[pycrypto] Wanted: Additional unit tests for Python 3.x porting effort
paul_koning at dell.com
Fri Dec 31 08:24:23 CST 2010
If I remember right, some versions of pgp used ElGamal, so that might serve as an open reference implementation. Not as good as test vectors but it's a widely used implementation.
On Dec 31, 2010, at 9:22 AM, Michael Rodler wrote:
> I'd volunteer to write the unit test for ElGamal, but since ElGamal is
> not standardized, it's pretty hard to find any public test vectors.
> (quick google search reveals nothing) So you probably need to generate
> them yourself and I'm not sure how to do that properly. (generate using
> another library i.e. bouncycastle or using some algebra system i.e.
> Mathematica, which is probably less error prone?)
> If anyone can point me into the right direction, I'd be happy to help.
> (Obviously not today, so cheers and happy new year ;)
> On 2010-12-31 06:11, Thorsten Behrens wrote:
>> Crypto.Random.random has unit tests now. A few notes on that:
>> - random.shuffle() was broken, it had a TypeError and an
>> IndexError. Not my doing, that's always been the case. That's been fixed.
>> - random.sample() does not work on 2.1. Not my doing, that's always
>> been the case. I will not fix this. Others are welcome to.
>> - random does not implement the full compliment of functions that
>> Python offers, it is a much smaller subset
>> - I somehow doubt anyone ever actually used this interface. We may
>> declare it "static for backwards compatibility" and leave well enough alone.
>> I will not write unit tests for qNEW, though others are welcome to.
>> We still need unit tests for ElGamal and AllOrNothing. Anyone?
>> On 12/29/2010 4:28 PM, Thorsten Behrens wrote:
>>> Now that we have a working pycrypto on Python 3.x, it's time to see what
>>> we don't have unit tests for. See
>>> https://bugs.launchpad.net/pycrypto/+bug/335995 for results.
>>> Currently, the following are not included in unit testing at all:
>>> I'd love some unit tests for those. With public test vectors, please,
>>> not test vectors derived from pycrypto.
>>> pycrypto mailing list
>>> pycrypto at lists.dlitz.net
>> pycrypto mailing list
>> pycrypto at lists.dlitz.net
> pycrypto mailing list
> pycrypto at lists.dlitz.net
More information about the pycrypto