Thanks for responding Raymond.

In my test the pkg-level JAXB annotations were left out of the app by
mistake, but I was surprised to find it worked with the old transform but
not the new......  which is why I put this post out there for others to
maybe learn from.

So that makes sense.  Unfortunately though, I don't have any good ideas on
how to deal with this and still benefit from the OMDataSource performance,
but at least mine was a corner case and not a top priority I would think.

Scott

On Mon, Jul 7, 2008 at 2:12 PM, Raymond Feng <[EMAIL PROTECTED]> wrote:

>  Hi,
>
> For the XML data transformation, sometimes we don't care about the element.
> For example, if we pass a Customer from JAXB to SDO, we only care about the
> type (content) of the data. To be tolerating the case that both target and
> source sides expect the same type but different elements, we use
> "adjustElementName".
>
> Ideally for JAXB2OMElement, we should do the same trick. But we have an
> issue with JAXB as follows:
>
> Some JAXB classes are generated for global elements with @XmlRootElement
> annotation and the element QName is built into the JAXB metadata. The
> JAXB2OMElement uses the OMDataSource to lazily serialize JAXB data into XML
> stream. It would be difficult to adjust the element name at the stream level
> especially considering the performance penalty. Do you have any good ideas?
>
> Thanks,
> Raymond
>
>  *From:* Scott Kurz <[EMAIL PROTECTED]>
> *Sent:* Sunday, July 06, 2008 5:32 PM
> *To:* [email protected]
> *Subject:* databinding-jaxb-axiom doesn't do an adjustElementName
>
> I'm noticing a difference in the JAXB->Axiom serialization produced by the
> new databinding-jaxb-axiom transform compared to the old transform
> (JAXB2Node, Node2XMLStreamReader,XMLStreamReader2OMElement).
>
> The difference is caused by the fact that the older transform (before
> databinding-jaxb-axiom) had a call out to DOMHelper.adjustElementName() to,
> like the method name suggests, adjust the element name to the one calculated
> by tuscany-interfacedef processing, rather than relying only on JAXB APIs to
> calculate the interface name.
>
> Before I go into detail with my test app, (which I think I want to change
> anyway), let me just ask this:  did we intend to make the same "adjustment"
> with databinding-jaxb-axiom?
>
> I'm not 100% sure of the reason for doing the adjustment in the older code,
> so I just wanted to start with that question before bothering with my test
> details.
>
> Thanks,
> Scott
>

Reply via email to