Mailing-List: contact dev-help@cxf.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cxf.apache.org
Received-SPF: neutral (athena.apache.org: local policy)
From: Daniel Kulp <dkulp@apache.org>
To: dev@cxf.apache.org
Subject: Re: Possible Issue in
 org.apache.cxf.jaxws.interceptors.SwAOutInterceptor
Date: Thu, 21 May 2009 16:30:58 -0400
User-Agent: KMail/1.11.3 (Linux/2.6.29-gentoo; KDE/4.2.3; x86_64; ; )
Cc: "Ault, Matt" <Matt.Ault@elavon.com>
References: <74FC5E5EFE88994CBE00BAEDEB2BE9FA3247236858@ATLMAIL.nova.prv>
In-Reply-To: <74FC5E5EFE88994CBE00BAEDEB2BE9FA3247236858@ATLMAIL.nova.prv>
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200905211630.58919.dkulp@apache.org>


Interesting.   Didn't know that ConcurrentHashMaps couldn't have null values.   
Definitely different than normal HashMaps.   Learn something new every day.  
:-0

Yea.  It's an issue.   Committed fix to trunk for 2.2.2.

Dan


On Thu May 21 2009 4:04:03 pm Ault, Matt wrote:
> I believe I have found a possible issue in the
> org.apache.cxf.jaxws.interceptors.SwAOutInterceptor.  The functionality
> appears to have changed between 2.1.3 and 2.2.1.  I'm not sure which is the
> correct functionality so I thought I'd post here first.
>
> This problem occurs when you have a JAXBContextImplementation that does not
> have a hasSwaRef method.  I found this issue when moving a service from
> Tomcat to WebSphere.  IBM uses a different implementation of JAXBContext by
> default and IBM's version doesn't have a hasSwaRef method on it.
>
> In 2.2.1 this results in a NullPointerException in this block of code:
>
>  1    private boolean callSWARefMethod(JAXBContext ctx) {
>  2        Method m = SWA_REF_METHOD.get(ctx.getClass().getName());
>  3        if (m == null &&
> !SWA_REF_METHOD.containsKey(ctx.getClass().getName())) { 4            try {
>  5                m = ctx.getClass().getMethod("hasSwaRef", new Class[0]);
>  6            } catch (Exception e) {
>  7                //ignore
>  8            }
>  9            SWA_REF_METHOD.put(ctx.getClass().getName(), m);
> 10        }
> 11        try {
> 12            return (Boolean)m.invoke(ctx);
> 13        } catch (Exception e) {
> 14            return false;
> 15        }
> 16    }
>
> This throws the NPE at line 9 SWA_REF_METHOD.put(ctx.getClass().getName(),
> m); because m is null.  In JDK 1.6.07 ConcurrentHashMap.put() throws a
> NullPointerException when the value is null.  This causes a SOAPFault to be
> returned to the client.
>
> I believe that this code should just return false if the hasSwaRef method
> doesn't exist (which is how this operates in 2.1.3).  If the put() wasn't
> called when m is null, then this code should return true if the method
> exists and false if it doesn't.
>
> Fortunately, I can work around this by setting the
> javax.xml.bind.JAXBContext system property to the Sun version
> (com.sun.xml.bind.v2.ContextFactory) which returns a JAXBContext that
> implements hasSwaRef().
>
> Let me know if you need any more information from me regarding this issue,
> I didn't post it to JIRA since I wasn't sure if the code was performing
> correctly yet.
>
>
> Matt Ault
> Systems and Technology
> Elavon
>
> ----------------------------------------- Note: The information
> contained in this e-mail and in any attachments is intended only
> for the person or entity to which it is addressed and may contain
> confidential and/or privileged material. Any review,
> retransmission, dissemination or other use of, or taking of any
> action in reliance upon, this information by persons or entities
> other than the intended recipient is prohibited. The recipient
> should check this e-mail and any attachments for the presence of
> viruses. Sender accepts no liability for any damages caused by any
> virus transmitted by this e-mail. If you have received this e-mail
> in error, please notify us immediately by replying to the message
> and delete the e-mail from your computer.  Elavon (formerly NOVA
> Information Systems)

-- 
Daniel Kulp
dkulp@apache.org
http://www.dankulp.com/blog