[#96] Handle encoding errors on harvest object endpoint

When parsing the harvest object content to see if it is an XML file,
etree.fromstring would fail id there are incorrect unicode errors.

ckanext/harvest/controllers/view.py

@@ -103,8 +103,10 @@ class ViewController(BaseController):
                     content = obj['extras']['original_document']
                 else:
                     abort(404,_('No content found'))
-
+                try:
                     etree.fromstring(re.sub('<\?xml(.*)\?>','',content))
+                except UnicodeEncodeError:
+                    etree.fromstring(re.sub('<\?xml(.*)\?>','',content.encode('utf-8')))
                 response.content_type = 'application/xml; charset=utf-8'
                 if not '<?xml' in content.split('\n')[0]:
                     content = u'<?xml version="1.0" encoding="UTF-8"?>\n' + content