Monday, October 18, 2004

Resolvers and Tests
Same old, same old. Part of the day was spent finishing up with the build bugs for the remote resolver. Once this was all working fine I moved onto the unit tests. Like the resolver itself, I had implemented this already for the old interface, so it was more a matter of patching it up to work with the new interfaces.

While I went through the test code again I was reminded that it uses an in-memory model. The behaviour of these objects may have changed since this code was written, so any problems may stem from this area. I suppose I'll find out when I get there.

I also spoke to SR about the transaction interface. He hasn't determined exactly how to handle certain situations for distributed transactions, so for the moment there is no need to implement this code. He was pleased that I'm using the logging DummyXAResource class, as this will help him debug as he starts to implement this on the server.

My last step of the day was to try running the tests. The test code is particularly complex, so I didn't really expect it to work straight away, and it hasn't. For the moment the problem is a classpath issue in some reflection code. I'll probably spend tomorrow morning tracking this down.

No comments: