Renaud Allard wrote:
>
>
> On 07/08/09 23:55, John W. Baxter wrote:
>>
>>
>>
>> On 7/31/09 2:29 AM, "Adam Funk"<a24061@???> wrote:
>>
>>> On 2009-07-30, Heiko Schlittermann wrote:
>>>
>>>> Adam Funk<a24061@???> (Do 30 Jul 2009 16:49:08 CEST):
>>>> ...
>>>>> curious about the second: can a message really get to the delivery
>>>>> point without a Message-ID header?
>>>>
>>>> If I remember well, only in submission-mode exim should do some
>>>> fixups (Message-ID, Date, Sender, ...).
>>>
>>> Yes, I remember setting that option a few years ago on my outgoing
>>> exim. ISTR (but I could be wrong) that I needed to do it because
>>> messages were getting rejected by subsequent mail servers for not
>>> having MIDs.
>>>
>>
>> [Catching up]
>> Unfortunately, having a Message-Id: header is still a SHOULD, even in RFC
>> 5322. So one really ought not to reject based (only) on their lack. It
>> would
>> be very nice if I could. (And if running a server only for myself, I
>> likely
>> would, with provision for an exception list.)
>>
>
> A SHOULD in the RFC, is about the same a MUST. That means, with SHOULD,
> you are not "forced" to implement it, but if you don't, expect to have
> problems.
>
> 3. SHOULD This word, or the adjective "RECOMMENDED", mean that there
> may exist valid reasons in particular circumstances to ignore a
> particular item, but the full implications must be understood and
> carefully weighed before choosing a different course.
>
On the servers I've setup, I set up rules that for each RFC the sender
breaks (whether a should or must) it scores a bunch of points. Score
too many points and either the sender get rejected at SMTP time or if
past that, either they go to into the bit bucket or quarantine (if the
client requests it)
About 80% of the crap floating around don't even make it halfway through
the rules before they have enough points to get rejected.
HTH
