Support trailing dots in DNS names. Fixes #782510. Will backport to 2.3.
diff --git a/Lib/encodings/idna.py b/Lib/encodings/idna.py
index 62aa450..37462db 100644
--- a/Lib/encodings/idna.py
+++ b/Lib/encodings/idna.py
@@ -150,10 +150,16 @@
raise UnicodeError, "unsupported error handling "+errors
result = []
- for label in dots.split(input):
+ labels = dots.split(input)
+ if labels and len(labels[-1])==0:
+ trailing_dot = '.'
+ del labels[-1]
+ else:
+ trailing_dot = ''
+ for label in labels:
result.append(ToASCII(label))
# Join with U+002E
- return ".".join(result), len(input)
+ return ".".join(result)+trailing_dot, len(input)
def decode(self,input,errors='strict'):
@@ -168,11 +174,17 @@
unicode(input, "ascii")
labels = input.split(".")
+ if labels and len(labels[-1]) == 0:
+ trailing_dot = u'.'
+ del labels[-1]
+ else:
+ trailing_dot = u''
+
result = []
for label in labels:
result.append(ToUnicode(label))
- return u".".join(result), len(input)
+ return u".".join(result)+trailing_dot, len(input)
class StreamWriter(Codec,codecs.StreamWriter):
pass
diff --git a/Lib/test/test_unicode.py b/Lib/test/test_unicode.py
index 8e1f0b1..6e40b9f 100644
--- a/Lib/test/test_unicode.py
+++ b/Lib/test/test_unicode.py
@@ -524,6 +524,10 @@
# * strict decoding testing for all of the
# UTF8_ERROR cases in PyUnicode_DecodeUTF8
+ def test_codecs_idna(self):
+ # Test whether trailing dot is preserved
+ self.assertEqual(u"www.python.org.".encode("idna"), "www.python.org.")
+
def test_codecs_errors(self):
# Error handling (encoding)
self.assertRaises(UnicodeError, u'Andr\202 x'.encode, 'ascii')