STL and explicit template instantiation

[Paolo Carlini]

Robert Schweikert wrote:

>>You can't, for the very same reason, only involving slightly 
>>different helper templates, that is the point, right?
>>    
>>
>touche
>  
>
Elegant ;) Really, my point is that this kind of issue is not new, has 
always been there, besided the specific case of __uninitialized_fill_n_a 
which are going to fix (and thanks for your report). As pointed out by 
Benjamin, we should definitely try to reassess whether we can inline 
more implicit templates, that will certainly help, but personally I'm 
not sure we can do that always and everywhere, I think it's a pretty 
strong requirement for such heavily templated bits of the library if one 
wants to fight at the same time code bloat.

Paolo.