Code: Select all

Vector2i vec(100, 100);
vec *= 0.5;
std::cout << "vec=\n" << vec << std::endl; // This gives [0, 0] instead of [50, 50]

To get the desired behavior, I need to do:

Code: Select all

Vector2i vec(100, 100);
vec = (vec.cast<double>() * 0.5).cast<int>();
std::cout << "vec=\n" << vec << std::endl; // This gives [50, 50]


Is there a simpler way of doing the above?

The Scalar operations defined in Eigen DenseBase all casts the provided scalar to the Scalar type of the DenseBase. Why not make these methods templated on arbitrary scalar types?

Because we do not allow mixing types implicitly: this kills vectorization and in most cases it means the user is doing something work like multiplying integers by 0.5 instead of dividing by 2.