RE: Activation Groups 102, the plot thickens....

[Buck Calabro <Buck.Calabro@xxxxxxxxxxxx>]

...talking about shared opens (i.e. OPNQRYF)

I had an issue with OPNQRYF long ago, and Scott made me recall it.  That is,
the overrides by default percolate from OPM to ILE NAG, but shared opens do
not.  While that seems desirable at first blush, look at the default value
for OPNQRYF: OPNSCOPE(*ACTGRPDFN).  This is exactly the same default as
OVRDBF, and has virtually the same help text wording.  Given that, I would
expect the OPEN to follow the same scoping rules as the override.

Here is some code that demonstrates this.  It uses the previously posted RPG
code.
(CL is OPM, RPG is in AG QILE)

This reads the wrong member:
 pgm (&file &lib &mbr)
 dcl &file *char 10
 dcl &lib  *char 10
 dcl &mbr  *char 10

 dcl &slt  *char 512
 chgvar &slt 'SRCSEQ *LT 7'

 ovrdbf QRPGLESRC share(*yes)
 opnqryf ((&lib/QRPGLESRC &mbr)) qryslt(&slt)
 call ag01a
 clof qrpglesrc

 endpgm

This reads the wrong member:
 ovrdbf QRPGLESRC share(*yes) opnscope(*actgrpdfn)
 opnqryf ((&lib/QRPGLESRC &mbr)) qryslt(&slt)
 call ag01a

This reads the wrong member:
 ovrdbf QRPGLESRC share(*yes)
 opnqryf ((&lib/QRPGLESRC &mbr)) qryslt(&slt) opnscope(*actgrpdfn)
 call ag01a

This reads the right member and does the selection:
 ovrdbf QRPGLESRC share(*yes)
 opnqryf ((&lib/QRPGLESRC &mbr)) qryslt(&slt) opnscope(*job)
 call ag01a

This reads the right member and does the selection:
 ovrdbf QRPGLESRC share(*yes) opnscope(*job)
 opnqryf ((&lib/QRPGLESRC &mbr)) qryslt(&slt)
 call ag01a

It would appear that only by specifying OPNSCOPE(*JOB) will the open get
shared between AG's, despite the wording of the help text for *ACTGRPDFN.
Did I misunderstand how this is supposed to work?

Manual reference: CL Programming
http://publib.boulder.ibm.com/pubs/html/as400/v5r1/ic2924/info/cl/opnqryf.ht
m
  --buck