fixup! Add a more general fix for #127 (CPy #20007) based on #136.

gsnedders · gsnedders · commit 3b3bc21bc5eb · 2015-12-12T03:12:22.000Z
diff --git a/html5lib/inputstream.py b/html5lib/inputstream.py
@@ -1,7 +1,7 @@
 from __future__ import absolute_import, division, unicode_literals
 
-from six import text_type, get_method_self
-from six.moves import http_client
+from six import text_type
+from six.moves import http_client, urllib
 
 import codecs
 import re
@@ -131,13 +131,15 @@ def _readFromBuffer(self, bytes):
 
 
 def HTMLInputStream(source, encoding=None, parseMeta=True, chardet=True):
-    if hasattr(source, "read"):
-        if isinstance(get_method_self(source.read), http_client.HTTPResponse):
-            # Work around Python bug #20007: read(0) closes the connection.
-            # http://bugs.python.org/issue20007
-            isUnicode = False
-        else:
-            isUnicode = isinstance(source.read(0), text_type)
+    # Work around Python bug #20007: read(0) closes the connection.
+    # http://bugs.python.org/issue20007
+    if (isinstance(source, http_client.HTTPResponse) or
+        # Also check for addinfourl wrapping HTTPResponse
+        (isinstance(source, urllib.response.addbase) and
+         isinstance(source.fp, http_client.HTTPResponse))):
+        isUnicode = False
+    elif hasattr(source, "read"):
+        isUnicode = isinstance(source.read(0), text_type)
     else:
         isUnicode = isinstance(source, text_type)
 
diff --git a/html5lib/tests/test_stream.py b/html5lib/tests/test_stream.py
@@ -4,7 +4,9 @@
 import unittest
 import codecs
 from io import BytesIO
+import socket
 
+import six
 from six.moves import http_client, urllib
 
 from html5lib.inputstream import (BufferedStream, HTMLInputStream,
@@ -175,14 +177,17 @@ def test_python_issue_20007_b(self):
         Make sure we have a work-around for Python bug #20007
         http://bugs.python.org/issue20007
         """
+        if six.PY2:
+            return
+
         class FakeSocket(object):
             def makefile(self, _mode, _bufsize=None):
                 return BytesIO(b"HTTP/1.1 200 Ok\r\n\r\nText")
 
         source = http_client.HTTPResponse(FakeSocket())
         source.begin()
         wrapped = urllib.response.addinfourl(source, source.msg, "http://example.com")
-        stream = HTMLInputStream(source)
+        stream = HTMLInputStream(wrapped)
         self.assertEqual(stream.charsUntil(" "), "Text")
 
 
