Compromise on Issue #84

["Hal Lockhart" <hlockhar@bea.com>]

In an effort to resolve issue 84, I will make a proposal which may address
Tony's concerns.

To review, I proposed that order of encryptions and signatures be solely
determined by the order of the headers. This allows processing without
lookahead. As a corollary I proposed that the decryption transform is not
needed and should be disallowed.

It is my understanding that Tony's concern is the decryption transform may
sometimes be required.

Although I cannot exactly think of a case where this is so, I would be happy
to allow the decryption transform to be used  as long as the header can be
processed linearly without lookahead.

Does that work for you Tony?

Hal