On 7. Jul 2024, at 08:40, Carsten Bormann <[email protected]> wrote:
> If the structures can contain floating point values (e.g., via a new registry 
> entry), some of 4.2.2 may be needed.
> (If deterministic serialization for general data structures is needed it 
> would be best to normatively reference draft-ietf-cbor-cde.  Do not restate 
> or second-guess.)

(I don’t understand Section 5.3 of draft-ietf-cose-key-thumbprint-04 [1].
How does the “defined”-ness of the format enter the thumbprint?
We seem to need Section 4 for each existing format [which all work just with 
Section 4.2.1 of RFC 8949].
Who writes the new subsection for the additional “defined” format?)

Grüße, Carsten

The cbor-diag form of the first example in Section 6 [2] breaks because 
1decade2facade3 is not an even number of hex digits.
Please use RFC 8792 for the long lines or break the h’…’ strings more readably 
(not in the middle of a byte!), like you did in the second example in Section 6.

The hex form of both examples uses an odd number of hex digits per line, 
leaving a line break in the middle of a byte.
That confuses tools.
Preferably use the default setting of xxd -u -p (60 columns).

Oh, and please do use proper section references.

[1]: 
https://www.ietf.org/archive/id/draft-ietf-cose-key-thumbprint-04.html#section-5.3
[2]: 
https://www.ietf.org/archive/id/draft-ietf-cose-key-thumbprint-04.html#section-6-3
_______________________________________________
COSE mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to