Re: Error handling while mixing ILE and OPM

Glenn Gundermann wrote:

...
But this program is not ILE so even if it's called by an ILE program, I
assume the error does not get percolated and the user will get an error
message immediately.

I think the same holds true if I write and register a condition handler.
It would not take affect because the program with the error is OPM.

Am I correct in my assumptions above?

Yes, that's correct.  The behaviour would be the same if the called
program was a DFTACTGRP(*YES) program.  Exceptions don't get percolated
across a program boundary, and there is an implicit program boundary for
OPM programs and DFTACTGRP(*YES) programs.

I wouldn't compile an ILE RPG version with FIXNBR, though.  That's only
marginally better than OPM's IGNDECERR(*YES); they both hide the fact
that an error occurred, making you continue processing with what is
probably bad data.  (It's "marginally better" because the results are
defined and predictable for FIXNBR, but undefined and unpredictable for
IGNDECERR.)