RE: QRPLOBJ Library

[Lurton Keel <Lurton.Keel@xxxxxxxx>]

If you compile an object (pgm) with replace yes, it will create a duplicate
of the object in QRPLOBJ with a strange name like Q34578ikl.  The original
object name will be in the text. So, if a program is running and you
recompile, the old program will still be in the job stack until the user
ends the program.   The next instances will always get the new program
because the rename process prevents them from calling the old.

                -----Original Message-----
                From:   Allen Overeem [mailto:AOvereem@stvgb.org]
                Sent:   Friday, January 28, 2000 9:53 AM
                To:     MIDRANGE-L@midrange.com
                Subject:        QRPLOBJ Library

                Help! I put a question to the list some time back about
this, but did not get the response I was hoping for.  Perhaps I need to
modify my question extensively.
                On our AS400e v4r4 we have a library called production.
Pgmrs have test libraries.
                We have a move process written in cl and cobol. Whenever a
program is moved from test into production, this move process is used. When
the move is done, the program to be moved is NOT running. During the move,
any version of this object in qrplobj is deleted, the existing production
program object is forced to QRPLOBJ library with the movobj command. The
test program object is then moved to production.
                We also have a special little recompile process written in
cl. When an object is recompiled: any version in qrplobj is deleted, the
existing object is first forced to QRPLOBJ with the movobj command, the
existing source is compiled into new pgm object in production.
                I had an incident recently where I made some changes to a
program that took a right justified field. left justified it and output it.
The program object was moved into production.  At that point, there are (
unless IBM stores them somewhere else on a movobj??) only two program
objects on the system: the one we forced into QRPLOBJ and the one we forced
from my test library into production. After starting program again, it
becomes obvious that the changes are not in affect; without a doubt. Doing a
wrkobj at this time shows that only two versions of object exist, as I
mentioned.  Looking in call stack for active program object shows that the
object running is the one in production.
                Doing a recompile (no changes to the source code) fixes the
problem.

                My Questions:

                1. Is there ever a time when the call stack will show that
version of object in one library is running when in fact it is the version
in QRPLOBJ that is running?

                2. Could the force of the object to QRPLOBJ system library
cause the kind of behavior I am describing?

                TIA

                +---
                | This is the Midrange System Mailing List!
                | To submit a new message, send your mail to
MIDRANGE-L@midrange.com.
                | To subscribe to this list send email to
MIDRANGE-L-SUB@midrange.com.
                | To unsubscribe from this list send email to
MIDRANGE-L-UNSUB@midrange.com.
                | Questions should be directed to the list owner/operator:
david@midrange.com
                +---
+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---