <mvo> but I'm fine looking at it
[10:16] <seb128> mvo: I'll do that sponsoring now
[10:16] <seb128> mvo: I want to change the scrollkeeper to rarian depends there too
[10:18] <mvo> seb128, what does the new version of the patch do? set the tab to insensitive? or set the hbox/vbox with the effects buttons insensitive?
[10:19] <seb128> mvo: set the tab to insensitive, do you think that's not the right way?
[10:20] <mvo> seb128, no sure, if the tab text "visual effects" can still be clicked and the content of the tab is still visible then I think that is the right fix (the content is all insensitive)
[10:21] <seb128> mvo: ideally it would give some clue about why it's insensitive too
[10:21] <mvo> seb128, just checked the new patch, I think its looks good
[10:22] <seb128> mvo: ok, so do you want to upload it or should I? ;-)
[10:22] <mvo> yeah, ideally it would offer to install compiz
[10:22] <mvo> go ahead, if you have it locally anyway
[10:22] <mvo> I would have to download it first etc
[10:23] <mpt> pitti, hi, did you write "Cowboy ui_download_start() to take an alternative title for the progress window, so that we can abuse it for other progress dialogs"?
[10:31] <slomo> mvo: did you fix gnome-app-install already? ;)
[10:35] <seb128> slomo: do you know if bug #119861 is a known issue?
[10:35] <seb128> slomo: the example attached to the bug still gives an error on intrepid
[10:37] <slomo> seb128: known problem, gstreamer can't handle swf... only very very few are working with gst-ffmpeg
[10:37] <pitti> seb128: shall I grab bug 245166 now, or are you working on a gdm ATM?
[10:37] <seb128> slomo: gst-ffmpeg bug?
[10:38] <seb128> pitti: feel free to do it, there is an another gdm sponsoring request waiting too
[10:38] <pitti> seb128: oh, bug 264834; will look at that as well
[10:38] <slomo> seb128: more a ffmpeg bug... i don't think gst will support swf in any usable way until someone makes a swfdec/gnash plugin for gstreamer :)
[10:38] <seb128> pitti: thanks
[10:39] <seb128> slomo: ok I'll reassign it then, thank you
[10:42] <mvo> slomo, yes, monday IIRC
[10:53] <asac_> lame question, but how can i make a screenshot of menus using the gnome user-experience ;)?
[10:54] <mvo> asac, use the screenshot thing and give it a delay (or gimp with a delay)
[10:54] <asac_> cant be the right approach to install ksnapshot to take a timed snapshot ;)
[10:54] <asac_> mvo: so gimp is the only way?
[10:54] <asac_> (do do that through ui)
[10:54] <asac_> s/do do/to do/
[10:55] <seb128> asac_: applications, accessories, screenshot?
[10:56] <asac_> heh
[10:56] <asac_> obvious
[10:56] <asac_> what i used was the snapshot applet ;)
[10:56] <seb128> there is a snapshot applet?
[10:56] <asac_> i think so. i dont have it anymore because i couldnt make it snap with a delay
[10:57] <mvo> seb128, not on windowmaker ;)
[10:57] <asac_> hmm ... seems to be not in the default applets anymore
[10:57]  * mvo hides from seb128 
[10:57]  * seb128 hugs mvo
[10:57] <asac_> sorry for the noise :)
[10:57]  * mvo hugs seb128 
[10:58] <seb128> mvo: did you fix the compiz focus bug yet btw? ;-)
[10:59]  * seb128 runs
[11:00] <mvo> *cough*
[11:00]  * mvo hides again
[11:01] <seb128> mvo: it seems to not happen immediatly but once I start getting the issue it's consistent, I just need to figure what triggers it
[11:01] <seb128> mvo: I usually get it quickly when normally use my laptop, but normal use is lot of basic actions so hard to figure which one triggers the bug
[11:02] <seb128> mvo: what there many changes between the snapshot we had and 0.7.8 which could create a such bug?
[11:02] <seb128> mvo: maybe I should try to bisect the changes
[11:02] <davmor2> asac: just drag the snapshot menu item to the panel works for me :)
[11:02] <mvo> seb128, could you try unloading the animation module and see if that helps?
[11:03] <asac> davmor2: hehe
[11:03] <mvo> seb128, I suspect its there somewhere, but I'm not sure
[11:03] <seb128> mvo: will do
[11:03] <asac> see how rudimentary my daily live is ;)
[11:03] <seb128> I'm on my desktop right now which doesn't run compiz
[11:03] <mvo> seb128, if we can isolate the plugin that causes it, half the work is done
[11:03] <asac> just terminal, terminal, firefox, nm applet :) ... all day long ;)
[11:05] <davmor2> asac: Seriously you just get an icon on the panel that act like the applet does but you can set timer ect
[11:05] <davmor2> s/ect/etc
[11:06] <asac> davmor2: i just didnt know about the menu item ;)
[11:06] <davmor2> :)
[11:11] <Ng> seb128: yeah that matches up with what I'm seeing, it doesn't happen to begin with, but then becomes pretty much constant.
[11:11] <Ng> seb128: out of interest, do you use sloppy focus?
