Re: ratio implementation

[Kevin Ryde]

Marius Vollmer <address@hidden> writes:
>
>       mpq_init (frac);
>       mpq_set_d (frac, SCM_REAL_VALUE (z));
>       q = scm_make_ratio (scm_i_mpz2num (mpq_numref (frac)),
>                         scm_i_mpz2num (mpq_denref (frac)));
>       mpq_clear (frac);

Yep, that's about it.

I guess the num and den from the mpq_t could be copied directly into
bignums and used as mpz_t's from then on, rather than initializing new
mpz_t's.

Just have to think a bit whether that would be compatible going
forward.  We already allow arbitrary mpz calculations on the two
parts, just have to decide if that can include clearing :).