Mailing-List: contact derby-dev-help@db.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Derby Development" <derby-dev@db.apache.org>
Received-SPF: pass (hermes.apache.org: local policy)
Date: Thu, 21 Apr 2005 20:40:59 -0700
From: David Van Couvering <David.Vancouvering@Sun.COM>
Subject: Re: Sun, Java and BigDecimal incompatible changes in J2SE 5.0
In-reply-to: <42681FB3.8010206@debrunners.com>
To: Derby Development <derby-dev@db.apache.org>
Message-id: <4268724B.9020908@sun.com>
MIME-version: 1.0
Content-type: text/plain; charset=us-ascii; format=flowed
Content-transfer-encoding: 7BIT
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.6)
 Gecko/20040113
References: <42681FB3.8010206@debrunners.com>

Hi, Dan.  Those of us on the list are pretty far away from these kinds 
of changes to J2SE.  That said, I do have some contacts, and will try 
and follow-up on this and get back to you.  Your open-ended question 
"are there any more incompatibilities like this that could break Derby 
code?" I think implies are there any more *known* incompatibilities. 
I'll try to find that out too.

David

Daniel John Debrunner wrote:

> I'm wonder if any of the Sun folks on the list can help out, J2SE 5.0
> has changed the behaviour of java.math.BigDecimal in incompatible ways
> from JDK 1.4.x, but these are not mentioned in
> 
> http://java.sun.com/j2se/1.5.0/compatibility.html
> 
> Is there any document that details all the ways J2SE 5.0's BigDecimal is
> incompatible with all earlier Java releases?
> 
> While we changed Derby to address the change in behaviour of
> BigDecimal.toString() a while ago, I just discovered another change that
> breaks Derby and potentially other applications. The method
> BigDecimal.scale() used to be
> 
> (JDK 1.4.1 javadoc)
> Returns the scale of this BigDecimal. (The scale is the number of digits
> to the right of the decimal point.)
> 
> but now it can return negative numbers
> 
> (J2SE 5.0 docs)
> Returns the scale of this BigDecimal. If zero or positive, the scale is
> the number of digits to the right of the decimal point. If negative, the
> unscaled value of the number is multiplied by ten to the power of the
> negation of the scale. For example, a scale of -3 means the unscaled
> value is multiplied by 1000.
> 
> 
> I discovered this when looking at Derby-225 and simple cleanup changes
> that should have not fixed the issue suddenly fixed it. Further
> investigation showed a scale of -302 which of course confused the
> existing code.
> 
> I wonder if there are any more incompatibilities in this class that
> could break Derby code?
> 
> Thanks,
> Dan.
>