Keysym event in the text protocol

[Bill Spitzak]

On 07/25/2014 11:59 AM, Daniel Stone wrote:
> Since you're just repeating yourself without taking on anything that's
> been said, so will I: that won't work.

This I am finding hard to believe.

My proposal is that the compositor maintain a xkb_state, and keystroke 
events have the result of xkb_state_key_get_syms(), 
xkb_state_key_get_consumed_mods(), etc added to them. Note the keycode 
is still there in the event.

For RDP this would be run on the actual machine that has the keyboard 
hardware and thus has access to the keyboard layout information.

Yes I know there are some details about modifiers and repeats, similar 
to the problems with the current scheme of syncing xkb_state between 
clients. But I hope that describes the basic idea. I would love to know 
why this "won't work" because I am just not seeing it.