On 2013-11-08, at 09:54, Frank Swarbrick wrote:
> Hmmm! You're right. It's not the symbolic that is not allowed. It's the
> comparison to non-numeric values that is not allowed. Interesting!
> ...
Is this documented as supported, or is one comparand required
to be a return code? Is it documented that comparing return
codes of two previous steps is supported?
> Wish I would have realized this a few years ago!
>
And the only use of the boolean primaries, TRUE and FALSE
documented as supported is as operands of relational operators.
So, my predecessor settled on:
//STEP0 EXEC PGM=IEFBR14
...
//IF ( STEP0.RUN EQ FALSE ) THEN
...
to unconditionally bypass a job step. I followed his lead.
This appears to employ only documented supported constructs.
But who can tell? The JCL Ref states that transgressions may
not be reported.
(And STEP0 circumvents the misbehavior of IF before the
first job step.)
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
