Adi Oanca <adioanca@xxxxxxxxxxxxx>: > > Mikael Jansson (mailing lists) wrote: > > Hello, > > > > I'm sorry for not contributing much of anything, but I have a > > question: > > In what way are you providing for a framework so that remote > > controlling through for example the VNC protocol can be made as > > fast as > > painlessly as possible? Today that involves doing a memcpy() from > > the > > pointer received from a BWindowScreen at a high rate, but this > > scheme > > could be vastly improved if the background buffer used for drawing > > could be shared and apps could subscribe to update messages, or > > through > > other means. > > Don't know... we'll think about that. > > Regarding the backbuffer, it appears it can be avoided... > Okay, sure, as long as it'll be possible to get at the contents at the same rate as they're being updated by the app_server. In Windows, the VNC servers are very fast. What does Windows do that makes it that much better for remote desktop-style apps? Maybe that could be a source of inspiration. I suppose looking at the sources for any Windows VNC server would suffice. -- Mikael Jansson http://mikael.jansson.be