On 20 Apr 2015, Derek Atkins wrote:
Specified Revokers use the (binary) full fingerprint, not the
(truncated) keyID.
I would motion for designated revokers to use (or include) the full
public key. This allows verification of a revocation signature in
combination with a designated revoker certificate *without* the
requirement to retrieve, parse, and verify an entire other key, allows
supporting designated revokers without requiring the possibility to
retrieve keys during import. There are two downsides to consider:
- increased packet size. not by an order of magnitude though, and if
this is a concern the designated revoker certificate can be published
together with the revocation only
- incomplete verification of the designated revoker's key. if we fetch a
key by fingerprint, it might have been revoked before the revocation
was issued, invalidating the revocation signature. this still leaves
the key in question in a very fishy state and it probably makes more
sense than not to consider it revoked.
Another alternative would be adding the public key to the revocation
certificate as a subpacket. Both options have the same downsides, and
none of them stands out as the natural choice to me.
- V
_______________________________________________
openpgp mailing list
openpgp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/openpgp