Issue #19543: Emit deprecation warning for known non-text encodings.

Backported issues #19619: encode() and decode() methods and constructors
of str, unicode and bytearray classes now emit deprecation warning for known
non-text encodings when Python is ran with the -3 option.

Backported issues #20404: io.TextIOWrapper (and hence io.open()) now uses the
internal codec marking system added to emit deprecation warning for known non-text
encodings at stream construction time when Python is ran with the -3 option.

1 files changed, 46 insertions, 2 deletions

diff --git a/Lib/test/test_codecs.py b/Lib/test/test_codecs.py
index de80b07..c7072a6 100644
--- a/Lib/test/test_codecs.py
+++ b/Lib/test/test_codecs.py
@@ -1395,14 +1395,14 @@ class EncodedFileTest(unittest.TestCase):
 class Str2StrTest(unittest.TestCase):
 
     def test_read(self):
-        sin = "\x80".encode("base64_codec")
+        sin = codecs.encode("\x80", "base64_codec")
         reader = codecs.getreader("base64_codec")(StringIO.StringIO(sin))
         sout = reader.read()
         self.assertEqual(sout, "\x80")
         self.assertIsInstance(sout, str)
 
     def test_readline(self):
-        sin = "\x80".encode("base64_codec")
+        sin = codecs.encode("\x80", "base64_codec")
         reader = codecs.getreader("base64_codec")(StringIO.StringIO(sin))
         sout = reader.readline()
         self.assertEqual(sout, "\x80")
@@ -1536,6 +1536,9 @@ broken_unicode_with_streams = [
 ]
 broken_incremental_coders = broken_unicode_with_streams[:]
 
+if sys.flags.py3k_warning:
+    broken_unicode_with_streams.append("rot_13")
+
 # The following encodings only support "strict" mode
 only_strict_mode = [
     "idna",
@@ -2135,6 +2138,47 @@ def test_main():
         # Missing "begin" line
         self.assertRaises(ValueError, codecs.decode, "", "uu-codec")
 
+    def test_text_to_binary_blacklists_binary_transforms(self):
+        # Check binary -> binary codecs give a good error for str input
+        bad_input = "bad input type"
+        for encoding in bytes_transform_encodings:
+            fmt = (r"{!r} is not a text encoding; "
+                   r"use codecs.encode\(\) to handle arbitrary codecs")
+            msg = fmt.format(encoding)
+            with self.assertRaisesRegex(LookupError, msg) as failure:
+                bad_input.encode(encoding)
+            self.assertIsNone(failure.exception.__cause__)
+
+    def test_text_to_binary_blacklists_text_transforms(self):
+        # Check str.encode gives a good error message for str -> str codecs
+        msg = (r"^'rot_13' is not a text encoding; "
+               r"use codecs.encode\(\) to handle arbitrary codecs")
+        with self.assertRaisesRegex(LookupError, msg):
+            "just an example message".encode("rot_13")
+
+    def test_binary_to_text_blacklists_binary_transforms(self):
+        # Check bytes.decode and bytearray.decode give a good error
+        # message for binary -> binary codecs
+        data = b"encode first to ensure we meet any format restrictions"
+        for encoding in bytes_transform_encodings:
+            encoded_data = codecs.encode(data, encoding)
+            fmt = (r"{!r} is not a text encoding; "
+                   r"use codecs.decode\(\) to handle arbitrary codecs")
+            msg = fmt.format(encoding)
+            with self.assertRaisesRegex(LookupError, msg):
+                encoded_data.decode(encoding)
+            with self.assertRaisesRegex(LookupError, msg):
+                bytearray(encoded_data).decode(encoding)
+
+    def test_binary_to_text_blacklists_text_transforms(self):
+        # Check str -> str codec gives a good error for binary input
+        for bad_input in (b"immutable", bytearray(b"mutable")):
+            msg = (r"^'rot_13' is not a text encoding; "
+                   r"use codecs.decode\(\) to handle arbitrary codecs")
+            with self.assertRaisesRegex(LookupError, msg) as failure:
+                bad_input.decode("rot_13")
+            self.assertIsNone(failure.exception.__cause__)
+
 
 if __name__ == "__main__":
     test_main()