🚀 go-pugleaf

In article <843440153snz@tsys.demon.co.uk>, Tom Wheeley
<tomw@tsys.demon.co.uk> writes

>Whilst developing my SMTP server, I finally tested it on the punts (after
>checking no important mail was waiting for me!).  Now, there were two
>messages on one of them, and these seemed to be sent in one session.

this is perfectly proper

>Now,  I had planned for this, but my program did not reset its MAIL FROM and
>RCPT TO buffers

keeping the old data is in fact wise, despite what the RFC may appear to
say. If you get a new MAIL FROM or RCPT TO (more likely) then overwrite
the old data.

> after a piece of mail was sent;

received ?

> I had expected (perhaps by
>misreading the RFCs) that the Clients would have sent a RSET.

no such obligation. 

>Answay, my server responded with a couple of error messages to both the new
>MAIL FROM and RCPT TO commands from the punt.

this was unwise

>  After the RCPT error, the
>punt sent RSET, then QUIT.

this is valid behaviour (the RSET isn't necessary, but that's what the
punts currently do. I think they think it is friendly.)

>Now here's the catch: I fingered post soon after, and the second piece of
>mail had disappeared.  

it was on its way back to the sender

>Now, am I right in saying that *shouldn't* happen?

nope. If you reject mail, it is sent back along the return path

-- 
richard                      richard.clayton    @    T U R N P I K E .com
                                                     tel: +44 1306 732300
"Assembly of Japanese bicycle require great peace of mind" quoted in ZAMM