SHA-3 (Keccak) about cryptography

[alex]

IMO we should wait until it's finalized (I assume that's why OpenSSL doesn't have it yet). Happy to heart other opinions though.

[lvh]

If we include hash functions where internal state != digest value, please also provide an API for getting and setting the internal state. I don't care if it optionally raises on some implementations, but just provide the API.

Without this, it's impossible to write half-decent hash chains using these hash functions.

(I think that's a misfeature anyway; the old way is only bad if you abuse hash functions and have no clue what you're doing.)

[dstufft]

Did you mean to close this @lvh?

[lvh]

Oops. I didn't know what you were talking about, I just didn't realize I had closed it. My bad :)

[tiran]

Interesting idea ... SHA-3 objects aren't pickle-able because the code doesn't expose their internal state. I never thought of that. I'm going to talk to A.M. Kuchling about it.

[alex]

I think we're going to close this until SHA-3 is a standard.