Derive Hash for Signature

[tcharding]

In preparation for deriving Hash in miniscript, derive Hash on the ecdsa::Signature.

ref: rust-bitcoin/rust-miniscript#226

[apoelstra]

I wonder what guarantees Hash is supposed to have. I guess, only that it's compatible with Eq...but in particular, this implementation will give different hash values when the library is compiled with a different endianness.

[apoelstra]

ACK ef7f197

but would like another ack before merging.

[tcharding]

Am I understanding you correctly that you think the following workflow is broken?

A: little-endian machine
B: big-endian machine

1. A: create data structure that includes a signature
2. A: hash data structure
3. A: serialize data structure 2
4. send over wire to B
5. B: deserialzie data structure 2 (as GOT)
6. B: create data structure 1 that includes a signature (using same data as A used)
7. B: hash data structure 1, put hash into data structure 2 (as WANT)
8. B: Expect GOT == WANT

If the assertion in (8) is not true doesn't that basically break all portability for any data structure that uses Hash and since its std::hash::Hash this would imply that code written with the Rust stdlib is non-portable? That seems like a pretty big deal, I'd like to think that Rust guarantees the assertion in (8)? Or am I being brain dead?

[tcharding]

Whatever direction we take, should this PR unify the derives on ecdsa::Signature and schnorr::Signature? The schnorr sig currently has no derives.

[elichai]

Due to differences in endianness and type sizes, data fed by Hash to a Hasher should not be considered portable across platforms. Additionally the data passed by most standard library types should not be considered stable between compiler versions.

From: https://doc.rust-lang.org/std/hash/trait.Hash.html#portability

[elichai]

ACK ef7f197

[tcharding]

Wow, thanks for the lesson.