Hi all,
Jon and I got on the phone today and discussed the AEAD chunk situation a
little more. The compromise we reached is:
(1) MUST support chunks up to and including 8KiB.
(2) SHOULD NOT emit chunks larger than 8KiB.
(3) SHOULD reject chunks larger than 8KiB.
Jon likes this because it doesn't fully shut the door on experimentation. I
like this because it strengthens the norm and puts the burden of justification
for violating the norm on the person asking for huge chunks rather than the
implementer. I expect (though please correct me if not) that Derek, Neil, and
others working on embedded or otherwise constrained system will like this
because the normative limit is low enough for most embedded/constrained systems
to do streaming when incremental no-release semantics are desired. And I hope
Werner likes this because GnuPG is already doing 8KiB chunks, so the work
involved in changing GnuPG's implementation should be minimal :)
Does anyone have any objections to or comments regarding this language?
Thanks,
Bart
signature.asc
Description: OpenPGP digital signature
_______________________________________________
openpgp mailing list
openpgp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/openpgp