Fixing the SortedRange

[Andrei Alexandrescu]

On 3/4/12 10:24 AM, Dmitry Olshansky wrote:
> Nah, that's not gonna work as !(b < a) is a >= b, not a > b, a kind of
> thing obvious on the second thought only. And anyway, looking at the
> implementation as lowerBound uses !pred(lhs, rhs) while upperBound uses
> pred(rhs, lhs). So I think it's okay and less fragile to always require
> both ways. It's all in regex pull btw, like I could fix one bug on it's
> own ;)

Yah, you'd need a == b too. I think it's reasonable to demand that both 
a < b and b < a work.

Thanks,

Andrei