as there's a bunch of good fortran code out there, i've been fiddling on wrapping fortran 77 subroutines, it's as simple as with C but compatible real & double precision must generated with gfortran, i.e.:
@gfortran -Ofast -c -fdefault-real-8 -fdefault-double-8 493.f
this ensures nim::float (==float64) compatibility,
generated fortran objects are lowercase with a trailing '_': 'rpoly' in this case:
so nim interface is:
proc poly_root_f(op: ptr cdouble, degree: ptr cint, zeror, zeroi: ptr cdouble, fail: ptr cint) {.importc: "rpoly_".}
full code for polynomial roots:
https://github.com/rforcen/nim/blob/main/rpoly.nim https://github.com/rforcen/nim/blob/main/493.f