Re: More ILE 101

Hi Joe,

> I've finally confused myself.  I compile a program with DFTACTGRP(*NO)
> ACTGRP(*CALLER).  Do bad things happen if:
>
> 1. I call from another ILE program
> 2. I call from an OPM program
> 3. I call it from the command line
> 4. I invoke it directly as my signon program
> 5. I invoke it as the first program in a submitted job

I believe that option 1 entirely depends on the nature of the program that
called it.

Options 2-5 are all the same.

Here's the way that I understand it:

Basically, when a program is compiled with *CALLER, and that caller is in
the default activation group (as is the case with items 2-5, and could
possibly be the case with item 1) the program will remain activated in
memory, regardless of the state of *INLR and regardless of RCLRSC.  If
INLR or RCLRSC is used, the program will re-initialize all of it's fields
back to blanks & zeroes and start over, but it will not unload from
memory.

If the same programs had been compiled with DFTACTGRP(*YES) instead of
*CALLER, then RCLRSC or *INLR=*ON would actually de-activate the program
from memory...  (That is, it would actually reclaim the resources!)

The one place that I know of where strange things happen is with service
programs that are compiled with ACTGRP(*CALLER). RCLRSC does not cause
static memory to be reinitialized in service programs.  That means that if
a service program has files open or has already resolved to a program for
a program call, doing a RCLRSC will cause bad things to happen.

> My guess is that 3, 4 and 5 should be very similar in nature.  For case
> one, I simply take on the ACTGRP of the ILE.  I don't even care to guess
> at the second option.

Since the command processor is an OPM program itself, options 2-5 should
all be the same.   Assuming that we're talking about programs rather than
service programs, the only big drawback to *CALLER being invoked from an
OPM program is that there's no way to free up it's storage (short of
ending the job)

That may not be a problem, depending on your circumstances.