I only got one error on Windows 10 – thanks for the fixes, Pascal and Gary!


The new error is:
  FileSystemUtilsTestCase.testGetFreeSpace_String:89
expected:<1.02861164E8> but was:<1.0286066E8>


I have:
               5 Dir(s)  104,991,649,792 bytes free

Test calls:

            final long bytes = FileSystemUtils.freeSpace("");
            final long kb = FileSystemUtils.freeSpaceKb("");
            assertEquals((double) bytes / 1024, kb, 256d);


Presumably something else on my machine downloaded 504 kB between the
two freeSpace calls – which is not much these days – perhaps one email
:-)

I changed the test to use instead a 1% delta, which should generally
be a considerable amount of disk space (1 GB in my case), which is
still small enough to detect the ~2.4% difference between a kilobyte
vs kibibyte (the legacy freeSpaceKb is misnamed, it actually returns
kibibyte instead of pre-1998 “kilobyte”)

Also added javadoc that it's actually kibibytes -- perhaps a bit too
late :-) -- IO-506 deprecated the remaining methods of FileSystemUtils
in favour of java.nio.file.FileStore -- I think then we can deprecate
the whole FileSystemUtils  class as well.


On 27 September 2017 at 20:24, Pascal Schumacher
<pascalschumac...@gmx.net> wrote:
> Great!
>
> That should reduce the number of random test failures.
>
> I was just taking a look at this and the usage patterns of the folder looked
> horrible. :/
>
> Cheers,
> Pascal
>
>
> Am 27.09.2017 um 21:19 schrieb Gary Gregory:
>>
>> I updated the tests to use JUnit's TemporaryFolder rule instead of the old
>> custom and _shared_ folder. No more random errors.
>>
>> Gary
>>
>> On Wed, Sep 27, 2017 at 1:05 PM, Pascal Schumacher
>> <pascalschumac...@gmx.net
>>>
>>> wrote:
>>> Am 27.09.2017 um 19:58 schrieb Gary Gregory:
>>>
>>>> I wonder if we need to force Maven to disable any concurrency when
>>>> running
>>>> tests?
>>>>
>>> That should not be necessary.
>>>
>>> Maven/Maven-Surefire-Plugin does not run test in parallel by default. The
>>> surefire configuration of commons-io does not specify the parallel
>>> parameter and the forkcount parameter is set to 1 so the tests should be
>>> executed sequentially.
>>>
>>> Cheers,
>>> Pascal
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
>>> For additional commands, e-mail: dev-h...@commons.apache.org
>>>
>>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
> For additional commands, e-mail: dev-h...@commons.apache.org
>



-- 
Stian Soiland-Reyes
http://orcid.org/0000-0001-9842-9718

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org

Reply via email to