for some reason, I had it in my head that the verbose form of send did not report much. It reports a lot with the -snoop and -watch flags set so here is what is always happening these days as seen by exim, itself and nmh. Some of us still use it;-)
I send a message and here's the log with comments. LOG: MAIN Warning: No server certificate defined; will use a selfsigned one. Suggested action: either install a certificate or change tls_advertise_hosts option #I did generate a cert and it is in /etc/exim4 and I still get this warning, but one thing at a time. LOG: address_rewrite MAIN "martin@localhost" from env-from rewritten as "[email protected]" by rule 1 #so far, so good. LOG: MAIN <= [email protected] U=martin P=local S=487 M8S=0 RT=0.009s LOG: MAIN delivering 1nolOX-0005He-Bd R: smarthost for [email protected] T: remote_smtp_smarthost for [email protected] Transport port=25 replaced by host-specific port=587 #Again, good. Connecting to smtp.mx-altice.prod.cloud.synchronoss.net [208.180.40.68]:587 ... connected SMTP<< 220 omta02.suddenlink.net ESMTP server (InterMail vM.8.04.03.22.02 201-2389-100-169-20190213) ready Wed, 11 May 2022 07:24:45 -0500 SMTP>> EHLO wb5agz SMTP<< 250-omta02.suddenlink.net 250-HELP 250-XREMOTEQUEUE 250-ETRN 250-AUTH=LOGIN PLAIN 250-AUTH LOGIN PLAIN 250-PIPELINING 250-DSN 250-8BITMIME 250-SIZE 52428800 250 STARTTLS #hinse forth, this will be #Banner. SMTP>> STARTTLS SMTP<< 220 Ready to start TLS #This looks right so far but What could possibly go wrong? SMTP>> EHLO wb5agz #Banner once again but the script marches on SMTP>> MAIL FROM:<[email protected]> SIZE=1524 I think we just saw what goes wrong. There probably should have been an "auth login" response sent followed by the authentication sequence but no such luck. SMTP<< 553 Authentication is required to send mail as <[email protected]> #exim doesn't realise that this isn't what should be sent and is proceeding as if all is well. SMTP<< 503 Bad sequence of commands (specify MAIL first) SMTP<< 503 No sender specified SMTP>> QUIT So, the question is, Does this look like I can rearrange things somewhere to correct what is happening? In the slightly less opaque world of text-based expect scripts and thelike, one just makes sure not to get ahead of the responses from the remote site so as to send the appropriate response or envoke a case or switch module. In this case, we seem to be totally missing what SMTP is asking for next. When trying to read the network traffic, all that key changing and restarting was the remains of the broken mail transfer sequence and smtp.suddenlink's server dropping the connection several times. When Suddenlink.net upgraded something around March 14, the banner changed and the server doesn't appear to use Port 465 in any way. I've fixed that situation in that I call for 587 in /etc/exim4/localmacros by editing the appropriate line as shown: MAIN_TLS_ENABLE = 1 REMOTE_SMTP_SMARTHOST_HOSTS_REQUIRE_TLS = * TLS_ON_CONNECT_PORTS = 587 REQUIRE_PROTOCOL = smtps Putting localmacros in or taking it out has no effect and, of course, I am running dpkg-reconfigure exim4-config each time after the change. I must admit that I strangely feel a bit satisfied that one can actually see the place where things go wrong. It's just as important to know what doesn't need tweaking as what does. Again thanks for any and all constructive ideas. Martin McCormick -- ## List details at https://lists.exim.org/mailman/listinfo/exim-users ## Exim details at http://www.exim.org/ ## Please use the Wiki with this list - http://wiki.exim.org/
