Re: [Cbor] Supporting IPv6 Link-Local with scope (was Re: Éric Vyncke's Discuss on draft-ietf-cbor-network-addresses-09: (with DISCUSS and COMMENT))

[Barry Leiba <[e-mail]>]

Let's try to keep the distribution list trimmed so these messages
don't keep going into moderation.  In this message I've eliminated the
people who are subscribed to the CBOR list.  As it's also going to the
IESG, we could remove the two ADs from the distribution as well,
though I'm leaving them on for now, thinking that having them on
directly might make it easier for them.

Thanks,
Barry

On Thu, Oct 7, 2021 at 7:50 AM Carsten Bormann <[e-mail]> wrote:
>
> On 2021-10-07, at 13:00, Rob Wilton (rwilton) <[e-mail]> wrote:
> >
> > Hi Carsten,
> >
> > Snipped most of it, just one comment inline ...
> >
> >> -----Original Message-----
> >> From: Carsten Bormann <[e-mail]>
> >> Sent: 07 October 2021 11:39
> >> To: Rob Wilton (rwilton) <[e-mail]>
> > <snipped>
> >>> - How often are zones actually used?  Would it not just be simpler to always
> >> use a string interface identifier?
> >>
> >> Ifindex and Ifname are two different concepts that should not be mixed.
> >> The textual nature of RFC 6991 makes it seem natural to always use a textual
> >> representation for the Ifindex; this is not the case here.
> >
> > Okay, so thinking about converting from YANG to this CBOR representation, then would the expectation be that it is okay to use a string that represents an integer?
>
> If I needed to write a converter from YANG ietf-inet-types, I would match on digits only and then create an unsigned integer, and use a text string otherwise.
>
> > And is that logically the same or different?  Or would the expectation be that the zone string should be converted (normalized) to an integer
> > if it can be?
>
> You’d need to be on the system in question, and the conversion result might be rather ephemeral...
>
> > Does any of this need to be specified?
>
> Depends on where the conversion from YANG ietf-inet-types to CBOR 52/54 should be specified.  With zero operational experience, that would seem a bit more speculative to me than I like.
>
> Grüße, Carsten
>