ws@kurt.tools.de (Wolfgang Solfrank) writes:

> 
> While the explicit declaration of p_nice as signed char would solve the
> problem, I'd much rather see either

I really don't see a plausible reason to object to it.

> 	p_nice declared as short

NO.

> 	p_nice left as a char, and have the NZERO code be fixed to allow
> 		and use a value other than 0 for NZERO (just like PZERO).

This would be fine, too.