I like & get what your saying, I'd like to add something too.

Can we separate the Service Interface from the Smart Proxy? So they're in separate artefacts /downloadable jar's?

Then the Client depends on the Service Interface artefact, but not the proxy implementation-dl.jar artefact, leaving the Service to use any proxy-dl.jar artefact it likes. (Via Entry)

Then the Client chooses the ServiceInterface artefact, which dictates the Service, while the Service implementation dictates the proxy artefact, for a total separation of concerns. Both the client and Proxy then depend on the ServiceInterface-dl.jar artefact, but not each other directly.

If we don't grant ServiceInterface-dl.jar any permissions whatsoever and place it in a Parent ClassLoader, then it can be visible to both the client and the proxy, and the Proxy and the Client will be otherwise isolated from each other, with their own permissions and avoid any risk of namespace conflicts.

That probably helps to make more sense of my previous message too.

Once the required classes are in place, after the proxy has been unmarshalled, the Remote calls marshalled streams can be annotated with the ServiceID, so the client platform knows which classloaders to use.

The separation of concerns, the Service Interface, from the proxy implementation, would allow us to provide a runtime dynamic service that advertises available Service Interfaces, in a GUI browser, including Javadoc for developers to discover and explore new and different Service Interfaces as they are created. This might be useful for Groovy script developers too, I might be getting ahead of myself here.

I'm starting to have a much more positive feeling about this project, we definitely seem to be making headway.

Cheers,

Peter.

Dennis Reedy wrote:
Why cant the client express a dependency on that associated service's -dl.jar 
file(s)?

The client depends on the Service Interface, not the -dl.jar, what you probably 
meant was Marshalled Objects depend on the -dl.jar during Unmarshalling.

No, not at all.

The issue one finds when developing services is that you need to have at build and 
deployment time the interfaces you require in order to build, test and deploy a 
complete application composed of services. Why change this when you move from 
development & test to production? Why introduce the frailty of lost codebase 
problem when you know a priori what DL jars you need to load classes from in order 
for the smart proxies services you have developed use?

We already know what the DL jars are, and what is needed in order to resolve 
compile and runtime dependencies. There is nothing to discover (other than the 
advertised service(s) of course).

The services that we make available on the network use the artifacts we create 
and bundle to publish smart proxies that are resolvable from DL jars. I suggest 
that we know what the dependencies are of these services in advance. We need to 
in order to develop test cases, to develop the fundamental interactions 
required to enable the semantics across services for the applications that we 
write.

This is not a matter of class loader hierarchy or Marshalled Objects, its 
simply getting the required jars local to the client (or service acting as a 
client) such that the client can load those classes locally instead of loading 
them via a codebase service. The only reason that we need dynamic class loading 
is if the required classes are not resolvable by the class loader. Why not 
provision them to make these issues go away?

It doesnt change the smart proxy semantics necessarily, it just changes 'where' 
the classes get loaded from. If we can provision the requisite artifacts to the 
client, then the classes can be loaded locally.


Yours and Patrick's ideas are good, including the codebase Entry idea, as you 
suggest a Service could pre-empt installation of the correct version *-dl.jar 
prior to unmarshalling objects.  That *-dl.jar could have dependency's it 
requires resolved by a frame work too, all prior to any unmarshalling.

Yep, these are known as transitive dependencies. All known based on convention, 
and easily navigatable.




Reply via email to