Re: Questions about primality certification in Pari

[Bill Allombert <Bill.Allombert@math.u-bordeaux1.fr>]

On Mon, Aug 31, 2009 at 07:05:40AM +0200, Karim Belabas wrote:
> * William Hart [2009-08-31 05:29]:
> > I have two questions, the second one regarding what may be a bug:
> 
> First a remark: your questions concern pari-stable, but you should
> rather study the current SVN code which is a bit different (and more
> efficient).
...
> > 
> > This code is used to determine if the BSW test is sufficient to guarantee primality, by testing whether the number is < 10^13.
> > 
> > But on a 64 bit machine, doesn't this always return 1 when x is a single limb integer? It looks like it was written for a 32 bit machine. Shouldn't it be checking if l == 3 on a 64 bit machine then checking if x < 10^13. 
> > My apologies if that is not right.
> 
> You are right: this one is a bug.

Does it affect PARI 2.3.4 ? Should we backport it ?

Cheers,
Bill.