On Sun, Apr 15, 2001 at 03:33:04PM -0500, Eric Haszlakiewicz wrote:
> so wouldn't the obvious solution be to change the char to signed char?

Um, sort of.

It'd be nice if there were a signed char typedef'ed in sys/types.h,
but there isn't (vaguely i386-centric to have a bunch of u*_ts and
no signed versions, but whatever).

Just throwing sign modifiers around on variables in kernel source
files strikes me as a bit shady.

Any comment from those with commit access to all this?

       ~ g r @ eclipsed.net