Current desktop detection / app access - take 2.

[Michael Meeks]

On Tue, 2004-05-18 at 11:40 +0100, Thomas Leonard wrote:
> And for ROX, we normally *only* do user installs (the only part of the
> install that requires writing to a non-user directory is adding ROX to the
> list of sessions on the login screen, and that's optional).

	Ok - so we should clearly walk the path - which ( I guess ) makes the
whole process somewhat expensive, seek-wise, but the path is there to be
walked I guess :-)

> I dislike this desktop detection idea in general. We had a lot of problems
> in the past with, eg, sawfish only supporting the extended window manager
> hints if it detected it was running under GNOME, which meant that programs
> that worked correctly under GNOME failed under ROX for no good reason.

	Interesting; that sounds like some pretty obvious brokenness in
sawfish :-) clearly the specification attempts to ensure that it is
obvious that there are a plurality of desktop environments based on
different underlying technology. Are there any other sub-dividable
pieces beyond toolkit/vfs that you can think of that we should be
looking at ? WM stuff (AFAIK) should be done by standard methods of
interogating the WM I guess.

	Regards,

		Michael.

-- 
 michael at ximian.com  <><, Pseudo Engineer, itinerant idiot