Peter M. Weiss <[log in to unmask]> writes:
> What is not clear to me is if the DISTRIBUTE MAIL bounces
> w/o a clear source of the problem child, why is it that the
> PROBE feature will be any better in revealing that same
> problem?  Won't the FROM: be just as easily managled,
> hidden? (As Mark indicated, this was a semi-automated
> method of REVEALing the bad address, not the be-all and
> end-all in list management).
 
  The DISTRIBUTE mail will all be "from" the submittor (or whatever they
put in the headers), so any error will go back to that address - in par-
ticular all errors will go to the same address, so you can't tell which
subscriber the error was for by the address it was sent to.
 
  With PROBEs, the "from" address (both envelope from and header from, I
think) will be something like owner-listname*[log in to unmask],
so that when the bounces come back, LISTSERV can determine what subscriber
they pertain to without trying to parse any other headers or the body of
the bounce.
 
> Successfully used this technique this a.m. to find some
> non-specific bounces.
 
  Yes, but wouldn't it be nice to be able to say:
 
probe listname [for pattern@pattern]
 
  or as the site admin:
 
probe * [for pattern@pattern]
 
  Everybody who wanted to do manual probes could go ahead and do them the
same way, but the rest of us could fire off a probe whenever we got a
bounce that didn't appear to pertain to any existing subscriber.
 
        Terry Kennedy             Operations Manager, Academic Computing
        [log in to unmask]      St. Peter's College, Jersey City, NJ USA
        +1 201 915 9381 (voice)   +1 201 435-3662 (FAX)