🚀 go-pugleaf

> I have a question about passing of array parameters in Oberon-2.
> Lets say I want to write a procedure to print a string passed in to
> it.  I would do something like:
> 
>         PROCEDURE PrintString(VAR s : ARRAY OF CHAR);
>         BEGIN
>            (* print string here *)
>         END Print String.
> 
> But then the following would not work.
> 
>         PrintString("Hello");
>         
> Because a constant is being passed in.  So instead, the PrintString
> routine would be:
> 
>         PROCEDURE PrintString(s : ARRAY OF CHAR);
>         BEGIN
>            (* print string here *)
>         END Print String.

Well here is one my pet weaknesses with PASCAL/M2/Oberon??. In parameter
passing the VAR (or its absence) is being forced into a dual service
of logically (from the programmers point of view) indicating whether a
parameter will be modified or not by the routine; and from the compiler
point of view, it is used for determining passing mechanism. Strictly
the programmer shouldnt have to worry about such implementation-dependent
detail but in practise the need for efficiency comes in. Many "real"
compilers use "pragmas" or "attributes" to convey that kind of extra
information to the compiler and personally I would like to see this practise
made consistant in the language. Ie allow optional attributes like 
[READONLY] on parameters.

Returning to your example, I dont see a problem with  
         PROCEDURE PrintString(VAR s : ARRAY OF CHAR);
         BEGIN
            (* print string here *)
         END PrintString.
Your compiler creates "hello" in a write-protected page (if your memory
hardware will support such a beast). It has an address like any other variable
which can be passed. If your routine tries to modify it, then a memory
exception will be generated at runtime. A lot of smarts needed to flag such
a problem at compile time though without the pragma/attribute approach.
	Phil