On Thu, Jul 03, 2025 at 05:48:21PM +0000, Sipos, Brian J. wrote:
> WG,
> 
> I was looking for, and have failed to find, any requirement in the base COSE
> specification [1] that if the protected header map is non-empty the
> associated algorithm must support additional authenticated data (AAD). The
> non-normative text and examples seem to support this but I don't see
> anything normative around this. Am I just missing something? Or does this
> seem like something that deserves a constraint?
> 

I guess this was what was intended — like intending to only have
authenticated symmetric encryption so to not need algorithm binding —
but it wasn't explicitly written anywhere. The section on AE certainly
forbids protected headers.




-Ilari

_______________________________________________
COSE mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to