Paul,
You are clearly rather new to this game - and you can be very, very thankful
for that. Some, perhaps many, perhaps most, who subscribe to this list/group
will remember the bad old days when if it needed "installing", it was
"installed" with the aid of the assembler macro language[1]. Very often,
when you wanted to "install" a product, you coded up one or, more likely,
many macros and submitted an assembler "compile". You then took the cards
out of the card punch stacker, took a couple of steps round the 2540 card
reader/punch and put the cards into the hopper. Then you went to the coffee
lounge, read the paper or had a chat and came back some time later to find
your product "installed" - with any luck.
Alternatively, you had to "generate" an NCP or a TCAM "network" where much
the same thing was done - although you probably had a good book to read in
the coffee lounge.
As a matter of interest, since I managed to avoid this particular task, when
was it no longer necessary to "generate" "OS", where "OS" refers to the
ancestors of z/OS?
[1] I did this twice in my time. Once was for a diagnostic package for
BTAM-based network programming where I managed to produce a sort in macro
language - for device addresses. Another time was for my "adaption layer"
product. Unfortunately the latter broke down a bit when one customer site
fooled it by having volume serial numbers actually starting with a number
and I was using the volume serial numbers as names - :-(
Chris Mason
----- Original Message -----
From: "Paul Gilmartin" <[EMAIL PROTECTED]>
Newsgroups: bit.listserv.ibm-main
To: <IBM-MAIN@BAMA.UA.EDU>
Sent: Sunday, 07 May, 2006 6:33 PM
Subject: Re: rexx or other macro processor on z/os?
> In a recent note, Gilbert Saint-Flour said:
>
> > Date: Sun, 7 May 2006 11:50:33 -0400
> >
> > REXX is probably your best bet, but in the 1970s and 1980s, REXX wasn't
> > available on MVS and I often used the AREAD capability of ASMH (also
> > available in HLASM) to generate job streams. See example below (it's
> > been a while, correct syntax is NOT guaranteed):
> >
> Having attempted this sort of tailoring variously in Rexx,
> POSIX shell from a here-document, and sed, I find Rexx is
> uniformly hardest to use; POSIX shell is best for simple
> operations (the sort of thing that could almost be done
> with JCL symbol substitutions except for various onerous
> restrictions), and sed is best for complex operations or
> any use of external prototype data.
>
> I haven't tried assembler, but it intrigues me. Does
> the "AREAD ... PUNCH" sequence perform symbol substitution,
> presumably of GBLC variables? (I assume that's the main
> point of the exercise.) This could well bypass the
> limitations of JCL symbol substitutions so often discussed
> here, and permit self-contained tailorable jobs with no
> external skeletons, filters, etc. Might be easier than
> Rexx.
>
> > MACRO
> > JPUNCH
> > .LOOP ANOP
> > AREAD &CARD
> > AIF ('&CARD' EQ '').EOJ
>
> No specific EOF test? Pity. What if the input data contain
> an actual empty line?
>
> > PUNCH '&CARD'
> > AGO .LOOP
> > .EOJ END
> > MEND ---------------
> > JPUNCH , <---------- execute macro and read data cards
> > PFX TUSER.HH
> > RST ADCD.A.CNTL
> > RST ADCD.LIB.JCL
> > /*
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
