Re: [GNU Crypto] Bug in gnu.util.Prime.passEulerCriterion(java.math.BigInteger)

[Casey Marshall]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

>>>>> "Eric" == Eric Gilbert <address@hidden> writes:

Eric> Hello, You may already know about this one, but I figured it
Eric> would be nice to pass it on if you don't.  I am integrating the
Eric> GNU Crypto primality testing code into my own project CatCon,
Eric> http://catcon.sourceforge.net/pub/.  The method
Eric> gnu.util.Prime.passEulerCriterion(java.math.BigInteger)wrongly
Eric> identifies the following numbers as composite:

Eric> 38737 61681 65537 229153 274177

Eric> These are all of the primes less than 300000 that
Eric> passEulerCriterion incorrectly labels as composite.  There are
Eric> more, I am sure, probably having some identifying
Eric> characteristic.  Just thought you ought to know.

Looks like you're right, and no, I didn't know this. I will remove
references to passEulerCriterion in isProbablePrime, and really I
don't see any benefit to using any tests other than Miller-Rabin.

Ugh, I'm getting increasingly tired of the Prime class; I'm
considering just punting this class and using
BigInteger.isProbablePrime instead.

- -- 
Casey Marshall || address@hidden
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.7 <http://mailcrypt.sourceforge.net/>

iD8DBQFAhLpGgAuWMgRGsWsRAgqvAJ4ymly0xVKtoN79zoMvKwO7jhkwrwCcDFgE
MS1vP+GpYWoBONimXUR+Xg8=
=/+GV
-----END PGP SIGNATURE-----