[11:12] <seb128> Ng: what is this sloppy focus thing? ;-)
[11:13] <Ng> seb128: hey, if you have time to click in windows before you use them, that's great ;)
[11:13]  * Ng ducks
[11:13] <Ng> I've always used sloppy focus and I am very protective of it :D
[11:13] <seb128> ah, I use the click focus option
[11:13] <seb128> I don't really get the point of the sloppy focus ;-)
[11:13] <Ng> it's exactly the same as click-to-focus, without the clicking :)
[11:14] <seb128> in fact I tend to not use the mouse to do workspace and applications switching, just shortcuts and alt-tab
[11:14] <Ng> ah
[11:14] <Ng> but that's good, that tells us the focus thing isn't just because of focus-follows-mouse
[11:14] <seb128> so having the mouse pointer doing changes when I move the mouse tend to annoy me
[11:14] <seb128> right
[11:14] <mvo> Ng, me too!
[11:15] <mvo> Ng, could you disable animation and see if that helps too? and maybe wall if animation does not help
[11:15] <Ng> sure
[11:17] <Ng> well disabling animation hasn't instantly fixed it, but I'm wondering if I should restart compiz.
[11:17] <Ng> hah, disabling wall leaves me unable to change workspace
[11:18] <mvo> Ng, right, there should be some replacements
[11:18]  * Ng nods. cube it is
[11:18] <mvo> there used to be plane as well, I think it got removed at some point
[11:19] <Ng> ok so switching to cube has made the focus start working again. it's "switch to face N" bindings conflicted with my previous "switch to workspace N" ones, so I kept mine.
[11:20]  * Ng switches back to wall to see if it breaks again
[11:20] <Ng> back with wall and it's still working, so I guess I'll stick with cube for a while and see if it goes wrong there
[11:22] <mvo> ok, thanks
[11:22] <mvo> you can tune cube to switch faster if the animation slows you down
[11:22] <Ng> I already did :)
[11:22] <Ng> I always tune the animations to be really fast :)
[11:26] <pitti> oh, it just occurred to me that we still launch trackerd and tracker-applet by default, although it's not supposed to
[11:26] <pitti> does anyone have an idea why this happens?
[11:26] <pitti> oh, /etc/xdg/autostart/trackerd.desktop
[11:26] <pitti> hmm
[11:29] <pitti> seb128: any idea why /etc/xdg/autostart/tracker-applet.desktop even exists? shouldn't that be handled by the panel itself, if you add the applet?
[11:30] <seb128> pitti: because it's a notification area icon
[11:35] <pitti> seb128: hm, just upgraded to new gvfs; this didn't only stop auto-mount, but apparently the complete nautilus handling; nothign happens now, I don't get that "what do you want to do" dialog
[11:35] <pitti> does that work for you?
[11:42] <seb128> pitti: hum, but the camera is listed and you can double click on it?
[11:42] <seb128> pitti: I didn't try, I don't have my camera here
[11:42] <pitti> seb128: I get it only in the computer place, no icon on the desktop (which would be the mount already, so that's ok)
[11:42] <seb128> and do you get something if you double click on the icon?
[11:46] <pitti> seb128: I get the nautilus browser
[11:46] <pitti> no question what to do with it
[11:46] <pitti> media settins says "Photos: Ask"
[11:46] <seb128> pitti: ok, I'll debug that later but I need to get my camera first
[11:47] <pitti> ok
[11:47] <pitti> seb128: shall I revert the f-spot change already, to talk driectly to libgphoto again?
[11:47] <seb128> yes please do
[11:47] <seb128> was the question working in hardy?
[11:48] <seb128> the issue is that when you don't automount you don't know the device content and that's used to figure if the device is a mtp player or a camera iirc
[11:49] <seb128> ok, I'm late for lunch I've to run
[11:49] <seb128> bbl
[11:51]  * mvo -> food
[12:21] <Laney> pitti: Thanks for gnome-terminal sponsoring \o/ my irssi loves you now
[13:11] <pitti> Laney: :)
[13:28] <seb128> re
[13:29] <mvo> alt-w again?
[13:29] <pitti> hey seb128
[13:29] <seb128> mvo: no, wrong click this time ;-)
[13:29] <seb128> pitti: I was thinking about the gphoto mount on my way to lunch, we used g-v-m again for that in hardy
[13:30] <seb128> pitti: gvfs need to mount the device to know if that's a camera
[13:30] <pitti> seb128: that's strange; nautilus already knows it's a camera even when unmounted
[13:30] <pitti> I get the camera symbol in computer:/// (it's in hal, after all)
[13:31] <seb128> pitti: is there an icon property for the device?
[13:31] <pitti> what is an icon property?
[13:32] <seb128> pitti: there is something which let you set an icon to use for devices in hal no?
[13:32] <pitti> seb128: hm, that might be for GNOME, but I never saw it
[13:33] <pitti> info.category = 'camera'  (string)
[13:33] <pitti> info.capabilities = {'camera', 'camera', 'camera', 'access_control'} (string list)
[13:33] <pitti> but I really guess nautilus is looking at those ones
[13:34] <seb128> pitti: I need to look at the code again but the issue comes from the fact that some devices are mtp player and cameras
[13:35] <seb128> pitti: the nautilus code is somewhat stupid it looks for a PICTURES directory on the mount
[13:35] <pitti> seb128: uh? that doesn't even exist on my cam, it's "DCIM" (and that's very common)
[13:36] <seb128> pitti: see bug #258936
[13:36] <pitti> seb128: hm, I had assumed it'd look at the hal properties
[13:36] <seb128> pitti: gvfs does that's why you get the correct icon
[13:37] <seb128> pitti: see libnautilus-private/nautilus-autorun.c _g_mount_guess_content_type()
[13:39] <seb128> pitti:
[13:39] <seb128> 				if (strcmp (stor_device_caps[n], "portable_audio_player") == 0) {
[13:39] <seb128> 					g_ptr_array_add (types, g_strdup ("x-content/audio-player"));
[13:39] <seb128> 				}
[13:39] <seb128> 				/* TODO: map other hal capabilities to x-content/ types */
[13:39] <seb128> 			}
[13:39] <seb128> pitti: adding some snippet there might work though
[13:40] <pitti> hm, I don't see that in n-autorun.c
[13:40] <pitti> is that in gvfs?
