mailscanner 3.15-3 proc dies without error/warning - tied to mailman?

[Peter Peters]

On Tue, 21 May 2002 09:49:37 -0400, you wrote:

>I received NO errors at all on the console, even when the Perl process
>died.  I sent a few emails back and forth which all worked just fine.
>Then, I sent an email to "test at dlugosz.net", a test Mailman list which
>only has two subscribers (myself and another remote account).  The email
>was scanned, passed to the Mailman wrapper script, then sent out again.
>Now, each of the 2 emails (one for each address on the list) are scanned -
>Boom - Perl process is gone, no errors.  The mail sits in the mqueue.in
>folder - They appear to have actually been scanned (they contain the
>mailscanner headers), but MS appears to have died before it moved them
>into the outgoing queue.

Could it be that the mailscanner header comes from the first scan
(leaves mailman those headers intact) and mailscanner doesn't now what
to do when he encounters an already scanned e-mail?

In what way is the output of mailman handed over to sendmail (and
mailscanner for that matter)?

--
Peter Peters
senior netwerkbeheerder,  Centrum voor Informatievoorziening,
Universiteit Twente,   Postbus 217,  7500 AE  Enschede
telefoon: +31 53 489 2301, fax:+31 53 489 2383, http://www.utwente.nl/civ