Comparison: when operator<() is better than opCmp()

[Giuseppe Bilotta]

On Saturday 8 September 2007 08:19, BCS wrote:

> Reply to Giuseppe,
>>
>> whose result will be compared with 0 to determine the outcome.
> 
> IIRC on most systems you can get "cmp to 0" with little more than a jump 
> so that one is a freebe.

That's true (assuming it holds for reals (potentially NaNs) too), but
this doesn't really reduce the disadvantage by that much.

-- 
Giuseppe "Oblomov" Bilotta