> > ... (register 0 would report keycodes, register 2 would hold modifier
> > key states and LED states - here's the long sought after capslock
> bit!).
>
> <drool>Aahhhh.... CapsLock bit... Aaaahhh...</drool>
>
> Sweet.
>
> Now, CapsLock can become a truly viable Control key!
I didn't say that. I didn't even mean to imply that.
> That will also mean that the code to fix capslock will be robust, and
> therefore viable for inclusing in the kernel... Which means I no longer
> have to apply my silly little patch each time I rsync ben's kernel!
The code would probably become real ugly - like, on each keypress while
capslok is down, read keyboard register 2 and update the capslock state
before processing the keycode that actually was received. That's from
inside the keyboard input handler. Yikes. Only polling for capslock state
changes would be uglier.
Be prepared to maintain your little patch a bit longer ...
Michael
