unittest.TestCase and functools.partial don't seem to mix

[Gabriel Genellina]

En Tue, 06 Oct 2009 18:01:34 -0300, Joel Smith <js-pythonlist at jk1.net>
escribió:
> Gabriel Genellina wrote:

>> Note that you don't *have* to use partial in this case, as you're  
>> building the suite yourself. Just create the TestCase instances  
>> manually:
>>
>>     suite = unittest.TestSuite([
>>       TestGenericWindow('testit', 'brown'),
>>       TestGenericWindow('testit', 'blue'),
>>       TestGenericWindow('testit', 'green')
>>       ])
>>     unittest.TextTestRunner().run(suite)
>
> Perfect!  This is exactly what I needed.  For some reason, I didn't
> understand that I could construct my TestCase objects directly... I
> thought that I had to allow the unittest framework to construct them for
> me.  This is straightforward, and does exactly what I need.

You're not alone :) The unittest design is rather convoluted, probably
because it aims to mimic the jUnit design as closely as possible, with all
its javaism and unpythonicity. I don't think one could understand how to
use the framework just by reading the package documentation; more
background info is needed. The "Dive into Python" book contains a chapter
on this topic.

-- 
Gabriel Genellina