RE: FOR op-code

Bob wrote:
>The DO operation is NOT upwardly compatible, is it?
>I mean, if you say:
>
>C        10      DO          1       Count
>C..........
>C                EndDO       -1
>
>The loop does not work as expected. You can't count down from 10 to 1.
>But the code does compile. So how do this port to FOR?

What I meant was that the FOR loop was meant to be largely
"functionally" upward compatible with DO.  But in this
particular case, the conversion is straight-forward:

   for count = 10 to 1 by -1;
      ...
   endfor;

and it would work the same way.  (That is, the loop would
not be performed.)

DO only handles upwards iteration.  FOR offers the
additional functionality of downwards iteration.

>I don't see how upward compatibility for the DO operation should have
>influenced the FOR? FYI, I too was surprised that "FOR X=10 downto 1 by
>-1" did not work. So I solve my personal issue, by leaving off the "BY"
>clause.

Why shouldn't compatibility with DO be a consideration?
Programmers have embraced and adopted the expression
statements in the language.  And so that's why FOR came up.
By making the FOR largely compatible with DO, RPG III to
RPG IV conversion utilities could add this to their
repertoire of conversions.

But as I suggested, there are a couple of points where FOR
is not functionally upward compatible.  The first is that
you can code an indexed array as loop index in DO, and you
can't today in FOR.  Because of side issues involved in one
particular enhancement we're working on, this might be
changed in a future release.  (Apart from that specific side
issue, I wouldn't have worried about this difference too
much.)

The other incompatibility is that an index must be coded
with FOR.  This is easy to get around, though - just code
an index.

>At least you didn't use colons to separate the values! Thank you!

LOL!

Lately, I've been playing around with Python.  It's a fun
language, but the one aspect of Python syntax that often
messes me up is I often forget to code the colon!

Cheers!  Hans

Hans Boldt, ILE RPG Development, IBM Toronto Lab, boldt@ca.ibm.com