diff --git a/Lib/test/test_xml_etree.py b/Lib/test/test_xml_etree.py
index baa4e1f5342..661ad8b9d4d 100644
--- a/Lib/test/test_xml_etree.py
+++ b/Lib/test/test_xml_etree.py
@@ -34,6 +34,7 @@ try:
except UnicodeEncodeError:
raise unittest.SkipTest("filename is not encodable to utf8")
SIMPLE_NS_XMLFILE = findfile("simple-ns.xml", subdir="xmltestdata")
+UTF8_BUG_XMLFILE = findfile("expat224_utf8_bug.xml", subdir="xmltestdata")
SAMPLE_XML = """\
@@ -1739,6 +1740,37 @@ class BugsTest(unittest.TestCase):
self.assertIsInstance(e[0].tag, str)
self.assertEqual(e[0].tag, 'changed')
+ def check_expat224_utf8_bug(self, text):
+ xml = b'' % text
+ root = ET.XML(xml)
+ self.assertEqual(root.get('b'), text.decode('utf-8'))
+
+ def test_expat224_utf8_bug(self):
+ # bpo-31170: Expat 2.2.3 had a bug in its UTF-8 decoder.
+ # Check that Expat 2.2.4 fixed the bug.
+ #
+ # Test buffer bounds at odd and even positions.
+
+ text = b'\xc3\xa0' * 1024
+ self.check_expat224_utf8_bug(text)
+
+ text = b'x' + b'\xc3\xa0' * 1024
+ self.check_expat224_utf8_bug(text)
+
+ def test_expat224_utf8_bug_file(self):
+ with open(UTF8_BUG_XMLFILE, 'rb') as fp:
+ raw = fp.read()
+ root = ET.fromstring(raw)
+ xmlattr = root.get('b')
+
+ # "Parse" manually the XML file to extract the value of the 'b'
+ # attribute of the XML element
+ text = raw.decode('utf-8').strip()
+ text = text.replace('\r\n', ' ')
+ text = text[6:-4]
+ self.assertEqual(root.get('b'), text)
+
+
# --------------------------------------------------------------------
diff --git a/Lib/test/xmltestdata/expat224_utf8_bug.xml b/Lib/test/xmltestdata/expat224_utf8_bug.xml
new file mode 100644
index 00000000000..d66a8e6b50f
--- /dev/null
+++ b/Lib/test/xmltestdata/expat224_utf8_bug.xml
@@ -0,0 +1,2 @@
+