Issue #4574: reading an UTF16-encoded text file crashes if \r on 64-char boundary.

Lib/io.py

@@ -1282,25 +1282,23 @@ class IncrementalNewlineDecoder(codecs.IncrementalDecoder):
     """
     def __init__(self, decoder, translate, errors='strict'):
         codecs.IncrementalDecoder.__init__(self, errors=errors)
-        self.buffer = b''
         self.translate = translate
         self.decoder = decoder
         self.seennl = 0
+        self.pendingcr = False
 
     def decode(self, input, final=False):
         # decode input (with the eventual \r from a previous pass)
-        if self.buffer:
-            input = self.buffer + input
-
         output = self.decoder.decode(input, final=final)
+        if self.pendingcr and (output or final):
+            output = "\r" + output
+            self.pendingcr = False
 
         # retain last \r even when not translating data:
         # then readline() is sure to get \r\n in one pass
         if output.endswith("\r") and not final:
             output = output[:-1]
-            self.buffer = b'\r'
-        else:
-            self.buffer = b''
+            self.pendingcr = True
 
         # Record which newlines are read
         crlf = output.count('\r\n')
@@ -1319,20 +1317,19 @@ class IncrementalNewlineDecoder(codecs.IncrementalDecoder):
 
     def getstate(self):
         buf, flag = self.decoder.getstate()
-        return buf + self.buffer, flag
+        flag <<= 1
+        if self.pendingcr:
+            flag |= 1
+        return buf, flag
 
     def setstate(self, state):
         buf, flag = state
-        if buf.endswith(b'\r'):
-            self.buffer = b'\r'
-            buf = buf[:-1]
-        else:
-            self.buffer = b''
-        self.decoder.setstate((buf, flag))
+        self.pendingcr = bool(flag & 1)
+        self.decoder.setstate((buf, flag >> 1))
 
     def reset(self):
         self.seennl = 0
-        self.buffer = b''
+        self.pendingcr = False
         self.decoder.reset()
 
     _LF = 1

Lib/test/test_io.py

@@ -679,8 +679,9 @@ class StatefulIncrementalDecoder(codecs.IncrementalDecoder):
     @classmethod
     def lookupTestDecoder(cls, name):
         if cls.codecEnabled and name == 'test_decoder':
+            latin1 = codecs.lookup('latin-1')
             return codecs.CodecInfo(
-                name='test_decoder', encode=None, decode=None,
+                name='test_decoder', encode=latin1.encode, decode=None,
                 incrementalencoder=None,
                 streamreader=None, streamwriter=None,
                 incrementaldecoder=cls)
@@ -840,8 +841,11 @@ class TextIOWrapperTest(unittest.TestCase):
             [ '\r\n', [ "unix\nwindows\r\n", "os9\rlast\nnonl" ] ],
             [ '\r', [ "unix\nwindows\r", "\nos9\r", "last\nnonl" ] ],
         ]
-
-        encodings = ('utf-8', 'latin-1')
+        encodings = (
+            'utf-8', 'latin-1',
+            'utf-16', 'utf-16-le', 'utf-16-be',
+            'utf-32', 'utf-32-le', 'utf-32-be',
+        )
 
         # Try a range of buffer sizes to test the case where \r is the last
         # character in TextIOWrapper._pending_line.
@@ -1195,56 +1199,84 @@ class TextIOWrapperTest(unittest.TestCase):
 
         self.assertEqual(buffer.seekable(), txt.seekable())
 
-    def test_newline_decoder(self):
-        import codecs
-        decoder = codecs.getincrementaldecoder("utf-8")()
-        decoder = io.IncrementalNewlineDecoder(decoder, translate=True)
+    def check_newline_decoder_utf8(self, decoder):
+        # UTF-8 specific tests for a newline decoder
+        def _check_decode(b, s, **kwargs):
+            # We exercise getstate() / setstate() as well as decode()
+            state = decoder.getstate()
+            self.assertEquals(decoder.decode(b, **kwargs), s)
+            decoder.setstate(state)
+            self.assertEquals(decoder.decode(b, **kwargs), s)
 
-        self.assertEquals(decoder.decode(b'\xe8\xa2\x88'), "\u8888")
+        _check_decode(b'\xe8\xa2\x88', "\u8888")
 
-        self.assertEquals(decoder.decode(b'\xe8'), "")
-        self.assertEquals(decoder.decode(b'\xa2'), "")
-        self.assertEquals(decoder.decode(b'\x88'), "\u8888")
+        _check_decode(b'\xe8', "")
+        _check_decode(b'\xa2', "")
+        _check_decode(b'\x88', "\u8888")
 
-        self.assertEquals(decoder.decode(b'\xe8'), "")
-        self.assertRaises(UnicodeDecodeError, decoder.decode, b'', final=True)
+        _check_decode(b'\xe8', "")
+        _check_decode(b'\xa2', "")
+        _check_decode(b'\x88', "\u8888")
 
