ERP5TypeFunctionalTestCase: fix decoding of HTML

If we use bytes in html parser, it will assume the encoding is latin1,
by passing str (unicode on python2), the parser does not need to deal
with encoding. This fixes a case of Jérome being decoded as Jérome

product/ERP5Type/tests/ERP5TypeFunctionalTestCase.py

@@ -360,8 +360,8 @@ class FunctionalTestRunner:
     iframe = etree.fromstring(
       browser.execute_script(
         "return document.getElementById('testSuiteFrame').contentDocument.querySelector('html').innerHTML"
-      ).encode('UTF-8'),
-        html_parser
+      ),
+      html_parser
     )
     return iframe