On the validity of the anomalous diffraction theory to light scattering by cubes