[13:40] <seb128> pitti: no, nautilus
[13:40] <pitti> grep -r portable_audio_player -> no result in nautilus
[13:41] <seb128> let me update, my checkout on this box is outdated
[13:42] <seb128> hum, they changed the code since
[13:43] <seb128> 	* libnautilus-private/nautilus-autorun.c:
[13:43] <seb128> 	* libnautilus-private/nautilus-autorun.h:
[13:43] <seb128> 	Use g_mount_guess_content_type() instead of using our own
[13:43] <seb128> 	sniffing code. Also provide an option for the user to select
[13:43] <seb128> 	an application to use (#532474).
[13:43]  * seb128 looks in glib
[13:44] <seb128> pitti: they moved the detection to glib which is lower than hal in the stack and use shared-mime-info
[13:44] <seb128> pitti: rather not lower but which doesn't depends on dbus or hal rather
[13:45] <pitti> ugh, how bad
[13:45] <seb128> that's probably why they use the directories
[13:45] <pitti> *headdesk*
[13:45] <pitti> that's soo backwards :(
[13:46] <pitti> but that wouldn't work for things like scanners at all
[13:46] <seb128> pitti: anyway in hardy I'm pretty sure that was similar, they needed the device to be mount to guess the content so we used gnome-volume-manager for cameras
[13:46] <seb128> pitti: you don't browse scanners, we are speaking file manager there
[13:46] <seb128> pitti: gvfs correctly detects the devices and applications would use gvfs
[14:16] <MacSlow> greetings everybody
[14:19] <pitti> hey MacSlow
[14:19] <pitti> seb128: where does davidz usually hang out? I'd like to talk with him about this camera thing
[14:19] <seb128> pitti: #gnome-hackers
[14:19] <pitti> on gimp.net?
[14:19] <seb128> pitti: it's still a bit early for him though
[14:20] <MacSlow> pitti, mvo, seb128: hey there
[14:20] <seb128> pitti: irc.gnome.org
[14:20] <seb128> hello MacSlow
[15:02] <seb128> pitti, tedg: one think which has been discussed to the "everybody should have an user switcher applet in the corner", is what we do for people who changed their config to have only one bottom gnome-panel for example (that's quite common since that's the default on some other well known os or some distros ;-) (I added a common on the bug about that)
[15:02] <seb128> "one thing which has not been discussed" rather
[15:02] <seb128> some users also have a floating bar and use autohiding, the corner way is nice when you have a corner
[15:05] <mvo> seb128: let me know if the scripts needs ajustments
[15:05] <tedg> seb128: This is precisely why the next-gen panel will be less configurable :)
[15:06] <mvo> seb128: just to confirm, you have click_to_focus, right?
[15:07] <mvo> Ng: and you have click_to_focus disabled?
[15:07] <seb128> mvo: will do, it could be smarter for quite some things but I think we did that too late in the cycle to get it rocking stable
[15:07] <seb128> mvo: right
[15:08] <mvo> seb128: I think we can do more in the script itself, just the stuff around it (like only show if there is actually a quit button etc) is a bit difficult to do because it requires changes in update-notifier
[15:08] <mvo> or in the infrastructure that picks up the note
[15:08] <Ng> mvo: correct
[15:08] <mvo> thanks ng and seb128
[15:08] <seb128> mvo: right, but right know it adds the applet if you don't have the applet installed for example
[15:09] <mvo> I'm talking to upstream currently, lets see what he thinks (I'm confident, compiz upstream are rockstars!)
[15:09] <mvo> seb128: it does that now? or it should do that?
[15:09] <seb128> mvo: again not sure when you did the previous snapshot and if lot of code changed but the snapshot didn't have the bug
[15:10] <seb128> mvo: it does that, it just modify the config but doesn't look to install packages before suggesting the change no?
[15:11] <mvo> seb128: right, you think of the case when fusa is not installed on the system?
[15:11] <seb128> mvo: correct
[15:11] <mvo> hmmm
[15:11] <mvo> we could add that I guess
[15:29] <mvo> seb128, Ng: could you please try http://people.ubuntu.com/~mvo/test/compiz-fusion-plugins-main_0.7.8-0ubuntu2_i386.deb ? it may fix the issue
[15:29] <seb128> mvo: ok
[15:29] <seb128> mvo: did anybody find an issue or just a random try?
[15:30] <mvo> seb128: this comes from upstream and he thinks it should fix it
[15:31] <seb128> mvo: ok, will let you know in some hours if I didn't the issue again
[15:35] <Ng> mvo: sure
[15:38] <Ng> hmm, compiz won't start for me now
[15:38] <Ng>  /usr/bin/compiz.real: symbol lookup error: /usr/lib/compiz/libresizeinfo.so: undefined symbol: getCompPluginInfo
[15:39] <mvo> Ng: urgs, woah
[15:39] <mvo> Ng: looks like I miscompiled it
[15:40]  * mvo scratches his head and rebuilds it in pbuilder
[15:44] <mvo> ping mvo_
[15:45] <mvo> ping mvo_
[15:46] <mvo> ping mvo_
[15:46] <mvo> ping mvo_
[15:47] <james_w> hey glatzor
[15:48] <glatzor> hello james_w and mvo !
