Make 'fortran float sin(), cos();' work with gcc by replacing it with