Fix GH-16890: array_sum() with GMP can loose precision (LLP64)

We must use `mpz_fits_si_p()` instead of `mpz_fits_slong_p()` since the
latter is not suitable for LLP64 data models.

Fix GH-16890: array_sum() with GMP can loose precision (LLP64)

We must use `mpz_fits_si_p()` instead of `mpz_fits_slong_p()` since the
latter is not suitable for LLP64 data models.

libgmp, however, does not define `mpz_fits_si_p()` (which is an mpir
addition), so we use `mpz_fits_slong_p()` there which should be fine.

Closes GH-16891.

show more ...