[15:48] <james_w> glatzor: did you see that I uploaded packagekit 0.3.6?
[15:49] <glatzor> james_w, not yet. I was offline the since yesterday :)
[15:49] <glatzor> s/the//
[15:49] <james_w> glatzor: there were some differences between bzr and the PPA
[15:49] <james_w> I went with the PPA and pushed the changes to bzr
[15:52] <mvo> hey glatzor!
[15:53] <mvo> Ng: I updated http://people.ubuntu.com/~mvo/test/compiz-fusion-plugins-main_0.7.8-0ubuntu2_i386.deb (md5sum 5c318f24c098356828eed0889efeb7a4 - just in case the proxy is playing silly bugger)
[15:55] <glatzor> james_w, right. I did not upload the latest version to the ppa.
[15:56] <glatzor> james_w, so, but this is ok. It was only a small patch to make the gstreamer plugin coexist with gnome-app-install
[15:56] <james_w> glatzor: hmm, so should the packages have a versioned conflicts on gnome-app-install or an un-versioned one?
[15:56] <james_w> glatzor: and should it add an alternative for gnome-codec install?
[15:56] <glatzor> james_w, a versioned one. mvo changed gnome-app-install to use an alternative for gstreamer-codec-install
[15:57] <pitti> seb128: indeed, I just have one panel, and I mentioned that in the bug
[15:57] <glatzor> james_w, gnome-codec-install is deprecated now
[15:57] <glatzor> james_w, there was a change in the gstreamer package
[15:57] <james_w> glatzor: ah, ok. I uploaded the wrong thing then, sorry.
[15:58] <glatzor> james_w, nobody was harmed :)
[15:58] <james_w> glatzor: is it ok like that, or do I need to fix something?
[16:00] <glatzor> james_w, it makes testing the codec installer more annoying since a tester would have to remove gnome-app-install, but this should be ok.
[16:01] <james_w> glatzor: I can upload a fix for that, no problem
[16:01] <glatzor> james_w, I should have uploaded the correct version to ppa
[16:01] <glatzor> james_w, if it is only a 5 minute thing I would welcome it
[16:01] <james_w> less than 5 :-)
[16:01] <glatzor> james_w, go for it :)
[16:03] <james_w> glatzor: should I drop the update-alternatives for gnome-codec-install as well?
[16:05] <pitti> tedg: your ppa package in bug 261084 at least seems to improve matters (although not fully fix them); do you think it's something worth uploading?
[16:06] <pitti> didrocks: how's bug 212098 coming along, BTW?
[16:06] <mvo> Ng: with new compiz-fusion-plugins-main now :) ?
[16:06] <tedg> pitti: Yes, I've got another couple bugs I'm looking at rolling into the package -- and then requesting sponsorship.
[16:07] <tedg> pitti: I think it doesn't harm things if nothing else.  Though, I think there is a lower level problem going on.
[16:07] <tedg> pitti: There are a lot of reports of repeated keys -- suspend, brightness, etc.  I think the new X Input stuff has broken some of the hotkey support.
[16:08] <Ng> mvo: I think it's the same, but I'm doing something else atm, will get back to you
[16:09] <mvo> Ng: "I think it's the same"> so the fix does not work?
[16:09] <mvo> Ng: sure, when you have time
[16:10] <pitti> tedg: "some" is an understatement...
[16:13] <glatzor> james_w, yes. you can safely drop it
[16:13] <glatzor> james_w, sorry I was away for some minutes
[16:13]  * tedg doesn't have hotkey support on Intrepid :(
[16:25] <didrocks> pitti: I do not understand the "it's not fixed in intrepid", did you see my tests and comments on a fresh ubuntu install ?
[16:25] <pitti> didrocks: I saw them, but they aren't the point of that bug report
[16:26] <pitti> didrocks: we only put the user into sambashare to workaround exactly this bug, but we don't actually like to :/
[16:26] <pitti> didrocks: and it does not help for non-admin users who want to share stuff
[16:26] <seb128> pitti: is there any reason we don't really like to? that provides a better user experience because in this case no session restart is required
[16:27] <didrocks> pitti: yes, but non-admin will not initiate the share, so, they will not have to figure about this "bug" (cf my use case)
[16:27] <pitti> seb128: we want to get rid of using groups altogether, since you need all this session restarts, and it isn't very flexible
[16:28] <pitti> didrocks: ok, true for non-admins
[16:28] <pitti> since they can't put itself into sambashare anyway (except an admin does it for them)
[16:29] <didrocks> pitti: so, for an non-admin user, the only "weird stuff" is when an admin switch the session to add them (without logged off) and so, they can't see their change when switch again to the user session
[16:31] <pitti> didrocks: right, isn't any different to an admin user in this case (needs to restart session)
[16:33] <didrocks> pitti: so, the only solution is either: 1/ change group dynamically (which is a hard thing to achieve)
[16:33] <didrocks> 2/ detect that an user has been added to a group but session information aren't reloaded
[16:34] <pitti> didrocks: 2) is easy, compare your process' groups and getpwent
[16:34] <pitti> didrocks: 1) currently doesn't work for existing processes, you'd need to restart everything relevant under a "newgrp" session (let's not think about it)
[16:35] <didrocks> ok, so, maybe 2) can be implemented to tell "ok, you have the right, but reload your session"
[16:36] <didrocks> pitti: I first just thought about a hardy SRU
[17:00]  * calc will be a part of desktop starting next monday :)
[17:02] <pitti> seb128: ok, seems we need to revert that gvfs change then; sorry
[17:03] <seb128> pitti: I'm reading #gnome-hackers
[17:03] <pitti> seb128: I'll work on that, and fix f-spot to unmount the gvfs at start; ok?
