Re: [exim] Bogus HELOs

Greg A. Woods wrote:

>However there's no magic about it -- just call getsockname() to find the
>local address the outgoing connection is bound to. Indeed some MTAs
>already include at least primitive support for doing this right
>(including Exim, IIUC), and hacking in getsockname() and gethostbyaddr()
>to find the name automatically would be trivial in most MTAs too
>(including Exim).
>
But in practice, what "would be trivial" is not going to happen unless
every vendor makes it so. This would include going to each customer
site and updating FOR them.

>However the last time I saw, or even heard of, a true multi-homed SMTP
>gateway (that was not a gateway to a private network (*)) was about 15
>years ago.
>
My server, inferno.inside: 192.168.2.4, is also...
x.x.177.3 (dns) (and ntp, but reverse is for dns)
x.x.177.16 (mail)
x.x.177.106 (www)
x.x.177.111 (webmail)
which isn't to say that I don't still use mail.corvu.com in my HELO
lines (although the default setup for Exim [used to?] allow mail to go
out on any bound IP with a suitable default gateway).
driver=smtp
interface=IP_ADDRESS_HERE

>In any case there really is no valid excuse for any MTA to utter the
>wrong hostname when it greets some remote SMTP server on the public
>Internet.
>
>
Valid excuse, as in - "I plugged it in and the default settings already
work?" (The mantra of the Windows admin).
Also, of course, assuming Exchange servers didn't exist...
EHLO srv_exchange.domain.com
...I would agree with you. They "should", but they don't, and that
won't go away for a long, long time.

My point remains. For a business, it's impractical to expect that
everyone else has done it the same way you do. But do what you want.