On 7/28/2015 4:27 PM, Frank Swarbrick wrote:
I found this to be interesting, but I don't quite understand it:
"z/OS V2.2 is designed to
support a new IEFOPZxx parmlib member in which you can specify pairs
of partitioned (PDS and PDSE) data sets. In each pair, you can specify
one that is to be searched ahead of its counterpart data set when
programs are fetched. This new function is intended to allow you to
insert program libraries ahead of others in the link list, in STEPLIB
and JOBLIB concatenations, and for LLA-managed libraries without making
JCL changes. A new DISPLAY IEFOPZ command is designed to display information
about existing pair definitions, and a SET IEFOPZ command is designed
to allow you to add, remove, or change pair definitions dynamically.
This is expected to be useful for activities such as converting application
program libraries from PDS to PDSE, as is necessary for converting
to COBOL V5 (5655-W32), without requiring JCL changes. Availability
of these functions is planned for December 2015 with the PTF for APAR
OA47689."
Can someone "in the know" give an example of how this would work with JOBLIB or
STEPLIB?
Thanks!
Frank
Frank,
IEFOPZxx allows you to define "pairs" of libraries, where the paired
library is concatenated ahead of the original. F'rinstance:
//STEPLIB DD DISP=SHR,DSN=GOODOLD.COBOLPDS.LOADLIB
In IEFOPZxx, we define DUMBNEW.COBOL5.PDSELOAD as a pair to
GOODOLD.COBOLPDS.LOADLIB. So whenever the system sees GOODOLD in a
STEPLIB or JOBLIB, allocation will automagically concatenate DUMBNEW
ahead of it, and VIOLA! Conversion to COBOL 5 PDSE's, without changing
the STEPLIB in thousands of jobs. No need to recompile everything at
once, you can phase in the new compiler. This also works great for new
ARCH levels, so if you have multiple CEC's, you can upgrade each CEC to
the new ARCH level, again without recompiling everything.
There are other uses, but if I told you, I'd have to kill you.
Regards,
Tom Conley
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
