Hi Amindri,
It could be good to also support some of the SCA domain things like
installed contributions and running composites, but you don't need to worry
about finding everything yet it might be easier to get going if you just
pick a small few and get something working for those as a first step.
I'm not sure i fully understood the comments about JSON and Wink? Was that
just in relation to access to the Message payloads?
...ant
On Sun, May 15, 2011 at 9:27 AM, Amindri Udugala
<amindriudug...@gmail.com>wrote:
> Hello,
>
> Here is an update of my work so far,
>
> I created a new jmx module in the modules folder. There I created MXBean
> interfaces for some runtime components,
> Here are the runtime components,
> EndpointReference
> Message
> ProviderfactoryextensionPointRuntimeComponentreference
> RuntimeComponentReference
> RuntimeComponentService
> RuntimeWireProcessorExtensionPoint
> RuntimeWireProcessor
> and all other interfaces implementing ProviderFactory interface
>
> (please mention if I need to add more runtime components)
>
>
> These MXBean interfaces are implemented in classes , in "mgt" sub package
> which I created in tuscany-spi-core,
> I've got few more questions before proceeding....
>
> When I was going through the Message interface, I understood that neither
> of two implementation classes set the messegeID
>
>
> public Object getMessageID() {
> return null;
> }
>
> public void setMessageID(Object messageId) {
> throw new UnsupportedOperationException();
> }
>
>
> Since many messages can be generated I think it is important to identify
> message uniquely. When implementing the JMX management console I propose to
> use JSON messages for communication (Wink can be used for this). If a user
> needs to change, set or view the message body, of a particular message how
> to uniquely identify the message?
>
> The same problem exists with other interfaces (eg EndpointReference) as
> well (No way to uniquely identify objects). This would not be a problem if
> only one object is created from each runtime components. Or may be that I
> misunderstood the whole procedure. Please correct me if I'm wrong
>
>
> --
> Thanks
>
> Amindri Udugala
> University of Colombo School of Computing,
> Sri Lanka.
>
>
