On Sat, May 03, 2003 at 10:56:54 +0200, Martin Husemann wrote:

> > We don't add any additional keyboard maps into kernel.
> 
> While for sure all maps need to be available as userland maps, I
> don't like the "no additional ones inside the kernel" dogma.  They
> are all optional,

Nit pick: they are not, b/c the keydescs are always compiled in.
I think we have a whopping 1.5KBs of them already. ;)


> I'm pretty glad my keyboard maps are there already, and for sure I
> use them in every single (non-headless) kernel I build. So when I'm
> in ddb I don't have to type the wrong keys.

I'd say that's the primary reason for in-kernel maps.  Anything with a
distinct primary layer layout should be known to the kernel.  OTOH,
the primary layer of the Russian layout is the same as the US keyboard
has, so Russian must be handled via wsconsctl(8).

I'd even say that swapctrlcaps should *not* go to the kernel either as
it's not an existing physical layout but a personal preference.  And
it is inconsistently applied to only few layouts in pckbd_keydesctab
anyway.


> Everyone using a keyboard with US layout please stay out of the
> discussion ;-)

On Jornadas of 600 and 700 series the US layout has back-tick,
tilde, and square and curly parens *not* on the primary layer. ;)
(but we handle that via MACHDEP entries, so those port-specific
layouts are not polluting MI code).

SY, Uwe
-- 
uwe@ptc.spbu.ru                         |       Zu Grunde kommen
http://www.ptc.spbu.ru/~uwe/            |       Ist zu Grunde gehen