[17:03] <crevette> hello
[17:03] <seb128> pitti: works for me, thank you
[17:03] <seb128> lut crevette
[17:03] <seb128> pitti: I'll revert the gvfs change
[17:04] <pitti> seb128: ok, thanks
[17:05] <crevette> hello seb128
[17:06] <seb128> switching to my laptop for some testing, brb
[17:06] <crevette> seb128: you're were talking about bluetooth device next day, most phone have bluetooth support, and a bluetooth adapter is really affordable, I bought mine 3 years ago for 8€
[17:06] <crevette> ah toolate
[17:06] <crevette> :)
[17:06] <asac> hmm ... my laptop battery is at 27% but tray icon is still 100% green (though mouse over tells me its 27%)
[17:06] <asac> known issue?
[17:07] <seb128> re
[17:07] <seb128> mvo:
[17:07] <seb128> /usr/bin/compiz.real: symbol lookup error: /usr/lib/compiz/libresizeinfo.so: undefined symbol: getCompPluginInfo
[17:07] <seb128> mvo: the current version is still buggy
[17:08] <seb128> and I didn't download the previous one so that's not a proxy issue
[17:29] <seb128> dinner time, bbl
[17:31] <crevette> diner time at 18:30 ?
[17:47] <mvo> asac: tedg may know about gnome-power-manager issues
[17:52] <tedg> asac: Will be fixed shortly.  PPA building, then I'll request sponsorship :)
[17:57] <Ng> mvo: so yeah, it does look like I'm still getting that symbol error
[17:58] <mvo> Ng: hm, preparing a new one yet again
[17:59] <Ng> (a quick scan of the updates available to me atm suggest there are no compiz ones, so I should have all the latest intrepid compiz jazz)
[18:01] <huats> crevette: remember that seb lives in the North
[18:04] <crevette> huats: yeah but I don't recall my gd parents (who lives nothern than him) eat so early :)
[18:04] <huats> :)
[18:04] <crevette> probably a local habits
[18:04] <mvo> Ng: ok, a new one with f827659f6e9fb36d5b011773f14ca520 is oneline
[18:08] <asac> tedg: thanks. will be out for a few now.
[18:09] <Ng> mvo: that may actually have been user error on my part, but I'm going to blame wget for not saying <blink>I saved that as .1 ha ha ha</blink> ;)
[18:09] <Ng> but it's running now
[18:09] <mvo> Ng: heh :)
[18:09] <mvo> Ng: thanks!
[18:10] <Ng> mvo: btw, one random other compiz thing while I think of it, windows (by which I mean X windows, so even things like the popup part of a GTK dropdown widget) seem to appear with random noise in them before being painted - is that known?
[18:11] <Ng> (it's been around for ages, so it's by no means a regression)
[18:12] <tedg> Ng: Nvidia?  I think that's an nvidia bug.
[18:12] <Ng> intel
[18:13] <Ng> X3100 (GM965)
[18:14] <mvo> Ng: I have seen that yes
[18:15] <pitti> oh argh, /usr/sbin/f-spot-import is totally broken
[18:15] <Ng> hardly an end-of-the-world kind of thing, it's just a bit weird looking. I'll add it to my list of things I should file bugs for :)
[18:15]  * pitti fixes that while he's touching the file anyway, to avoid this silly device dialog f-spot gives you
[18:16] <Ng> mvo: I tend to notice it because my firefox sessions are insanely huge and it has to think for about 15 seconds after it owns its window before it paints anything ;)
[18:45] <tedg> Okay, so I have a package in my PPA that fixes several bugs.  Should I open up a new one for sponsorship?  Request on each of them?
[18:45] <tedg> Bat my eyes and look beautiful?
[19:40] <tedg> bug #283951
[19:41] <asac> tedg: how about a diff?
[19:41] <tedg> asac: between what and what?
[19:42] <asac> tedg: between currnet intrepid version and the one to sponsor
[19:42] <asac> tedg: or if its just a single fix that is included just the patch you applied for that
[19:42] <asac> helps to do review if i can just see the diff ;)
[19:46] <tedg> bzr diff -r 88.1.2..98 lp:~ted-gould/gnome-power/ubuntu-packaging-2.23
[19:46] <tedg> asac: ^
[19:49] <asac> tedg: i did a debdiff and that diff and the outcome isnt identical ;)
[19:49] <asac> but maybe its just me
[19:50] <tedg> asac: Probably the paths are different?
[19:50] <tedg> I think that bzr makes everything relative to the base of the repository.
[19:50] <asac> tedg: i used -p1 for interdiff
[19:51] <asac> and my bzr diff generates prefixes
[19:54] <tedg> asac: Hmm, that might be because of the fix in the source.  Since I moved it out into a formal patch.  I didn't ever do it in the source.
[19:54] <tedg> I think it must be from 0ubuntu4
[19:55] <asac> tedg: maybe. doesnt the "idle" risk of not accepting anything when system is under high load?
