[dns-operations] EDNS1 tests in the ednscomp tool

[Chris Thompson]

There seems to  be something a bit amiss in the results reported for the
EDNS1 tests. See, for example this test on ui-dns.biz:

  https://ednscomp.isc.org/ednscomp/0d84de4b75

The edns1 and edsn1opt tests for some servers report a failure as "status,soa",
i.e. status wrong and SOA unexpectedly present. But doing the dig's by hand,
it seems to me that the servers do give BADVERS as they should, and the only
thing wrong is that they also copy the query section in the response.

Is this even categorically wrong? I couldn't find anything by casual inspection
of RFC 2671 or RFC 6891. Even if it is technically wrong, I don't see how
any resolver is going to take the wrong action as a result of receiving such
a response.

-- 
Chris Thompson
Email: cet1 at cam.ac.uk