Re: [MLS] [interim_transcript_hash]

Hi Pascal,

Note that it is a technicality due to a set of successive change that lead to
the current design, our proofs don’t rely on such complicated GroupContext
and interim transcript but only on the hash of the transcript itself.

I promised I would spend the time on fixing this but I haven’t find it yet.
I’ll do a PR to fix it within the next few weeks.

B.

> On Oct 11, 2019, at 3:53 PM, Richard Barnes <rlb@ipv.sx> wrote:
> 
> Because if you did that, then you would have a circular dependency:
> 
> interim_transcript_hash
> <= MLSPlaintext.confirmation
> <= confirmation_key
> <= GroupContext (via DeriveSecret)
> <= interim_transcript_hash
> 
> This was a problem in earlier versions!
> 
> On Thu, Oct 10, 2019 at 10:02 Pascal Junod <pascalj=40snap.com@dmarc.ietf.org <mailto:40snap.com@dmarc.ietf.org>> wrote:
> Hi !
> 
> Just wondering whether there is a reason why the interim_transcript_hash value (necessary to build Welcome messages) value is not contained into the GroupContext structure. 
> 
> Best, 
> 
> Pascal
> _______________________________________________
> MLS mailing list
> MLS@ietf.org <mailto:MLS@ietf.org>
> https://www.ietf.org/mailman/listinfo/mls <https://www.ietf.org/mailman/listinfo/mls>
> _______________________________________________
> MLS mailing list
> MLS@ietf.org
> https://www.ietf.org/mailman/listinfo/mls