>> Rasops seems to use "long" to store character attributes.
> The reason is that there might be more complex display adapters where
> more than a bitmask is needed to represent a color, or perhaps some
> dynamic ressource allocation is needed.  In that case, a pointer
> could be used as handle.

...and in that case, shouldn't it be using something like a union of an
integral type and a void *?  I realize that this is the kernel, and a
certain amount of relaxation of the integral/pointer divide is to be
expected, but this is one of the last places I would expect it to show.
(I expect it in things like the pmap, or the VM subsystem.  Not places
that just *use* pointers.)

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B