Stopped HTMLParser from raising an exception in very rare cases of

  bad markup. [bug=1708831]

2 files changed, 18 insertions, 3 deletions

diff --git a/bs4/builder/_htmlparser.py b/bs4/builder/_htmlparser.py
index 67890b3..71604c5 100644
--- a/bs4/builder/_htmlparser.py
+++ b/bs4/builder/_htmlparser.py
@@ -64,7 +64,18 @@ class BeautifulSoupHTMLParser(HTMLParser):
         # order. It's a list of closing tags we've already handled and
         # will ignore, assuming they ever show up.
         self.already_closed_empty_element = []
-    
+
+    def error(self, msg):
+        """In Python 3, HTMLParser subclasses must implement error(), although this
+        requirement doesn't appear to be documented.
+
+        In Python 2, HTMLParser implements error() as raising an exception.
+
+        In any event, this method is called only on very strange markup and our best strategy
+        is to pretend it didn't happen and keep going.
+        """
+        warnings.warn(msg)
+        
     def handle_startendtag(self, name, attrs):
         # This is only called when the markup looks like
         # <tag/>.
diff --git a/bs4/builder/_lxml.py b/bs4/builder/_lxml.py
index d2ca287..3439271 100644
--- a/bs4/builder/_lxml.py
+++ b/bs4/builder/_lxml.py
@@ -5,9 +5,13 @@ __all__ = [
     'LXMLTreeBuilder',
     ]
 
+try:
+    from collections.abc import Callable # Python 3.6
+except ImportError , e:
+    from collections import Callable
+
 from io import BytesIO
 from StringIO import StringIO
-import collections
 from lxml import etree
 from bs4.element import (
     Comment,
@@ -58,7 +62,7 @@ class LXMLTreeBuilderForXML(TreeBuilder):
         # Use the default parser.
         parser = self.default_parser(encoding)
 
-        if isinstance(parser, collections.Callable):
+        if isinstance(parser, Callable):
             # Instantiate the parser with default arguments
             parser = parser(target=self, strip_cdata=False, encoding=encoding)
         return parser