bpo-30566: Fix IndexError when using punycode codec (GH-18632)
Trying to decode an invalid string with the punycode codec
shoud raise UnicodeError.
diff --git a/Lib/encodings/punycode.py b/Lib/encodings/punycode.py
index 66c5101..1c57264 100644
--- a/Lib/encodings/punycode.py
+++ b/Lib/encodings/punycode.py
@@ -143,7 +143,7 @@
digit = char - 22 # 0x30-26
elif errors == "strict":
raise UnicodeError("Invalid extended code point '%s'"
- % extended[extpos])
+ % extended[extpos-1])
else:
return extpos, None
t = T(j, bias)
diff --git a/Lib/test/test_codecs.py b/Lib/test/test_codecs.py
index 3aec34c..8d9cb90 100644
--- a/Lib/test/test_codecs.py
+++ b/Lib/test/test_codecs.py
@@ -1343,6 +1343,18 @@
puny = puny.decode("ascii").encode("ascii")
self.assertEqual(uni, puny.decode("punycode"))
+ def test_decode_invalid(self):
+ testcases = [
+ (b"xn--w&", "strict", UnicodeError()),
+ (b"xn--w&", "ignore", "xn-"),
+ ]
+ for puny, errors, expected in testcases:
+ with self.subTest(puny=puny, errors=errors):
+ if isinstance(expected, Exception):
+ self.assertRaises(UnicodeError, puny.decode, "punycode", errors)
+ else:
+ self.assertEqual(puny.decode("punycode", errors), expected)
+
# From http://www.gnu.org/software/libidn/draft-josefsson-idn-test-vectors.html
nameprep_tests = [