On Thu, 7 Jun 2018 09:27:26 -0500, John McKown wrote:

>I've already written a utility program which is designed to run as a UNIX
>command, mainly interactively from a UNIX shell, but it works from TSO
>OSHELL and in batch via BPXBATCH or Co:Z launcher just fine. So, from a
>"practical" viewpoint the following is not really necessary.
>
>But I am considering making a "version 2.0" which will run "natively" in
>the various environments. In particular, the environments I envision are:
>
>1) batch - directly from an EXEC PGM=UTILITY
>2) REXX via IRXJCL - "pure" non-TSO REXX
>3) native IKJEFT01 - a true TSO CP (getline/putline with TSO CP command
>parameters)
>4) REXX under IKJEFT01 - a TSO REXEC environment
>5) native UNIX - a UNIX command via a shell
>​6) REXX under UNIX - program invoked with a UNIX shell script written in
>REXX​
> 
>Have I missed some other "normal" (not CICS, DB2 stored proc, IMS, etc)
>z/OS environment. 
>
I have one similar.  Fewer environments: 2, 4, 5, 6 plus ISPF Edit command
line as a macro.  It needs a job ID argument which it sometimes gets (thanks,
Pedro Vera, for the suggestion) by scraping an SDSF screen -- just type its
member name on the command line.

> ... In a REXX environment (#s 2, 4, & 6) I plan to have a
>"STEM varname." type option such as with EXECIO. 
>
Input?  Output?  Kinda like BPXWUNIX?  (I'm surprised how many programmers
are unaware of or eschew its ddname form for stdin/stdout/stderr, given that
they work beautifully.  I can run BPXWUNIX with stdout and stderr directed to
SYSOUT and tail them with SDSF in real time -- no buffer latency.)

> ...In all environments, the
>default output would be to the normal output (in batch, I consider
>//SYSPRINT to be the "normal" place for program output)

-- gil

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

Reply via email to