[RFC] Two versions of struct _Rb_tree_impl... or not?!

[Paolo Carlini]

Dhruv Matani wrote:

>Yes, but just to be sure, just run it once more, because I think that
>someone may have touched rb_tree.h after I submitted that patch.
>
Ok, thanks.

Anyway, I'm very unhappy about these numbers:

3.3.3
-----
Size of std::map is: 12
Size of std::map is: 12
Size of std::map is: 12
Size of std::map is: 12
Size of std::map is: 12
Size of std::map is: 12
Size of std::map is: 12
Size of std::map is: 12

3.4/mainline
------------
Size of std::map is: 24
Size of std::map is: 24
Size of std::map is: 24
Size of std::map is: 24
Size of std::map is: 24
Size of std::map is: 24
Size of std::map is: 24
Size of std::map is: 24

3.4/mainline patched (Dhruv patch)
----------------------------------
Size of std::map is: 20
Size of std::map is: 20
Size of std::map is: 20
Size of std::map is: 20
Size of std::map is: 24
Size of std::map is: 24
Size of std::map is: 24
Size of std::map is: 24

Icc8.0
------
Size of std::map is: 12
Size of std::map is: 12
Size of std::map is: 12
Size of std::map is: 12
Size of std::map is: 16
Size of std::map is: 16
Size of std::map is: 16
Size of std::map is: 16

So, even with Dhruv patch, we would still regress *a lot* wrt
what we had in 3.3.3.

What we have gained in the meanwhile so important to justify
this horrible (yes, seems horrible to me, frankly) size
regression??? Can we fix it???

Thanks,
Paolo.