Re: [openpgp] Embedded TPK subpacket

[Justus Winter <[e-mail]>]

Vincent Breitmoser <[e-mail]> writes:

>> My proposal is ment to obsolete the existing mechanisms.  The fact that
>> we now have multiple incompatible mechanisms is a bit sad, and I'm
>> trying to extend OpenPGP so that we can have interoperable
>> implementations again.
>
> So what your proposal brings to the table is in-band key distribution without
> MUA involvement, but hinges on the use of signed-only mails.

Why would it be restricted to sign-only messages?  My proposal also
works with OpenPGP's usual sign-then-encrypt messages.  Marcus'es point
was about it not working with encrypt-only messages.

>> For example, if you look at Autocrypt, implementing it means that the MUA
>> needs to do a lot of low-level key manipulations.
>
> Can you elaborate on this? We designed Autocrypt to be as agnostic of OpenPGP
> implementation details as possible, especially for public key management it can
> get away with treating keys as opaque bytes blobs. IINM the required API from an
> OpenPGP implementation should be complete with just "get minimal own public
> key", "check TPK integrity", and "encrypt to keys (given as blobs)". In practice
> OpenPGP support in MUAs tends to be more involved than that, but I don't think
> there is an actual "need" for that.

"get minimal own public key" according to
https://autocrypt.org/level1.html#openpgp-based-key-data seems pretty
involved to me.  I'd be surprised if one can even implement that using
the various OpenPGP implementations out there.  Same for the filtering
of keys to be gossiped.

(I just noticed that I cannot do Autocrypt with my key because my
primary key is not signing-capable...)


Justus