[Tutor] unit testing

[Tom Plunket]

I have just been getting into Python, but one of the most
infuriating things that I've come across is that I cannot for the
life of me figure out how to easily write unit test code that
automatically gets loaded and imported at runtime.  As a C++ pro,
CppUnit did exactly what I wanted it to do, and there was never
any question.  PyUnit, however...

My test.py is pretty much ripped off from the regrtest.py found
in Py22's test unit (although I also don't understand why I can't
import test.regrtest):

#!/usr/bin/python, but I'm on Windows.  :)

import unittest
import os

if __name__ =3D=3D "__main__":
    names =3D os.listdir('.')
    tests =3D []
    for name in names:
        if name[:5] =3D=3D "test_" \
           and name[-3:] =3D=3D os.extsep+"py":
            modname =3D name[:-3]
            tests.append(modname)
    tests.sort()

    for test in tests:
        module =3D __import__(test, globals(), locals(), [])

    unittest.main()

Now if I get rid of all of this hooplah and manually=20

import unittest
from test_something import *
from test_somethingelse import *

if __name__ =3D=3D "__main__":
    unittest.main()

then everything works.  Why isn't my code pulling my tests into
my local namespace?  I really don't want to have to maintain this
(with CppUnit I could have test files that had no external
dependencies, I just had to link them into the project and
everything would start getting tests run.  I want to just have
all tests in every test_* file run automatically.)

Additionally- why is it that unittest throws an exception when it
ends properly?  Why on earth can't it be nice and procedural and
just allow execution to return normally?  (Mind you, I am
probably not "getting" Python yet, but that will come!)

thanks.
-tom!