Functions
Fortran:
C++
Class
Conversion
Asign a double to a float should trigger a warning. Due to the reduced precision of float. Asign a integer to a float should also warn. There exist big integers which cannot stored in float without loose of the last digits.
There are 4 different cases:
1) integer constant to float
2) integer variable to float
3) double constant to float
4) double variable to float
Fortran:
integer :: i
real(4) :: a
real(8) :: b
a = 1
a = 100000001a
a = i
a = 0.1234567890123_8
a = b
C++
int i;
float a;
double b;
a = 1;
a = 100000001;
a = i;
a = 0.1234567890123;
a = b;
gfortran fortran.F90 -Wconversion -Wconversion-extra
fortran.F90:20.8:
a = 1
1
Warning: Conversion from INTEGER(4) to REAL(4) at (1)
fortran.F90:21.8:
a = 100000001
1
Warning: Conversion from INTEGER(4) to REAL(4) at (1)
fortran.F90:22.8:
a = i
1
Warning: Conversion from INTEGER(4) to REAL(4) at (1)
fortran.F90:23.8:
a = 0.1234567890123_8
1
Warning: Change of value in conversion from REAL(8) to REAL(4) at (1)
fortran.F90:24.8:
a = b
1
Warning: Conversion from REAL(8) to REAL(4) at (1)
Initializer lists
Fortran:
Module TestModule
contains
subroutine func(list)
implicit none
integer, intent(in) :: list(:)
write(*,*) list
end subroutine
end module
program testprogrm
use TestModule
implicit none
call func( (/1,2,3/) )
end program
$ gfortran -Wall -Wextra -fcheck=all fortran.F90
$ ./a.out
1 2 3
C++
#include < iostream >
#include < initializer_list >
void func(std::initializer_list list) {
for(int x: list) std::cout << x << ' ';
}
int main() {
func( {1,2,3} );
}
$ g++ -Wall -Wextra -std=c++11 cpp.cpp $ ./a.out 1 2 3
(Strongly typed) enumerations