[Esd-l] SECURITY_NOTIFY_RECIPIENT
Paul Thomas
paul at cuenet.com
Mon Sep 10 12:13:01 PDT 2001
John,
I don't know what your plans are for the SECURITY_NOTIFY_RECIPIENT
feature and you have your hands full enough as it is, but this
is what seems to work for my situation, at least to start on for
the moment:
In html-trap.procmail I have changed:
TO=$LOGNAME
:0
* LOGNAME ?? ^root$
{
# If $LOGNAME is root, we're probably running as a gateway filter:
# get the "real" to name(s) out of the message headers.
:0
* ^To: +\/.*
{
TO="$MATCH"
}
}
to this:
TO=$LOGNAME
:0
* LOGNAME ?? ^(root$|daemon$)
{
# If $LOGNAME is root, we're probably running as a gateway filter:
# get the "real" to name(s) out of the message headers.
:0
* ^To: +\/.*
{
TO="$MATCH"
}
}
because I'm running as daemon in my configuration. Perhaps something
can be done with daemon_bias to handle various daemons?
Then after that I added:
RECIPIENT="$TO"
So then hop down a few lines to:
:0
* SECURITY_NOTIFY_RECIPIENT ?? [^ ]
{
LOG="${NL}NOTIFY RECIPIENT${NL}"
# We could stuff this directly into $DEFAULT but then
# we'd have to worry about generating Message-ID and Date headers...
:0 h ci
| ( \
echo "To: <$LOGNAME>";\
and change the above line to:
echo "To: <$TO>";\
I know Cc: needs to be checked and even if To: & Cc: are checked and
combined, only one of the addresses is the actual destination address
if any of them are, and if the recipient is listed in Bcc: welll we
miss it altogether.
In my Sendmail style headers quite often is something like:
for <recipient at here.com>
which reflects the address of the actual recipient delivery is being
accepted for. Maybe a testing of these fields and headers in order of
importance would give some functionality to the SECURITY_NOTIFY_RECIPIENT
feature.
Thanks again for your work!
--Paul T.
--
"So enjoy yourselves, dabble in sound, paint your pictures, turn
on your mind and gently float downstream..." -- George Martin,
Beatles's Producer
More information about the esd-l
mailing list