Re: Odd CALLPRC behavior when using RPG indicator with RTNVAL

Charles Wilt wrote:

In both cases, the code that breaks wasn't "correct" in the first
place. In may have worked, but that was basically luck.

What wasn't correct to begin with was the fact the CL and RPG weren't compatible! If they wanted to fix something, they should've fixed THAT and eliminated the need for the ExtProc(*CL) kludge.

When your files are properly tagged with CCSID, CPYxxxIMPF works fine.
When you use the *CL keyword added in v5r1, everything's fine too.

Well, that's sort of true. The problem is, the CL/RPG incompatibility existed before V5R1! It was a problem in V3R2, V3R7, V4R1, V4R2, V4R3, V4R4, and V4R5. Our only option during that time was to use the 2-byte field workaround (and that workaround was given to us by IBM!) Folks may have many thousands of programs that need to be changed, possibly breaking compatiblity with THEIR customer's programs, etc, etc, since there was a long period of time between V3R2 and V5R1.

Granted, David said it's not a big deal for him. It's also not a big deal for me, I have only a few things to change.

But if you're going to put the entire IBM i community through the hassle of finding and fixing all of the places this happened, it sure would've been nice to fix the underlying issue as well, instead of just breaking the workaround.