Re: CRTDUPOBJ an ILE program

[Gary Guthrie <garyguthrie@xxxxxxxxxxx>]

Comments in-line.

Gary Guthrie



> >
> > Scott, thanks for you response.  Sounds like the short answer is No, you
> > don't want to compile all your programs with ACTGRP(*NEW).  If every program
> > started it's own activation group, you could use considerably more memory
> > and get unintended results (loss of overrides... etc).
> >
> 
> That's not what I said at all...     The "loss of overrides" business is
> somewhat true, but only until you learn how activation groups work, and
> do your overrides using the new paradigm.

It's more than somewhat true. It's absolutely true. Only the most recent
job-level overrides survive amongst programs when each of these programs
has its activation group attribute set to value *New. 

 
> The rest is completely off-base.   I agree wholeheartedly with what
> Barbara and Joe have said, use ACTGRP(*NEW) for the first program called,
> and ACTGRP(*CALLER) for each program that's called by that program.

I don't understand how folks can come to such conclusions in such a
void. While it might be true that a considerable portion of your
applications may adhere to such a scheme, it's not wise to blindly
follow such a rule of thumb. Proper assignment of activation groups
requires thought - you need to survey your applications and application
goals and make your decisions based on specific needs rather than based
on some "rule of thumb" general scheme. This becomes increasingly true
when you integrate applications with each other.