🚀 go-pugleaf

>>>>> On Mon, 11 Jan 1993 14:54:08 -0500, Keith Moore <moore@cs.utk.edu> said:

moore> FYI, MIME does *NOT* use content-length, or anything like it.  The
moore> content-length header is completely nonstandard.

Actually Content-Length may be X.400, which is as much a standard as
the RFC's.  I seem to recall that AT&T Mail has been using
Content-Length since they started, and they are X.400.

moore> The ">From" hack is ugly, no doubt about it.  However, using a
moore> header field to delimit messages within a mailbox strikes me as
moore> even worse, particularly when it breaks interoperability with
moore> other systems using NFS as a mail access protocol.

Using a header field to tell you when the message ends seems to be the
only sane thing to do in a multi-part multimedia world.
Interoperability with NFS mounted maildrops could be a problem, but
(thankfully) it's not ours.


moore> However, it seems like MH wants to support the local mailbox
moore> format, whatever that is...so it might do well for a future
moore> version of MH to be able to grok Content-length headers on
moore> local mailboxes only on Solaris systems.

I've now done that hack.  I'll try to figure out to whom I should pass
it back.

moore> But I'd just as soon they be REMOVED from all outgoing mail.
moore> The presence of Content-length headers in Internet email is
moore> just going to add to the general confusion about what is
moore> standard and what isn't.

That battle was lost long ago.  Lots of X.400->Internet gateways leave
their non-822 headers in the messages.  Sendmail is real happy about
munging and rewriting headers, but I don't know how well it likes to
remove them.  Just ignore them.  :^|.

Andy
--
Andy Sherman
Salomon Inc  -  Unix Systems Support  -  Rutherford, NJ
(201) 896-7018  -  andys@sbi.com or asherman@sbi.com
"These opinions are mine, all *MINE*.  My employer can't have them."