Hi I understand the assertions in Analyzer to enforce 'final' on those methods, but I wanted to ask why do we care if a user's code does not declare them final? Why fail the tests on it? Can we change the assertion to fail if the code from o.a.l?
The thing is, we run tests w/ -ea to catch all sorts of errors (real errors) Lucene may detect. Having some custom analyzers non-final is something we may choose to live with, yet our tests keep failing. I don't want to argue if we should or shouldn't make them final - just ask why does Lucene code cares if the application code may not follow 'best practices'. Is there real danger in having my analyzer not declaring these methods final - something that can affect Lucene code for example? Or am I only risking my code? Running w/o -ea is not an option, so please avoid suggesting it :). Shai
