From owner-freebsd-java  Fri Jul  9 15:52:32 1999
Delivered-To: freebsd-java@freebsd.org
Received: from ns.mt.sri.com (unknown [206.127.79.91])
	by hub.freebsd.org (Postfix) with ESMTP
	id 71E3014E10; Fri,  9 Jul 1999 15:52:27 -0700 (PDT)
	(envelope-from nate@mt.sri.com)
Received: from mt.sri.com (rocky.mt.sri.com [206.127.76.100])
	by ns.mt.sri.com (8.8.8/8.8.8) with SMTP id QAA16326;
	Fri, 9 Jul 1999 16:52:21 -0600 (MDT)
	(envelope-from nate@rocky.mt.sri.com)
Received: by mt.sri.com (SMI-8.6/SMI-SVR4)
	id QAA10373; Fri, 9 Jul 1999 16:52:19 -0600
Date: Fri, 9 Jul 1999 16:52:19 -0600
Message-Id: <199907092252.QAA10373@mt.sri.com>
From: Nate Williams <nate@mt.sri.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
To: Martin Dieringer <dieringe@zedat.fu-berlin.de>
Cc: Nate Williams <nate@mt.sri.com>,
	Stephen McKay <syssgm@detir.qld.gov.au>, freebsd-java@FreeBSD.ORG,
	java-port@FreeBSD.ORG
Subject: Re: strange java calculation errors 
In-Reply-To: <Pine.BSF.4.10.9907092353330.44163-100000@ThinkPad.nowhere.local>
References: <199907091530.JAA05956@mt.sri.com>
	<Pine.BSF.4.10.9907092353330.44163-100000@ThinkPad.nowhere.local>
X-Mailer: VM 6.34 under 19.16 "Lille" XEmacs Lucid
Sender: owner-freebsd-java@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.org

> Ok, I set up a sample program. I think it needs to be an awt-application.
> I don't get any errors without setVisible(), but with a visible frame
> I get errors every few seconds. It has to be a Thread, too, I think.
> Probably you know how to track this down further.

I just spent a bit of time with this, and this is indeed a weird one.
It seems to depend on how the storage is allocated, since if I calculate
the value of denom as part of the local heap, it seems to work.  Either
that or I'm just getting lucky, because the code works if I calculate
the exact same code and store it in a 'local' variable.  (However, local
variables are treated almost exactly the same as global variables in
java.)

In any case, it is indeed a bug, and one that I don't see on Solaris.
Whether it's a bug in some FreeBSD library or in the way the VM operates
is still unknown.  If I had to make a guess, I'd start looking at the C
interpreter, given that we are one of the only platforms that use it,
while others use the assembly version.  However, it's just a wild guess.

Thanks again Martin!


Nate

> //BugTest.java ----------------------------------------------
> 
> import java.awt.*;
> import java.util.Random;
> 
> public class BugTest extends Frame implements  Runnable
> {
> 
>     Random random = new Random();
> 
>     double weight;
>     double denom;
>     double meatEnergy;
>     double weightgain;
> 
>     public BugTest(){
> 	setSize(100,100);
>  	setVisible(true);
>     }
> 
> 
>     public void run(){
> 	while (true){
> 	    try{
> 		calc();
> 	    } catch (ArithmeticException a) {
> 		System.err.println("weight: "+ weight + "\ndenom: "+ denom);
> 		a.printStackTrace();
> 	    }
> 	}
>     }
> 
>     void calc() throws ArithmeticException {
> 	weight =  random.nextDouble()*10.;
> 	meatEnergy= 0.029994563969828213;
> 	denom = (.0836 * Math.pow(weight, 1.37));
> 	weightgain = .2849372 * meatEnergy * weight / denom;
> 	if (weightgain > 1000.) 
> 		throw new ArithmeticException("weightgain "+weightgain);
>     }	
> 
>     public static void main(String args[]){
> 	BugTest bug = new BugTest();
> 	Thread bugThread = new Thread(bug);
> 	bugThread.setPriority(Thread.MIN_PRIORITY);
> 	bugThread.start();
>     }
>     
> }
> 
> //------------------------------------------------------
> 
> 


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-java" in the body of the message