I haven't been quite able to get my setup working in a foolproof manner, so I
was hoping someone here could point me in the right direction...
I'm using
http://razor.sf.net/ to filter out known spam. I have my first
director setup to check that if the received protocol doesn't match "razor",
then to accept that message and use my razor_transport. razor_transport pipes
the mail to a perl script that calls razor and either dumps the message, or
returns it back to exim, by invoking a new exim and setting the protocol
to "razor".
I want to set this up so that regardless of the error or response from the pipe
command, the message is _NEVER_ returned to the sender (bounced).
I setup the temp_errors on the pipe transport to contain the errors that I know
that the command can give, but I occassionally get other errors from the
command.
I'm looking for a way to in effect set temp_errors to "*" so that any error
from the command is treated as a temporary error and the message is deferred
and delivery attempted later.
Or even better, if the director/transport combination could be configured so
that if the transport of the razor director failed, then exim would treat this
as if the director did not match, and continue looking for a matching director
for the message. (in effect, skip razor director if for some reason its
transport fails).
Thanks for any help, tips, pointers...
Padraic
