Flatten test setup.

We can tell `setuptools` to use `pytest_runner` when setup.py is
invoked.

setup.py

 import os
-import sys
 from setuptools import setup
-from setuptools.command.test import test as TestCommand
-
-tests_path = os.path.join(os.path.dirname(__file__), 'tests')
-
-
-class PyTest(TestCommand):
-    user_options = [('pytest-args=', 'a', "Arguments to pass to py.test"), ]
-
-    def initialize_options(self):
-        TestCommand.initialize_options(self)
-        self.pytest_args = []
-
-    def finalize_options(self):
-        TestCommand.finalize_options(self)
-        self.test_args = []
-        self.test_suite = True
-
-    def run_tests(self):
-        # import here, cause outside the eggs aren't loaded
-        import pytest
-        errno = pytest.main(self.pytest_args)
-        sys.exit(errno)                                # pragma: no cover
 
 
 def read(fname):
     return open(os.path.join(os.path.dirname(__file__), fname)).read()
 
+setup_requires = [
+    'pytest_runner',
+    ]
+
 install_requires = [
     'setuptools',
     ]
@@ -80,13 +61,13 @@ setup(
     ],
     include_package_data=True,
     zip_safe=False,
+    setup_requires=setup_requires,
     install_requires=install_requires,
     tests_require=tests_require,
     extras_require=dict(
         tests=tests_require,
         docs=docs_require,
         ),
-    cmdclass={'test': PyTest},
     entry_points={
         'console_scripts': [
             'diceware = diceware:main',