Re: primetables SEGV

[Ilya Zakharevich <ilya@math.ohio-state.edu>]

On Thu, Dec 05, 2002 at 08:57:07PM -0500, Igor Schein wrote:
> Hi,
> 
> % gp -q -f -p 2156858852
> Segmentation fault
> 
> 1 extra check is needed.
> 
> That's with 32bit binary.

The current logic is flawed above 2^31.  Actually, I have a patch for
several days already.  ;-) But I'm still documenting the
cache-optimization policy...

BTW, the 1.07 * ... estimate is *way* above the actual pi(x).  And the
miniscule (about 1/1e7 ;-) number of primes which require two bytes
does not matter either.

Ilya