Re: Parameter passing RPV III <--> RPG IV vs. RPG IV <--> RPG IV

[Alan <steelville@xxxxxxxxxxx>]

This is not the only situation where you would think something should 
have bombed and didn't. 

More than once I have wracked my brain looking for a problem because 
something has the exact, precise same code as always and the last time 
it was changed was eight years ago, and yet some new data value or some 
new situation, or who knows what, was different today.

- Alan


Joep Beckeringh wrote:

> Charles,
>
> I think you were lucky that it worked. In the caller you have field 
> P@ERR defined as 1 byte. When you call CADATE you are passing a 
> pointer to this field. Now when CADATE puts '123' in the memory that 
> the pointer points at, the '1' ends up in field P@ERR, but '23' ends 
> up in whatever variable happens to sit next to P@ERR. It is quite 
> possible that in the memory layout that the RPG III compiler generated 
> this did no harm, while in the RPG IV layout it did.
>
> Joep Beckeringh