Re: [gtk-vnc-devel] Supporting VMware VNC extensions in gtk-vnc

[Anthony Liguori <anthony codemonkey ws>]

Philip Langdale wrote:
> > Hi Ramesh,
> >
> > I hope this mailing list is still live.  I noticed in the most recent 
> > RFB spec that the encoding range used by VMware is now reserved.  I'd 
> > like to add support for these extensions to gtk-vnc.  I'm wondering if 
> > there's more definitive documentation than [1].  Specifically, for the 
> > new client/server messages, does the server support a negotiation 
> > mechanism yet as recommended in the RFB spec?
> >
> > [1] http://wiki.multimedia.cx/index.php?title=VMware_Video
> >     
>
> Hi Anthony,
>   

Hey Phil,

> Ramesh is on vacation right now but is fairly responsive. The wiki page is
> the most definitive documentation out there - and probably the most
> definitive
> documentation full stop, apart from the source of our implementation. If you
> have any questions about it, please ask and we'll try to get them answered.
>   

Yeah, I've implemented what's on the wiki page.  What's missing from it 
is the definitions of the two additional client message types which I 
believe are to send raw scancodes and to send relative pointer events.

> There is a negotiation process but you'll have to wait for Ramesh to explain
> how it works - I actually thought the mechanism came from the RFB spec but
> your question suggests that said spec doesn't actually define it.
>   

There is a negotiation mechanism for psuedo-encodings (the SetEncodings 
message) but not necessarily for using new client messages.  I was 
thinking though that for the relative pointer events, it may be 
sufficient to just wait for the first cursor state message before 
sending any relative pointer events.

I'm not so sure though what you wait for before using the raw scancode 
interface.  It looks like a keyboard info message is usually sent early 
on so perhaps that could be used.

Regards,

nNthony Liguori

> --phil
>