2nd Try: --fd-status and stderr msgs from gpg

Joerg Toellner toellner@oss-gmbh.de
Thu May 23 14:15:01 2002

Hello List,

as im not having received a line for my problem posted a week ago, please
allow me to do a second try. I cant believe, that nobody here has handled
the --status-fd and can point me in the right direction.

Summary of my problem as posted last week:
I use gpg in our mailserver-gateway for our lan in an unattendet environment
from within some perl scripts. All outgoing mails should be encrypted by gpg
for every email-adress (used as the uid) where a public key is present in
the pubring.

To decide (in my perlscript which opens gpg where necessary) whether the
operation was succesful or not i capture --fd-status and the stderr from
gpg. I listen on the --status-fd, get there the well formatted gpg-msgs
(described in File DETAIL) and interpret this msgs to decide what to do

The problem is, that there are errors that are NOT reported on
the --status-fd but at the stderr. And this msgs are really ugly to
interpret (ok they are more human readable but this is not my intension).

if i decrypt i file i got alot of msgs (DECRYPTION_STARTED, DECRYPTION_OKAY
or DECRYPTION_FAILED a.s.o.) from the --status-fd prefixed with [GNUPG: ]
into my perlscript. All okay here.

But if i export a key in armor mode for sending it via email to someone and
the key is not present (for what reason ever), i got *NO* msgs on the
--status-fd but on the stderr there is a output that says: blah blah key not

The operation has failed and no msg on --status-fd appears in this case
indicating the error :-(. Even as there is a msg for this case
for --status-fd (see File DETAILS) named NO_PUBKEY.

Do i sth. wrong. I thought its enough to listen on --status-fd to know all
about gpgs actual workflow. I really dont want to interpret the stderr
output in addition to the --status-fd if i can avoid it.

Please! This is my second try to get an answer. Be kind and give me some
lines back this time. Im sure there are some good guys in this list that
have managed the error-interpretation for gpg in their scripts and can give
me a little hint.

Joerg Toellner