nazar at mokrynskyi.com
Wed Oct 4 19:33:24 PDT 2017
I've tested cacophony vectors from Alex's repo initially and even have support for protocol_name field already, but like you've said, PSK vectors were failing, so tests are running against those in noise-c's repository for now.
Rhys Weatherley is working on updates in noise-c already (https://github.com/rweather/noise-c/issues/24), I'll make necessary updates shortly after noise-c.
Sincerely, Nazar Mokrynskyi
On 10/5/17 5:14 AM, Trevor Perrin wrote:
> On Wed, Oct 4, 2017 at 11:30 AM, Nazar Mokrynskyi <nazar at mokrynskyi.com <mailto:nazar at mokrynskyi.com>> wrote:
> I've worked on test vectors today and now noise-c.wasm passes all of the vectors included in noise-c repository, so in terms of spec compliance (v30 according to noise-c's website at the moment) it should be on par with noise-c.
> Hi Nazar,
> That's great! Thanks for doing that.
> You might also test with the latest cacophony vectors (from Noise wiki, or Alex's repo). But I think the only differences are:
> * The test vector format (it now parses the protocol name, instead of having separate JSON fields for DH/cipher/hash)
> * PSK handling, which changed in revision 32
> (We should probably version the "cacophony" vectors, like include the creation date in the filename or something).
> I think Noise-C is up-to-date with revision 31. The substantive changes since then are PSK handling and the addition of "rekey" functionality.
> It would be great to update Noise-C for these, since WireGuard uses PSKs, and other people might use rekey .
> But I'll link your implementation on the Website in mid-October in either case (I encourage other people to take a look and send feedback before then).
>  https://github.com/Katzenpost/docs/blob/master/specs/wire-protocol.txt
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Noise