CODEWITHSUNDEEP
c++
bcf
bcf

Reputation: 2134

Adding doubles and complex numbers in C++

Consider this bit of code: 

#include <iostream>
#include <complex>

int main()
{
  std::complex<double> z1 = 5;
  std::cout << z1 - 1 << "\n";  // must change to z1 - 1.0 to compile

  std::complex<int> z2 = 5;
  std::cout << z2 - 1.0 << "\n";  // must change to z2 - 1 to compile
}

This produces a compilation error, as no operator- is found for types in the expressions z1 - 1 or z2 - 1.0. On the other hand, changing these expressions so that the base types match works fine.

Naively, for z1 - 1 I would expect the int 1 to be promoted to a double, and expected the z2, with base type int, in z2 - 1.0 to be promoted to a complex<double>. What's going on?

Upvotes: 7

Views: 2385

Answers (2)

Praetorian
Praetorian

Reputation: 109289

The operator- you're trying to invoke is a function template with a single type template parameter.

template< class T >
complex<T> operator-( const complex<T>& lhs, const T& rhs);

Neither of the template parameters in the two function parameters appears in a non-deduced context, so template argument deduction is performed on both parameters individually, and this results in the T for lhs being deduced as double, while that for rhs is deduced as int.

Due to this mismatch between the deduced types, template argument deduction fails, and your code doesn't compile.

Upvotes: 6

rlbond
rlbond

Reputation: 67849

The compiler will not automatically convert types from A to B to C if A is integral and B is floating-point. You want the compiler to convert from int -> double -> complex<double>, but it won't do that.

For the second example, you would need to promote the complex<int> to a complex<double>, which is not automatically handled by the complex class template.

Upvotes: 1

Related Questions