-        decoder.setstate((b'', 0))
-        self.assertEquals(decoder.decode(b'\n'), "\n")
-        self.assertEquals(decoder.decode(b'\r'), "")
-        self.assertEquals(decoder.decode(b'', final=True), "\n")
-        self.assertEquals(decoder.decode(b'\r', final=True), "\n")
-
-        self.assertEquals(decoder.decode(b'\r'), "")
-        self.assertEquals(decoder.decode(b'a'), "\na")
-
-        self.assertEquals(decoder.decode(b'\r\r\n'), "\n\n")
-        self.assertEquals(decoder.decode(b'\r'), "")
-        self.assertEquals(decoder.decode(b'\r'), "\n")
-        self.assertEquals(decoder.decode(b'\na'), "\na")
-
-        self.assertEquals(decoder.decode(b'\xe8\xa2\x88\r\n'), "\u8888\n")
-        self.assertEquals(decoder.decode(b'\xe8\xa2\x88'), "\u8888")
-        self.assertEquals(decoder.decode(b'\n'), "\n")
-        self.assertEquals(decoder.decode(b'\xe8\xa2\x88\r'), "\u8888")
-        self.assertEquals(decoder.decode(b'\n'), "\n")
+        _check_decode(b'\xe8', "")
+        self.assertRaises(UnicodeDecodeError, decoder.decode, b'', final=True)
 
-        decoder = codecs.getincrementaldecoder("utf-8")()
-        decoder = io.IncrementalNewlineDecoder(decoder, translate=True)
+        decoder.reset()
+        _check_decode(b'\n', "\n")
+        _check_decode(b'\r', "")
+        _check_decode(b'', "\n", final=True)
+        _check_decode(b'\r', "\n", final=True)
+
+        _check_decode(b'\r', "")
+        _check_decode(b'a', "\na")
+
+        _check_decode(b'\r\r\n', "\n\n")
+        _check_decode(b'\r', "")
+        _check_decode(b'\r', "\n")
+        _check_decode(b'\na', "\na")
+
+        _check_decode(b'\xe8\xa2\x88\r\n', "\u8888\n")
+        _check_decode(b'\xe8\xa2\x88', "\u8888")
+        _check_decode(b'\n', "\n")
+        _check_decode(b'\xe8\xa2\x88\r', "\u8888")
+        _check_decode(b'\n', "\n")
+
+    def check_newline_decoder(self, decoder, encoding):
+        result = []
+        encoder = codecs.getincrementalencoder(encoding)()
+        def _decode_bytewise(s):
+            for b in encoder.encode(s):
+                result.append(decoder.decode(bytes([b])))
         self.assertEquals(decoder.newlines, None)
-        decoder.decode(b"abc\n\r")
+        _decode_bytewise("abc\n\r")
         self.assertEquals(decoder.newlines, '\n')
-        decoder.decode(b"\nabc")
+        _decode_bytewise("\nabc")
         self.assertEquals(decoder.newlines, ('\n', '\r\n'))
-        decoder.decode(b"abc\r")
+        _decode_bytewise("abc\r")
         self.assertEquals(decoder.newlines, ('\n', '\r\n'))
-        decoder.decode(b"abc")
+        _decode_bytewise("abc")
         self.assertEquals(decoder.newlines, ('\r', '\n', '\r\n'))
-        decoder.decode(b"abc\r")
+        _decode_bytewise("abc\r")
+        self.assertEquals("".join(result), "abc\n\nabcabc\nabcabc")
         decoder.reset()
-        self.assertEquals(decoder.decode(b"abc"), "abc")
+        self.assertEquals(decoder.decode("abc".encode(encoding)), "abc")
         self.assertEquals(decoder.newlines, None)
 
+    def test_newline_decoder(self):
+        encodings = (
+            'utf-8', 'latin-1',
+            'utf-16', 'utf-16-le', 'utf-16-be',
+            'utf-32', 'utf-32-le', 'utf-32-be',
+        )
+        for enc in encodings:
+            decoder = codecs.getincrementaldecoder(enc)()
+            decoder = io.IncrementalNewlineDecoder(decoder, translate=True)
+            self.check_newline_decoder(decoder, enc)
+        decoder = codecs.getincrementaldecoder("utf-8")()
+        decoder = io.IncrementalNewlineDecoder(decoder, translate=True)
+        self.check_newline_decoder_utf8(decoder)
+
+
 # XXX Tests for open()
 
 class MiscIOTest(unittest.TestCase):

Misc/NEWS

@@ -45,6 +45,9 @@ Core and Builtins
 Library
 -------
 
+- Issue #4574: reading an UTF16-encoded text file crashes if \r on 64-char
+  boundary.
+
 - Issue #4223: inspect.getsource() will now correctly display source code
   for packages loaded via zipimport (or any other conformant PEP 302
   loader). Original patch by Alexander Belopolsky.