Michael Bluejay wrote:
Curiously, this error does not appear when matching mail from Mike
Schuppe.
Nothing surprising about it. You're feeding an entire message to
formail -r without -k, a command which accepts only the head. So when
the body is short enough to fit into the same bufferful that was needed
for to finish feeding the head, procmail doesn't notice a write error.
But when the body is large enough to make procmail start another
bufferful, the write error shows up. The messages from Mike Schuppe are
short; thsoe with the viral payload are much longer.
So add either the `h' flag (so that procmail will feed it only the head)
or the `i' flag (so that procmail will know to ignore the apparent write
error) to the recipe that invokes formail.
However, there's another aspect: virus-laden messages always have false
return addresses, so there's no gain in bouncing them, plus there is the
possible damage of sending the bounce to an innocent party whose address
was forged on the infected message that you're bouncing. If you can't
reject it during the SMTP dialogue, just send it to /dev/null.
____________________________________________________________
procmail mailing list Procmail homepage: http://www.procmail.org/
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail