OK, I will find out which variant seems to fit best and fix this.
Also I am going to answer the bug report.
Volker
2012/12/4 Raymond Toy <toy.raymond at gmail.com>:
>>>>>> "Volker" == Volker van Nek <volkervannek at gmail.com> writes:
>
> Volker> About (float n) in get-one-factor-pollard:
>
> Volker> (float n) for numbers greater or equal to 2^1024 is definitely not a
> Volker> good thing - regardless of whether we want to factor such a number or
> Volker> not. If you don't expect that (log n) will work it will be possible to
> Volker> use numbers as 'terms' which have about the same size as (log n).
> Volker> Maybe something like (integer-length n) or (ash (integer-length n)
> Volker> -1). Maybe (isqrt (integer-length n)) is better. Let's see if I find
> Volker> some time in the next weeks to check this out.
>
> I think any of the integer-length variants will work. Or maybe even
> (floor (* (log 2) (integer-length n))), which is a rough approximation
> to the original (ceiling (log (float n))).
>
> Ray
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima