Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!snorkelwacker!bloom-beacon!eru!luth!sunic!dkuug!freja.diku.dk!skinfaxe.diku.dk!jensting
From: jensting@skinfaxe.diku.dk (Jens Tingleff)
Newsgroups: comp.lang.modula2
Subject: Re: FOR loops
Message-ID: <1990Jul4.074634.645@diku.dk>
Date: 4 Jul 90 07:46:34 GMT
References: <INFO-M2%90070316593556@UCF1VM.BITNET>
Sender: news@diku.dk (The Netnews System)
Organization: Department Of Computer Science, University Of Copenhagen
Lines: 116

ROSS@UCF1VM.BITNET (Bri) writes:

[..]

>Anyone else have any answers as to why Wirth designed such a restriction?

could be a continuation of the restriction from Pascal, i.e. a FOR
loop index MUST be a locally declared simple integral variable. Having
the syntax say ``Ident'' instead of ``Designator'' hints at this 
restriction (if not enforcing it, UGH! semantics in grammmmar are ugly..).

The reason for demanding a local variable is (or could be ?) that a
local variable can be detected to be non-aliased. I.E. that no context
switch can alter the loop control variable (a context switch such 
as an interupt). Since only global procedures are allowed as co-routines
(see p. 167 of PIM2 ed 3), it's impossible to change the value of 
a FOR loop index with a co-routine. Modula-2 also requires a local 
var, incidently, PIM2 ed 3 p. 158.


The following is thus illegal (pseducode..)

PROCEDURE lotawork();

  VAR controlvar, bar : INTEGER;

  PROCEDURE Dosomework():INTEGER;
  VAR foo : CARDINAL;
  BEGIN foo := 0;
    FOR controlvar := 0 TO 700000 DO		(* Not allowed %1	*)
      foo := foo + 1;
    END;
    RETURN foo;
  END Dosomework;

  PROCEDURE Stompit;
  BEGIN
    controlvar := 1000000; 
    ContinueAfterInterupt();
  END Stompit;

BEGIN (* lotawork *)
  SetToInteruptEverySplitSecond(Stompit);	(* Not allowed, %2 *)
  bar := Dosomework();
END lotawork;

where the value of `bar' just might be weird.
%1	Nonlocal variable leads to disaster for any call to Stompit.
%2	local procedure an actual for PROCESS formal parameter, not allowed.
So, this mayhem is illegal (phew).


This is, however, legal (???)

MODULE globalmodule;

FROM SYSTEM INPORT ADR;

TYPE intptrtyp = POINTER TO INTEGER;
VAR intptr : intptrtyp;

  PROCEDURE Stompit;
  BEGIN
    intptr^ := 10000000;
    ContinueAfterInterupt();
  END Stompit;

  PROCEDURE notsosafe;
    VAR i : INTEGER;
  BEGIN
    intptr := ADR(i);		(* Boo, aliasing *)
    FOR i := 0 TO 700000 DO 
      whatever();
    END;
  END notsosafe;

BEGIN (* globalmodule *)
  SetToInteruptEverySplitSecond(Stompit);
  notsosafe;
END globalmodule;

This last example, possible aliasing of `i', CAN be detected by the 
compiler, at which point a warning may be printed out.


ANYWAY. A slightly more levelheaded reason for requiring the FOR control
var to be simple and local is that it's easier to decide to make 
the loop-cntr. var a register var...... .

By The Way, new question :
  Is the FOR loop control var defined after the FOR statement ? Wirth
  (PIM2 ed 3 p 158) doesn't say not, in fact the paragraph reads

 	"FOR v := S TO B BY C DO SS END
	 expresses repeated execution of the stat. seq. SS with
	 v successively assuming the values A, A+C, A+2C, ..,
	 A+nC where A+nC is the last term not exceeding B."

  It says "v successively assuming the values", but not what v is AFTER
  the statement. Hmmm.

  Modula-3 (see page 25 of the revised report) says

	"FOR id := first TO last BY step DO S END;
	 .. The identifier id denotes a readonly variable whose
	 scope is S and whose type is the common basetype of first & last."
	 ^^^^^^^^^^
  right on. Likewise Occam-2 (and algol68 ?)


Jens

Jens Tingleff MSc EE, Institute of Computer Science, Copenhagen University
Snail mail: DIKU Universitetsparken 1 DK2100 KBH O
"It never runs around here; it just comes crashing down"
	apologies to  Dire Straits