Re: BPCS SQL programs Compilation Parameter Question

["Genyphyr Novak" <novakg@xxxxxxxx>]

Hello,

You don't mention what release of BPCS we are talking about, and the answer
is 'it depends'. You are correct in your doubts if you are on a BPCS release
below 6100. I would suggest running future ideas of this nature past the SSA
GT AS/400 technical team if you have OGS support, before attempting anything
else suggested by a non-SSA GT person regarding changing any compile options
in BPCS programs for performance or other reasons.

In this case, the advice was incorrect and potentially damaging to your
system. Don't do it, or various program errors will result from cursors left
open between program calls, making the position in the file random, and
potentially causing data corruption along with definite program errors. This
scenario has been tested by SSA GT and does NOT work on releases prior to
6100.

Versions of BPCS starting with 6100 were developed specifically with
corrections for cursor handling which had resulted in the errors on older
releases when using *ENDJOB. All programs on these newer releases were
compiled with *ENDJOB and verified in testing and were delivered in the base
release and forwards with the *ENDJOB option.

Note that in older releases there are a few programs that are verified to
work properly with *ENDJOB. Due to performance BMRs, the default compile
options from AS/SET were changed for those few programs in 6002 and 6004 to
automatically compile the programs with this setting. It is OK to run these
few programs with *ENDJOB in the older releases, and in fact they are
delivered that way from SSA GT.

Thanks,

Genyphyr Novak
SSA GT


----- Original Message -----
From: <FWSoftware@aol.com>
To: <bpcs-l@midrange.com>
Sent: Monday, May 20, 2002 3:32 PM
Subject: BPCS SQL programs Compilation Parameter Question


> Hello,
>
> Recently a Client of mine was told by IBM to recompile BPCS SQL programs
specifing CLOSSQLCSR(*ENDJOB) instead of the current *ENDPGM, for
performance reasons.
>
> I have my reservations about doing this because specifing *ENDJOB will
maintain the cursor between calls. I believe that may have adverse effects
on BPCS program logic.
>
> Thoughts?
>
> Thanks