[Oberon] Refine *Mail*.Mod

[W B Hacker]

easlab at absamail.co.za wrote:
> edgarschwarz wrote:

*trimmed a bit...

> # telnet smtp.absamail.co.za 25
> Trying 196.35.40.10...
> Connected to smtp.absamail.co.za.
> Escape character is '^]'.
> 220 absamail.co.za Novonyx SMTP ready $Revision:   3.22.1.16  $
> ehlo

Should furnish "ehlo <hostname/domain>.<tld>" if acting as a peer MTA
Submission client string is less rigourous, but should arrive on port 587

> 250-absamail.co.za Pleased to meet you
> 250-HELP
> 250-EXPN
> 250-PIPELINING
> 250-8BITMIME
> 250-DSN

> 250-AUTH LOGIN   <--- the Tx-authentication type available [per M$!]
Actally that one is the standard

> 250-AUTH=LOGIN

The one with embedded '=' sign is the WinWOES smudge
Not all hosts support it (none of our MTA do)

*trim*

> pasted. Importantly [AFAIK] telnet protocol expects CR/LF,
> which linux > telnet does.  I wonder if N-O telnet does too ?

Yes. All the over-the-wire protocols were originally based on
what various teletype machines used, mostly KSR-15 ASR-33/37

These needed *both* a CR AND an LF, sometimes used separately
for rudimentary X-Y positioning.

Unix OTOH, was developed in a DECWriter & Line Printer environment,
which had a combined-action 'newline'. The dot-matrix DECWriter
had separate X-Y positioning capability, could do graphics.

The 'newline' was all you had on drum and chain 'line' printers,
'coz there was no 'head'- just a hammer at each column position.

> 
> Since the telnet access to pop and smtp servers is so simple and
> exactly as per rfc, I'm wondering how it is that my '90's V2.3.2
> Mail.Panel would 'not connect'. I remember suspecting some subtle
> timing problem[s], which seemed to be confirmed when the DOSbased 
> V2.3.6 did connect without any apparent change in the source code ?
> 
> .. to be continued ....as a proposal HOW to fix ETH-mail...
> 
> 
> == Chris Glur.
> 

I can advise on setting up a local 'dummy MTA' with Exim to run
tests against if it would help.

Most other MTA are harder to 'instrument' so you can see what's up.

With a *BSD/Linux test box, it can be done with two IP aliased onto
the same NIC, doesn't even need a cable.

Otherwise, second local box, internal LAN, dummy DNS in ~/hosts.

That way you could see both sides of the connection
and also alter parameters to emulate different MTA 'strictness'

Too much mucking about during testing might otherwise get you
blacklisted with your ISP.

;-)

No need to hammer the outside word until it is working,
but when you reach *that* stage, I have MTA-test R&D
servers online also.

Bill Hacker