RE: Activation groups for beginners

[bmorris@xxxxxxxxxx]

>From: "Smith, Nelson" <NSmith@lincare.com>
>Date: Thu, 7 Mar 2002 16:32:01 -0500
>
>Isn't that more due to setting on *INLR?  Wouldn't an OPM program work the
>same if you return with *INLR off?  I'm trying to understand what the
>activation group has to do with static variables.

Nelson, with regards to static storage, OPM programs do work the same.
Where activation groups come in is that each activation group has its
own copy of static storage.  Mostly this doesn't make any difference.
Where it does make a difference is

- if the program is in the *NEW activation group, each call to the program
  gets its own copy of static storage, and you'd see 1 11 1 11 each time

- if the program is running in *CALLER, and it gets called by programs
  in different activation groups, the results would depend on the
  AG.  For example if that program was called from AG1 AG2 AG1, the
  results would be 1 11 1 11 then 1 11 1 11 then 2 12 1 11.

- in error cases, where a program corrupts its own static storage (for
  example by a parameter-passing mismatch), it's possible for it to
  corrupt so much storage that it runs into the static storage of other
  programs.  If a program is running in an activation group, it can only
  corrupt the storage of other programs running in the same activation
  group.  (I just mention this because it's an area where the scope of
  static storage is a factor; I don't particularly think it's a reason
  for choosing an activation-group strategy.)

Barbara Morris