Fix tests on URL path encoding for links from latin1 document

UTF-8 is to be used for path component.
Page encoding only affects query part of URLs.

Author: redapple

tests/test_linkextractors.py

@@ -80,8 +80,8 @@ def test_extraction_encoding(self):
         ])
 
         self.assertEqual(lx.extract_links(response_latin1), [
-            Link(url='http://example.com/sample_%%3Cspan%20class="x x-first x-last">F1.html', text=''),
-            Link(url='http://example.com/sample_%%3Cspan%20class="x x-first x-last">E1.html', text='sample \xe1 text'.decode('latin1')),
+            Link(url='http://example.com/sample_%%3Cspan%20class="x x-first x-last">C3%B1.html', text=''),
+            Link(url='http://example.com/sample_%%3Cspan%20class="x x-first x-last">C3%A1.html', text='sample \xe1 text'.decode('latin1')),
         ])
 
     def test_matches(self):