🚀 go-pugleaf

In <9305101915.AA12181@crystal.wonderworks.com> kyle@crystal.wonderworks.com (Kyle Jones) writes:
>Joe Kelsey writes:
> > VM needs to change its algorithm.  Among other things, SVR4 mail uses
> > a Content-Length: header to specify the mail body-length and DOES NOT
> > put blank lines between messages.

>To say that I am unimpressed with System V's past and present
>mail handling capabilities would be a marvel of understatement.
>Besides the faft that System V does this, is there another reason
>VM should bother with this content length header?

Well, If you ever want to support full MIME-compliance, then you
definitiely have to change the algorithm.  When you want to parse MIME
messages, especially ones with message/rfc822 bodies, then your
current algorithm fails completely.

Simply because you don't like System V does not mean they haven't
done a good job implementing something.  Using a Content-Length:
header provides a simple, reliable way to separate messages.  Relying
on a rather arbitrary and capricous mail system which puts blank lines
between messages leads to lots of problems actually finding message
boundaries.  My code for VM 4.xx to scan forward considerably
simplifies the lisp code and makes it much more understandable and
readable.

The future of SMTP-based mail is MIME.  MIME support allows simple,
reliable message encapsulations and inter-operation with other mail
transports (X.400 (ugh!), etc.)

Check out MIME (RFC 1341, 1342) and the metamail implementation from
N. Borenstein (get the mm distribution from
thumper.bellcore.com:pub/nsb).  The file ``mailers.txt'' in the mm2.4
distribution contains code to patch VM4 and VM5 to allow a very
primitive interaction with metamail.  With a little more work, VM can
easily parse the MIME headers and become a leader in the multi-media
mail future.

Anyway, sniping at System V does nothing to imporve the situation for
people who use that system.  A truly useful mail reading program needs
to try to accept a wide variety of inputs, not just those that the
author finds somehow more acceptable than others.

I want to implement MIME parsing in VM.  Do you want to discuss it, or
possibly di it yourself?

/Joe