vector<> can probably never grow to it's maximum size!

[Paolo Carlini]

Dhruv Matani wrote:

>The part which handles this case is within the
>__mt_alloc::allocate(n,hint,hint_passer) overload.
>
>Look at the case where __n*sizeof(_Tp) > 128. There is a series of try
>catch which use the _M_at_least field of the hint_passer to detect if
>the allocator can if the __n failed, allocate at least _M_at_least
>objects. If so, then it will succeed.
>  
>
You never answered my first objection: if, for whatever reason, 
::operator new starts
throwing and every time you allocate only 1 position more, the container 
doesn't
grow anymore with the speed mandated by the standard. The only way to be 
correct
is multiplying the previous size by factor bigger than 1, typically 1.5-2.0.

Also, please, don't send every time a whole patch including 
*everything*: you
understand that we are definitely at the general design stage, so many 
things and
details have to be sorted out and discussed before something like this 
finds its
way in the repository...

And, please, again, let's discuss one issue at a time.

Paolo.