_Rb_tree + allocator observations

[Benjamin Kosnik]

>Why not
>derive from the allocator object itself somehow to take advantage of the
>empty base class optimization?

You might want to look through this thread from earlier in the year:
http://gcc.gnu.org/ml/libstdc++/2002-11/msg00176.html
http://gcc.gnu.org/ml/libstdc++/2002-11/msg00165.html

No doubt, the allocator bits could be done differently, and more
efficiently. If you want to work on this, it would be appreciated.

There's been some work on alternate allocators on mainline. 

-benjamin