I belong to the school which says that, in the case of multiple "exp="
modifiers, no PermError should be issued, but that the ambiquity of both
"exp=" modifiers means they must both be created as if absent. Because this
is not the sort of syntax error that warrants a permanent failure. You can
remove both from scope, without detracting from the "essential nature" of
the SPF record.
Sure, if it does not violate the standard, it is OK. Let me emphasize
again that I suggested PermError according the old (?) SPF draft
version I mentioned.
As Roger pointer out, this rule should be clearly indicated in the SPF
draft then. It is not a good idea to leave the decision to the
developer.
Offtopic question: which version of the draft the developers should
implement *now* if draft-ietf-marid-protocol-00 is outdated?
Peter