FIX15M3 is on its way (also a FIX15M3 ASM file for the OCO=NO sites):
 
- REXX error  during INSTALL  ('Routine not  found' on  'Call LSVDBASE'):
  This  was due  to the  use of  CARD/SP3 under  CMS4. IF  YOU HAVE  BEEN
  AFFECTED, YOU --> MUST <-- GET YOURSELF THE LATEST VERSION OF CARD.
 
- LSVDBASE: Fix for the REXX error in the VCPU routine. I had been talked
  into changing a  'QUERY TIME' output parsing routine to  diag X'0C' for
  better compatibility  with CP  moded sites, which  sounded like  a good
  idea, but I had  forgotten that the C2D function will  crash if you ask
  it to  output something larger than  NUMERIC DIGITS. I don't  like this
  notion at all (after all it's  a function not an arithmetic operation),
  especially as REXX doesn't  crash if I ask it to display  2**64 or if I
  ask for 10000000000000000000000/2 with  the standard NUMERIC DIGITS: it
  merely  chops off  the  digits  beyond the  significance  point it  can
  handle. But well that's the way it is, I'll be more careful next time I
  use C2D.
 
- LSVDBNB: Fix for the REXX error  that occured in database searches when
  the target  database is set to  "Notebook= Yes,etc" but no  notebook is
  available  at the  time the  query is  issued. Previously  the database
  could  be opened  and the  number of  entries was  set to  0, but  some
  routines in LSVDBASE could not handle it so now the open gets rejected.
 
- LSVDIST2:
 
  a) Now uses diag X'0C' in the same way as LSVDBASE.
 
  b) The recipient duplication problem has been fixed.
 
  c) Code has been added to prevent more  than one copy of the file to be
     sent to any  given recipient (during the  comparison, the local-part
     case is respected but the domain-part  case is ignored). This is not
     just to  avoid the  problem we recently  experienced: it  will avoid
     duplication of files  if more than one occurence  of the destination
     appears  in the  input  parameters,  which may  happen  when a  dumb
     program generates them.
 
  The problem with DIST2 tests is that usually everything works just fine
  with 1 server,  and you start getting  into problems with 2  or more. I
  think I have learnt a lesson  there: now that we *eventually* have disk
  space, I will be able (ie  allowed) to create a second LISTSERV machine
  (on a  second RSCS)  and make  some more  local tests  without wreaking
  havoc on the network :-(
 
- LSVNBFL: Fix  for Jim's VM READ  problem, which was probably  due to an
  invalid value in  the filemode section of the  'Notebook=' keyword (the
  links are presently down so I can't check).
 
Eric