Re: Is the progam OPM or ILE?

On 15 October 2015 at 19:13, Vicki Wilson <VWilson@xxxxxxxxxxxxx> wrote:

I wrote a generic service program for some error handling. It works as expected if calling program is OPM. It does not behave as desired if the calling program is ILE - some of us are writing more programs by declaring MAIN.

One option is to change the behavior if the calling program is ILE. How would I determine if the calling program is OPM or ILE?

If you are curious - I wrote a CHECK_SQL_STATE which uses QMHSNDPM to issue a CPF9898 when there is an sql error. Basically folks were used to seeing a hard-halt on I/O errors and since the CPF9898 wasn't handled we ended up with an inquiry message RNQ0202.

This comes up somewhat frequently. The tl;dr version is that ILE adds
to the call stack, so tossing a message up one level doesn't do what
it used to do. Scott Klement illustrates it here:
http://archive.midrange.com/rpg400-l/200303/msg00261.html

There are several approaches to the situation:
Bruce interrogates the call stack:
http://archive.midrange.com/rpg400-l/201410/msg00092.html
Rory passes the message's destination to the sndMsg() procedure:
http://archive.midrange.com/rpg400-l/201005/msg00185.html As did
Simon: http://archive.midrange.com/rpg400-l/201005/msg00178.html
David Morris used a single program queue to pass the messages around:
http://archive.midrange.com/rpg400-l/200204/msg00764.html
And this thread from last year has interesting ideas from Carsten:
http://archive.midrange.com/rpg400-l/201410/msg00012.html

Essentially what I am trying to do is re-create a hard-halt in ILE. Which is sooo wrong. And definitely not the end game.

There's nothing inherently wrong with hard halts, especially if that's
what the rest of the system expects.
--buck