Double factorial and primorial

[Torbjorn Granlund]

nisse at lysator.liu.se (Niels Möller) writes:

  I don't know how important that is. Long-term, I think gmp should at
  least export some sieving functionality which us suitable as a building
  block for that.
  
I agree we want a shared prime sieve, at least for internal use.  It
would be nice not to have two (as today).

Perhaps it makes sense to have two, but then we should have a clear idea
why both are really needed.

As to a public interface, I am more hesitant, but vote me down if you
feel the need.

-- 
Torbjörn