Re: [Rats] eat-claims definition

I think you are right and we should change to:

eat-claims = {
   ? ueid-claim,
   ? nonce-claim,
   ...

LL

> On Nov 6, 2020, at 5:02 AM, Thomas Fossati <tho.ietf@gmail.com> wrote:
> 
> Hi Laurence, all,
> 
> I'm giving the latest EAT version a go and I got slightly puzzled by the
> fact that the CDDL seems to allow claim duplication (at least
> theoretically):
> 
> ~~~
> eat-claims = { ; the top-level payload that is signed using COSE or JOSE
>    * claim
> }
> 
> claim = (
>    ueid-claim //
>    nonce-claim //
>    origination-claim //
>    oemid-claim //
>    security-level-claim //
>    secure-boot-claim //
>    debug-disable-claim //
>    location-claim //
>    age-claim //
>    uptime-claim //
>    submods-part //
>    cwt-claim //
> ;    generic-claim-type //
> )
> ~~~
> 
> Now, CBOR does not allow duplicate keys in a map, and that is re-stated
> in Section 4.4.1 of EAT - BTW, I guess a ref to 7049 should be enough,
> rather than duplicating text?
> 
> With JSON the situation is slightly more nuanced (8259 a "names SHOULD
> be unique"), but still it makes total sense to avoid duplicates if one
> wants to maximise interop across different stacks.
> 
> Besides, if we aim at isomorphism between serialisation formats we need
> to go for the LCD here.
> 
> So, wouldn't something like:
> 
> ~~~
> eat-claims = {
>    ? ueid-claim,
>    ? nonce-claim,
>    ...
> }
> ~~~
> 
> more precisely describe what we actually want?
> 
> cheers!
> -- 
> Thomas
> 
> _______________________________________________
> RATS mailing list
> RATS@ietf.org
> https://www.ietf.org/mailman/listinfo/rats
> 