[noise] More Implementation Pitfalls: ReadMessage() Error Conditions

[Alex]

On Mon, 16 Nov 2015 20:16:15 +0100
"Jason A. Donenfeld" <Jason at zx2c4.com> wrote:

> I'd suggest the other implementations also aim for an idempotent
> ReadMessage() implementation too. At the least, it prevents people
> from shooting themselves in the foot. At the most, it lets these
> implementations be used in UDP-based protocols like WireGuard.
> 

This is not a problem with cacophony because all variables are
immutable in Haskell. As a result, every function must return both
the ciphertext/plaintext plus a fresh altered state[1].

-- 
Alex

[1] http://hackage.haskell.org/package/cacophony-0.3.0/docs/Crypto-Noise-Handshake.html#v:readMessage