On Thu, 10 Oct 2019, Ladislav Lhotka wrote:
Wouldn't this define what you need, without hardcoding all the valid
values from the snapshot of the IANA registry?
This would instruct the implementor to go to the IANA registry, notice
there what is obsoleted/deprecated, and they will know they will have
to check IANA when doing a release update.
Am I misunderstanding something?
This may be acceptable for humans but tools cannot do much with it. One benefit
of an explicit enumeration is that tools can generate sensible code from it,
e.g. to provide the user with the list of available choices.
You can use tools to parse the IANA registry, but yes. The work _should
not_ be to blindly pick up an obsoleted/incomplete set of data that is
deemed valid a few years ago in time. People will read your enumeration
and start coding support for all those things because those are
"implemented" by you and then we see people implemented things like
the MD or MF or A6 RRTYPEs just because your yang model listed these
obsoleted record types.
So I would argue this is a feature, not a bug.
Paul
_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop
