From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.3 required=5.0 tests=BAYES_00,INVALID_MSGID
	autolearn=no autolearn_force=no version=3.4.4
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 103376,86fd56abf3579c34
X-Google-Attributes: gid103376,public
X-Google-Thread: 10db24,77f71d0bde4c5bb4
X-Google-Attributes: gid10db24,public
From: dewar@cs.nyu.edu (Robert Dewar)
Subject: Re: What good is halting prob?
Date: 1995/04/19
Message-ID: <dewar.798337304@gnat>#1/1
X-Deja-AN: 101283273
references: <dewar.798172270@gnat> <cppD786FM.1u9@netcom.com>
 <1995Apr19.204744.21914@wdl.loral.com> <3n429h$pvu@acmex.gatech.edu>
organization: Courant Institute of Mathematical Sciences
newsgroups: comp.lang.ada,comp.edu
Date: 1995-04-19T00:00:00+00:00
List-Id: <comp.lang.ada>

It is easy to construct examples where the effect depends on the parameter
passing mechanism:

   procedure X (a, b : in out string) is
   begin
      a(1) := 'x';
      if b(1) = 'x' then raise program_error; end if;
   end x;

   G : string := "abc";
   H : string := "def";
   X (G, H);

will raise program error if the strings are passed by reference but not
if they are passed by copy. In Ada 83 this program execution is erroneous
but effect is not very well defined and so in Ada 95, a simpler definition
is made that this program is simply non-deterministic (i.e. it may do one
of two well defined things, depending on the parameter passing method).