In message <563bc83c.3050...@bellis.me.uk>, Ray Bellis writes: > > > On 06/11/2015 00:54, Niall O'Reilly wrote: > > > Wouldn't it be a simpler clarification to say that a client MUST NOT > > depend on the order of the RRsets in an answer? > > That's what I meant (blame jetlag). > > Ray
The RFC 1034 actually says: If the data at the node is a CNAME, and QTYPE doesn't match CNAME, copy the CNAME RR into the answer section of the response, change QNAME to the canonical name in the CNAME RR, and go back to step 1. Otherwise, copy all RRs which match QTYPE into the answer section and go to step 6. Now I fail to see how one can "copy the CNAME RR" on pass one and then "copy all RRs which match QTYPE" on a later pass and end up with the "second set of records before the CNAME". This is not "add to a list to be later copied to the answer section". This is do it right now. Now I know that many implementations just add records to lists and then render them all at once but if that does not produce the same results as just copying them when instructed then the implementation is broken. I also know that resolver should be more robust. I also know that there are a many magnitudes more resolvers than servers and that fixing (or declaring the servers broken and then fixing for Andrew) is the best thing to do. Mark > _______________________________________________ > DNSOP mailing list > DNSOP@ietf.org > https://www.ietf.org/mailman/listinfo/dnsop -- Mark Andrews, ISC 1 Seymour St., Dundas Valley, NSW 2117, Australia PHONE: +61 2 9871 4742 INTERNET: ma...@isc.org _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop