Tim M

I've been thinking about playing around with bouncy castle for a while now. I ran some analysis on V3 Sign/Verify and it looks like the bulk of allocations (and presumbaly...

> I wonder if we can reduce the method complexity of those intrinsics ones by breaking them apart. > I've split Intrinsics up by chunk size and altered `ChaCha64` and...

I've experimented with a [refactor](https://github.com/TimothyMakkison/NaCl.Core/tree/experimental_intrinsics_refactor) that treats `ChaCha20Base` and `Salsa20Base` as a proxy for the internal `...Core` and `...CoreItrinsics` methods. It removes the intrinsics checks and preprocessor spam but arguably...

Thanks for the review. I got sidetracked and forgot about this PR so it's unfinished. > I wonder if we do need the preprocessor directive since we are checking if...

> Yeah, I agree with that. It added a lot of maintainability complexity, if we limit the scope of the supported .NET versions we can stick with the first logic...

Conflicts should be resolved 🤞 and I've added some documentation to explain the different encoding operations. I'm a bit concerned with the Argon2id encode/decode, running all tests used to take...

Looks like the problem might be caused by the difference in Argon2id implementations. PBKW takes the arguments `password`, `memory cost`, `time cost` and `parallelism degree`. The [php algorithm](https://www.php.net/manual/en/function.password-hash.php) uses `memory...

Don't know if my earlier comment is right but looking at `TestPwVectors` in PaserkTests.cs ``` C# // Decode paserk to verify decoding works var decoded = Paserk.Decode(test.Paserk, test.Password); decoded.Key.Span.ToArray().Should().BeEquivalentTo(pasetoKey.Key.ToArray()); ```...

- Changed Paserk.ArgonEncode to take bytes instead of KiB. Not sure about this change as most Argon2id implementations measure memoryCost in KiB, on the other hand Paserk might specify that...

> Sorry for looking at the PR until now, I was a bit overwhelmed by work. No worries, I'm in no rush, this pr could probably do with more time...