Adds some test coverage to pydoc.

Hi,

Here are some small comments to your otherwise good to have patch:

-- assertEquals has been deprecated in favor of assertEqual, and usually it's great to be consistent across the test suite
-- likewise maxDiff should be max_diff mainly because of consistency and because of pep8's naming styles
-- self.temp_dir is never cleaned up (on tearDown maybe?) and gettempdir() doesn't really look like a good option; maybe you wanted to create a new temporary directory rather than pointing to the filesystem's tmp dir ?
-- the patch doesn't apply cleanly, at least for me - the ACKS hunk is failing for me

Sure thing, ill make the improvements and upload a new patch. Thanks!

--Changes assertequals to assertequal
--Removes maxdiff as it should not really be there in the first place
--Creates an explicit testdir, and cleans it up
--Last patch did not actually apply cleanly... I reverted and applied this one and this seems to work

Thanks for the patch.

Rather than create and destroy a directory for every test (setUp/tearDown), it is possible to use the test.support.temp_cwd context manager to create and destroy one inside the single tests that need it.

A nit: we prefer to keep the line length to <80 (that is, max 79).

In test_getting_all_methods_from_class, why is 'method_returning_true' not in the list?  I also worry that this test is a bit fragile, but I don't have a good suggestion for how to fix that, so we'll probably just have to live with it.

Thanks for the feedback :-D

Changes:

Utilize test.support.temp_cwd()
Removed imports that this made unnecessary
 Awesome trick!

Cut line length down. Don't worry about nitpicking. Anything to get this to be as good as possible.

I added an explanation to the test class explaining why test_getting_all_methods_from_class does not appear. I also agree this is fragile. I have an idea for how to fix this and will upload it as a separate patch (in this thread) because I am not sure what people will think.

Eh, sorry... I tried to do a less fragile way of generating the configuration dict. Tried a few things but it very quickly got messy and even worse started to smell like testing the method with... the method being tested.

I am open to ideas though.

New changeset 474f078ec958 by R David Murray in branch 'default':
#17464: improve pydoc test coverage.
http://hg.python.org/cpython/rev/474f078ec958

As discussed with Matt, I deleted the allmethods tests, since the current behavior is actually a bug.  pydoc itself doesn't use allmethods, but I did find that numpy at least does so.  A separate issue will be opened for that bug.