[Bug c++/51083] gnu++0x: resolving overloaded functions belonging to different namespaces

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51083

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |paolo.carlini at oracle dot
                   |                            |com

--- Comment #6 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-11-10 18:46:24 UTC ---
Paolo, in <complex> this will match anything:

  template<typename _Tp, bool = std::__is_integer<_Tp>::__value>
    struct __promote
    { typedef double __type; };

  template<typename _Tp>
    struct __promote<_Tp, false>
    { typedef _Tp __type; };

  template<typename _Tp>
    inline typename __gnu_cxx::__promote<_Tp>::__type
    imag(_Tp)
    { return _Tp(); }

Shouldn't it be restricted to *only* match long double, double, float and
integral types?


e.g. by changing the non-integer specialization

  template<typename _Tp>
    struct __promote<_Tp, false>
    { };

and adding explicit specialization 

  template<>
    struct __promote<long double>
    { typedef long double __type; };

  template<>
    struct __promote<double>
    { typedef double __type; };

  template<>
    struct __promote<float>
    { typedef float __type; };

that allows the attached testcase to compile