On Dec 18, 2007, at 6:55 AM, Frank Ellermann wrote:
Jim Fenton wrote:
My suggestion: "non-compliant"/"compliant".
If "non-compliant" is actually a "Resent-* as specified in 2822upd
with a twist" (signature didn't survive it), then that's rather
strong.
Any changes to message content might cause a signature to be invalid.
An invalid signature may cause a message to appear "exceptional" when
messages from the domain are asserted as "all" being signed. As with
any retrofit, exceptions _must_ be permitted.
Why not FAIL ? FAIL is short, neutral, and some folks are used to
the idea that FAIL is a defined term.
"FAIL" says little about an underlying cause. Was the signature
valid, but "on-behalf-of" a header other than "From", or by a
different domain, when "strict" had been asserted by the From domain?
A breakdown of causes with respect to actual "exceptions" would help
clarify the different causes that _will_ legitimately exist.
-Doug
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html