Re: Panel behaviour under Xinerama

[Havoc Pennington <hp redhat com>]

On Thu, May 08, 2003 at 03:35:19PM +0100, Mark McLoughlin wrote:
> Hey,
> 
> On Thu, 2003-05-08 at 15:08, Jeff Waugh wrote:
> 
> > Perhaps each panel ought to have "span displays" prefernce, available only
> > when Xinerama is present? It would probably be better to figure out a "just
> > works" behaviour...
> 
> 	It struck me that something like Metacity's maximising behaviour on
> Xinerama would be appropriate (don't expand, expand to just this
> monitor, expand to the full screen) i.e. its a tri-state thing when
> Xinerama is available ...
> 

metacity does not do that. ;-) 

I actually had an argument about this with one of the X people (yeah,
I know, me in an argument surprises everyone). Xinerama gets
overloaded; I would say most people want to use it for multihead where
you can move windows between screens. This is when you have just a
couple of monitors. In this case, you want to maximize to each screen,
center dialogs on one screen, etc.

Sometimes you use Xinerama to convert a 4x4 array of monitors into one
huge monitor.

My argument was that if you convert a 4x4 array into one huge monitor,
the X server should actually report that Xinerama is disabled. Because
there's no reason any app needs to know about that. Xinerama should
only report monitors that people want to be treated as separate
monitors.

But in any case, you have to decide what your Xinerama setup
represents; the desktop can't just waffle between these two views of
the matter.

Havoc