On Fri, 5 Apr 2013 09:10:14 -0500, John McKown wrote:
>OK, I'll byte (ah, bite). Can the DD be allocated to a UNIX file? I don't
>see why not, but then ... .
>
>//parmdd DD PATH='/home/myid/my-parm-file.txt',
>// FILEDATA=TEXT,
>// RECFM=VB,LRECL=32756,BLKSIZE=32760
>// PATHOPTS=(ORDONLY)
>
(or even FILEDATA=BINARY)
John Eells's SHARE presentation said:
A SEQUENTIAL DATA SET OR A MEMBER OF A PDS OR PDSE WILL WORK AS WELL.
May I assume "sequential" includes tape data sets? But it was noncommital
about UNIX files. Why should that make a difference? For a Horrid
Counterexample, see ISPF LMINIT (which also prohibits tapes).
>On Fri, Apr 5, 2013 at 8:28 AM, Peter Relson wrote:
>
>> Informally, for those curious, the processing does what you might expect it
>> to do.
>> It allocates system subpool storage of approximately the maximum size so
>> that the PARMDD data can be read in, up to the maximum size.
>>
It can make a difference, if only to a fanatic. In JCL PARM='string', the
coded string may far exceed 100 characters (I've tried it with thousands)
_provided_ that after symbol substitution the resultant PARM does not
exceed 100 characters. So, is there a size limit imposed on PARMDD before
substituting symbols (and before stripping sequence numbers and blanks)?
(You know the Black Team and their notional successors will try it.)
>> What I don't recall anyone inquiring about is
>>
>> -- what columns of the PARMDD data are used?
>> Answer: all, unless for fixed-block the last 8 characters are
>> decimal-numeric in which case those are treated as a sequence number and
>> are stripped out
>>
o Does this apply alike for FB, LRECL other than 80?
o Is this rule applied individual line-by-line, or like Rexx does it
decide based on the last 8 characters of the first line whether
to strip the last 8 characters of every line?
>> -- what about trailing blanks (including those preceding a stripped out
>> sequence number)?
>> Answer: they are stripped out
>>
This makes it tricky to end a PARM from PARMDD with trailing blanks,
but it can be done with symbol substitution.
I'm basically a minimalist when it comes to lexical munging. I'd
prefer to see no stripping of trailing blanks or digits. If the
programmer enters the data, respect it. It makes the rules
simpler.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
