I would use UTF-8 strings to identify the keys. This would make it
relatively obvious how to assign new ids to new keys, and make it
possible for a program to report an intelligible error message when it
does not recognize a key, and allow key assignment APIs to work with
unknown new keys. I think also this would assist greatly in allowing an
application to migrate between screens on different devices with
different keyboards.
This is different than the "text" that the key produces. The "text" is
computed by the input method, which honestly I'm not sure where it
resides (I may try to write something about that next).
Key names might be "T", "5%", "Enter", "F0", "Keypad Enter", "Left
Shift", etc. If there is concern about the lookup time then perhaps a
hash method can be defined and this hash integer of the name is included
in the event, but I really don't see that as being necessary.
This id is much more like an X keysym than a keycode. But really this is
necessary if an appliation is to be able to migrate to different
keyboards. The X server atop Wayland should translate these id's to a
fake set of "keycodes" and provide the X app a keymapping table that
turns those keycodes back into the closest matching keysyms.
_______________________________________________
wayland-devel mailing list
wayland-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/wayland-devel
