Re: bug with the Message-ID ?

[Brian Stafford <brian stafford uklinux net>]

On Mon,  3 December 12:27 M. Thielker wrote:

> Basically, RFC 2822 does not require a message id on a message that is ...

 From RFC 2822

3.6.4. Identification fields

    Though optional, every message SHOULD have a "Message-ID:" field.

In an RFC, statements like this really mean :-

    Every message must have a "Message-ID:" field.

When SHOULD is used in upper case, it means there must be a *very* good reason 
to disregard the requirement.  It is only slightly weaker than MUST and does 
not imply an optional feature.  In this case I read the requirement as meaning 
a message generated by legacy software should not be rejected as mal-formed if 
Message-ID is missing, i.e. it allows old compliant clients to be considered 
to be still compliant post-RFC 2822 publication.  It does not permit new 
implementations to omit the Message-ID header from messages.

> ... _submitted_ via SMTP and Balsa, by definition, does not relay.

RFC 2822 is the message format standard.  It does not discuss message 
submission or relay.

> So, if it's a confusing checkbox that you want to avoid, just have Balsa 
> never generate message id's. That would make it work with all mail systems, 
> including CIS.

I disagree with this point of view, it disregards the RFC 2822 directive.

--
BCS