On Tue, Sep 18, 2001 at 11:55:29AM +0200, Peter Palmreuther wrote:
> Also referring to RFC 2392 the message-id has to be in syntax a
> 'url-addr-spec' like described in RFC 822. There the syntax of 'url-addr-spec'
> is specified as
You're referring to the wrong RFC (2392). It has nothing todo with
message-id as a *header* of message.
> Unless the 'MyUsername' as wished by Roelof ain't a hostname it would not be
> 'formal correct' to use the period as delimeter. A second '@' ain't valid too
> so unless his e-mail-address-domain ain't 'MyUsername.MyDomain.MyTLD' it would
> not be RFC following to use it.
>
> Of course I _do_ know nobody would feel disturbed if he does nevertheless but
> as a good MUA tries to follow RFCs one can not expect it to (re)write message-id
> in this 'false' way :-)
First, even the appearance of a message-id in the message is NOT a must.
But, if a message-id exist, then it has to guarantee a uniqeness. That's all.
Try to read (a new) RFC2822 carefully.
Then <[EMAIL PROTECTED]> is a valid message-id.
> So the only solution I see is an own MTA that filters incoming mails from
> internal for 'message-id' in header and rewrites it, but unhappily I don't
> know of any that does this.
> I could imagine of 'work arounds' or better 'expansions to existing MTA' e.g.
> qmail but I don't think it's _really_ worth the effort :-)
The MTA that do not alter any header in the message is a good practice(tm).
Because RFC(2)822 state that a message SHOULD has a message-id, then some
MTA that I know of (sendmail, qmail, exim, postfix, zmailer, smail and mmdf)
only add a uniq message-id if and only if such a message doesn't have a
message-id. If it already has, then those MTA would lovely to use the
existing one.
Regards,
Flash
--
________________________________________________________
Archives : http://tbudl.thebat.dutaint.com
Moderators : mailto:[EMAIL PROTECTED]
TBTech List: mailto:[EMAIL PROTECTED]
Unsubscribe: mailto:[EMAIL PROTECTED]
Latest Vers: 1.53d
