Re: Cool V5R2 courtesy of IBM(?)

Tom,
When IBM introduced the V5R3 CL enhancements they announced, at the same
time, some PTFs for V5R3 AND V5R2 that, as you have discovered, would allow
developers to write CLs with the new instructions and compile them with a
TGTRLS of V5R2. The exception is, if I remember correctly, the use of multiples
DCLFs (not allowed for V5R2).
Best Regards,
Luis Rodriguez

midrange-l-request@xxxxxxxxxxxx wrote:------------------------------

message: 3
date: Tue, 03 Apr 2007 17:50:26 -0700
from: qsrvbas@xxxxxxxxxxxx
subject: Cool V5R2 courtesy of IBM(?)

Here's a simple CL program:

pgm

dcl &uint *uint
dcl â?« *int

chgvar â?« ( 0 )

dofor &uint from( 1 ) to( 10 )
select
when ( &uint *eq 1 ) +
dmpclpgm
when ( &uint *eq 10 ) +
dmpclpgm
otherwise +
chgvar â?« &uint
endselect
enddo
return

endpgm

Rather pointless except that it uses various fancy things that IBM
made available in V5R3... _and_ it compiles back to and runs on V5R2!

How? By compiling it on our *V5R4* system with TGTRLS(V5R2M0). Simple.

While looking for something (which I promptly forgot about), I
noticed that the QSYSV5R2M0 library on one of our V5R3 machines was
different from the library on our V5R4 system. Wondering if the
implication held up, I threw together the above, made sure it
'worked' on V5R4, recompiled back to V5R2, FTPd it via savefile to
one of our V5R2 systems, restored... and Voila! It worked.

Nothing earth-shattering, I know. But when you have a bunch of
systems that you'd like to manage with general control procedures,
this can make a very nice difference.

Am I just now learning something the world has known for years?
(Wouldn't be the first time.) This is 'normal' for V5R2, right?

I've often thought that CL "internal" commands such as DO, SELECT,
DCL, etc., that are only valid in compiled programs should be easier
to create back-level, but I never really expected to see it.

Tom Liotta