Thanks for confirming that, Knut. Are the logs for compatibility test online when they are run nightly? It would have been nice for these startup/ping/shutdown to be part of the actual test so that manual inspection of the log is not required.
I will soon post the updated testScript.xml which covers all the releases so far so that all the combinations can be tested. It will definitely increase the time it takes to finish the test so may be we want to pick only the highest point release of each branch! I will post that script to DERBY-4786. thanks, Mamta On Thu, Sep 23, 2010 at 3:37 AM, Knut Anders Hatlen <[email protected]> wrote: > Mamta Satoor <[email protected]> writes: > >> I was curious how was DERBY-3644 was discovered. If I understand it >> right, the results of server start, ping and shutdown for the >> compatibility test are all done in the >> org\apache\derbyTesting\functionTests\tests\junitTests\compatibility\testScript.xml >> which means there is no way to verify if those commands are behaving >> the way they should be. The rest of the compatibility testing happens >> in CompatibilityCombinations.java and in problems there are probably >> caught by that test. By I do not think any problems with >> start/ping/shutdown done through the testScript.xml will be caught >> anywhere. Is my understanding correct? > > Yes, I think so. From the bug description it sounds like this was > something Ole had found when inspecting the logs, and the tests didn't > actually fail. I looked at the history for the compatibility tests, and > there were no failures reported around that time. > > -- > Knut Anders >
