Re: [gtk-vnc-devel] PATCH: Implement cursor encoding

[Anthony Liguori <anthony codemonkey ws>]

On Thu, 2007-08-30 at 22:50 +0100, Daniel P. Berrange wrote:
> On Thu, Aug 30, 2007 at 05:09:37PM -0300, Jonh Wendell wrote:
> > Em Qui, 2007-08-30 às 20:50 +0100, Daniel P. Berrange escreveu:
> > > The OSX-VNC server violates the RFB spec by never drawing a remote cursor
> > > itself. So if you connect to it with GTK-VNC you never see a cursor on the
> > > remote framebuffer. The server stupidly assumes that all clients will 
> > > implement the rich cursor extension, even if the client doesn't advertise
> > > this fact. I can't find anyway around this, so I implemented the rich
> > > cursor extension
> > > 
> > > Dan.
> > 
> > Well, the patch did not work to me.
> > I connect into a UltraVNC server on windows xp and get no cursor after
> > applying this patch.
> 
> The VNC cursor extensions are fun in that there are two different ones.
> One call Rich Cursor, the other called XCursor. I implemented the former,
> and I'd guess UltraVNC wants the latter. Sucks. Guess I'll have to set
> up a virtual machine runnings Windows & UltraVNC to test.
> 
> > BTW, Why did not you commit yet your last patch?
> 
> I was waiting to see if anthony had any comments on it.

Are we talking about the tunneling API?  I'm at KVM forum right now so I
don't think I'll get to look at it before next week.

The fundamental question in my mind about this patch is why not just
exec() and pass an fd to open_fd()?

Regards,

Anthony Liguori

> Dan