2008/03/27/#ubuntu-desktop.txt

[02:52] <ember> asac there?
=== crd1b is now known as crdlb
[07:12] <dholbach> good morning
[07:23] <crevette> hello
[07:37] <crevette> hey seb128
[07:37] <seb128> lut crevette
[07:37] <crevette> seb128 https://bugs.launchpad.net/ubuntu/+source/empathy/+bug/207508 ; I would tend to lose it as Invalid and say to use the session manager
[07:37] <ubotu> Launchpad bug 207508 in empathy "no way to stop empathy startup with session" [Undecided,New]
[07:38] <seb128> crevette: right
[07:38] <crevette> =okay
[07:38] <crevette> thanks
[09:07] <huats> morning evryone
[09:08] <seb128_> hey huats
[09:08] <lmanul> Morning everyone!
[09:15] <huats> lut seb128
[09:46] <seb128_> ok
[09:46] <seb128_> I'm going to do a totem-plugins-extra
[09:46] <seb128_> which will not be on the cd
[09:47] <seb128_> and contain the mythtv, lirc and gromit plugins
[09:47] <seb128_> anybody has an objection to the change?
[10:23] <MacSlow> kwwii, the new tooltip-icons look good sofar
[10:25] <seb128_> hey MacSlow
[10:25] <MacSlow> seb128_, salut
[10:25] <pochu> hi all
[10:26] <crevette> guen morgen
[10:26] <seb128_> MacSlow: do you intend to look at all the desktop effects tab bugs assigned to you for months before hardy?
[10:26] <crevette> guten
[10:27] <MacSlow> seb128_, atm I'm putting the last touches on profiles and shortcuts... keybuk wants me to work on the face-browser right after that and skip bug-squashing this cycle
[10:28] <MacSlow> seb128_, if you want to change my schedule persuade him... not me :)
[10:28] <kwwii> MacSlow: cool :-)
[10:41]  * kwwii heads out for a longer lunch
=== asac_ is now known as asac
=== Hobbsee` is now known as Hobbsee
[11:50] <lapo> ciao
=== Baba-B00ie is now known as MBR
=== MBR is now known as Baba-Booie
=== huats is now known as huats_
[14:39] <vuntz> seb128: my preferred packager
[14:39] <seb128> hey vuntz ;-)
[14:39] <vuntz> seb128: question from a packager point of view
[14:39] <vuntz> (for a panel patch)
[14:39] <vuntz> for the new gnome-session, we need to install gnome-panel.desktop in /usr/share/applications and in another directory
[14:40] <vuntz> would packagers be happy with a symlink?
[14:40] <vuntz> or is it better to have the same file twice?
[14:40] <seb128> what other directory?
[14:41] <vuntz> /usr/share/gnome-session/default or something like this
[14:41] <seb128> I guess a symlink is alright in any case
[14:41] <vuntz> ok
[14:41] <vuntz> I thought it could cause some problems, but if it's fine, good :-)
[14:42] <seb128> vuntz: btw I'm not sure I understood what will be responsive to store applications coordinates, workspaces, etc
[14:42] <seb128> vuntz: symlinks are usually no issue
[14:42] <vuntz> it should be the apps
[14:42] <vuntz> (via gtk+)
[14:42] <seb128> they can go cross mounts, etc
[14:42] <seb128> vuntz: any reason it can't be done on the wm side?
[14:43] <vuntz> seb128: because it's a horrible hack? :-)
[14:43] <vuntz> what happens if you change the wm?
[14:43] <seb128> I don't like the idea to force every application to do the same things
[14:44] <seb128> well, yours mean that you will never have non gtk applications behave correctly
[14:44] <seb128> users don't change wm
[14:44] <seb128> and if they do they can redo placement
[14:44] <vuntz> seb128: fd.o spec, etc.
[14:44] <vuntz> there's clearly work to do to have this correctly work, though
[14:45] <seb128> you will never have 10 years old applications to follow fd.org spec
=== dholbach_ is now known as dholbach
[14:45] <seb128> it would be so much easier to let the wm do the work
[14:45] <seb128> change one thing
[14:45] <seb128> rather than patching every cycle application in the world to use the same api
[14:46] <seb128> s/cycle/single
[14:46] <seb128> well, if you think that's the way to go
[14:46] <seb128> but I've some doubt on the fact that we will have something working correctly any time soon if every application need to be modified to work
[14:47] <vuntz> seb128: the wm way can only be hacky, so it's not good in the long-term
[14:48] <seb128> yours is not good either
[14:48] <seb128> the only solid way is to never shutdown
[14:48] <seb128> always use suspend to ram or disk and reload from that
[14:49] <seb128> and don't crash
[14:51] <seb128> I just don't think that having complex logic in applications to store the placement will scale
[14:51] <seb128> it means complicate code, bugs, extra work, etc
[14:53] <vuntz> seb128: why?
[14:54] <seb128> why what?
[14:54] <vuntz> why would it be more complex?
