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: 103376,16a35419d117fb15
X-Google-Attributes: gid103376,public
From: rracine@draper.com (Roger Racine)
Subject: Re: discriminant
Date: 1999/07/12
Message-ID: <rracine.22.000F0325@draper.com>#1/1
X-Deja-AN: 500209823
Sender: nntp@news.draper.com (NNTP Master)
References: <3789bfc2.97977684@news.dsuper.net>
 <rracine.21.0008309A@draper.com> <378c1097.184220259@news.dsuper.net>
Organization: Draper Laboratory
Newsgroups: comp.lang.ada
Date: 1999-07-12T00:00:00+00:00
List-Id: <comp.lang.ada>

In article <378c1097.184220259@news.dsuper.net> fluffy_doo@dsuper.net writes:
>On Mon, 12 Jul 1999 13:11:20 GMT, in comp.lang.ada you wrote:
>>After that, I created a test, which worked fine on the latest 
>>GNAT (3.12a) on Windows NT.
>>
>>What version of what compiler are you using on what host?

>Aonix :         ObjectAda for Windows V7.1.105 (special edition).
>OS :            Windows 95 (4.00.950)
>machine :       P150

>>Can you provide a complete test that fails?

>There's no failure, and there *should* be one.  I was using the
>debugger while checking something else when I noticed the problem I've
>described (which made it difficult or impossible to continue the
>original check with the debugger I was trying to make).  Maybe it's a
>debugger problem.  Maybe it's just not showing me the actual value for
>the variable I'm looking at.

>Anyway, I'll make some changes so the var names and comments
>correspond to English terms and I'll make another post with the
>attachment containing the function that returns a T_Number type.  I
>should probably also include the function that makes the reciprocal
>conversion ( T_Number => String ) so that you can at least output to
>the screen.

>Keep in mind though that both functions have errors in them, they're
>earlier versions.  The errors are not related to the problem I'm
>discussing here.  Also, in the current versions there is no longer a
>"Point" field with the "kind = REAL" discriminant.  The whole point of
>this post and my previous one is not the actual value returned or
>output of my String => T_Number function but what goes on inside it:
>assigning a new value to the "Power_R" field affects an unwanted and
>unexplained change in the value of the "Point"  field, and the new
>value of the "Point" field is *outside* the field's range.  When I
>then try to assign a new value to Point, a few lines of code down, its
>value doesn't change, like it's locked in, but the debugger doesn't
>complain.  This is what I see with the debugger.

>I've got to go to school now.  I'll do it tonight.

>Thanks for helping.

It looks like my test was similar to your code.  I simply tried to assign a
value to the Power_R field without changing anything else.  I did it three
times in a row without any problems using GNAT.  

I just used the AONIX compiler (same version as you) to see 
what happens, and your behavior is seen (the value of Point is changed 
to a garbage value, at least as seen in the debugger).

It looks like it is an AONIX bug.  

Roger Racine