c++0x implementation and compatibilty thoughts

[Douglas Gregor]

On Apr 27, 2007, at 7:30 AM, Benjamin Kosnik wrote:
>> C++0x behavior from components in namespace std, TR1 behavior from
>> components in namespace std::tr1.
>
> I'm still a little leary of mixing items with the same name and  
> different behavior in a namespace (std) and a nested namespace  
> (std::tr1).
>
> We should definitely try it though.

It occurred to me afterward that it still makes sense to put the C+ 
+0x-specific versions into namesspace std::__cxx200x and then use a  
strong using directive to pull them into namespace 'std'. That way,  
we don't pollute the linker namespace 'std' with experimental C++0x  
stuff.

	- Doug