Path: csiph.com!fu-berlin.de!uni-berlin.de!not-for-mail
From: Peter Otten <__peter__@web.de>
Newsgroups: comp.lang.python
Subject: Re: python unit test frame work
Date: Sat, 12 Dec 2015 14:49:43 +0100
Organization: None
Lines: 98
Message-ID: <mailman.179.1449928202.12405.python-list@python.org>
References: <CACT3xuUOrbdkiEDkDMcVxCO+td4setJMb2TamW5te3FQK8oL0g@mail.gmail.com> <n4c706$g88$1@ger.gmane.org> <CACT3xuUyYSU94EfmHBHf=RvPG-x+P=TYS8cQGR5ZneGT_60LHg@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 7Bit
User-Agent: KNode/4.13.3
Precedence: list
Xref: csiph.com comp.lang.python:100334

Ganesh Pal wrote:

> On Thu, Dec 10, 2015 at 9:20 PM, Peter Otten <__peter__@web.de> wrote:
>> Ganesh Pal wrote:
>>
> 
>> I recommend that you reread the unittest documentation.
>>
>> setUpClass() should be a class method, and if it succeeds you can release
>> the ressources it required in the corresponding tearDownClass() method.
>> As written the flags and the setUp()/tearDown() seem unnecessary.
>>
> 
> Thanks to peter , Cameron and  Ben Finney , for replying to my various
> question post . I needed a hint  on the below
> 
> 
> 1. If there is a setUpClass exception or failure , I don't want the
> unittest to run ( I don't have teardown ) how do I handle this  ?
>     The traceback on the console  looks very bad  it repeats for all
> the test cases  , that means if I have 100 testcases if setup fails .
> I will get the failure for all the test cases
> 
> #c_t.py
> EEEE
> ======================================================================
> ERROR: test01: test_01_inode_test
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "c_t.py", line xx, in setUp
>     self.setupClass()
>   File "c_t.py", line xxx, in TestSetup
>     self.TestSetup()
>   File "c_t.py", line xx, in corruptSetup
>     sys.exit("/tmp is not mounted ...Exiting !!!")
> SystemExit: /tmp is not mounted ...Exiting !!!
> ======================================================================
> ERROR: test02
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "c_t.py", line 162, in  test_02_hardlink_test
>     self.inject_failures['test02']))
> KeyError: 'test02'
> 
> Ran 2 tests in 0.003s
> FAILED (errors=2)

Don't invoke sys.exit(), raise a meaningful exception instead. Then in 
setUpClass() you can catch the expected exceptions and raise a SkipTest. 
Example:

$ cat mayfail.py
import os
import unittest

def setup_that_may_fail():
    if "FAIL" in os.environ:
        1/0

class MyTests(unittest.TestCase):
    @classmethod
    def setUpClass(cls):
        try:
            setup_that_may_fail() # placeholder for your actual setup
        except Exception as err:
            raise unittest.SkipTest(
                "class setup failed") # todo: better message

    def test_one(self):
        pass

    def test_two(self):
        pass

if __name__ == "__main__":
    unittest.main()

When the setup succeeds:

$ python mayfail.py
..
----------------------------------------------------------------------
Ran 2 tests in 0.000s

OK

Now let's pretend a failure by setting the FAIL environment variable.
(In your actual code you won't do that as you get a "real" failure)
$ FAIL=1 python mayfail.py
s
----------------------------------------------------------------------
Ran 0 tests in 0.000s

OK (skipped=1)

There's one oddity with this approach -- only one skipped test is reported 
even though there are two tests in the class, neither of which is run.