> Is there a reason we can't provide an inline/optimized isctype? What we'd need is a very strong reason to provide such a nonstandard interface. There is no public interface that enshrines the notion of ctype classifications being represented by a bitmask.