With SKS 1.1.5 keyservers return the full fingerprint in response to a search of
a v4 key.
According to http://tools.ietf.org/html/draft-shaw-openpgp-hkp-00#section-5.2
this is ok. But gpgme internally only reserves 16 byte for the keyid that it
parses from the keyserver response. If the response is longer it is ignored and
the key lookup fails.
As the 16 byte restriction is part of _gpgme_subkey changing it would break ABI
compatibility. As a workaround I would propose to just use the last 8 characters
of the fingerprint as Keyid in gpgme. (See attached patch)
Werner do you think this would be ok? (Maybe as a fix until a ABI breaking
version is released?)