[hotkey-setup] Extra keyboard buttons

Bastien Nocera hadess at hadess.net
Sun Apr 29 15:26:16 PDT 2007


On Sun, 2007-04-29 at 22:39 +0100, Richard Hughes wrote:
> On Sun, 2007-04-29 at 22:29 +0100, Paul Sladen wrote:
> > 
> > These files are sourced every time an event occurs.  Doing the
> > translation
> > once per boot (at FDI load time) is unlikely to be slower.  The
> > crucial
> > difference between numbers and strings is that the human-readable
> > constants
> > are *maintainable*. 
> 
> But we're not just doing variable substitution - we've got to do hash
> matching of each KEY_xxx keyword, i.e. parse several strings at startup
> when we could just add a note in the fdi file <!-- KEY_EMAIL --> and
> have no run-time overhead (think embedded machines)... I'm trying to
> make this as fast as possible, and I don't think there'll be significant
> changes in existing keymaps.
> 
> IMO, A simple howto can show the user how to go from a warning in dmesg
> to working buttons (and a patch to hal-devel). Why do you think the
> constants are more maintainable than a simple constant with a suitable
> comment? Thanks for your help with this BTW.

Because there's less chance of a mismatch between comment and "code".

I've been following this conversation, and I would still want to know
from kernel people, especially people familiar with the input layer to
let us know whether this is the best course of action, and whether we'd
be able to fix this more cleanly, and without hacks, in the kernel.

-- 
Bastien Nocera <hadess at hadess.net> 



More information about the hal mailing list