From: Nadeem Vawda <nadeem.vawda@gmail.com>
Date: Wed, 4 Dec 2013 22:29:51 +0000 (+0100)
Subject: Closes #19839: Fix regression in bz2 module's handling of non-bzip2 data at EOF.
X-Git-Tag: v3.4.0b2~340
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=ae02d07bc2939f2f9e7a82543be3bb832ad8b29d;p=python

Closes #19839: Fix regression in bz2 module's handling of non-bzip2 data at EOF.

Also fix an analogous bug (not a regression) in the lzma module.
---

ae02d07bc2939f2f9e7a82543be3bb832ad8b29d
diff --cc Lib/bz2.py
index 6bc611e71f,1de8f3ca7d..6f47bfaf6c
--- a/Lib/bz2.py
+++ b/Lib/bz2.py
@@@ -204,11 -196,18 +204,18 @@@ class BZ2File(io.BufferedIOBase)
                      raise EOFError("Compressed file ended before the "
                                     "end-of-stream marker was reached")
  
 -            # Continue to next stream.
              if self._decompressor.eof:
 +                # Continue to next stream.
                  self._decompressor = BZ2Decompressor()
- 
-             self._buffer = self._decompressor.decompress(rawblock)
+                 try:
+                     self._buffer = self._decompressor.decompress(rawblock)
+                 except OSError:
+                     # Trailing data isn't a valid bzip2 stream. We're done here.
+                     self._mode = _MODE_READ_EOF
+                     self._size = self._pos
+                     return False
+             else:
+                 self._buffer = self._decompressor.decompress(rawblock)
              self._buffer_offset = 0
          return True
  
diff --cc Lib/lzma.py
index 011ad271fb,1a89887b39..f1d3958179
--- a/Lib/lzma.py
+++ b/Lib/lzma.py
@@@ -225,13 -224,18 +225,20 @@@ class LZMAFile(io.BufferedIOBase)
                      raise EOFError("Compressed file ended before the "
                                     "end-of-stream marker was reached")
  
-             # Continue to next stream.
              if self._decompressor.eof:
+                 # Continue to next stream.
                  self._decompressor = LZMADecompressor(**self._init_args)
- 
-             self._buffer = self._decompressor.decompress(rawblock)
+                 try:
+                     self._buffer = self._decompressor.decompress(rawblock)
+                 except LZMAError:
+                     # Trailing data isn't a valid compressed stream; ignore it.
+                     self._mode = _MODE_READ_EOF
+                     self._size = self._pos
+                     return False
+             else:
+                 self._buffer = self._decompressor.decompress(rawblock)
 +            self._buffer_offset = 0
 +        return True
  
      # Read data until EOF.
      # If return_data is false, consume the data without returning it.
diff --cc Lib/test/test_bz2.py
index 8d93e2d88f,98557f2510..cd58f319b7
--- a/Lib/test/test_bz2.py
+++ b/Lib/test/test_bz2.py
@@@ -50,9 -49,10 +50,10 @@@ class BaseTest(unittest.TestCase)
      TEXT = b''.join(TEXT_LINES)
      DATA = b'BZh91AY&SY.\xc8N\x18\x00\x01>_\x80\x00\x10@\x02\xff\xf0\x01\x07n\x00?\xe7\xff\xe00\x01\x99\xaa\x00\xc0\x03F\x86\x8c#&\x83F\x9a\x03\x06\xa6\xd0\xa6\x93M\x0fQ\xa7\xa8\x06\x804hh\x12$\x11\xa4i4\xf14S\xd2<Q\xb5\x0fH\xd3\xd4\xdd\xd5\x87\xbb\xf8\x94\r\x8f\xafI\x12\xe1\xc9\xf8/E\x00pu\x89\x12]\xc9\xbbDL\nQ\x0e\t1\x12\xdf\xa0\xc0\x97\xac2O9\x89\x13\x94\x0e\x1c7\x0ed\x95I\x0c\xaaJ\xa4\x18L\x10\x05#\x9c\xaf\xba\xbc/\x97\x8a#C\xc8\xe1\x8cW\xf9\xe2\xd0\xd6M\xa7\x8bXa<e\x84t\xcbL\xb3\xa7\xd9\xcd\xd1\xcb\x84.\xaf\xb3\xab\xab\xad`n}\xa0lh\tE,\x8eZ\x15\x17VH>\x88\xe5\xcd9gd6\x0b\n\xe9\x9b\xd5\x8a\x99\xf7\x08.K\x8ev\xfb\xf7xw\xbb\xdf\xa1\x92\xf1\xdd|/";\xa2\xba\x9f\xd5\xb1#A\xb6\xf6\xb3o\xc9\xc5y\\\xebO\xe7\x85\x9a\xbc\xb6f8\x952\xd5\xd7"%\x89>V,\xf7\xa6z\xe2\x9f\xa3\xdf\x11\x11"\xd6E)I\xa9\x13^\xca\xf3r\xd0\x03U\x922\xf26\xec\xb6\xed\x8b\xc3U\x13\x9d\xc5\x170\xa4\xfa^\x92\xacDF\x8a\x97\xd6\x19\xfe\xdd\xb8\xbd\x1a\x9a\x19\xa3\x80ankR\x8b\xe5\xd83]\xa9\xc6\x08\x82f\xf6\xb9"6l$\xb8j@\xc0\x8a\xb0l1..\xbak\x83ls\x15\xbc\xf4\xc1\x13\xbe\xf8E\xb8\x9d\r\xa8\x9dk\x84\xd3n\xfa\xacQ\x07\xb1%y\xaav\xb4\x08\xe0z\x1b\x16\xf5\x04\xe9\xcc\xb9\x08z\x1en7.G\xfc]\xc9\x14\xe1B@\xbb!8`'
      EMPTY_DATA = b'BZh9\x17rE8P\x90\x00\x00\x00\x00'
+     BAD_DATA = b'this is not a valid bzip2 file'
  
