On Mon, 8 Dec 2014 07:59:53 -0500, Peter Relson wrote:
>
>... At the time that PARMDD is (and
>has to be) processed (under the initiator task), various i/o operations
>are not supported (apparently including those needed to write to a the
>intended log). Changing that could be very costly, thus requiring
>significant justification.
> 
Might it be more readily be reported in JESYSMSG?  It would seem that
since allocation messages appear there that data set is writeable at the
needed time (whether it's in scope for the initiator is a different question).

Treatment of PARMDD that refers to a SYSIN with SYMBOLS= is so
more complicated than EXEC ...,PARM= that reporting is correspondingly
more valuable.

It's wrong to quietly ignore an unsupported construct.  Logging-DDname
in a SYSIN data set mentioned in PARMDD should result in a JCL syntax
error and failure of the job before execution.

And another idiosyncrasy I suspect.  From:

http://www-01.ibm.com/support/knowledgecenter/SSLTBW_2.1.0/com.ibm.zos.v2r1.ieab600/iea3b6_parm_string_requirements_parmdd.htm
z/OS 2.1.0>z/OS MVS>z/OS MVS JCL Reference>EXEC statement>PARMDD 
parameter>Parameter string requirements
z/OS MVS JCL Reference
SA23-1385-00 

After concatenation, the parameter string that is passed to the job step 
program is examined for double ampersand character (&&) sequences. Double 
ampersands are converted to single ampersands in the same way that double 
ampersands are converted to single ampersands by PARM= processing. 

I guess (I haven't tested) that if I code:

//  SET TWOAMP='&&'
//  EXPORT TWOAMP
//STEP  EXEC  PGM=WOMBAT,PARM='&TWOAMP'

... the parm passed to WOMBAT is two ampersands (there is no rescan).
However, in:

//  SET TWOAMP='&&'
//  EXPORT TWOAMP
//STEP  EXEC  PGM=WOMBAT,PARMDD=P
//P  DD  *,SYMBOLS=JCLONLY
&TWOAMP

... PARMDD processing converts the two ampersand to a single ampersand
which is passed to WOMBAT.  Ouch!  FSVO "... in the same way ...".  The
effect of passing a symbol via PARM= may differ unexpectedly from passing
it via PARMDD=.

I suspect that when the LONGPARM Requirement was submitted the
requestor expected processing consistent with that of PARM=, but with the
100-character limit relaxed (substitution of system symbols would be an
added value).  What was provided was significantly different.

-- gil

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to