Re: RPG overloading (was RPGV?)

[Barbara Morris <bmorris@xxxxxxxxxx>]

Bob P. Roche wrote:
> Isn't one of the features of prototyping calls that always gets pointed 
> out, "the compiler checks your data to make sure you don't have a 
> mismatch"? If the compiler is just going to change my data to whatever it 
> thinks it needs, how is that helping me not make a mistake and reverse 
> parms? I know it will get caught in testing, but prototyping was supposed 
> to save me from this kind of mistake at compile time, unless I 
> specifically told it not to by using CONST.

I guess if all the parameters had slightly different types, you could 
get that kind of parameter-order check.

But to ensure that check was reliable, you'd have to avoid using CONST 
and VALUE and you'd have to give a slightly different data type to each 
parameter.  Those seem like awkward restrictions to deal with.

Prototyping was supposed to save you from making type and size mistakes 
with individual parameters, but there was no particular intention that 
RPG programmers would use the parameter checks to make sure the 
parameters were passed in the right order.

But ... that's one of the scenarios that would have to be considered 
when deciding whether there would be a new keyword or whether the 
behaviour would just change.

This reminds me of a few complaints when the LR indicator was made 
optional on the READ opcode.  People complained that the editor didn't 
forced them to code the indicator, so they also forgot to code the eof 
check.