Re: Odd CALLPRC behavior when using RPG indicator with RTNVAL

[Tom Liotta <qsrvbas@xxxxxxxxxxxx>]

Barbara Morris wrote:

>   If 
> either RPG+COBOL or CL had changed how they handle 1 byte 
> character/indicator return values, it would have required a recompile of 
> all existing programs.

I've felt that CL could be more flexible than that. We've seen 
plenty of new commands and command parameters over the years, not to 
mention new data types. I wouldn't have been at all surprised to see 
a new parm added to CALLPRC that simply returned the value differently.

Or a new RtnVarAs() parm that defaulted to *PREVIOUS behavior, but 
allowed a new value of *RPGV6R1 (or whatever) to described _how_ a 
value should be returned. Developers could choose how/when to 
convert existing code.

Anyway, it seems to me that many possibilities could have handled a 
transition with minimal need for recompilation. Or is the ILE CL 
compiler too different from OPM to allow it?

Tom Liotta