On May 24, 2:44 pm, Roy Smith <[EMAIL PROTECTED]> wrote: > In article > <[EMAIL PROTECTED]>, > > Fuzzyman <[EMAIL PROTECTED]> wrote: > > Also, like others, I have had wonderful experiences of trying to track > > down test failures that depend on the order that tests run in. Having > > interdependencies between tests is a recipe for madness... > > I agree that tests should not depend on each other, but sometimes it's > still useful to have the tests run in a certain order for reporting > purposes. > > If you're doing requirements tracking, it's nice to have the tests execute > in the same order as the requirements are listed. It makes interpreting > the output easier. Sure, you could give the test cases names like > "test_fr17.3a" and write your own getTestCaseNames(), but just putting them > into the file in the order you want them to run is easier. And making > things easy is what this is all about.
Whilst I understand your point, I think the danger is that you end up with hidden dependencies on the test order - which you're not aware of and that the tests never expose. Certainly layout your tests in a logical order within the file, but I think you risk potential problems by controlling the order they are run in. Other frameworks specifically provide test order randomizers for this very reason. A worthwhile question for the OP - your patch seems fairly simple. Is it easy for you to extend unittest for your own testing needs by subclassing? Unittest should definitely be easy for people who *want* this to add it to their own testing environment. All the best, Michael Foord http://www.ironpythoninaction.com/ -- http://mail.python.org/mailman/listinfo/python-list
