[Openicc] xrandr output properties / icc profiles

[Tomas Carnecky]

Graeme Gill wrote:
> Tomas Carnecky wrote:
>> Graeme Gill wrote:
> 
>>> can figure out what to use. It's really bad to claim XRandR V1.2
>>> though, and not actually implement it (nVidia and ATI).
>> Well, XRRQueryVersion() returns the server-side version and there is no 
>> way to find out whether the drivers support it or not. 
> 
> Why not ? The server should call down and return what the driver supports,
> just like any other extension.

should != does

See the server-side implementation of XRRQueryVersion():
http://gitweb.freedesktop.org/?p=xorg/xserver.git;a=blob;h=5525427f667dcfb0e5b8b3cb8a2a99c4623c6df8;hb=55644d2f1cddf1777587f213ed581b6087d9fdad;f=randr/rrdispatch.c#l37

This is one of the shortcomings of randr and the issue was brought up on 
the xorg mailinglist not too long ago:
http://lists.freedesktop.org/archives/xorg/2008-July/037090.html
(Hey, it was even you :) )

All of the extensions I've looked at report a fixed server side version 
number. How the actual extension is then implemented, whether it uses 
the driver/hardware or falls back to software, is then decided at 
runtime. So there never really was the need to ask the driver whether it 
supports the features or not.
The version doesn't say which features the server supports, but which 
protocol version it understands. So it's more along 'Here is the server, 
  I understand XRandR 1.2 requests', but whether 1.2 functions are 
implemented in a reasonable way isn't specified.

tom