[RFC PATCH libinput 1/2] Add a string-based input API

[Bill Spitzak]

Shouldn't this use the same api as the input methods, since it is in 
fact an input method?

On 12/07/2014 10:05 PM, Peter Hutterer wrote:
> Some devices don't rely on keycodes + xkb layout but rather send a specific
> keysym (or multiple) in response to physical button presses. This is the case
> for chorded keyboards for example.
>
> This adds a new type of key event that provides UTF8 strings. The
> press/release pair applies to this type as well, the exact behavior of when
> the release event is sent is implementation dependent.
>
> Signed-off-by: Peter Hutterer <peter.hutterer at who-t.net>
> Cc: Samuel Thibault <samuel.thibault at ens-lyon.org>
> ---
> Not intending to merge this yet, this is supposed to spark some discussion.
> We don't have real support for this in the kernel and I don't have access to
> any device that really generates keysysms. PoC is in Patch 2, but in the
> real world the device will likely be hooked up through some other means than
> evdev.
>
> Until we have reliable testing and know what the details need to be for
> device support, this remains an RFC (but now archived on the list instead of
> just locally).