mpi larger than indicated length

[David Shaw]

On Fri, Jan 27, 2006 at 05:27:40PM -0600, david.t.kerns at us.hsbc.com wrote:

> $ gpg -d -vvv  IFO.SECURE.PGP
> gpg: using character set `iso-8859-1'
> gpg: armor: BEGIN PGP MESSAGE
> gpg: armor header: Version: PsypherOPS 4.30.00 - www.primefactors.com
> :pubkey enc packet: version 3, algo 16, keyid 9E1BA0486180F04C
>         data: [2047 bits]
>         data: [2045 bits]
> gpg: public key is 6180F04C
> gpg: using subkey 6180F04C instead of primary key CDCF6506
> gpg: public key encrypted data: good DEK
> :pubkey enc packet: version 3, algo 16, keyid D2DDC51A2F833978
>         data: [1024 bits]
> gpg: mpi larger than indicated length (0 bytes)
>         data: [MPI_NULL]
> 
> Can anyone shed some light on this?
> Do I need to regenerate my key and send my counter part the new public key?
> Or have I run into a compatibility issue with PsypherOPS?

As you noted, this is unlikely to be related to pks.  It looks like
some interaction between GnuPG and PsypherOPS.

The message was encrypted to two keys, 9E1BA0486180F04C (you) and
D2DDC51A2F833978 (someone else).  The encrypted session key for the
second key (the one who isn't you) appears malformed.  Specifically,
the error says that the pubkey enc packet contained two MPIs, the
first of which seems to be 1024 bits (too small), and the second of
which is missing altogether.  Anyway, that's breaking the decryption.

David