Question

I would like to be able to cast a class A<double> to A<float>. Here is what I tried:

#include<iostream>

template<class T1>
class A {
  template<class T0>
  operator A() {
    std::cout << __PRETTY_FUNCTION__ << "\n";
  }
};

template<class T0,class T1>
void bar( const A<T0>& a, const A<T1>& b )
{
  reinterpret_cast< const A<T0> >(b);
}


int main()
{
  A<float> a_f;
  A<double> a_d;
  bar(a_f,a_d);
}

The compiler spits: error: invalid cast from type ‘const A<double>’ to type ‘const A<float>’.

Is a conversion operator converting to a class differing only in the template argument possible? If yes: How?

Answer 1

template<class T0> operator A<T0> ...

Without the template argument, its a conversion to A<T1>, i.e. a no-op.