RE: VNC plug-in architecture (wishlist)

Quentin Stafford-Fraser ()
Thu, 25 Jun 1998 11:18:21 +0100

This sounds like a nice model, but it would only allow you a single
connection to the server at any one time. It's also a bit harder to do
things like this on Win32, where sockets and file descriptors are very
different things.

We agree, though, that the authentication is a separate issue from the VNC
protocol itself, and there should, perhaps, be a separate well-defined API
to a "VNC transport" layer, whose negotiation would happen before the real
VNC session started.

Quentin
----
Dr Quentin Stafford-Fraser
The Olivetti & Oracle Research Lab

>
> One of the easiest things is the following "API":
>
> Let us get rid of any authentication in the VNC protocol and use only
> predefined (well, passed as command line arguments? :)
> pair of filedescriprtors, like stdion/stdout,
> existing during the whole lifetime of the server process.
>
> This way you can set any authentication daemon on top of the server
> process, that communicates with the server through a socket and
> accepts/rejects outside connections requests based on any suitable
> protocol(s) over any suitable channels.
>
> Regards,
> --
> Ivan Popov <>
> Systemman, Driftavdelningen, Matematiska institutionen, Chalmers TH
>
>
>

This archive was generated by on Wed Feb 03 1999 - 15:34:53 GMT