[PATCH weston v2 1/3] toytoolkit: Allow operation without a keymap

[Bill Spitzak]

If clients use the input method for all the text input, exactly how much 
is left of xkb keyboards? It looks like all this compiling is so that 
dead keys and all the shift states work for producing different 
characters. That can all be done by the input method and thus only the 
input method has to compile xkb, if anybody.

Once all text comes from the input method, it seems to me that a useful 
api is going to be a lot more like the proposed gamepad api. I'm only 
interested in what physical keys are held down or transitioning. In X 
software I have written I actually have to go through some pain to 
remove the effects NumLock and Shift have on the keypad keys.

Also RDP delivers events that look much more like keysyms than scan codes.

Rob Bradford wrote:
> From: Matt Roper <matthew.d.roper at intel.com>
> 
> In preparation for upcoming changes, we want to make sure that apps
> written with the toy toolkit continue to operate properly if no XKB
> keymap is received.  If there's no XKB keymap, then we shouldn't
> try to figure out keyboard modifier states (since we probably don't
> even have equivalents of PC-style modifiers).