Re: Binary variables in RPG

JHHL,

> I've recently observed that a 2-byte binary variable that is a field
> within a data structure apparently gets clamped to 4 digits, even though
> the natural floor and ceiling for a 2-byte signed binary are -32768 and
> +32767.

The solution is simple.  Don't use the "B" data type.

You want true binary integers?  Use the "I" data type (or "U" if you want
unsigned integers).

The "B" data type is nasty legacy of RPG III that we have to keep for
backward compatibility. All math operations on type "B" are done in PACKED
(despite that integer math is faster!) and therefore the "B" type must be
capped-off at decimal digit boundaries -- because packed is capped off at
those same boundaries.

The "I" and "U" data types in RPG IV are true binary integers that work
just like the ones you're used to in other programming languages. Not only
do they hold the full range of values, they perform much better!


> Am I seeing things, or is this how it works in RPG?
>
> And assuming I'm not seeing things, can anybody explain WHY RPG does this,
> when no other programming language that I'm aware of (and anybody who's
> read one of my diatribes advocating mixed-language programming knows I'm
> "aware of" a great many, and fluent to varying degrees in most of them)
> does this?

All I know is that RPG IV has to do it because RPG III did it.  And they
added the "I" and "U" data types to RPG IV because the "B" data type
couldn't be changed without causing problems.

Why they original designed the "B" data type that way in RPG III is
completely beyond me.