Hi,
I don't see a conflict of interests here. The proposed spec changes would
make some of the aspects configurable in the SCDL to provide the runtime
more information. The SPI discussion is for how Tuscany supports these
requirements. IMO, we can use the binding.jms use cases to validate the SPI
design.
Thanks,
Raymond
From: ant elder
Sent: Tuesday, September 09, 2008 12:16 AM
To: [email protected]
Subject: Re: Request Response Binding - was: Re: Tuscany data binding
framework enhancements
On Mon, Sep 8, 2008 at 7:33 PM, Raymond Feng <[EMAIL PROTECTED]> wrote:
I'm thinking about reusing the Interceptor SPI. We can pass in the native
message and context using the combination of the Message body and header
(to
be added).
Thanks,
Raymond
Just checking i'm not misunderstanding...this is bigger than just some SPI
changes isnt it? From the spec discussion emails eg
http://lists.oasis-open.org/archives/sca-bindings/200808/msg00036.html:
<binding uri="xs:anyURI"? name="xs:NCName"?
requires="list of xs:QName"? policySets="list of xs:QName"?/>*
<databinding />?
<operationSelector />?
</binding>
so there are new elements <databinding /> and <operationSelector />, and i
think there's been talk of a fault handler too. Note databinding has
subsequently been renamed wireFormat but i'll used the old name here for
consistency with the email link.
So we'd need to update our model to support those new elements everywhere,
have some default impls etc. I wondered about having some binding neutral
impls like <databinding.jaxb>, <databinding.xmltext> etc which then could be
used by specific bindings to get a native message so <databinding.jmsobject>
would use <databinding.jaxb> and <databinding.jmsxmltext> would use
<databinding.xmltext> etc.
Is this anything like what others are thinking?
...ant
