Tony Hansen wrote:
Jim Fenton wrote:
>>Is the missing semicolon at the end of the "z" tag incorrect? From
>>the ABNF, it looks like the final semicolon is optional. What are we
>>dealing with here, pascal or C?
>>
>>
>William didn't include the entire header when he quoted the example.
>The b= tag/value will always be the last one, so the z= value must
>terminate with a semicolon.
The b= tag/value is not required to be the last one. The only
requirement is that the signature be computed as if the b= value were
empty. The order of the tags is not mandated anywhere in this version of
the dkim-base draft (or earlier ones).
I don't know where I got the notion that b= had to be last; thanks.
There is one ordering rule: the v= tag/value in the key record must be
first.
-Jim