Re: Fullscreen and Xinerama

[Havoc Pennington <hp redhat com>]

On Fri, 2005-06-17 at 16:35 +0200, Lubos Lunak wrote:
>  I wonder if anybody has any good idea about this: Xine apparently supports 
> command-line switches -f and -F, with -f meaning showing in fullscreen, -F 
> meaning showing across all Xinerama screens (or in whichever way one 
> configures it). For which we clearly don't have any support in the spec, 
> since _NET_WM_STATE_FULLSCREEN simply means whichever fullscreen the WM 
> things it should be.
> 
>  Does any of your WMs have any special code for this? I have here a patch for 
> KWin that tries to guess on which Xinerama screens the fullscreen should be, 
> depending on the window geometry, but it's hackish and I don't like it. 
> _NET_WM_STATE_FULLSCREEN is not supposed to have anything to do with the 
> normal window geometry, and it also would break with Xinerama-unaware apps.
> 

I'm pretty sure this has been discussed in the past... metacity right
now I think will fullscreen to a single monitor in the xinerama case.

Generally there are two cases with xinerama: 
 1) each monitor separate, but unlike multihead you can move 
    windows between monitors
 2) the real goal is one big monitor made up of multiple 
    monitors (e.g. video wall)

My argument in the past has been that "xinerama awareness" is #1, and
the ideal fix for #2 is that the X server does not report a xinerama
extension (or reports one big monitor, even though there are really
more)

I'm not sure apps should ever fullscreen spanning monitors in case #1,
really

Havoc