      def setUp(self):
 -        self.filename = TESTFN
 +        self.filename = support.TESTFN
  
      def tearDown(self):
          if os.path.isfile(self.filename):
@@@ -76,21 -72,32 +77,22 @@@
                  ret = bz2.decompress(data)
              return ret
  
 -    else:
 -        # bunzip2 isn't available to run on Windows.
 -        def decompress(self, data):
 -            return bz2.decompress(data)
  
  class BZ2FileTest(BaseTest):
 -    "Test BZ2File type miscellaneous methods."
 +    "Test the BZ2File class."
  
-     def createTempFile(self, streams=1):
+     def createTempFile(self, streams=1, suffix=b""):
          with open(self.filename, "wb") as f:
              f.write(self.DATA * streams)
+             f.write(suffix)
  
      def testBadArgs(self):
 -        with self.assertRaises(TypeError):
 -            BZ2File(123.456)
 -        with self.assertRaises(ValueError):
 -            BZ2File("/dev/null", "z")
 -        with self.assertRaises(ValueError):
 -            BZ2File("/dev/null", "rx")
 -        with self.assertRaises(ValueError):
 -            BZ2File("/dev/null", "rbt")
 -        with self.assertRaises(ValueError):
 -            BZ2File("/dev/null", compresslevel=0)
 -        with self.assertRaises(ValueError):
 -            BZ2File("/dev/null", compresslevel=10)
 +        self.assertRaises(TypeError, BZ2File, 123.456)
 +        self.assertRaises(ValueError, BZ2File, "/dev/null", "z")
 +        self.assertRaises(ValueError, BZ2File, "/dev/null", "rx")
 +        self.assertRaises(ValueError, BZ2File, "/dev/null", "rbt")
 +        self.assertRaises(ValueError, BZ2File, "/dev/null", compresslevel=0)
 +        self.assertRaises(ValueError, BZ2File, "/dev/null", compresslevel=10)
  
      def testRead(self):
          self.createTempFile()
@@@ -718,29 -731,29 +738,37 @@@ class CompressDecompressTest(BaseTest)
          text = bz2.decompress(self.DATA * 5)
          self.assertEqual(text, self.TEXT * 5)
  
+     def testDecompressTrailingJunk(self):
+         text = bz2.decompress(self.DATA + self.BAD_DATA)
+         self.assertEqual(text, self.TEXT)
+ 
+     def testDecompressMultiStreamTrailingJunk(self):
+         text = bz2.decompress(self.DATA * 5 + self.BAD_DATA)
+         self.assertEqual(text, self.TEXT * 5)
+ 
  
  class OpenTest(BaseTest):
 +    "Test the open function."
 +
 +    def open(self, *args, **kwargs):
 +        return bz2.open(*args, **kwargs)
 +
      def test_binary_modes(self):
 -        with bz2.open(self.filename, "wb") as f:
 -            f.write(self.TEXT)
 -        with open(self.filename, "rb") as f:
 -            file_data = bz2.decompress(f.read())
 -            self.assertEqual(file_data, self.TEXT)
 -        with bz2.open(self.filename, "rb") as f:
 -            self.assertEqual(f.read(), self.TEXT)
 -        with bz2.open(self.filename, "ab") as f:
 -            f.write(self.TEXT)
 -        with open(self.filename, "rb") as f:
 -            file_data = bz2.decompress(f.read())
 -            self.assertEqual(file_data, self.TEXT * 2)
 +        for mode in ("wb", "xb"):
 +            if mode == "xb":
 +                unlink(self.filename)
 +            with self.open(self.filename, mode) as f:
 +                f.write(self.TEXT)
 +            with open(self.filename, "rb") as f:
 +                file_data = self.decompress(f.read())
 +                self.assertEqual(file_data, self.TEXT)
 +            with self.open(self.filename, "rb") as f:
 +                self.assertEqual(f.read(), self.TEXT)
 +            with self.open(self.filename, "ab") as f:
 +                f.write(self.TEXT)
 +            with open(self.filename, "rb") as f:
 +                file_data = self.decompress(f.read())
 +                self.assertEqual(file_data, self.TEXT * 2)
  
      def test_implicit_binary_modes(self):
          # Test implicit binary modes (no "b" or "t" in mode string).