Re: Rejecting null sender from my clients

Bill Cole wrote:

Not necessarily. For example, the Sendmail "vacation" program can be
configured to use <> as the SMTP envelope sender for its out-of-office
messages. As long as you put your real address in the From: header,
I don't see how you can call that fraudulent.

I don't think 'vacation' really qualifies as a MUA,

Why not? One definition of MUA (from Wikipedia:
http://en.wikipedia.org/wiki/Email_client) is as follows:

"An e-mail client, also called a mail user agent (MUA), is a computer
program that is used to read and send e-mail."

vacation "reads" the mail itself, and certainly sends mail. :-)

In any event, there are (traditional) MUAs out there that include
vacation's functionality.

and I am also not
convinced that its behavior is definitively good just because it is
ancient and distributed with Sendmail.

The reason you'd want to use <> is if you want to avoid a bounce in
response to your message being non-deliverable or delayed. There
are many legitimate reasons for this, quite apart from DSNs.

[...]

The use of the null envelope sender for 'vacation' replies is somewhat
reasonable, because the program is designed to sit in the delivery chain
and require the envelope sender in a "From " line (NOT a "From: "
header) so that a fully informed decision can be made about how and
whether to send a response. A post-delivery tool (like MailWasher, te
best-known implementation of this misfeature) cannot reliably determine
what the SMTP envelope sender was, and so cannot reliably generate a
correct DSN. All too often, they also end up generating fake their
DSN's so that they seem to have come from the last SMTP hop rather than
where they really originate, essentially forging a DSN appearing from
the mail provider that the mail provider knows nothing about.

Well, then you should block messages from misconfigured auto-responders
(which should be reasonably easy to detect) rather than imposing a
blanket ban on <> as an envelope sender. You seem to equate "Envelope
sender is <>" with "Message is a DSN", which is not necessarily the
case. (Kari Hurtta posted a counter-example on another branch of this
thread.)

Regards,

David.
.

Relevant Pages

  • Re: Rejecting null sender from my clients
    ... As long as you put your real address in the From: header, ... One definition of MUA (from Wikipedia: ... The reason you'd want to use is if you want to avoid a bounce in ... correct DSN. ...
    (comp.mail.sendmail)
  • Re: Imagination
    ... I know of no reason I need to adjust my post to the level ... To limit spam harvesting it is not in the posting header. ... post exact recipes. ... But there is a real issue--- Diabetes has damaged my hands ...
    (alt.support.diabetes)
  • Re: offsetOf
    ... in that file (only using pointers, the header with the struct definition ... When you use opaque types, translation units which work with that type ... single common header file. ... a normal consequence of that bad decision, and is in fact the reason why ...
    (comp.lang.c)
  • Re: where is a copy of atlstr.h ??
    ... > compilers, there is no reason, at all, for stdio.h to be a file. ... > with one previous version of Visual Studio, MFC and ATL headers were in ... A header is useless in isolation. ... > 1) If your position on missing header locations is so reasonable, ...
    (microsoft.public.vc.language)
  • Re: One to One - Multilple tables for large field qty
    ... If I add the third one-to-one, the primary key ... If you were advised to normalize your data, there's a reason. ... On the header table there is a form ID (primary ... joining the same FormID. ...
    (microsoft.public.access.tablesdbdesign)