OK, fair enough. Just as long as we understand and properly record the
design decision that was made here:
I.e. we're more afraid of the negative consequences of software/OSes
that don't treat null addresses reasonably (i.e. pointless/doomed
retries, possible self-looping) than we are of the negative consequences
of software that is slow or defective in their adoption of the new magic
meaning of root-name MX targets (i.e. pointless/doomed A/AAAA queries of
the root name).
- Kevin
On 7/23/2014 12:16 PM, Tony Finch wrote:
Kevin Darcy <k...@chrysler.com> wrote:
But if we're going to assign "magic meaning" to something, why not
assign "magic meaning" to the null address
*specifically*in*the*context*of*SMTP*message*delivery*strategy*, i.e.
auto-fail messages destined for the null address and never retry them?
Because that will have horrible fallout for existing software that does
not undestand the magic meaning.
@ MX 0 . already has the correct semantics; the only magic in the nullmx
spec is to suppress the A and AAAA lookups because they will always fail.
"must not be used as the destination address of IPv6 packets or in IPv6
Routing headers"
I haven't gone back to see if the IPv4 null address has been similarly
clarified/redefined, because, who still uses IPv4 anyway? :-)
Yes, 0.0.0.0 is only for use as a source address.
Tony.
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop
