VNC patch for Xserver 1.3

[Colin Guthrie]

Alan Hourihane wrote:
> On Tue, 2007-10-23 at 16:27 +0200, Stefan Dirsch wrote:
>> On Tue, Oct 23, 2007 at 01:42:08PM +0300, Daniel Stone wrote:
>>> On Tue, Oct 23, 2007 at 12:34:34PM +0200, Stefan Dirsch wrote:
>>>> On Mon, Oct 22, 2007 at 10:12:27PM +0200, Stefan Dirsch wrote:
>>>>> AFAICS the pInfo structure is filled correctly in xf86rfbMouseInit(),
>>>>> but the content gets lost again in InitInput() before calling
>>>>> xf86ActivateDevice(). :-(
>>>> Obviously type_name/always_core_feedback, conf_idev/drv, and
>>>> options/history_size are swapped on x86_64.
>>> Sure it's being built with _XSERVER64?  All external modules require
>>> xorg-server.h.
>> Meanwhile I figured out that it is *not* set. A patch, which fixes
>> this issue for me, is attached. Some background:
>>
>> * fixes segfault on 64bit during Xserver start; make sure to define
>>   _XSERVER64 by having HAVE_DIX_CONFIG_H defined and therefore
>>   including dix-config.h, so Atom is CARD32 instead of unsigned long
>>   before and no longer messes up the pInfo structure in
>>   xf86rfbMouseInit/xf86rfbKeybInit
> 
> Thanks Stefan.
> 
> Committed.

I should point out that there is a current issue that I've not had time
to look into just yet:

"There's one minor problem: it's including xorg/xf86Cursor.h from
/usr/include instead of using the file from the compilation tree,
thus I can't pass it through our build-system.

(remove x11-xserver-devel and try to compile it)"

As I say I've not had time to fix this, but obviously it needs to be
fixed otherwise there is a bit of a chicken and egg scenario :)

If someone spots it please feel free to save me the trouble of creating
a patch ;) (I'm kinda rushed right now)

Col