Frank Ellermann wrote:
Alessandro Vesely wrote:
I would not call that a "fix".
Forwarding mail from server scripts often results in messages
transferred with an empty MAIL FROM, RFCs notwithstanding.
Forwarders are unrelated to RFC 3834
Yes, I just meant to give one more example of empty sender, apparently
violating rfc2821[bis].
Some even hypothesize a BCP candidature for such "modernism"...
I have never heard of that idea, and don't think that it is good,
I heard the idea in discussions about the
MAIL FROM: <> AUTH=original(_dot_)sender(_at_)example(_dot_)com
The rationale is that, even if the target server does not understand
the provided AUTH field, an empty sender is more convenient for both
parties. While I think it may break the original SMTP design, I must
acknowledge that spam drives people crazy.
but of course adding "or BCP" to "standards-track" would be okay
in 4.5.5.
Even if that is fine, it doesn't address the real problem. I know that
the decision of not addressing the spam issue was made much ahead.
However, it may make sense to try and assess its possible
consequences, now that the process is nearly but not yet concluded.