[dns-operations] NXDOMAIN vs NODATA for suffixes of existing name

[Edward Lewis]

Ah, right.  I read too quickly.

Then, part a applies - "copy all RRs which match QTYPE into the 
answer section" - since the number of matching RRs is 0, there is an 
empty answer section.  However, there is no error (i.e., return code 
remains "no error", = 0).

You wouldn't set the return code to name error because it is only set 
that way in "c", and, as you note, you don't get there.

At 14:27 +0200 4/14/06, Florian Weimer wrote:
>* Edward Lewis:
>
>>  RFC 1034, 4.3.2, step 3, part c, 2nd paragraph, 2nd sent.:
>>
>>  "If the name is original, set an authoritative name error in the
>>  response and exit."
>
>I don't think this part of the algorithm is reached in my example.
>See the first graf of part c:
>
>| If at some label, a match is impossible (i.e., the corresponding
>| label does not exist), look to see if a the "*" label exists.
>
>Apparently, the issue boils down to the question whether the label "b"
>exists under example.net. if there are are records for
>a.b.example.net.  I think it does (and BIND 9 happens to agree with
>me, although this behavior changed in the past IIRC), but I can't find
>a definite reference.

-- 
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Edward Lewis                                                +1-571-434-5468
NeuStar

Nothin' more exciting than going to the printer to watch the toner drain...