On Feb 14, 2005, at 9:57 AM, Ceki G�lc� wrote:
As long as the test for stripDuplicateBackslashes exists and gets executed frequently, then I think the details should be left to the player doing the footwork -- that means you in this case.
If you are in addition saying that the many of the test in log4j are too coarse-grained, then I agree. It would be really nice if we could devise smaller tests. I'd appreciate your input on this matter.
The stripDuplicateBackslashes exists and does get executed so it effectively tests the implementation of FileAppender.stripDuplicateBackspaces. I put it in the most convenient existing test case that I could find, but that results in it running as part of the AsyncAppender test case though the method under test is particularly related to AsyncAppender's.
It is a "fine-grained" test, but there isn't an obvious home for "fine-grained" tests within the current test structure. What I was wondering if we should provide a home for such tests and if so, how to do it. I don't think we want to add an Ant target for each class in log4j. Maybe add a new target for method and class level tests which runs a corresponding test suite class. So afterwards if I want to add a unit test for an arbitrary class, I create the test class and add it to the suite and not have to touch the ant script.
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
