I asked about how the b= is handled in my first post commenting
about the DKIM draft:
Why isn't the signature data provided in its own separate
header field to avoid having to extract out the sig data
first and dealing with ambiguities of whitespace? For example,
is the whitespace before and after the "b=" tag also removed,
or only the whitespace after (or before)?
I never saw an answer to this question.
I suspect the answer is that we wish to maintain the capability for multiple
signatures in a message - or - at least - we don't wish to require the
stripping of existing signatures out of messages. Multiple signatures, each
split into multiple headers, present some parsing complexity while a single
header places everything about a signature nicely in one place. I much
prefer that approach myself and don't see what is gained by using multiple
headers per signature. If there are problems with b= and canonicalization
we can fix them without having to resort to using a separate header for b=.
*in detail*, especially wrt to "removing" b= and how surrounding
whitespace and/or semicolon is handled.
Agreed. Can you propose some language to solve this problem?
--
Arvel