[MRG+1] Fix part of issue #3128 - None should not be a valid type for 'url' in Response.follow (#3131)

* fix one issue of issue#3128

because @kmike posted: 'If url is '', Scrapy should follow the same page, this is an intended behavior.'

*  fix one issue of issue#3128

because @kmike posted: 'If url is '', Scrapy should follow the same page, this is an intended behavior.'

scrapy/http/response/__init__.py

@@ -120,6 +120,8 @@ class Response(object_ref):
         """
         if isinstance(url, Link):
             url = url.url
+        elif url is None:
+            raise ValueError("url can't be None")
         url = self.urljoin(url)
         return Request(url, callback,
                        method=method,

tests/test_http_response.py

@@ -155,6 +155,10 @@ class BaseResponseTest(unittest.TestCase):
         self._assert_followed_url(Link('http://example.com/foo'),
                                   'http://example.com/foo')
 
+    def test_follow_None_url(self):
+        r = self.response_class("http://example.com")
+        self.assertRaises(ValueError, r.follow, None)
+
     def test_follow_whitespace_url(self):
         self._assert_followed_url('foo ',
                                   'http://example.com/foo%20')