[patch] Improving shared_ptr for C++0x

[Paolo Carlini]

Jonathan Wakely wrote:
> In another thread Paolo pointed out DR402, which I hadn't thought of
> when implementing _Sp_counted_ptr_inplace.  The two uses of placement
> new with the embedded type should be qualified in case the type
> defines its own operator new.  The attached testcase verifies this and
> I'll add it with the other parts.
>   
Good ;)

So... do we have other comments about your work? I think it's time to
commit. If nobody else replies during the next 24 hours, please do it!

In case something weird will shows up (on less widespread targets?) your
help will be appreciated.

Thanks again,
Paolo.