New code for primality testing

[Torbjörn Granlund]

paul zimmermann <Paul.Zimmermann at inria.fr> writes:

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

Doing that wouldbe a betrayal of the promised interface, I think.

I like Marco's approach of skipping the first 24 M-R iterations.  People
that are likely to distrust BPSW are also likely to use a over-large M-R
count, so will remain reasonably happy.

-- 
Torbjörn
Please encrypt, key id 0xC8601622