From: Dmitry Antipov <dmitry.antipov@...>
Date: 2002-06-09T12:12:00+09:00
Subject: Re: Another implementation of Bignum [tarball attached]

-----Original Message-----
From: "Christoph" <chr_news@gmx.net>
To: <ruby-core@ruby-lang.org>
Date: Sun, 9 Jun 2002 09:28:53 +0900
Subject: RE: Another implementation of Bignum [tarball attached]
 
> One comment, did you consider implementing the shift methods
> with the mpn methods mpn rshift and mpn lshift? 

Yes (for now, shifts by n a naive multiplication/division by 2^n :-()

> What about the ``extended'' gcd (mpz gcdext), which is
> probably more useful the gcd itself - lets say
> 
>     Integer#gcd ext(other int) -> [gcd,m,n] such that
>     self*m + other int*n = gcd (of self and other int)

May be. In any case approval from matz is required.
 
> You surely are aware the gmp has a probabilistic primality test.
> Anyway, what about using a free implementation based on 
> Lenstra's elliptic curve Method - see
> http://www.loria.fr/~zimmerma/records/ecmnet.html

Thanks for URL - I'll take a look.

Dmitry