On Sun, 26 Jul 1998, Ian Brown wrote:
We have had some discussion on OpenPGP of a flag, perhaps in the literal
packet, which would indicate that messages are in this form. Perhaps a
future version will have a clean way of doing this.
This would also be nice for super-encrypting messages, for something
like IDEA(CAST(plaintext))
Perhaps have an 'n' for nested mode in Literal packets that tells
clients to treat the contained data as another PGP data structure.
Several months ago there was an offline discussion on this topic. PGP
uses the literal packet as a "stop processing here" construct, though that
is not clear from the spec.
The context of the original discussion was: Why have *nested* one pass
sigs, when you can encapsulate a signed message in a literal packet. The
answer is that PGP uses the literal packet as an end-of-algorithms marker,
i.e. dearmor->decrypt->decompress->checksig&extract(stop).
For the next revision of this, I am going to strongly advocate a 'n'
literal data type to add to t and b to tell a PGP implemetation to process
the result of stripping the literal information off.
--- reply to tzeruch - at - ceddec - dot - com ---