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

[Paolo Carlini]

Gabriel Dos Reis wrote:

>And I ask again tha very question I asked earlier: You, Paolo, what do
>you believe it is *useful* for to follow the current wording which is
>questionable?  What arguments are you going to oppose to moving this
>issue at the next meeting, which is in two weeks?  Are you going to
>say that if after 5 years, it is still open it must serious and we
>should leave it in open status?
>
I'm not going to oppose any argument, FWIW my personal opinion. Because,
basically, I agree with the proposed resolution. But this will happen in
2 weeks, if everything goes well, not *now*. Now I see an *open* DR,
which, if implemented, leads to a different and also surprising behavior
(/a/g/a/i/n/, /s/e/e /m/y /m/e/s/s/a/g/e/s about the implementation of
insert, and user espectations, and throwing comparisons) if compared to
the current one.

Paolo.