Issue #12016: Multibyte CJK decoders now resynchronize faster

They only ignore the first byte of an invalid byte sequence.

For example, b'\xff\n'.decode('gb2312', 'replace') gives '\ufffd\n' instead of
'\ufffd'.

1 files changed, 13 insertions, 8 deletions

diff --git a/Lib/test/test_codecencodings_cn.py b/Lib/test/test_codecencodings_cn.py
index dca9f10..ee3d165 100644
--- a/Lib/test/test_codecencodings_cn.py
+++ b/Lib/test/test_codecencodings_cn.py
@@ -15,8 +15,8 @@ class Test_GB2312(test_multibytecodec_support.TestBase, unittest.TestCase):
         # invalid bytes
         (b"abc\x81\x81\xc1\xc4", "strict",  None),
         (b"abc\xc8", "strict",  None),
-        (b"abc\x81\x81\xc1\xc4", "replace", "abc\ufffd\u804a"),
-        (b"abc\x81\x81\xc1\xc4\xc8", "replace", "abc\ufffd\u804a\ufffd"),
+        (b"abc\x81\x81\xc1\xc4", "replace", "abc\ufffd\ufffd\u804a"),
+        (b"abc\x81\x81\xc1\xc4\xc8", "replace", "abc\ufffd\ufffd\u804a\ufffd"),
         (b"abc\x81\x81\xc1\xc4", "ignore",  "abc\u804a"),
         (b"\xc1\x64", "strict", None),
     )
@@ -28,8 +28,8 @@ class Test_GBK(test_multibytecodec_support.TestBase, unittest.TestCase):
         # invalid bytes
         (b"abc\x80\x80\xc1\xc4", "strict",  None),
         (b"abc\xc8", "strict",  None),
-        (b"abc\x80\x80\xc1\xc4", "replace", "abc\ufffd\u804a"),
-        (b"abc\x80\x80\xc1\xc4\xc8", "replace", "abc\ufffd\u804a\ufffd"),
+        (b"abc\x80\x80\xc1\xc4", "replace", "abc\ufffd\ufffd\u804a"),
+        (b"abc\x80\x80\xc1\xc4\xc8", "replace", "abc\ufffd\ufffd\u804a\ufffd"),
         (b"abc\x80\x80\xc1\xc4", "ignore",  "abc\u804a"),
         (b"\x83\x34\x83\x31", "strict", None),
         ("\u30fb", "strict", None),
@@ -42,11 +42,14 @@ class Test_GB18030(test_multibytecodec_support.TestBase, unittest.TestCase):
         # invalid bytes
         (b"abc\x80\x80\xc1\xc4", "strict",  None),
         (b"abc\xc8", "strict",  None),
-        (b"abc\x80\x80\xc1\xc4", "replace", "abc\ufffd\u804a"),
-        (b"abc\x80\x80\xc1\xc4\xc8", "replace", "abc\ufffd\u804a\ufffd"),
+        (b"abc\x80\x80\xc1\xc4", "replace", "abc\ufffd\ufffd\u804a"),
+        (b"abc\x80\x80\xc1\xc4\xc8", "replace", "abc\ufffd\ufffd\u804a\ufffd"),
         (b"abc\x80\x80\xc1\xc4", "ignore",  "abc\u804a"),
-        (b"abc\x84\x39\x84\x39\xc1\xc4", "replace", "abc\ufffd\u804a"),
+        (b"abc\x84\x39\x84\x39\xc1\xc4", "replace", "abc\ufffd9\ufffd9\u804a"),
         ("\u30fb", "strict", b"\x819\xa79"),
+        (b"abc\x84\x32\x80\x80def", "replace", 'abc\ufffd2\ufffd\ufffddef'),
+        (b"abc\x81\x30\x81\x30def", "strict", 'abc\x80def'),
+        (b"abc\x86\x30\x81\x30def", "replace", 'abc\ufffd0\ufffd0def'),
     )
     has_iso10646 = True
 
@@ -74,9 +77,11 @@ class Test_HZ(test_multibytecodec_support.TestBase, unittest.TestCase):
          '\u5df1\u6240\u4e0d\u6b32\uff0c\u52ff\u65bd\u65bc\u4eba\u3002'
          'Bye.\n'),
         # invalid bytes
-        (b'ab~cd', 'replace', 'ab\uFFFDd'),
+        (b'ab~cd', 'replace', 'ab\uFFFDcd'),
         (b'ab\xffcd', 'replace', 'ab\uFFFDcd'),
         (b'ab~{\x81\x81\x41\x44~}cd', 'replace', 'ab\uFFFD\uFFFD\u804Acd'),
+        (b'ab~{\x41\x44~}cd', 'replace', 'ab\u804Acd'),
+        (b"ab~{\x79\x79\x41\x44~}cd", "replace", "ab\ufffd\ufffd\u804acd"),
     )
 
 def test_main():