Re: TFRGRPJOB vs ACTGRP

Dave:

The library list is an attribute of the job, not an attribute of each activation group.

It is fairly trivial to create a simple test to observe this. Just create one or two small CLLE programs that change the library list, and create each one to run in its own named AG. For example:

member AG1:

PGM
/* CRTBNDCL PGM(QGPL/AG1) DFTACTGRP(*NO) ACTGRP(AG1) */
ADDLIBLE QTEMP
MONMSG CPF0000
ENDPGM

member AG2:

PGM
/* CRTBNDCL PGM(QGPL/AG2) DFTACTGRP(*NO) ACTGRP(AG2) */
RMVLIBLE QTEMP
MONMSG CPF0000
ENDPGM

To test:

1. run DSPJOB OPTION(*LIBL)
2. run CALL PGM(QGPL/AG1)
3. runDSPJOB OPTION(*LIBL)
4. run CALL PGM(QGPL/AG2)
5. runDSPJOB OPTION(*LIBL)

Repeat steps 2 through 5 as often as desired to convince yourself of what is going on.

Activation groups are mainly about controlling the scope of resources, especially open files and open data paths and dynamically allocated storage. Study the ILE Concepts manual (SC41-5606) for more details.

I think you will need to use group jobs to achieve what you are describing.

Hope that helps,

Mark S. Waterbury

> On 7/24/2012 8:55 AM, Dave wrote:

Hi,

Been looking at the possibilities of running a program using a different
library list than the caller. Particularly as the program is a large OPM
calling lots of others all copybooked in, I've been looking at TFRGRPJOB. I
have to submit my analysis, and I'm pretty sure I'll be told to convert
everything to ILE and use an ACTGRP to keep things isolated from the
calling program. I'm hard put to think of an advantage of running the
program in an activation group rather than a group job. Here's one: when
using STRDBG you can step into the called program.
Any more?
Thanks!