On Mon, Mar 16, 2015 at 11:25:33AM +0000, Tony Finch wrote:
respond with a CNAME answer if there is one, or respond with NOERROR / NODATA - which is exactly the response you would give to a CNAME query.
This is elegant, except for the base case of what record type to synthesize in the NOERROR/NODATA case. Robert Edmonds pointed out that using NULL is not a perfect solution here. On the other hand, I am not convinced that creating a completely new record type is better than NULL for achieving interoperability with broken implementations. These solutions are both likely to lead to problems. Because NOERROR/NODATA are cached, these should be avoided. But are problems with NULL or a new data type really a good tradeoff against the undesirable behaviour of clients retrying at other servers? -- Andras Salamon [email protected] _______________________________________________ DNSOP mailing list [email protected] https://www.ietf.org/mailman/listinfo/dnsop
