On 05/21/2015 09:46 AM, Shmuel Metz (Seymour J.) wrote:
> In <[email protected]>, on 05/20/2015
>    at 09:35 AM, Joel Ewing <[email protected]> said:
> 
>> My recollection is that the immediate bailout from the CLIST and 
>>from batch TSO on a non-zero TSO command processor RC only 
>> occurred for commands running "directly" under the batch TMP TCB
> 
> That's a different issue from another TCB being involved; both CLIST
> and REXX will ATTACH a command, creating a separate TCB for it. The
> difference between your CLIST and REXX code lies in the default ERROR
> handling; You can use the CLIST ERROR statement to get behavior more
> like REXX. IMHO, IBM was remiss in not suggesting that you try ERROR
> and in not listing the change to SEND as a migration issue.
>  
> 

This was an interesting and somewhat obscure problem, so I kept a copy
of the original 2006 ETR discussion with IBM, which I have just
re-found.  Here is IBM's response from Adam Nadel at TSO/E Level 2:

"... This is indeed a question of how CLISTs behave while executing
under IKJEFT1A [batch TSO job step].  When executing a CLIST, the CLIST
is placed on the TSO STACK and PUTGETs are used to execute each CLIST
statement.  If the statement is a COMMAND ... that command is returned
to IKJEFT02, which then attaches IKJEFT09 to ATTACH the command.  That
is, the command is invoked 'directly from the TMP'.  But from APPENDIX A
in TSO/E Customization, any command invoked directly from the TMP that
returns a non-0 return code causes the TMP to end."

"So under entry points IKJEFT1A and IKJEFT1B, any cmd invoked from a
CLIST that returns with a non-0 return code causes the CLIST to be
flushed and the JOB [step] terminates."

He then proceeded to suggest either using alternative entry point
IKJEFT01, or as we had done, using a REXX exec, "because then the
command is invoked by the REXX exec itself, not by the TMP".  The TSO/E
User's Guide and Customization manuals were revised to clarify the
behavior of IKJEFT1A with CLISTs and cmds with non-0 return codes.

Note the issue is not one of CLIST default error handling, but of the
environment in which the CLIST is invoked.


-- 
Joel C. Ewing,    Bentonville, AR       [email protected] 

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

Reply via email to