> I'm interested in helping out with a MacOS X version of the VNC server.

OSXVnc:

http://homepage.mac.com/muhaque/

v0.7 includes Tight and Zlib encodings.


> I believe a major stumbling point is that the PPostEvent() and
> CursorDeviceMoveTo() functions are not available in MacOS X (Carbon).
> The exiting MacOS Classic version of VNC Server uses these functions to
> input the remote mouse clicks, keystrokes etc into the local system.
> Anyone know how to do this on MacOS X?

OSX has completely different ways to accomplish this - see the OSXVnc
code for examples.


I do find that OSXVnc is not as capable as ChromiVNC across a slower
connection - not that surprising, since the OSX interface is *much* more
'interesting' (Hmmm...?)

But there are some tricks in ChromiVNC that would be worth including into
OSXVnc (and the servers for other platforms too! -Those who work on 
these,
please take note...)

In particular, the "on-the-fly encoding comparison" might be worth 
looking
into (believe it or not, there are times when (Co)RRE really is a better
encoder for a particular rect, and using anything else is a waste...)

Anyone got a bit of time to take the relevant ChromiVNC code and throw it
into the OSXVnc code? (Shouldn't be more than a couple of days work to
translate from C++ to C and add the relevant calls to the main body...?)

It would probably also benefit from some sort of update rect buffering
and merging/splitting/eliminating, as found in my vncPatches68k.
I'll have to see if I can look into that one...

Finally, I'd like to see a better algorithm for merging/splitting of
buffered update rects based on content, so that the on-the-fly encoding
decision can be used to even greater effect.

(Ideally I would also like to see an extension to the standard (Co)RRE
and Hextile encodings to allow for a "transparent" colour. This would
mean that certain portions of an update rect, or particular tiles of a
Hextile rect, say, can be "left alone" [and hence do not need to be sent]
because their content has not changed - it would then also be possible to
design much better algorithms for update rect merging.)

Bye!

Adrian
_______________________________________________
VNC-List mailing list
[EMAIL PROTECTED]
http://www.realvnc.com/mailman/listinfo/vnc-list

Reply via email to