🚀 go-pugleaf

In article <1992Nov19.104407.338@jyu.fi>, sakkinen@jyu.fi (Markku Sakkinen) writes:
> Fortran, I think, has a rule stating that if a formal parameter is
> modified within a subprogram, the corresponding actual parameter must
> not be the same as any other actual parameter in the call.
> Does Modula-2 or Oberon have a similar rule?  (Value parameters need
> not be considered, but constant parameters would if passed by reference.)

I'm not familiar with Fortran but neither Modula-2 nor Oberon have
a rule which forbid to pass the same actual parameter to two or more
parameters, i.e.

   PROCEDURE p(VAR x, y: INTEGER); BEGIN ... END p;
   (* ... *)
   p(i, i);

is legal.

> Of course, those breaches of the rule that cannot be detected at compile
> time could be too expensive to catch at run time.

That's the problem and I still don't see a better solution by
use of constant attributes for parameters.

-- 
_______________________________________________________________________________

Andreas Borchert, University of Ulm, SAI, D-W-7900 Ulm, Germany
Internet: borchert@mathematik.uni-ulm.de