Hi  Marek:
According to Lang's suggestion, I think DH and DSS is a better choice. Thanks for your generous. DH and DSS is strange to me. I need to to learn more about this. I accept it 's difficult for me to absorb your proposal . it's better for me to get this step by step. As i said in my gsoc
proposal I'm interested in this. joining the wine  project is the best for me. 

thank you.


2013/5/10 Marek Chmiel <kcmark@gmail.com>
Hey Kaiyi Zhang,

Welcome to wine dev! If you decide to move forward with this(DH and DSS), I wish to share with you all that I have finished up to now. Using a few things from RSAENH I have all the related algorithms functioning. Also using LibTomCrypt I have partially implemented DSA but you would have to choose on the PRNG that DSA will use. My plan was to use RC4 but have not gotten to test that. I have yet to touch anything with DH_EPHEM (Ephemeral) or DH_SF (Store and Forward).  If you have any questions shoot them to Juan or me (Juan is the better choice). Best of luck!

-- 
Marek

>Hi Kaiyi Zhang,

>
>(or is Zhang Kaiyi?)
>
>I think there's a little misconception in your proposal. Bcrypt the
>algorithm is not the same as the BCrypt functions in crypt32. I believe
>that Microsoft redesigned their CryptoAPI and more or less renamed their
>functions BCrypt*. I don't believe they have any relationship to the Bcrypt
>algorithm by Niels Provos and David Mazieres. I don't know how many
>applications depend on the Bcrypt* functions in newer crypt32
>implementations, vs. the the Crypt* functions that are present in all
>versions of crypt32.
>
>If you want feedback on Wine crypto stuff, I'm probably the best to help
>you with that. Sorry if I've been a little absent, I'm pretty busy between
>my family and my day job of late.
>
>I have another idea for you that I think is tractable, and perhaps up your
>alley. The DH and DSS algorithms are still unimplemented in Wine. We
>already have tests for them, so I think it should be possible to contribute
>implementations. See the tests in dlls/dssenh/tests.
>
>I hope that helps.
>--Juan