Re: Wanna see something scary?

Thanks, Brian, I want to check that out on 7.1. That was one of my thoughts, that somehow the prototype might be the issue.

<checking...>

Okay, adding a prototype requires the DFTACTGRP as you noted. (You shouldn't need a named PI). I got TEMPVAR: [Data ]. I removed the prototype (but left the DFTACTGRP) and got blanks again. So, yes, the prototype changes the behavior of the program.

It's especially weird that the lack of a prototype allows the program to compile without DFTACTGRP. Obviously there's way more going on than meets the eye here.

Joe

The plot thickens, Joe.

I tried your original example on a V7.1 m/c. Same results as you: DSPLY
TEMPVAR: [ ]

Copied code to a V6.1 m/c and made the modifications necessary to get it to
compile, (additional PR definitions and named PI definitions).

H DftActGrp(*No) ActGrp(*New)

d MyConstant c 'Data'

D DoProc Pr
D MyProc Pr
D MyParm 10A Const

/free
DoProc();
*inlr = *on;
/end-free

p DoProc b
d DoProc pi
/free
MyProc( MyConstant);
/end-free
p e

p MyProc b
d MyProc pi
d MyParm 10A const
d tempvar s 10A
/free
tempvar = MyParm;
dsply ('TEMPVAR: [' + Tempvar + ']') '*REQUESTER';
/end-free

This compiled and produced DSPLY TEMPVAR: [Data ]

Copied above example BACK to V7.1 m/c and compiled. Results were DSPLY
TEMPVAR: [Data ] !!

So, explicit PR/PI definitions - though not necessary - cure the problem.
PTF required, methinks.

Brian.