[getdns-api] Section 5.3 Types pf RDATA Returned in the API

Paul Hoffman paul.hoffman at vpnc.org
Sun Dec 1 05:14:00 CET 2013


On Nov 28, 2013, at 3:39 AM, Willem Toorop <Willem at NLnetLabs.nl> wrote:

> I have gone through the list of RR types and have a few remarks/suggestions.
> 
> - APL rdata should be a list of dicts containing
>      address_family (an int), prefix (an int), n (an int),
>      and afdpart (a bindata)
> - NSEC3 - remove salt_length (it is already in salt.size)
>        - remove hash_length (it is already in hash.size)
>        (those length/data wire format combinations have a single
>         key in other RRs (TXT, TSIG, TKEY))
> - NSEC3PARAM - remove salt_length (it is already in salt.size)
> - HIP - remove pk_length (it is already in public_key.size)
>      - remove hit_length (it is already in hit.size)
>      - turn rendezvous_servers in a list of bindata (which are dnames)
> - L32 - Replace locator32_msbs and locator64_lsbs with locator32
>        (the msb/lsb indication in the RFC are because of line break)

All of those seem fine to me. I had included the lengths because they were in the RRtype definitions, but I like the idea of just using the API's knowledge of the data it is holding.

> - Why add the _obsolete and _unknown rdata?
>  Is it different from rdata_raw?

It was to show in the API documentation that those RRs are obsolete or that the definitions are unknown. I admit that was probably overloading the names a bit...

> - Add EUI48 and EUI64 (rfc now)

Sure.

> - remove IXFR and AXFR. (The zone is indicated by the owner/query name).

Those were there for completeness. It seems fine to remove them.

--Paul Hoffman



More information about the spec mailing list