On Sun, Oct 25, 2026 at 9:17 AM John Levine <[email protected]> wrote:

> It appears that Wei Chuang  <[email protected]> said:
> >There is some difference in how this multiple signature support might
> >happen.  If I understood correctly, during Interim -06, Richard suggested
> a
> >more feature-rich syntax built into the "s=" tag's separators.  I
> interpret
> >this to mean that it would allow '|' to indicate an OR evaluation and '+'
> >to indicate an AND evaluation, presumably in sequential order.  I propose
> a
> >simpler approach using the "f=" flag tag to slightly simplify evaluation
> >and improve readability.  By default, multiple signatures evaluation is
> >OR'ed.  If an "f=" tag has a value of "and", then the multiple signatures
> >evaluation is ANDed.  The reason for the default ORing and flagged ANDing
> >is to improve the readability of the AND case by making it explicit.
>
> Hi.  For the umpteenth time, what is a validator supposed to do when it
> gets
> an AND of a scheme it understands and one it doesn't?
>

For interoperability, a receiver can treat an unknown algorithm signature
as local policy passing i.e., with a SHOULD and Authentication-Result of
"policy".  If a vulnerability has been announced or for some operator
defined reason, that same local policy can set the Authentication-Result to
"softfail".
-Wei


> This really looks to me like a fatal flaw that will prevent algorithm
> rotation from working.
>
> R's,
> John
>
_______________________________________________
Ietf-dkim mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to