Hi everybody, I'd like to start by saying that Jackrabbit is just great.
We're integrating Jackrabbit in a J2EE application at my company and I'm afraid we've found few bugs. I've managed to create 2 tests cases showing the issues. Basically the problem occurs when we're using XPath queries with child axis predicates like //*[a/b/c/d/e/@prop='something']. With this kind of query the results may contain too many nodes or not enough nodes. The attachment contains the source code for the JUnit tests (I apologize but the test data are in French). Any help would be really appreciated. Don't hesitate to ask for more information if needed. Best regards, Frederic Laugier.
