Pete Resnick wrote:
On 6/12/08 at 4:05 PM +0200, Arnt Gulbrandsen wrote:
Pete Resnick wrote:
You will note that 2822 requires you to accept messages with *none*
of these fields. (See section 4.)
I don't see that. Elaborate?
I do see that 2822 section 4 requires me to accept lots of additional
"syntax forms", but I do not see where it requires me to accept an
empty string and interpret that as an email message.
I guess I will back off of that statement: I had always assumed that
since section 4.5 relax the restriction on multiple occurrences of
fields, it also relaxed the minimums. But it never says that, and unlike
section 3.6, it does not add a new table and it does not make any
statement about which (if any) fields are required.
That said, I will stick by the statement that 822 (and 2822 for that
matter) make no claims about what a receiver ought to do when it
receives a message without the minimum required fields.
Well, there is an implied semantic with the term "minimum requirement".
But the issue is not when there are no x822 header - *none*. That's
simple. Local systems always handle this one way or another i.e. smtp
servers acting a "posting" device.
The issue in production/practice is two folds:
- incomplete, and
- supporting 2822 vs 822
and I guess I can add the "renewed" focus and attention towards proper
compliance in an era of high abuse/spam.
In addition, in our case:
- Filling in the missing header, versus
- Writing a new required set of header.
Anyway, I missed 2821 6.3 (thanks SM) and it covers everything I was
thinking, the concerns, the issues, the realities.
--
Sincerely
Hector Santos, CTO
http://www.santronics.com
http://santronics.blogspot.com