On Mon, 28 Jul 2008 10:55:40 -0500, Paul Gilmartin <[EMAIL PROTECTED]>
wrote:
>I'll sometimes disable a lot of JCL with IF. Which is why I regret
>that "IF FALSE" is documented as not supported and unpredictable
>in behavior (although no error is reported and the construct has
>(almost) the intuitive effect).
>
If one takes the JCL reference literally, most of the examples they give are in
fact "not supported and unpredictable". Why? The manual states;
"A relational-expression consists of:
Comparison operators
Logical operators
NOT (¬) operators
Relational-expression keywords."
A value, such as "8" in the relationship-expression "RC = 8" is none of these.
I recently opened an ETR to clarify this, and was told that what they test is
relationship-expressions of the form "keyword comparison-operator value", in
that order. Yet "value" is not listed as one of the allowed components of a
relationship expression. While the documentation does not rule out such
expressions as "STEP1.RC = STEP2.RC" (and it does work), I was told that this
was not something that it is not in fact supported and might not work in the
future.
What prompted my ETR was that I had encontered some installation JCL from
CA that contained a construct something like
// SET VAR1 = 1
intervening JCL
// IF &VAR1 = 1 THEN
more JCL
// ENDIF
which works, but CA-JCLCHECK flagged this as an error.
After some arguement with CA-JCLCHECK support, where I argued &VAR1 is
not a keyword (by the time the system evaluates the truth of the expression,
it has been replaced by the value "1"), and that the JCL Reference was clearly
incomplete in its definition, I expressed my observations to IBM in an ETR.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