[19:55] <asac> bug #261084
[19:56] <tedg> asac: Perhaps, it would have to be increadibly high load.  I mean to the point of process starvation.
[19:57] <chrisccoulson> ping tedg
[19:58] <tedg> chrisccoulson: Hello.
[19:58] <chrisccoulson> hi, following on from our brief discussion yesterday - i tried the build of GPM from your PPA
[19:59] <chrisccoulson> i get the same behaviour when I press the power button on my desktop though
[19:59] <chrisccoulson> every running instance of GPM calls a session dialog, even for inactive users.
[20:00] <chrisccoulson> shouldn't GPM check to see if it is on the active console before responding to the power button event?
[20:00] <tedg> GPM isn't doing that.  GPM is sending a message to GNOME Session Manager, it then shows the dialog.
[20:00] <asac> gnoem bug #50739
[20:01] <asac> gnome bug #507393
[20:01] <chrisccoulson> ah, ok. but one of those should decide to take no action, based on the fact that it isn't on the active console
[20:01] <chrisccoulson> asac - thanks for that
[20:02] <tedg> asac: That patch is in the 0ubuntu6 in my PPA.  Well, cleaned up so that it'll work on Xubuntu and others.
[20:02] <asac> tedg: why do you use egg_warning and not gpm_warning?
[20:02] <chrisccoulson> this 60 second time out is a pain, bacause if you accidentally hit the power button when you have more than 1 user logged in, then your machine will shut down regardless of whether the active user cancels
[20:03] <tedg> chrisccoulson: There should be no way for the message to get to the other active users.
[20:03] <tedg> is there a gnome-session process for all the users?
[20:03] <tedg> asac: Because after the code freeze gpm changed all the gpm_warnings to egg_warnings.
[20:03] <chrisccoulson> yeah, there is one gnome-session per user. thats similar to there being one gnome-power-manager per user
[20:04] <tedg> chrisccoulson: And there's and X server instance per user, right?
[20:04] <chrisccoulson> tedg - that's correct
[20:05] <asac> tedg: ok looks good then.
[20:06] <asac> tedg: only thing you could enlight me with is how the XSMP fallback works
[20:06] <tedg> chrisccoulson: I'm not sure how that could be happening.... is there a seperate dbus session for all the users?
[20:06] <chrisccoulson> there is a separate session bus for each user
[20:07] <tedg> asac: Basically it looks for gnome session on dbus, if it can't find it, then it uses the GNOME Client code to do XSMP.
[20:07] <chrisccoulson> there's one of pretty much everything for each user isn't there (except HAL and the system bus)
[20:07] <tedg> chrisccoulson: Could you install d-feet and see if you can call org.freedesktop.PowerManagement then look at the function "Shutdown" and see what happens?
[20:07] <asac> tedg: well. i see the fallback. but i didnt see the XSMP call
[20:08] <asac> but apparently nome_client_request_save triggers that
[20:08] <tedg> asac: Yes, it should.  But, you shouldn't see it if you have gnome-session installed.
[20:09] <tedg> GNOME Session should be generating the XSMP messages in that case.
[20:09] <asac> k
[20:09] <chrisccoulson> tedg - i'm not quite sure what you want me to do with d-feet. it only lists the objects and supported interfaces doesn't it?
[20:09] <tedg> chrisccoulson: Yes, but then you can run the functions also.  If you click on a function you can execute it.
[20:09] <chrisccoulson> ah, i didn't realise i could send messages with d-feet
[20:10] <chrisccoulson> well, i already know that calling org.freedesktop.PowerManagement.Shutdown will shutdown my machine (I have a script which does that)
[20:11] <tedg> And if you call the gnome session manager for logout, does it show the dialog for all the users?
[20:11] <asac> tedg: err, i dont see the fix for the battery :/
[20:11] <asac> (which is why i was looking at this in the first place)
[20:12] <chrisccoulson> tedg - do you want me to call org.gnome.SessionManager.Logout?
[20:14] <tedg> asac: Well, at least for other folks, that was the removal of patch 19.  Did it not help?
[20:14] <tedg> asac: Hmm, looks like that was in 0ubuntu3.
[20:14] <tedg> chrisccoulson: Yes.  That's the function that GPM calls.
[20:14] <chrisccoulson> it wants an argument, but i'm not sure what to send it
[20:15] <asac> tedg: not sure. let me check whast on that system
[20:15] <tedg> chrisccoulson: Sorry, we're calling "Shutdown".
[20:16] <chrisccoulson> the version in your PPA calls Shutdown
[20:16] <chrisccoulson> the official one still calls Logout I think
[20:17] <tedg> chrisccoulson: It uses XSMP, but basically yes.
[20:17] <chrisccoulson> right, calling Shutdown only causes the session dialog to appear on my own desktop
[20:17] <tedg> chrisccoulson: Did calling that bring up the dialog for all the users?
[20:17] <chrisccoulson> it only brought it up for me, which is expected I think, as i can only connect to my own session bus
[20:18] <asac> tedg: nope. i had ubuntu5
[20:18] <chrisccoulson> the problem seems to be that every user has one instance of GPM and gnome-session. when someone presses the power button, every instance of GPM calls Shutdown in it's own session, regardless of whether it is on the active console or not
[20:18] <asac> tedg: will see if i can still reproduce
[20:19] <tedg> chrisccoulson: Hmm, that's basically what GPM is calling.  So the power button must be getting sent to all of the X servers and sent to all the GPM instances.
