[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Epsilons update (long)

We don't have Symbolics DECnet here, so I'm just going to theorize
about why the Symbolics DNA mail server is doing what it is doing.

My suspicion is that VMS machines generally don't use RFC-822 format
messages among themselves.  The DECnet mail protocol sends the header
data separately, and the text is assumed to be the message body, with
no provision for nonstandard headers.  If this is so, the Symbolics
mailer is doing the right thing.

However, in your environment your VAX is also connected to the
Internet, so it receives RFC-822 headers.  It includes them in the
message text that is sent to the Lispm.

I suspect you would be seeing the same behavior if you were forwarding
your mail to another Vax instead of a Lispm.  The difference is that
Zmail makes use of header fields in order to parse the message body,
so it is significant whether certain lines are in the header or the

I'm not sure that there's a good fix to this.  It isn't necessarily
appropriate for the Lispm to try to parse the beginning of a DNA
message as RFC-822 headers.  What if a Vax user were sending you mail
(which wouldn't have RFC-822 headers), and the first few lines of his
message happened to be formatted like RFC-822 headers (RFC-822 is
purposely designed to look like "memo" headers).  It would not be
correct for these lines to be treated as RFC-822 headers.

What's needed is for the Vax mailer to somehow indicate to the Lispm
that the message originated from an RFC-822 environment rather than
from a DEC mail system.  In such a case, the header information passed
as part of the DNA mail protocol would not be used to cons a fake
header ahead of the message text.  This would require an extension to
the DNA mail protocol, though.