> To all UCP-L members:
>
>    I have received the following undelivered mail messages this past
>    week which seem to stop the rest of the list from getting the mail.
>    If anyone has any ideas why, I would be very happy to hear them so
>    we can get on with discussions.
>    Thanks,
> Jim...
> - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
> From: Network Mailer <MAILER@AWIWUW11>
>
> Mail to MAILER   is not allowed.
> Error loop detected.
> Another network mailer sent an invalid file.
>
> Validity check failed.
 
    Alas, I know too well what  your trouble is.  It has become a
bothersome  point  for  many  of  us, due  to  the  explosion  of
LISTSERVs around the globe.
 
    The  "Validity  check failed"  message  is  generated by  the
Crosswell Mailer,  run at most  VM sites on BITNET.   This mailer
purports to  know the names  of all  other mailers, and  thus can
decide on the  truth or falsehood of the From:  and Sender: lines
in the message header.
 
    When this mailer receives a file from a user@node that is not
in the  Known Mailers  list, it checks  that address  against the
From:  and Sender:  fields.  If  there  is no  match, the  Mailer
returns the file to it's source rather than forward it.  When the
file arrives back home, one of several things may happen.  If you
also run the Crosswell Mailer, the file will return to it, and be
transferred to  your Postmaster, with  the message that  "Mail to
MAILER  is not  allowed.  Error  loop detected.   Another network
mailer sent an invalid file." If  you don't run this mailer, then
the file probably  returned to LISTSERV.  Eric  Thomas, God bless
'im, has  code in LISTSERV  that recognizes these  "Returned Mail
Files",  and transfers  them to  the LISTSERV  Postmaster, rather
than send them out to the  whole list, as some other list servers
do.
 
    The solution to  your problem is twofold.  First,  if you use
the Crosswell Mailer, change your  MAILER PROFILE to send mail to
users on the rejecting systems  directly, thus bypassing the sick
mailer.  To do so, add an  entry to your MAILER MAILLCL file, for
each of these sick nodes, which says:
 
<nodename> <nodename> ? DEFRT 1
 
    Then  run  the MG  program  to  rebuild MAILER  PROFILE,  and
restart MAILER.
 
    Second, and more important, contact the postmaster at each of
these nodes (their  names, mail addresses, and  phone numbers are
in BITEARN NODES) and tell, do not  ask, them to get a fresh copy
of XMAILER  NAMES from their favorite  server (NICSERVE@BITNIC or
NETSERV@lots-of-places)  and  to  rebuild their  MAILER  PROFILE.
Once the've done so, you can try sending mail to someone on their
system and  see if  the problem persists.   If not,  delete their
entry  from your  MAILER MAILLCL  file, and  rebuild your  MAILER
PROFILE, so that your mailer will send via theirs.
 
    Please feel free to contact me if  you need any help, or if I
seem incoherent.
 
Ross Patterson
Rutgers University
Center for Computer and Information Systems
ATT et. al.: (201) 932-3435
BITNET     : A024012@RUTVM1
ARPANet    : A024012%RUTVM1.BITNET@WISCVM
Snail Mail : PO Box 879, Piscataway, NJ 08854