Four binary functions greatly simplify the notation for ordinary scalar multiplication and division:
template < typename compo_a, typename scalar_t = compo_a, typename alloc_a = std::allocator <compo_a> > var_matrix<compo_a,alloc_a> operator*= ( var_matrix_base<compo_a,alloc_a> const & mat_a, scalar_t const scalar_v ) { struct mul_asgn { compo_a operator() (compo_a & a, scalar_t const b) { return a *= b; } }; return scalar_ret_var < mul_asgn, compo_a, scalar_t, alloc_a > (mat_a, scalar_v); }
template < typename compo_a, typename scalar_t = compo_a, typename compo_z = compo_a, typename alloc_a = std::allocator <compo_a>, typename alloc_z = std::allocator <compo_z> > var_matrix<compo_z,alloc_z> operator* ( sub_matrix_base<compo_a,alloc_a> const & mat_a, scalar_t const scalar_v ) { struct mul { compo_z operator() (compo_a const a, scalar_t const b) { return a * b; } }; return scalar_ret_con < mul, compo_a, scalar_t, compo_z, alloc_a, alloc_z > (mat_a, scalar_v); }
For example, this program:
#include "SOME_DIRECTORY/mat_gen_dim.h" using namespace mat_gen_dim; form const frm {blt{1,3},blt{2,6},blt{5,7}}; int main () { std::cout.setf (std::ios::fixed, std::ios::floatfield); std::cout.precision (3); { cout << "\n\n -- multiply a matrix by a scalar, return a copy of it"; var_matrix<double> mat_a5 {sub_matrix_linear (frm,1.0)}; cout << "\n mat_a5 before: " << mat_a5; double const s {5.0}; cout << "\n scalar: " << s; var_matrix<double> mat_z5 {mat_a5 *= s}; cout << "\n mat_a5 after: " << mat_a5; cout << "\n mat_z5 == what is returned: " << mat_z5; } { cout << "\n\n -- multiply a copy of a matrix by a scalar, return that copy"; var_matrix<double> mat_a6 {sub_matrix_linear (frm,2.0)}; cout << "\n mat_a6 before: " << mat_a6; double const s {6.0}; cout << "\n scalar: " << s; con_matrix<double> mat_z6 {mat_a6 * s}; cout << "\n mat_a6 after: " << mat_a6; cout << "\n mat_z6 == what is returned: " << mat_z6; } return 0; }yields this output:
-- multiply a matrix by a scalar, return a copy of it mat_a5 before: ( 1 2 5 ) = 1.125 ( 1 2 6 ) = 1.126 ( 1 3 5 ) = 1.135 ( 1 3 6 ) = 1.136 ( 1 4 5 ) = 1.145 ( 1 4 6 ) = 1.146 ( 1 5 5 ) = 1.155 ( 1 5 6 ) = 1.156 ( 2 2 5 ) = 1.225 ( 2 2 6 ) = 1.226 ( 2 3 5 ) = 1.235 ( 2 3 6 ) = 1.236 ( 2 4 5 ) = 1.245 ( 2 4 6 ) = 1.246 ( 2 5 5 ) = 1.255 ( 2 5 6 ) = 1.256 scalar: 5.000 mat_a5 after: ( 1 2 5 ) = 5.625 ( 1 2 6 ) = 5.630 ( 1 3 5 ) = 5.675 ( 1 3 6 ) = 5.680 ( 1 4 5 ) = 5.725 ( 1 4 6 ) = 5.730 ( 1 5 5 ) = 5.775 ( 1 5 6 ) = 5.780 ( 2 2 5 ) = 6.125 ( 2 2 6 ) = 6.130 ( 2 3 5 ) = 6.175 ( 2 3 6 ) = 6.180 ( 2 4 5 ) = 6.225 ( 2 4 6 ) = 6.230 ( 2 5 5 ) = 6.275 ( 2 5 6 ) = 6.280 mat_z5 == what is returned: ( 1 2 5 ) = 5.625 ( 1 2 6 ) = 5.630 ( 1 3 5 ) = 5.675 ( 1 3 6 ) = 5.680 ( 1 4 5 ) = 5.725 ( 1 4 6 ) = 5.730 ( 1 5 5 ) = 5.775 ( 1 5 6 ) = 5.780 ( 2 2 5 ) = 6.125 ( 2 2 6 ) = 6.130 ( 2 3 5 ) = 6.175 ( 2 3 6 ) = 6.180 ( 2 4 5 ) = 6.225 ( 2 4 6 ) = 6.230 ( 2 5 5 ) = 6.275 ( 2 5 6 ) = 6.280 -- multiply a copy of a matrix by a scalar, return that copy mat_a6 before: ( 1 2 5 ) = 2.125 ( 1 2 6 ) = 2.126 ( 1 3 5 ) = 2.135 ( 1 3 6 ) = 2.136 ( 1 4 5 ) = 2.145 ( 1 4 6 ) = 2.146 ( 1 5 5 ) = 2.155 ( 1 5 6 ) = 2.156 ( 2 2 5 ) = 2.225 ( 2 2 6 ) = 2.226 ( 2 3 5 ) = 2.235 ( 2 3 6 ) = 2.236 ( 2 4 5 ) = 2.245 ( 2 4 6 ) = 2.246 ( 2 5 5 ) = 2.255 ( 2 5 6 ) = 2.256 scalar: 6.000 mat_a6 after: ( 1 2 5 ) = 2.125 ( 1 2 6 ) = 2.126 ( 1 3 5 ) = 2.135 ( 1 3 6 ) = 2.136 ( 1 4 5 ) = 2.145 ( 1 4 6 ) = 2.146 ( 1 5 5 ) = 2.155 ( 1 5 6 ) = 2.156 ( 2 2 5 ) = 2.225 ( 2 2 6 ) = 2.226 ( 2 3 5 ) = 2.235 ( 2 3 6 ) = 2.236 ( 2 4 5 ) = 2.245 ( 2 4 6 ) = 2.246 ( 2 5 5 ) = 2.255 ( 2 5 6 ) = 2.256 mat_z6 == what is returned: ( 1 2 5 ) = 12.750 ( 1 2 6 ) = 12.756 ( 1 3 5 ) = 12.810 ( 1 3 6 ) = 12.816 ( 1 4 5 ) = 12.870 ( 1 4 6 ) = 12.876 ( 1 5 5 ) = 12.930 ( 1 5 6 ) = 12.936 ( 2 2 5 ) = 13.350 ( 2 2 6 ) = 13.356 ( 2 3 5 ) = 13.410 ( 2 3 6 ) = 13.416 ( 2 4 5 ) = 13.470 ( 2 4 6 ) = 13.476 ( 2 5 5 ) = 13.530 ( 2 5 6 ) = 13.536