20 May
2008
20 May
'08
11:11 p.m.
Right, and it also reduces the number of template instantiations needed?
Yep, it instantiates almost half the number of templates instantiated by the first implementation. About inlining considerations: I don't think that the compiler cares about the deepness of function calls it has to inline. I see optimization and inlining as a recursive algorithm that does its job exactly the same way however deep you are. But I'm not a compiler expert... Anyway your solution is really good and I'll sure use it in the next implementation. Thanks! Bruno