Cesar Eduardo Barros

Sorry, but formatting this code with rustfmt requires strategically-placed `rustfmt_ignore` annotations. Some of the constants in this code are 4x4 (or other size) matrices, and were manually formatted to make...

@tarcieri Sure, you can go ahead! This is exactly what was done for his `blake2` crate, which is (with my blessing) a lightly modified copy of my `blake2-rfc` crate.

I took a quick look at the link you provided, and it seems that a little more is needed. In particular, it uses a 64-bit counter and a 64-bit IV,...

According to the error message, the issue is that the C compiler (in this case, `arm-none-eabi-gcc`) could not be found. That is not something which could be fixed by this...

Thanks for the tip! That crate also uses the same `#[inline(never)]` trick I used, plus some other more advanced "confuse the optimizer" tricks, though it doesn't do the only really...

I have a working implementation of the RFC 7539 ChaCha20-Poly1305 AEAD as the [chacha20-poly1305-aead](https://crates.io/crates/chacha20-poly1305-aead/) crate.

This is also the default configuration for encrypted Fedora installations, or at least used to be back when it used swap on disk instead of zram. Upgrading an encrypted Fedora...

As explained in the large comment within the fallback implementation, `black_box` is not used alone as the optimization barrier; it also uses `#[inline(never)]` for that function. Unfortunately, that's the best...

Since late last year, I have been working on making `constant_time_eq` even more robust against potential optimizer shenanigans (a "sufficiently smart optimizer" could notice that the accumulator would no longer...

> Let's see what Cesar thinks of that PR I just put up. Sorry for the delay. Looks fine to me. Merged and published `constant_time_eq` 0.3.1 with only that change.