std::map and std::set based on AVL, not RB trees.

[Paolo Carlini]

Jonathan Wakely wrote:

>I don't think anything strange is going on, I just wonder if it would be
>better if we implemented the proposed resolution (as Dinkum already
>does) because
>  
>
But I have to remind you that, as a policy we do **not** implement
resolutions in the "Open" status. What if the final resolution is different?

>2) we would still support code written to the old meaning of the hint.
>   The traditional meaning, used by the HP and SGI implementations, was
>   that insertion before the hint is constant. There could be lots of code
>   written using that guarantee.  Our docs still describe that behaviour.
>  
>
Then the docs must be fixed.

Paolo.