On 09/04/2015 02:18 AM, Paul Gilmartin wrote:
Perhaps they read the manual? People are often telling each other to do that.
If you want users to restrict themselves to the SUBMIT command, you should
submit RFEs to:
o Relax the fixed-80 restriction imposed by SUBMIT but not by INTRDR.
o SUBMIT from DDNAMES, UNIX files, and POSIX pipes, all of which are
supported by INTRDR but not by SUBMIT.
Well, after the conversation I just had over JCL "errors" that
aren't, I can understand why one would use INTRDR rather than SUBMIT.
It seems that TSO SUBMIT *assumes* statements such as IF, ELSE,
and ENDIF as operator commands!! It does not assume that for DD,
JOB, and EXEC statements (z/OS TSO/E Customization V2R1
SA32-0976-00, p. 300).
After the discussion that I've had with JES C/I (which is where
we thought this problem was), it seems that TSO built SUBMIT with
a rudimentary parse/scan, which INTRDR did not implement.
TSO confirmed that they do a certain amount of parse/scan, and
then tells us that "TSO" does not want to get into doing C/I
work. Yet that *is* what they effected in years gone by.
This then produces a different handling of IF, ELSE, ENDIF,
INCLUDE, SET, etc. as they would be seen as an operator command
if it ends with a "," (per the above cited doc -- which TSO cited
to me when I pointed out these conflicts).
This behavior by IBM effectively deprecates SUBMIT.
Regards,
Steve Thompson
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
