RANDR 1.2 vs libxinerama?

[Andy Ritger]

On Mon, 16 Apr 2007, Matthias Hopf wrote:

> On Apr 13, 07 15:45:48 -0300, Keith Packard wrote:
>> Xinerama is deprecated only in that it doesn't support everything that
>> RandR 1.2 can (mostly it only supports a single screen). However, the
>> RandR 1.2 implementation in the server also provides an implementation
>> of the Xinerama extension for compatibility with existing applications.
>>
>> This RandR-based Xinerama implementation works fine for me.
>>
>>> screen 0 [ 0/0 - 1024/768 ]
>>> screen 1 [ 0/0 - 1400/1050 ]
>>
>> Xinerama screens are equivalent to RandR CRTCs. Xinerama should have
>> used a term other than 'screen' which is already defined for core X.
>
> But behavior obviously changed with RandR 1.2, because both KDE and
> Gnome used libXinerama for multiscreen detection before.
>
> And in that particular case, only a single monitor was connected, so the
> "screen"s in libXinerama now represent CRTCs and not active CRTCs.
>
>>> now, on nvidia twinview libxinerama seems to be working allright, and xrandr
>>> returns nonsense:
>>
>> I don't know what nVidia's driver is doing these days.
>
> Sure :)
> But maybe Andy can comment on that.

NVIDIA's X driver does not yet provide RandR 1.2 support, so I think the
behavior you are seeing is a result of NVIDIA's TwinView implementation
reporting the "bounding box" of its MetaModes to the core X server.

My hope is that in the worst case, X clients would just see the MetaMode
as one big RandR 1.2 output/CRTC.

I hope that helps,
- Andy Ritger


> Thanks
>
> Matthias
>
> -- 
> Matthias Hopf <mhopf at suse.de>      __        __   __
> Maxfeldstr. 5 / 90409 Nuernberg   (_   | |  (_   |__          mat at mshopf.de
> Phone +49-911-74053-715           __)  |_|  __)  |__  R & D   www.mshopf.de
> _______________________________________________
> xorg mailing list
> xorg at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/xorg
>