[cp-patches] The CORBA status

[Meskauskas Audrius]

>Another question: Do you 'only' implement the CORBA framework? Or do 
you work on a complete CORBA implementation?

We work on the complete CORBA implementation.

> If the latter, how far is it?

The CORBA part, supported by Sun's JDK API 1.3, should work and, if 
needed, talk with another CORBA, running on Suns virtual machine up till 
JDK 1.4 inclusive.

>Maybe you have some testapps for showing what already works?

There are examples in the folder examples/CORBA/SimpleCommunication. 
These include a CORBA server and CORBA client. Client invokes remote 
methods of the object, located on the server side. The parameters are 
passed and returned. The exceptions, thrown by remote method, are 
re-thrown on a client side.

As the example must be simple to install, the Demo.java starts both 
client and server threads on the same virtual machine. However it is 
possible to start the server (executable class comServer) on one virtual 
machine and client (executable classes DirectTest, RequestTest) on 
another virtual machine, using another CORBA implementation. Hence it is 
simple to test interoperability with Suns JDK 1.4 and be sure in correct 
understanding of the OMG specifications.

We also already have CORBA naming (CosNaming) service that returns the 
object address in response to the object name. The example in 
examples/CORBA/NamingService shows how to use it. Our implementation is 
also talking rather friendly with the Suns jdk 1.4 naming services, both 
transient and persistent.

Also, there are numerous CORBA tests in the Mauve, including the third 
party tests from the former cost.omg.org project (Free CORBA testing). 
In these tests, server and client also starts on the same virtual 
machine. However I always verify the functionality starting one side on 
Suns and another on ours CORBA implementation.

The more detailed summary of the implemented features is placed on the 
NEWS file on the CVS.

Cheers
Audrius