> You said it yourself, unicode doesn't have representation for all keyboard keys. > As you have seen in my previous reply, I am looking at the problem from an > "input-method" level. And I don't know if it's a good idea to depend on X11/gdk > keysyms in general for Spice. VNC use that for 20 years now, so I guess the idea is not so bad. > > Else I need to listen on two different input channels, and mix them somehow. > > That does > > not make any sense to me. > > Indeed, that might be a problem since input channel and vdagent/main channel > are not synchronized. > > Could we have the utf8 input message on the input channel, and multiplex it with > the rest of vdagent messages, so at least the remote/guest receives event in > order (assuming input channel waits for vdagent before processing key event)? Sorry, I do not really understand that. I already send a patch extending the input channel, but nobody answered. That is why I rewrote the whole thing to use vd_agent instead. IMHO, extending the input channel keyboard message format would be the right thing to do. Simply send: - scancode - keysym - modifier key state inside a single message. _______________________________________________ Spice-devel mailing list Spice-devel@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/spice-devel