I don't know if has been mentioned -- I admit I have kind of tuned out as this 
thread has dragged on -- but it may be relevant to Gils' conjecture to note 
that PARMDD is quite analogous to the OPTFILE option of the XL C/C++ and 
Enterprise COBOL compilers, both of which are "roll your own" solutions by the 
compiler writers to the constraints of the 100-byte limitation on PARM=.

For those unfamiliar with the option, both compilers offer the user the ability 
to specify PARM='FOO,OPTFILE,BAR' whereupon the contents of DD SYSOPTF are 
logically inserted into the PARM= string between FOO and BAR. C offers slightly 
more flexibility in that one may code OPTFILE(filespec) and avoid the tyranny 
of DD SYSOPTF, and even allowing the contents of multiple, different files to 
be logically inserted at different spots in the PARM= string.

Charles

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf 
Of Paul Gilmartin
Sent: Sunday, April 07, 2013 10:58 AM
To: [email protected]
Subject: More PARMDD: Processing sequence?

More questions keep occurring to me.

When is the data set associated with PARMDD read?

o Is it early enough in processing that:

  //         SET DSNAME=MY.DATA.SET
  //UNAUTH  EXEC  PGM=MYPGM,PARMDD=PARMS
  //PARMS   DD   *,SYMBOLS=(JCL)
  //    ... &DSNAME ...

  ... can be effective?

o Is it late enough in processing that PARMDD can refer to a
  temporary data set passed from an earlier job step?

(Are both even logically possible?)

o In the absence of SYMBOLS=JCL, or if PARMDD is a data set
  external to the job's JCL, will symbol references appearing in
  that data set be resolved by JCL processing?

Questions such as above impel me to wonder, why was PARMDD introduced, rather 
than simply allowing a longer PARM= string?  This might be continued over 585 
lines, if necessary to accommodate 32760 characters, possibly read from JCLLIB. 
 But the facility to access the PARM from an external file may have its own 
peculiar added value.

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

Reply via email to