New code for primality testing

[paul zimmermann]

       Dear Marco,

> > if you have already implemented BPSW, probaby the simplest is to make
> > mpz_probab_prime_p use BPSW (removing the reps argument), and have another
> 
> We did. The newly implemented test was the main reason for this thread.
> 
> Currently we did not change the interface. BPSW is used as a first step in
> mpz_probab_prime_p, and substitute the first "number" repetitions, then
> the remaining iteration of MillerRabin are used. This is intended for
> GMP-6.2.
> 
> Do you have suggestions for that "number"? Then, I agree, for GMP-7 we
> will have to redesign the interface...

since no counter-example is known for BPSW, you could substitute all REPS
iterations.

On the other hand, you could leave all REPS iterations, and hope a
counter-example is found by GMP (in that case, make sure to send the
counter-example to the gmp-devel list).

Paul