Re: Reading an RPG dump

Hi, Pete:

It's been a really long time since I saw any BPCS stuff, but I seem to recall some programs required certain overrides to compile; they used to create "compile CLs" for those ... a small CL program with the needed overrides followed by the compile command. Look for such a CL with the name of the program it creates as part of the name; I seem to recall something like ("CL" + pgmname) for the naming convention for those.

Compile it in your own private library, set as *CURLIB, so it is always at the top of the *LIBL, and your *CURLIB should be TYPE(*TEST).

Use the UPDPROD(*NO) feature of the STRDBG command -- you would of course need to copy any files that it needs to open for output into your *TEST *CURLIB.

HTH,

Mark S. Waterbury

> On 9/29/2010 1:35 PM, Pete Helgren wrote:

Ok. Great suggestions all. First I am reluctant to run the program
because I do not know enough about what it will impact since I am not
familiar with the program and function. Yes, I could run it in debugt
but since I am a consultant I'd hate to make this my last day by running
something that causes downstream problems. I guess I could try to
compile. The only thing that may come back to haunt me is that I don't
know if there is a "tool" I need to use to compile the source in BPCS.

I was hoping that I could figure out the issue, which I am certain is a
data issue (although the program should anticipate the possible data
issue) without running the program but I'll see if I can compile it and
get it to run in debug. No easy way out I guess.

Thanks for all the suggestions.

Pete Helgren
Value Added Software, Inc
www.valadd.com
www.opensource4i.com


On 9/29/2010 10:05 AM, Pete Helgren wrote:

Amazingly, in my 20+ years of RPG programming I have never had to dump
an RPG program that was halted with an error. The software I work with
is so familiar to me that IF I get an RPG error, I usually know where to
look to fix it without much research.

However, I have taken an interim consulting gig with a company that runs
BPCS (for reasons other than running their BPCS installation). It is an
old version (at least 10 years old) and I have never had any BPCS
experience so when an error occurred, I had to start from scratch in
figuring out what went wrong.

The statement number in the RPG program doesn't resolve to a valid
statement number in the source referenced by the program (It has been
converted to ILE RPG). So, when the error occurred, I took the option
to dump the RPG program to figure out what went wrong.

The error is: The target for a numeric operation is too small to hold
the result. So I know what the problem is. Unfortunately the program
isn't kind enough to says: The target (MYFIELD) for a numeric operation
is too small to hold the result. So I was hoping to find out what
*field* the program was complaining about.

So (this is NOT a BPCS question): How do I determine from the RPG dump
which field is too small? Since I cannot reconcile the statement number
to the source, I was hoping the dump would illuminate the problem.
After looking at the dump, there is so much information I can't sort out
where to begin looking.

Suggestions?