In article <[EMAIL PROTECTED]>, John Roth <[EMAIL PROTECTED]> wrote:
> > Does the following patch has a chance of being introduced in the > > standard python distribution? > > I certainly hope not! I think you're being overly negative here. Antoon went to the trouble to read the sources and post a diff. At the very least, he deserves a more polite response. But, more than that, I think Antoon's idea has some merit. I understand that the mantra in unit testing is that the tests should be able to be run in any order. Still, it's the job of a library package to make it easy to do things, not to enforce policy. If somebody (for whatever reason) has a need to run their tests in a certain order, why is it our job to make it hard for them to do that? In fact, unittest.defaultTestLoader *already* sorts the tests into alphabetical order by their name. So, if somebody wanted to get the tests run in order, they could just name their tests "test0001", "test0002", etc. In fact, I've done that in the past when I had some (long forgotten) reason why I wanted to run a bunch of things in order. Allowing the tests to be sorted by line number order instead of by name just makes it a little easier to do the same thing. If somebody wants that functionality, and is willing to put in the effort to write the code to do it, and contribute that back to the community, I don't see any reason why it shouldn't be considered. It would have to be done in a way that doesn't change the current behavior (perhaps by shipping a subclass of TestLoader which users could use instead of the default). I'm not saying that we *need* to include it, just that it's not such a bad idea that it deserves responses like "I certainly hope not!" -- http://mail.python.org/mailman/listinfo/python-list