[picocontainer-dev] NanoContainer getComponentInstance() ???

Hi all,

 

I was surprised to find that the following groovy script didn't work:

 

 

def CUSTOM_OBJECT_KEY = 'SpecializedInstanceKey'

 

pico = builder.container(parent:parent) {

      //..

}

 

def customObject = parent.getComponentInstance(UNIT_OF_WORK_FACTORY_KEY)

if (customObject == null) {

      println("Using Default Unit Of Work Factory")

      customObject = parent.getComponentInstance(OriginalClass)

}

 

pico.registerComponentInstance(CUSTOM_OBJECT_KEY, customObject);

 

 

 

The culprit was the fact that the line:

pico = builder.container(parent:parent)

 

returns a NanoContainer.

 

And NanoContainer doesn't have 

 

getComponentInstance(Object key)

 

 

I'd really like to not have to do in my script:

 

def picoToUse;

if (parent instanceof NanoPicoContainer) {

     picoToUse = ((NanoPicoContainer) parent).getPico();

} else {

      throw new IllegalStateException("Pico returned from container builder
wasn't NanoPicoContainer")

}

 

def customObject = parent.getComponentInstance(UNIT_OF_WORK_FACTORY_KEY)

if (customObject == null) {

      println("Using Default Unit Of Work Factory")

      customObject = parent.getComponentInstance(OriginalClass)

}

 

pico.registerComponentInstance(CUSTOM_OBJECT_KEY, customObject);

 

Since the NanoContainer  Interface allows registration by key of type
Object, couldn't it have a getComponentInstance(Object key) as well?

 

The reason for something like this is that I'm deploying subcomponents
through Deployer and I would like to have the capability to define a
separate DataSource per-component if necessary yet if that custom datasource
doesn't exist, simply utilize the default one defined in the parent
container.

 

Thanks, 

 

 
-Mike