Re: Out of order responce from Smail3 when PIPELINING

[Eugene Crosser <crosser@online.ru>]

> > I am getting sessions of the kind as below.  Order of responces from the
> > remote seems inadequate.  Can someone confirm that this is a bug in
> > Smail? 
> 
>      Nope, it works exactly as specified.
>      (I have narrowed the transcripts somewhat)

The result is that the sender gets non-delivery report with transcript:
"503 Need MAIL FROM command first."
which it totally confusing...  Not a clue about the real source of the
problem.  Not good.

> > #       (Connecting to `mail.glasnet.ru' [193.124.5.37]  ...
> > r       220-mail.glas.apc.org Smail-3.2.0.96 (#11 DST-Aug-25) ...
> > r       220 ESMTP supported
> > w       EHLO deimos.sovam.com
> > r       250-mail.glas.apc.org Hello deimos.sovam.com ...
> > r       250-EXPN
> > r       250-SIZE 3145728
> > r       250-8BITMIME
> > r       250-PIPELINING
> > r       250 HELP
> > w       MAIL From:<guard133@online.ru> BODY=8BITMIME SIZE=3200904
> > w       RCPT To:<bpl@bpl.ru>
> > w       DATA
> 
>      In PIPELINING the preceeding 3 lines are buffered (in stdio!)
>      before they are flushed into the SMTP socket.
>      The flush-point is here at sending DATA command.
> 
>      A long latency connection will thus get MAIL/RCPT+/DATA
>      commands sent, and replies to them received in overlapping
>      fashion.  It will start to study the incoming responses now.
> 
> > r       552 message too large
> 
>           ("MAIL FROM:<..> .. SIZE=3.2 MB")
>      In fact the smtp-ta should not send this message there at all,
>      as it can do the comparison too from the EHLO response.
>      A minor (?) omission to fix regarding SIZE processing.
> 
> > r       501 RCPT TO must be preceded by MAIL FROM command.
> 
>           ("RCPT TO:<..>")
>      It was sent in pipeline, and because MAIL FROM reply was
>      processed delayed, so was this one.
> 
> > r       503 Need MAIL FROM command first.
> 
>           ("DATA")
>      Like above.
> 
> > w       RSET
> 
>      Per specification.
>  
> > Eugene
> 
> /Matti Aarnio <mea@nic.funet.fi>