Re: Named activation groups vs *CALLER (was: H ActGrp('QILE') vs H ActGrp(*caller))

We've been following this thread with great interest.  We're faced with the
'OPM calling ILE' situation, and we've been playing around with some small
programs to see how different scenerios act.  Here's what we've come up
with:

First situation:  We have an OPM CL program that does an override to a
specific member of a database file, override scope is *actgrpdfn.  An ILE
program (compiled with an activation group *new) is called.  The ILE program
calls a *srvpgm to retrieve a record from the database file.  When we look
at the open files, we see the file opened to the specific member, in the
*new activation group.

We were under the impression that the override would not be in effect in the
ILE program because the OPM program performed the override and the ILE
program was in a new  activation group (which would be a call boundry).

Second situation:  OPM CL program does an override to a specific member of a
database file, override scope is *actgrpdfn, and calls another OPM program
that opens/reads the file.  An ILE program (compiled with an activation
group *new) is called from the second OPM program.  When we look at the open
files, we see the specific member opened in the default activation group and
the same member opened in the *new activation group.

The big question is, why is the override still in effect?  .

Thanks for any light you can shed on this!