XI2 = moving target?

[Matthew Garrett]

On Sun, Jul 26, 2009 at 09:10:52AM -0400, Jim Gettys wrote:

> Embedded systems, however, are just as likely to expose raw scancodes; 
> their evdev driver will be having to stand on their head to fit the PC 
> model, which we may or may not consider to be a 'good thing'.

It's the kernel's responsibility to provide appropriate abstraction for 
the hardware. One of the consequences of this is that Linux input 
devices are always expected to generate keycodes that correspond to 
those found in linux/input.h. X keymappings are then expected to be 
limited to user preferences on top of the existing abstraction, such as 
qwerty/azerty shifts. There's nothing inherent to the PC model here.

Having hardware-specific keymaps in X made sense when X was responsible 
for most of the abstraction of the keyboard hardware. The historical 
disconnect between kernel and X keymaps has finally been fixed and 
that's a *good* thing, but it does mean that we need to revisit the 255 
keycode issue.

-- 
Matthew Garrett | mjg59 at srcf.ucam.org