[20:19] <chrisccoulson> tedg - that is what I think happens.
[20:19] <asac> tedg: anyway. uploaded
[20:19] <asac> will give you an update if i still see this
[20:19] <asac> (have to wait until a bit of battery is used)
[20:19] <tedg> asac: Thank you, tell me about the icon.
[20:20] <tedg> chrisccoulson: Can you get the verbose output of one of the GPM's?  See where it's getting the power key from?
[20:20] <tedg> chrisccoulson: gnome-power-manager --no-daemon --verbose
[20:21] <chrisccoulson> no problem
[20:21] <chrisccoulson> do i need to kill the existing daemon?
[20:25] <tedg> chrisccoulson: Yes.
[20:26] <chrisccoulson> tedg - http://paste.ubuntu.com/58007/
[20:26] <chrisccoulson> seems it gets it straight from hal
[20:26] <chrisccoulson> does that sound right?
[20:29] <tedg> chrisccoulson: Can be, depends on the hardware, drivers, etc.
[20:29] <james_w> yeah, it comes from hal, via acpi I believe
[20:29] <chrisccoulson> i can monitor the system bus and see the signal from hal when i press the button
[20:29] <tedg> Hmm, so HAL must be sending Power to all the instances of GPM.
[20:29] <james_w> there is one hal, so the event gets sent to all gpms that are listening
[20:30] <chrisccoulson> and that's where the problem is;)
[20:30] <james_w> something needs to decide if it's active or not
[20:30] <james_w> I think it should be g-p-m
[20:30] <tedg> james_w: You're probably right, but I like blaming other software better :)
[20:31] <james_w> heh :-)
[20:31] <chrisccoulson> lol
[20:31] <chrisccoulson> i'm not sure how it would be implemented in gpm
[20:31] <tedg> Hmm, I don't think GPM is linking to console kit...
[20:31] <chrisccoulson> it's not
[20:32] <chrisccoulson> but gpm knows whether it is active or not via calls to policykit i think
[20:32] <tedg> Active is not the issue as much as "on top" -- that's console kit's job.
[20:32] <james_w> policykit uses consolekit, so that could work
[20:33] <james_w> they both live on dbus, which gpm is obviously using
[20:33] <tedg> I don't think policy kit has domain over signals that are sent on teh bus, I believe it's only actions that can be called.
[20:34] <chrisccoulson> anyway, i'll leave you to think about it for a bit. i have to dash off ;)
[20:34] <james_w> tedg: correct, but gpm can query consolekit
[20:35] <tedg> james_w: Yes, I think it'll need to.
[20:35] <james_w> is there a bug open in GNOME about this?
[20:35] <chrisccoulson> i haven't seen one yet
[20:35] <james_w> or do we differ from upstream in gpm or gnome-session about something here?
[20:35] <chrisccoulson> i've not opened one on LP yet either
[20:38] <james_w> tedg: is there a gpm channel?
[20:38] <tedg> james_w: Not that I know of.  You can ping hughsie in #gnome-hackers
[20:44] <tedg> Looks like it should be added to gpm-button.c:emit_button_pressed -- I think blocking all the HAL button events makes sense.
[20:44] <chrisccoulson> it only sees a HAL button event in my case though, I think
[20:48] <asac> tedg: sparc ftb
[20:48] <asac> and hppa, but thats expected
[20:48] <tedg> asac: What's ftb?
[20:49] <asac> fails to build ;)
[20:49] <asac> short for ftbfs ;)
[20:49] <tedg> Failure to Build File system?
[20:49] <tedg> Is there a log?
[20:50] <james_w> the last few uploads at least have
[20:50] <asac> tedg: didnt you get a mail?
[20:50] <asac> tedg: http://launchpadlibrarian.net/18578657/buildlog_ubuntu-intrepid-sparc.gnome-power-manager_2.24.0-0ubuntu6_FAILEDTOBUILD.txt.gz
[20:50] <asac> tedg: should be an easy alignment fix
[20:52] <tedg> Yes, I did get it.
[20:54] <james_w>     *cur = *((int *) prop);
[20:54] <james_w> prop is unsigned char
[20:54] <james_w> prop is unsigned char * sorry
[20:54] <tedg> asac: So, it should be making the variable an "int *" and then casting it to a "char *" when it needs it that way?
[20:54] <james_w> cur is guint *
[21:00] <asac> tedg: what is "variable"?
[21:00] <asac> prop?
[21:01] <tedg> asac: james_w:  Like this?  http://pastebin.ubuntu.com/58021/
[21:01] <tedg> Basically allocate at the larger alignment.
[21:01] <james_w> possibly
[21:02] <asac> tedg: or cast the unreffed char maybe?
[21:03] <asac> err dereffed ;)
[21:03] <asac> e.g.  *cur = (int) ...
[21:04] <tedg> asac: Sure, probably that as it's only a one line change.
[21:04] <asac> tedg: maybe test a small code snippet on a sparc dev machine (if we have one :))
[21:06] <tedg> Yeah, I'm a touch concerned about sign extension in that case.
[21:06] <tedg> I think as long as it gets casted to an unsigned int it should be okay.
[21:10] <tedg> No, I think it has to be the first way.  Because if we're calling a function that wants to use it as an int (presumably doing a cast itself) we would need a pointer initially created with the right alignment.  Or else, there'll be a crash in that function.
[21:11] <mvo> Ng: any news from the proposed fix?
