Axis2-1.2 - ServiceClient configuration - question on the design of ServiceClient.configureServiceClient method

[Shantanu Sen <ss...@pacbell.net>]

In the ServiceClient.configureServiceClient method it assumes that the ConfigurationContext will not have the AxisService that is target service for the current invocation as shown in the following code snippet:

---
private void configureServiceClient(ConfigurationContext configContext, AxisService axisService)
            throws AxisFault {
        initializeTransports(configContext);
        // save the axisConfig and service
        this.axisConfig = this.configContext.getAxisConfiguration();
        if (axisService != null) {
            this.axisService = axisService;
        } else {
            this.axisService = createAnonymousService();
        }
        if (this.axisConfig.getService(this.axisService.getName()) == null) {
            this.axisService.setClientSide(true);
            this.axisConfig.addService(this.axisService);
        } else {
            throw new AxisFault(Messages.getMessage(
                    "twoservicecannothavesamename",
                    this.axisService.getName()));
        }
----
What is the rationale for this code i.e. throwing an exception if the service is already there in the configuration?

If I am invoking multiple operations of the same service (or the same operation repeatedly multiple times), why cannot I re-use a ConfigurationContext that I have created previously?

I want to avoid the performance penalty of creating a new configuration context for the client (i.e. reading the axis2.xml, loading the modules and so forth) every time I invoke a service. So, I was thinking of creating a ConfigurationContext and storing it in a ThreadLocal, so that multiple invocations of the same service in the same thread can use the cached ConfigurationContext.

Thanks for any info on this.

Shantanu Sen