Bug#339785: gnome-keybinding-properties: environment not set for apps launched through keybindings

[Sjoerd Simons]

On Sun, Aug 19, 2007 at 08:48:50PM +0200, Sven Arvidsson wrote:
> On Tue, 2007-05-29 at 18:13 +0200, Josselin Mouette wrote:
> > Le lundi 28 mai 2007 à 12:30 +0200, Sjoerd Simons a écrit :
> > > > That's what I supposed, but is there any way to have things work
> > > > correctly?
> > > > Further more, I see seahorse-agent is now started by dbus:
> > > >   /usr/bin/dbus-launch --exit-with-session /usr/bin/seahorse-agent --execute x-session-manager
> > > > 
> > > > I guess this can help?
> > > 
> > > Good point. If it was the other way around stuff would just work.. So either
> > > seahorses xsession should be later  or dbus-launch's should be earlier.
> > > 
> > > I guess moving dbus-launch from 75  to 60 and seahorse from 60 to  65 would be
> > > the nicest solution..
> > 
> > I'm not sure this is going to work, because seahorse itself is using
> > dbus, so it needs to be started after dbus.
> > 
> > It's kind of an egg-and-chicken issue.
> 
> Hi,
> 
> I haven't looked at this in detail, so I could be wrong, but isn't this
> related to this bug recently fixed in Ubuntu?
> https://bugs.launchpad.net/ubuntu/+source/control-center/+bug/62163
> 
> and the upstream bug; http://bugzilla.gnome.org/show_bug.cgi?id=360475

Their ``fix'' is to no longer have the xession.d script, but let the dbus
session bus be started by gnome-session. Which is just plain wrong (It's fixing
the symptomps not the issues). One of the many reasons for it being wrong is
that we can't assume the users X session will contain gnome-session.

Basically assuming magic environment variables are available
when started by the dbus session bus is just plain wrong. Starting the bus
daemon later in the session just works around this...

  Sjoerd
-- 
Doubt isn't the opposite of faith; it is an element of faith.
		-- Paul Tillich, German theologian.