[14:54]  * vuntz thinks seb128 should be able to read his mind instead ;-)
[14:54] <seb128> because you have to use this api
[14:54] <seb128> which is higher complexity than do nothing and let the wm do the work
[14:54] <vuntz> seb128: no. It would just work: gtk_application_new("myname") and that's all
[14:55] <seb128> that should be part of gtk_init() then ;-)
[14:55] <seb128> well, that means you transform gtk to a window manager
[14:55] <seb128> it has to know on what screen, viewport, positions, etc the dialog is now
[14:56] <seb128> where the wm already has all those informations
[15:42] <pochu> asac: thanks for the liferea fix :)
[15:43] <asac> np :)
[15:44] <asac> but see how botty i am ... i even forgot which function call i added 10 seconds after editing the xul patch :) ... its GTKEmbedGlueStartupInternal() not GREStartGlueInternal
[15:46] <asac> s/botty/buggy/
[16:44] <pitti> seb128: doing MacSlow's libwnck sponsoring now
[16:48] <seb128> pitti: danke
[16:49] <pitti> MacSlow: hm, I don't notice any difference (new libwnck), but at least my panel still works :) where are the missing icons?
[16:50] <seb128> pitti: workspace switcher tooltip I guess
[16:50] <pitti> looks exactly the same to me
[16:50] <seb128> pitti: if you use compiz
[16:51] <pitti> ah
[16:51] <seb128> do you use compiz?
[16:51] <pitti> desktop == metacity
[16:51] <pitti> on the laptop, yes, but it's off ATM
[16:51] <seb128> under compiz you have extra actions there
[16:51] <seb128> the scale and expo ones
[16:51] <pitti> actually I switched my desktop back to metacity because compiz' session management is broken
[16:52] <pitti> but now that metacity's session management regressed to a comparable dysfunctionality, it doesn't actually matter any more...
[16:52]  * pitti sighs
[16:52] <nxvl> seb128: i have notice some problem when i don't use compiz
[16:52] <pitti> (that's not what I meant with "compiz' session management shuold be as good as metacity's")
[16:53] <nxvl> seb128: for example if i click on a link on evolution, firefox goes to the desktop where evolution is, instead of open it in the one it is opened
[16:53] <seb128> nxvl: that's a known issue
[16:55] <nxvl> seb128: and that's a bug on firefox or in metacity/compiz?
[16:56] <seb128> not sure
[16:56] <seb128> I think the firefox guys blame the wm
[16:56] <walters> http://bugzilla.gnome.org/show_bug.cgi?id=482354
[16:56] <ubotu> Gnome bug 482354 in general "gtk_window_present() causes full applications to move workspaces" [Normal,New]
[16:56] <walters> i have the fedora metacity patched
[16:58] <seb128> elijah wrote that the hack is not a nice one
[16:59] <walters> virtual desktops themselves are a gigantic hack
[16:59] <seb128> did you change it to special case firefox?
[16:59] <walters> no
[16:59] <walters> they didn't suggest a way to do that
[16:59] <walters> i suppose i could look at the kwin code
[16:59] <seb128> I'm just wondering why it's going to break
[16:59] <seb128> according to elijah it's going to break things
[16:59] <walters> i wouldn't say it "breaks" anything
[17:00] <seb128> which is not really useful comment
[17:00] <walters> it changes the behavior of some cases
[17:09] <MacSlow> pitti, do you have compiz running and enabled expo- and scale-plugins?
[17:09] <MacSlow> pitti, if not you'll not see anything :)
[17:10] <pitti> right
[17:10] <MacSlow> pitti, have a look at http://people.ubuntu.com/~mmueller/tooltips.png
[17:11] <MacSlow> pitti, there used to be zoom too... but since that is usually only triggered via a mouse-binding (and those are not exposable via gtk-accelerators) I skipped it.
[17:17] <Amaranth> seb128: that's not nice to do (that patch)
[17:17] <seb128> Amaranth: what patch?
[17:17] <Amaranth> the metacity one
[17:17] <seb128> ah
[17:17] <seb128> why?
[17:18] <Amaranth> if you read the whole bug you see there are several distinct use cases here
[17:18] <seb128> MacSlow, (pitti since you do sponsoring): did you talk to ted btw? he's working on moving those items to the context menu rather than the tooltip as decided some time ago
[17:18] <Amaranth> sometimes you want it to move, sometimes you want to move to it, sometimes you want it to flash in the taskbar
[17:19] <Amaranth> the spec needs to be updated to include a hint windows can set
[17:20] <seb128> Amaranth: right, I think everybody agrees on that
[17:21] <seb128> Amaranth: but meanwhile firefox looks broken and we want to get that fixed this cycle
[17:21] <Amaranth> i will have no part in 'fixing' this for compiz
[17:22] <Amaranth> you're breaking my uses to fix firefox
[17:23] <seb128> Amaranth: well, the upstream bug mentions special casing firefox
[18:03] <kagou> hi
[20:30] <calc> anyone around here that knows about mime types?
[20:30] <calc> i need help asap
[20:31] <calc> i have to update the mime types in OOo but i don't know what to run against example files to generate the official mimetype name to use in /usr/lib/mime and /usr/share/applications/foo.desktop MimeType=
[20:41] <Amaranth> calc: gvfs-info foo.odt | grep content
[20:44] <calc> also found the mimetype program :)
[20:44] <calc> Amaranth: thanks