Update library: html5lib

libs/html5lib/__init__.py

@@ -20,4 +20,4 @@ from .serializer import serialize
 
 __all__ = ["HTMLParser", "parse", "parseFragment", "getTreeBuilder",
            "getTreeWalker", "serialize"]
-__version__ = "0.99"
+__version__ = "0.999"

libs/html5lib/inputstream.py

@@ -1,5 +1,6 @@
 from __future__ import absolute_import, division, unicode_literals
 from six import text_type
+from six.moves import http_client
 
 import codecs
 import re
@@ -118,7 +119,11 @@ class BufferedStream(object):
 
 
 def HTMLInputStream(source, encoding=None, parseMeta=True, chardet=True):
-    if hasattr(source, "read"):
+    if isinstance(source, http_client.HTTPResponse):
+        # Work around Python bug #20007: read(0) closes the connection.
+        # http://bugs.python.org/issue20007
+        isUnicode = False
+    elif hasattr(source, "read"):
         isUnicode = isinstance(source.read(0), text_type)
     else:
         isUnicode = isinstance(source, text_type)

libs/html5lib/treewalkers/lxmletree.py

@@ -87,10 +87,6 @@ class FragmentWrapper(object):
             self.tail = ensure_str(self.obj.tail)
         else:
             self.tail = None
-        self.isstring = isinstance(obj, str) or isinstance(obj, bytes)
-        # Support for bytes here is Py2
-        if self.isstring:
-            self.obj = ensure_str(self.obj)
 
     def __getattr__(self, name):
         return getattr(self.obj, name)
@@ -143,7 +139,7 @@ class TreeWalker(_base.NonRecursiveTreeWalker):
         elif isinstance(node, Doctype):
             return _base.DOCTYPE, node.name, node.public_id, node.system_id
 
-        elif isinstance(node, FragmentWrapper) and node.isstring:
+        elif isinstance(node, FragmentWrapper) and not hasattr(node, "tag"):
             return _base.TEXT, node.obj
 
         elif node.tag == etree.Comment: