Hi Jim, > -----Original Message----- > From: Jim
> On Wed, Aug 27, 2003 at 04:43:41PM -0400, Larry Gilson wrote: > > And shouldn't the first received line indicate > > that the host that sent the message? > > Not necessarily, > > for example, I use a single copy of Mutt to send mail from > address is several domains. Each outbound message will use > the same HELO name because that is configured in the MTA, but > the Message-Id (generated by the MUA) is based on the domain > name of the e-mail address I happen to be using at the time. But the Message-Id is guaranteed by the host and not the MUA based on sending domain. Regardless of which domain you are sending from, you Message-Id should contain the sending host (@SHOST). RFC2822 indicates that intent of the message is considered when modifying the Message-Id and not on the origination. As Dave pointed out to me though, the RFCs in this area are recommendations only. Adhering to a recommendation does not constitute compliance just like not adhering to a recommendation does not constitute non-compliance. > I suspect similar problems would arise with an ISP scenario, > where they could have clients with virtual domains sending > mail out through their SMTP server. Even if the server is virtual hosting, the Message-Id would be better described as @SHOST.domain regardless of the sending domain. Uniqueness would not be guaranteed by the domain any better than the host and basing it on the host would better tie in with the audit trail. --Larry ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk