15 May
2008
15 May
'08
8:27 a.m.
That's more or less what I did in the Sandbox version, but Hervé Brönnimann's solution looks more elegant maybe?
Yep that's exactly what I was just answering :-) It's much smarter. I'll be off until sunday evening unfortunately, but since I don't have trunk access I wouldn't be of much help I guess. I can test performances on my computer when I come back though (I remember the fact to add or remove a specialization could have big effects on perfs for some exponents). BTW, with this implementation we should be able to remove the specialization in <2>, I remember it was there only to avoid a compile-time infinite loop due to the explicit call to positive_power<2>. Bruno