Re: creating a new procedure thatcallanexistingprocedurewithoneadditional parm

Paul Morgan wrote:
> 
> Is this requirement to check %Parm more to do with ILE or with RPG?  Why
> wouldn't the call using *NOPASS work like using *OMIT?
> 
> Use the example Charles gave in another message why should:
> 
> MyProc(parm1:*omit);
> MyProc(parm1)
> 
> be any different?
> 

*NOPASS works for parameters passed by value as well as parameters
passed by reference.  A null pointer couldn't be passed for say an
integer parameter.  If a null pointer was passed for const or reference
*NOPASS parameters, it would make the %PARMS count incorrect.

If you mean "why doesn't the system put some default value where the
parameter would be if it were passed", that's an ILE thing I guess, but
I don't think any system does that, passing "something" for unpassed
parameters.  Most of the OPTIONS in RPG allow increased flexibility for
calling the procedure at the cost of increased complexity in the
procedure itself.  They allow the caller to bypass the normal rules, but
the programmer of the procedure then has to check whether the rule has
been bypassed.  (I think it's a fair trade-off, since the
complex-to-write procedure is only written once, while the
easier-to-call prototype might be used for thousands of call
statements.)

Farfetched scenario: The difference between those two calls of yours
might be important to some procedures.  Say your first version of MyProc
only had one parameter, and your second version has new *OMIT
parameter.  You want old callers to still be able to call your
procedure, so you also make it *NOPASS.  Here's the farfetched part:
maybe your procedure wants to be able to tell whether it has an old
caller or a new caller.