Darwin default_device breaks module tests

[JonnyJD]

On Darwin the default_device is no (platform dependent) constant, but we actively search for a suitable device on the system.

This will break when there is no drive (or maybe also when there is no disc?).
Tested on a virtualbox machine. That might also be some change on my end, since I updated virtualbox/kernel and many other things in between. I also can't get isrcsubmit to work and that did work before..

However, we should probably remove the default_device test on darwin.

[JonnyJD]

The error message in python setup.py test -vv is:

FAIL: test_default_device (test_discid.TestModule)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/Users/jonnyjd/git/python-discid/test_discid.py", line 32, in test_default_device
    self.assertTrue(device, "No default device given")
AssertionError: No default device given

[JonnyJD]

Maybe we should even remove that test altogether, since it doesn't test much.
It only tests that a default device is available and that obviously fails on any platform that doesn't implement this.
It is part of the "platform dependent" API.

So we should test this only on the platforms where it is supported, which is something python-discid shouldn't be bothered with.

-> Remove that test altogether

We might want to include that in a "platform test" suite. That suite should be excluded on python setup.py test, though.

[JonnyJD]

Platform dependent test implementation is handled in #18

We might just comment that part out in some way for now or start imlementing #18 a bit already.

[JonnyJD]

http://tickets.musicbrainz.org/browse/LIB-28 is relevant here.
That is basically about adding support for numbers as drives on Mac and using "1" as the (static) default.

However, there will always be the old libdiscid versions (starting with libdiscid 0.2.1) that have the "dynamic" behavior.