Jeff Boehlke wrote:
> Perhaps I do not understand what the purpose of
>
> tls_advertise_hosts = *
>
> If this is not possible, then that is what I need to know. I have a sight
> that is using a product called Communigate Server, they are trying to get
> that server to transfer mail between them and my site over SSL and it is not
> working. The company that makes the product states that they support SSL,
> but perhaps they do not correctly send EHLO, because it appears to send HELO
> and then try to use SSL.
Running Exim 4.20 as a daemon in debug mode:
exim -bd -d -q15m
or whatever your queue time is
will show exactly what is happening. I.e., whether Communigate is
sending an ehlo. You should run this in an X-terminal that can roll back
at least 4096 lines, as the output is verbose. Alternatively, you can
redirect the output to a file (exim -bd -d -q15m > /tmp/exim.out 2>&1)
and "less" or "more" that when the session is over.
It would surprise me immensely if any (non-Microsoft, non-NetWare,
non-Lotus etc. though specifically-Unix) commercial MTA did not obey
rfc2821 recommendations - even though the word "should" is used for ehlo
in that spec. It would not surprise me at all if it were a product for
Microsoft etc. servers. It's only by sending "ehlo" that the client MTA
can "see/sense/learn" what the server MTA can offer. That's what the
whole "ehlo" exchange is all about.
Exim's 'tls_advertise_hosts = *' is to let *all* clients (also MUAs)
know that Exim can offer STARTTLS.
Read, mark, learn and inwardly digest just *two* rfcs: 2821 and 2822.
They'll always strengthen your case in any discussion.
Best,
Tony
--
Tony Earnshaw
Looking backwards is always easy with hindsight
http://www.billy.demon.nl
Mail: tonni@???
