On 2026 Nov 17, at 12:45, Anton Sokolov <[email protected]> wrote:
> 
> Hi Carsten,
> 
> Because the serialization draft says so.

Thanks, that was the answer to the information question I was posing.
Maybe we (CBOR WG) can still get this fixed.

> No implementation problem led to it.

Good to know.

> The one NaN surprise was in our tests: struct in Python 3.13 drops the 
> payload of a half-precision NaN, and in 3.14 keeps it, so the reference code 
> decides from the encoded bits.

Interesting.
Fortunately, the operations that are really needed in a CBOR implementation to 
support NaNs are just narrowing and widening, for which simple reference code 
is available.

Grüße, Carsten

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

Reply via email to