Not sure if I am doing everything right.

Code: Select all

#include <Eigen/Core>
using namespace Eigen;
typedef const Ref<const MatrixXd>&       mat_const_ref_d;
typedef const Ref<const MatrixXf>&       mat_const_ref_f;
void foo(mat_const_ref_d) {}
void foo(mat_const_ref_f) {}

int main()
{
    MatrixXd D(10, 10);
    MatrixXf F(10, 10);
    foo(D);
}


I am trying to overload based on scalar type double / float. But somehow Visual Studio compiler doesn't like it. It tells me

Code: Select all

'foo' : ambiguous call to overloaded function

Thanks.

I replicated the problem on linux and gcc-4.7.2.

Fixed in devel branch, see: http://eigen.tuxfamily.org/bz/show_bug.cgi?id=969
and backported to 3.2: https://bitbucket.org/eigen/eigen/commits/2ecad781e8a3/

That sounds good.

I have a followup question.

As I understand Matrix::swap is efficient when used with dynamic size matrices because no data is copied. I am 99% sure about that, but still would like here a confirmation from the developers.

Thanks.

yes, in A.swap(B), if A and B are Matrix<> object of the exact same type, then swap simply swaps the pointers and sizes.