Re: Odd CALLPRC behavior when using RPG indicator with RTNVAL

David Gibbs wrote:

I'm trying to call a RPG procedure in a CL program ... the RPG
procedure returns an indicator (N) data type.

The normal way of handling this would be to specify a 2 character
field as the RTNVAL parameter on the CALLPRC statement, and then test
the first character of the field for '1'.

This is at least partly my fault, more likely all my fault. What you're doing used to work prior to V6R1, and it's probably due to something I posted that this is the normal way of handling this. (I believe it was even in the RPG manual before EXTPROC(*CL) was added in V5R1.)

What I should have suggested back then was to have a wrapper for the RPG procedure that returned a 2 byte variable. That way there would have been no lying to the compilers. And no dependency on an undocumented feature of the system.

Join me in the green-on-black ribbon campaign against lying to the compiler. 8<
I will not lie to the compiler.
I will not lie to the compiler.
I will not lie to the compiler.
(I'm not using cut and paste, honest)
I will not lie to the compiler.