[noise] New draft: "noh2" branch

[Trevor Perrin]

On Mon, Aug 31, 2015 at 3:17 AM, Jason A. Donenfeld <Jason at zx2c4.com> wrote:
> I don't risk truncation attacks anyway, because of the poly1305
> authenticator tag. You mentioned that file formats which rely on EOF will
> need a termination mark, but I don't see why: if a noise-encrypted file is
> truncated, won't the poly1305 tag be incorrect too?

No, I'm not talking about truncating a message to a partial message.
I was talking about truncating a stream of messages to a shorter
stream.

Trevor