Thanks for looking at this. I agree with you about the impenetrability of the test code - that is why I'm working by modifying existing tests rather than just writing the tests I need from scratch.
If this becomes the limiting factor on a release, I'll check in the patch without a QA test. However, I am strongly in favor of QA tests for the sake of regression testing. I want Hudson to tell us if some future change brings back the vulnerability to listener exceptions. Patricia On 4/15/2011 11:21 AM, Christopher Dolan wrote:
Patricia, I looked at the existing AddNewDiscoveryListener and agree with your assessment. But the test code is nigh impenetrable, I have to say. It would be beneficial in the long run if some of the common code in LookupDiscovery and LookupLocatorDiscovery were broken apart into separately testable units... I tried writing my own test case, but it doesn't seem possible to mock any of the providers, so to test it you need an actual registrar, it appears. For what it's worth, today I wrote my own private test that runs in my djinn and verified the patch. Chris -----Original Message----- From: Patricia Shanahan [mailto:[email protected]] Sent: Thursday, April 14, 2011 7:41 PM To: [email protected] Subject: Test anomaly I'm attempting to construct QA tests for the RIVER-395 by copying existing tests of event delivery and adding a special listener that throws exceptions. If the patch is correct, the modified test should pass with all expected events going to the existing listeners despite the exceptions. I'm running my tests with 'com.sun.jini.qa.harness.level = FINEST' logging, so that I can assure myself that the exceptions are being thrown and events delivered after an exception. I tried to base a test on com/sun/jini/test/spec/discoverymanager/AddNewDiscoveryListener.td, and got no events. I've since run the complete discoverymanager category. From the logs, I got no change events, and only 4 tests appeared to have non-zero numbers of initial events. Could someone else take a look at this and see if this is as expected? It looked to me from the comments in AddNewDiscoveryListener.java as though processing a non-zero number of events is important to the test. Patricia
