RE: URGENT! - Program Error

["Wills, Mike N. (TC)" <MNWills@xxxxxxxxxxxxxx>]

All fields look good no corruption from what I can tell. No fields have
changed at all.

The other weird thing is that it is now happening after the last statement
of the program ends and returns to the CL.

Library SPLTOOL removed from library list.
Pointer not set for location referenced.
Internal failure in compiler or subroutine.
Pointer not set for location referenced.
Internal failure in compiler or subroutine.
Function check. RNX9998 unmonitored by CUGPRCSMON at statement *N,
  instruction X'0000'.
RNX9998 received by procedure CUCPRCSMON. (C D I R)

-----Original Message-----
From: Buck Calabro [mailto:Buck.Calabro@commsoft.net]
Sent: Friday, September 13, 2002 5:03 PM
To: rpg400-l@midrange.com
Subject: RE: URGENT! - Program Error


>And the caller and this program all
>runs in batch.

That makes no difference in memory corruption issues.

There is a very high probability that you are seeing memory corruption,
which is why Scott asked about APIs.  It is very possible to write a program
that runs fine when tested and also in production, but when you make a small
change - even unrelated to the API call - it falls over.

Imagine allocating some memory like this:
var1 100
var2 100
var3 100

Get a pointer to var1.  Add 101 to it, either deliberately or by mistake.
You are now touching var2.
Get a pointer to var3.  Add 101 to it.  What are you touching now?  Could be
internal storage used by the compiler, right?  Or maybe not, depending on
the release, PTF and so on at the time you compile it.  Now add some
storage, maybe before var1.  Re-compile the program.  With the same code as
before adding 101 to prt->var3, are you touching the exact same storage as
before?  Who knows?  A change in a distant part of the program made the
reference to var3 fail, even though you never made any changes that refer to
var3.

A very common way to see this happen is some reference like MOVE *BLANKS to
a field the caller defined differently than the callee.  If the callee
defines the field larger than the caller, then the space it wants to clear
goes past the end of the storage allocated by the caller.

Hope this helps.  I'm going out roving in the ARRL microwave contest and
won't be seeing email until Monday.
  --buck
_______________________________________________
This is the RPG programming on the AS400 / iSeries (RPG400-L) mailing list
To post a message email: RPG400-L@midrange.com
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/cgi-bin/listinfo/rpg400-l
or email: RPG400-L-request@midrange.com
Before posting, please take a moment to review the archives
at http://archive.midrange.com/rpg400-l.