For MARID, I think it is extremely important for reasons Pete write
between the lines that the MX and the new RR type have exactly the same
owner, so they exist by definition in the same zone.
paf
On May 19, 2004, at 06:15, Greg Connor wrote:
--Pete Resnick <presnick(_at_)qualcomm(_dot_)com> wrote:
So let's say that all SPF TXT records start with "_spf", like
"_spf.example.com". Now let's say that I've got a domain that has
names
like "mail1.sales.example.com", "mail2.sales.example.com",
"unix.support.example.com", "mail.marketing.example.com", and that's
what
a recipient will be using for the SPF lookup. So I want an SPF record
that will match "*.sales.example.com". How do I make such a record?
Matthew's flippant response aside, I agree with Pete, this is a
problem. Creating extra labels is a great way to avoid conflicts with
existing labels, but it totally defeats wildcards. I'm not quite
ready to take wildcards off the table just because of someone else's
preference.
Wildcards may turn out to be important, because we haven't yet ironed
out the relationship between domains and their subdomains, LMAP-wise.
Does the LMAP record for mycorp.com also protect mail from
@www.mycorp.com? How about demon.co.uk and fruitbat.demon.co.uk?
Even if there is some kind of inheritance (which there very well might
be, but it has to be done carefully) -- even with that, there are
going to be cases where you might want to wildcard your entries...
like, if you already have a wildcard A record or MX record and you
want those made-up names to have a different LMAP info from the base.
My feeling is that reusing TXT (or SRV or A records for that matter)
is a bit of a hack. It is acceptable, but not optimal. Shifting the
label to add _spf or _ep or whatever is even more of a hack. BOTH of
these are considerably-less-than-perfect replacements for getting our
own record type, which is IMO the RIGHT answer.
--
Greg Connor <gconnor(_at_)nekodojo(_dot_)org>