Made conversion of markup to Unicode the responsibility of the builder, not the BeautifulSoup class itself. lxml uses Unicode, Dammit; html5lib uses its internal algorithms.

1 files changed, 4 insertions, 0 deletions

diff --git a/beautifulsoup/builder/__init__.py b/beautifulsoup/builder/__init__.py
index cf5e6c6..5bf5929 100644
--- a/beautifulsoup/builder/__init__.py
+++ b/beautifulsoup/builder/__init__.py
@@ -25,6 +25,10 @@ class TreeBuilder(Entities):
     def feed(self, markup):
         raise NotImplementedError()
 
+    def prepare_markup(self, markup, user_specified_encoding=None,
+                       document_declared_encoding=None):
+        return markup, None, None
+
     def test_fragment_to_document(self, fragment):
         """Wrap an HTML fragment to make it look like a document.