Defer to html5lib's Unicode converter rather than using Unicode, Dammit. The lxml treebuilder still uses UD.

1 files changed, 15 insertions, 0 deletions

diff --git a/beautifulsoup/builder/lxml_builder.py b/beautifulsoup/builder/lxml_builder.py
index 9ced9f0..a1f8c1e 100644
--- a/beautifulsoup/builder/lxml_builder.py
+++ b/beautifulsoup/builder/lxml_builder.py
@@ -1,6 +1,7 @@
 from lxml import etree
 from beautifulsoup.element import Comment, Doctype
 from beautifulsoup.builder import HTMLTreeBuilder
+from beautifulsoup.dammit import UnicodeDammit
 
 class LXMLTreeBuilder(HTMLTreeBuilder):
 
@@ -11,6 +12,20 @@ class LXMLTreeBuilder(HTMLTreeBuilder):
         self.parser = parser_class(target=self)
         self.soup = None
 
+    def prepare_markup(self, markup, user_specified_encoding=None,
+                       document_declared_encoding=None):
+        """
+        :return: A 3-tuple (markup, original encoding, encoding
+        declared within markup).
+        """
+        if isinstance(markup, unicode):
+            return markup, None, None
+
+        try_encodings = [user_specified_encoding, document_declared_encoding]
+        dammit = UnicodeDammit(markup, try_encodings, isHTML=True)
+        return dammit.markup, dammit.originalEncoding, dammit.declaredHTMLEncoding
+
+
     def feed(self, markup):
         self.parser.feed(markup)
         self.parser.close()