At 06:50 AM 5/10/2005 -0400, Mark Shewmaker wrote:
On Mon, May 09, 2005 at 04:19:17PM -0700, David MacQuigg wrote:
> At 05:46 PM 5/9/2005 -0400, Mark Shewmaker wrote:
> >
> >Since if you trust your incoming mail server you can always trust
> >top-most authentication results in example A but not in B, IMHO it makes
> >sense as a blanket rule for MTAs to always add their authentication
> >headers so they appear above their added Received: header.
>
> Don't we have the same fundamental problem drawing the line of trust at a
> Received header?
No, because you can always trust the topmost received line of your trusted
incoming mail servers.
All proper mail servers add a topmost received line, but not all proper
mail servers add a topmost authentication header, so without additional
user-side configuration or other information outside the content of the
mail message itself, the only things MUAs can in all cases trust are
things in and above the topmost received header.
I'm assuming a situation where all trusted forwarders add a standard
authentication header, even if the authentication they do is only capturing
an IP from the previous forwarder. A forwarder that doesn't do at least
that will not have a good reputation anyway. I wouldn't re-order the
headers to accommodate non-compliant forwarders, but again, this is a minor
preference that I wouldn't hold up a standard over.
The header-scanning algorithm I have in mind is:
for hop = 1 to MAXHOPS:
domain = ( extract from next Authent: header )
if domain is a trusted forwarder:
continue with next hop
else:
break
query for reputation on domain
--
Dave
************************************************************ *
* David MacQuigg, PhD email: dmquigg-spf at yahoo.com * *
* IC Design Engineer phone: USA 520-721-4583 * * *
* Analog Design Methodologies * * *
* 9320 East Mikelyn Lane * * *
* VRS Consulting, P.C. Tucson, Arizona 85710 *
************************************************************ *