On Wed, Apr 12, 2006 at 11:00:37PM +0100, Stephen Farrell allegedly wrote:
Hi Murray,
What would be wrong with the option of using the first N bytes of the
actual signature value?
If that were ok (and I've no idea really), then presumably the shortest
N that disambiguates the signatures could be used.
That seems like a reasonable strategy to me. I believe that the domain
is also included in Authentication-Results, so if the disambiguation
selection matches on the d= domain + unique signature bits, then most
of the time one could expect that the number of disambiguation bits
will normally be be zero. The exception being the presence of two
signatures with a common d= domain.
Mark.
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html