On Sat, 5 Dec 2009 23:57:23 -0800, Edward Jaffe wrote:
>
>The kludge in this case is the hard-coded checking for PGM=IEFBR14 with
>the presumption that this program is always the IBM-supplied utility;a
>reliance on common practice rather than certainty; the establishment of
>(what I believe to be) a new class of restrictions for the z/OS BCP:
>i.e., reserved program names on the JCL EXEC card; and no mechanism for
>enforcement of the those restrictions.
>
I understand. But would you propose an alternative?
o A new command name (NOOP?) alternative to EXEC, with the
semantic: "Perform allocations on DD statements, taking
shortcut to DELETE", but execute no program.
o A new keyword for the EXEC command, with similar semantics,
mutually exclusive with PGM and PRC.
o A new value for the COND keyword, with similar semantics.
o A new keyword on the DD statement; The impracticality of
making "UNIT=(,,DEFER)" has been well explained here. In
any case, it would be a change to existing semantics.
o Existing facilities sufficed; no innovation was necessary
or even desirable.
What existing facilities? Would you counsel?:
//DELE EXEC PGM=IKJEFT01,PARM='delete ''&DSN'''
//SYSTSIN DD *
/* Can't put command here because '&DSN' */
/* won't be substituted. */
...
//* This step does nothing, but provides needed serialization
//* and meets JES3 setup requirements.
//*
//ENQ EXEC PGM=IEFBR14,COND=(0,LE)
//DDN DD DISP=(MOD,DELETE),DSN='&DSN'
Of course, a separate IKJEFT01 step may be needed for each data
set to be deleted because of other syntactic restrictions,
well-known and not to be discussed here.
-- gil
----------------------------------------------------------------------
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
