round-off error in std::pow(std::complex<T>, double) in C++11

[Paolo Carlini]

.. oops, I just noticed what Marc suggested in the PR about changing 
pow(const complex<>&, const _Tp&) to *actually* forward to the complex 
builtins when not fast math. That seems a good idea, because we are 
normally *not* forwarding directly to the builtins in this specific 
mixed case, sorry I was wrong. I'd like to ear Gaby about this, however, 
because even for not fast math it seems a bit weird to not exploit the 
knowledge that the exponent is real...

Paolo.