Merge pull request #238 from gsnedders/inputstream_update

Update prescan to scan 1024 bytes; r=nobody!

Author: gsnedders

html5lib/inputstream.py

@@ -421,7 +421,7 @@ def __init__(self, source, encoding=None, parseMeta=True, chardet=True):
         # Encoding Information
         # Number of bytes to use when looking for a meta element with
         # encoding information
-        self.numBytesMeta = 512
+        self.numBytesMeta = 1024
         # Number of bytes to use when using detecting encoding using chardet
         self.numBytesChardet = 100
         # Encoding to use if no other information can be found
@@ -430,6 +430,7 @@ def __init__(self, source, encoding=None, parseMeta=True, chardet=True):
         # Detect encoding iff no explicit "transport level" encoding is supplied
         if (self.charEncoding[0] is None):
             self.charEncoding = self.detectEncoding(parseMeta, chardet)
+            assert self.charEncoding[0] is not None
 
         # Call superclass
         self.reset()