Tony Arcieri

> But it could be worth to postpone it until digest v0.10 release. I definitely think we should land #217 first and do another release of the current `blake2` crate...

I'm weakly in favor of merging the crates for the following reasons: 1. It will provide the simplest transition path for existing users of the `blake2` crate 2. It leaves...

As of fe77461, this PR implements what is effectively the sum of the `blake2` and `blake2b_simd`+`blake2s_simd` APIs. I'm not removing WIP/draft yet because I want to add the full set...

Took a look at the `clippy::ptr_offset_with_cast` warningss... they all originate in `arrayref`, which is using `offset` with a `usize` casted to an `isize` instead of `add`. Either `arrayref` needs to...

Removing WIP/draft cc @oconnor663 This now passes all of the original tests from `blake2` v0.9 and in theory should be a fully API compatible replacement. There's one remaining item on...

As of 2f45194 I have fixed the build warnings on `no_std` targets by gating all of the parallel implementations to only work on `x86`/`x86_64` platforms for now. On the `blake2b_simd`/`blake2s_simd`...

> But where does the runtime panic happen? I was referring to this panic: https://github.com/RustCrypto/hashes/pull/228#pullrequestreview-583526150 However I fixed the issue in 2f45194713e5c9beb56191ca65c3bd0cec91e6cb by gating all of the parallel implementations to...

I think based on https://github.com/RustCrypto/hashes/commit/2f45194713e5c9beb56191ca65c3bd0cec91e6cb the codepaths were dead/unreachable. The gating I did in that commit was necessary to eliminate all of the warnings on `thumbv7em-none-eabi` (since we check `-Dwarnings`...

Looks quite similar to 2f45194, no?

Aah, interesting. I was still getting warnings having done similar changes to yours.