Re: QCMDEXC question

[Peter Coffin <phcoffin@xxxxxxxxxx>]

Kilgore writes:

----------------quote--------------
A few years back there were trade rag articles about REXX ("The OS without a
brain" or some such) and for those of us that have a S/36 or S/38 background we
are accustomed to writing some form for procedural driver to a command that
executes the REAL programs.  Some may use REXX in the same context as I'm sure
the majority use CL, while the C programmers out there are probably scratching
their heads wondering "What do we need another language for?"

It may be time for us to look at another "callable" for the command processor
and CL may not be the best suited for the ILE environment.
---------------unquote-------------

You're not the only one who's using REXX by any stretch of the imagination.

Programs that report on other objects on the system that are going to become
hardcopy reports are so easy to write in REXX. The ability to pass commands
directly down to CL, default output to a spooled file, and REXX's wonderful
string handling features... Sure, the same thing could be done in RPGLE, but
for me, doing it in REXX takes about half the time.

I would like to see the ability to create REXX *MODULEs and bound REXX
programs, though. They wouldn't have to be actual compiled programs; a wrapper
of code that starts the REXX interpreter to execute the REXX procedure that's
_imbedded_ in the object (just copy it from the source member when building
the *MODULE or *PGM) would do nicely. The reason to imbed it is that then the
execution doesn't depend on the source member staying exactly where the *PGM
expects it. It moves with the object, and is saved and restored with the
object.


Peter H. Coffin
phcoffin@us.ibm.com
+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to "MIDRANGE-L@midrange.com".
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---