Add a better test for encoding prescan length

gsnedders · gsnedders · commit c5a800779b5a · 2016-05-05T00:12:52.000+01:00
diff --git a/html5lib/tests/test_encoding.py b/html5lib/tests/test_encoding.py
@@ -12,6 +12,15 @@
 from html5lib import HTMLParser, inputstream
 
 
+def test_basic_prescan_length():
+    data = "<title>Caf\u00E9</title><!--a--><meta charset='utf-8'>".encode('utf-8')
+    pad = 1024 - len(data) + 1
+    data = data.replace(b"-a-", b"-" + (b"a" * pad) + b"-")
+    assert len(data) == 1024  # Sanity
+    stream = inputstream.HTMLBinaryInputStream(data, chardet=False)
+    assert 'utf-8' == stream.charEncoding[0].name
+
+
 def runParserEncodingTest(data, encoding):
     p = HTMLParser()
     assert p.documentEncoding is None