[21:11] <mvo> seb128: I put a new deb up, if you are interessted in the fix
[21:11] <mvo> (or rather in the possible fix)
[21:12] <seb128> mvo: same location?
[21:12] <mvo> yes
[21:13] <seb128> mvo: will install it later and try tomorrow, I'm working on my desktop right now
[21:13] <seb128> mvo: but if you are confident it's not going to break things you should just upload to intrepid ;-)
[21:15] <asac> tedg: if we dont use a function there then it shouldnt matter. bug fixing the alignment upfront certainly has its advantages
[21:15] <mvo> seb128: yeah, I think I should do that
[21:16] <seb128> mvo: I'll work on the laptop tomorrow and confirm if the fix work or now for me so you can wait tomorrow morning too
[21:16] <mvo> seb128: ok, I commit it into bzr now, but wait until tomorrow with the upload
[21:17] <asac> tedg: ok the icon appears to be fixed. thanks
[21:17] <tedg> asac: Great.
[21:17] <tedg> BTW, do ftb's make bugs?  Is there some way that I should mark it in the changelog?
[21:19] <asac> tedg: i would suggest to just clearly say: "fix build-failure on sparc; tiny alignment fix in xxxx"
[21:19] <asac> tedg: no need to open a bug if no bug was opened by a user imo
[21:46] <seb128> mvo: ok, I found a way to trigger the compiz bug I think and your update seems to fix it
[21:46] <seb128> mvo: do you have a non patched version running to try? ;-)
[21:47] <mvo> seb128: I have
[21:47] <mvo> seb128: that sounds like excellent news
[21:47] <mvo> seb128: did I mention that compiz upstream (hello onestone) are rockstars :) ? seriously, they are!
[21:47] <seb128> mvo: assign alt-something shortcuts to different workspaces, let's say to workspace 6 and 7
[21:47] <mvo> done
[21:48] <seb128> mvo: usually I switch between workspace not relaxing alt, because I just look quickly on IRC for example and come back to the other workspace
[21:48] <seb128> mvo: now, do alt-tab once, don't unpress alt and do the key to switch to the other workspace
[21:49] <seb128> it'll go back to the workspace where you were
[21:49] <seb128> and now the bug happens every time you switch workspaces
[21:49] <seb128> ie, try alt-keybinding between those
[21:50] <seb128> not sure if the description is clear
[21:50] <seb128> mvo: let me know if that's not clear
[21:51] <seb128> mvo: when you switch to other workspace after the alt-tab stop touching the keyboard to stay there
[21:51] <mvo> yep, works!
[21:51] <mvo> or rather, triggers the bug
[21:51] <seb128> good ;-)
[21:51] <seb128> and your update fix it there apparently
[21:51] <seb128> mvo: good job!
[21:52]  * mvo dances
[21:52] <mvo> seb128: well, really all the praise needs to go to danny from compiz
[21:52] <seb128> mvo: I guess the way I trigger it is sitting on the alt key, I tend to do that a lot
[21:52] <seb128> I do alt-tab and switch workspaces using it
[21:52] <mvo> seb128: right, I use ctrl to switch so I did no notice
[21:53] <seb128> ctrl might have the same issue if you ctrl-tab to cycle
[21:53] <mvo> righ, I use alt-tab, but ctrl-1,2 :)
[21:53] <seb128> mvo: anyway was is the upstream change supposed to fix exactly? they made it after you described the issue or it was already there?
[21:54] <seb128> s/was/what
[21:54] <mvo> seb128: I described the issue, he looked at the code and shuffled a "+    ws->focusDefault = TRUE;" around
[21:54] <seb128> mvo: looks like they have a clue about their code ;-)
[21:55] <mvo> yes :)
[21:56] <seb128> mvo: thanks for bringing the issue upstream and backport the fix btw ;-)
[21:56] <seb128> mvo: I own you a $beverage
[21:56] <mvo> seb128: cheers, thanks for reporting it!
[21:56] <mvo> it was difficult to describe at first
[21:57] <seb128> it was difficult to figure what action was triggering it too
[21:57] <seb128> I never do this alt-tab and switch workspace which brings you to where you were
[21:58] <seb128> I just did it accidentally now while trying to trigger the bug
[21:58] <seb128> there is likely other similar ways to trigger the bug
[22:01] <mvo> right
[22:02] <seb128> mvo: and feel free to upload the fixed version ;-)
[22:05] <mvo_> hm, the notebook just froze (overheated?) - I guess this is a signal to stop working :)
[22:06] <seb128> mvo_: hey, would be a reasonable hour
[22:06] <asac> tedg: ok it failed on ia64 too
[22:07] <asac> same alignment issue.
[22:07] <asac> tedg: if you have something i should upload let me know
[22:07] <tedg> asac: I'm checking on my PPA right now :)
[22:08] <asac> tedg: how can you check on your PPA if it succeeds on sparc ;) ... i want that ppa too
[22:08] <tedg> asac: Heh, I wish.  I'm half considering trying to get a pbuilder setup in QEMU for Sparc...
[22:08] <asac> heh ;)
[22:09] <seb128> is there any desktop sparc user?
[22:09] <mvo_> seb128, you know me, I like to keep reasonable hours!
[22:09]  * asac is lucky to have a gentoo porter for all crazy archs in the mozillateam channel ;)
[22:09] <seb128> mvo_: I should take example on you ;-)
[22:12]  * mvo_ waves
[22:23] <tedg> seb128: If anyone from Sun is reading this backlog, I COULD be a desktop Sparc user :)