On 2002-02-06 at 17:11 -0500, Greg A. Woods wrote:
> Nope, not better -- bogus and broken by definition. In such an scenario
> you MUST have a PTR for every hostname that points to a given address.
>
> I.e. your example is missing the following critical PTRs:
>
> 3.2.1.123.in-addr.arpa. IN PTR mail.bar.net.
> 3.2.1.123.in-addr.arpa. IN PTR mail.foo.org.
>
> (yes, you really "MUST" have multiple PTRs in such a scenario!)
Interesting.
Reference for the mandate, please?
> That's it. Nothing more, nothing less. Any attempt to define A RRs for
> every virtual mail domain is literally stupid and self-limiting. Same
> for virtual HTTP domains -- just use the 'Host:' header! (there is no
> such thing as a "virtual" HTTPS domain -- you need separate IP#s for
> them, so no worries there)
Uhm, off-topic for exim-users, but I think that somewhere in the newer
specs there's something about virtual HTTPS stuff, so long as there
exists matching forward and reverse DNS; at the moment, I don't recall
where I discovered this and I don't have time to dig, sorry.
--
Hal, open the file
Hal, open the damn file, Hal
open the, please Hal
