added test for _encode_utf16_str

olefile/olefile.py

@@ -833,6 +833,7 @@ def __init__(self, entry, sid, ole_file):
  log.debug(' - sect: %Xh' % self.isectStart)
  log.debug(' - SID left: %d, right: %d, child: %d' % (self.sid_left,
  self.sid_right, self.sid_child))
+ log.debug(' - name_utf16: %r' % self.name_utf16)
 
  # sizeHigh is only used for 4K sectors, it should be zero for 512 bytes
  # sectors, BUT apparently some implementations set it as 0xFFFFFFFF, 1
@@ -1255,7 +1256,12 @@ def _encode_utf16_str(self, s, errors='replace'):
  # an encoding has been specified for path names:
  s = s.decode(self.path_encoding, errors)
  # Else we assume that s is a unicode string
- return s.encode('UTF-16LE', errors)
+ s_utf16 = s.encode('UTF-16LE', errors)
+ # remove the BOM if present (2 bytes FFFE in little-endian):
+ # => not needed with UTF-16LE codec, only UTF-16 adds the BOM
+ # if s_utf16.startswith(b'\xff\xfe'):
+ # s_utf16 = s_utf16[2:]
+ return s_utf16
 
 
  def open(self, filename, write_mode=False):

tests/test_olefile.py

@@ -232,6 +232,21 @@ def test_write_sect_too_large(self):
  with self.assertRaises(ValueError):
  ole.write_sect(ole.nb_sect - 1, b"x" * size)
 
+ def test_encode_utf16_str(self):
+ "test _encode_utf16_str"
+ # open OLE file with no path encoding (=Unicode)
+ with olefile.OleFileIO(self.ole_file, path_encoding=None) as ole:
+ # Test encoding various unicode strings
+ # Note: there should be no BOM (FFFE) in the encoded string
+ s = ole._encode_utf16_str(u'a')
+ self.assertEqual(s, b'a\x00')
+ # a Unicode code point that is encoded with 2 bytes in UTF-16:
+ s = ole._encode_utf16_str(u'\u265E') # 265E=BLACK CHESS KNIGHT
+ self.assertEqual(s, b'\x5E\x26')
+ # a Unicode code point that is encoded with 4 bytes in UTF-16:
+ s = ole._encode_utf16_str(u'\U0001F609') # 1F609=WINKING FACE
+ self.assertEqual(s, b'\x3D\xD8\x09\xDE')
+
 
 class FileHandleCloseTest(unittest.TestCase):
  """Test file handles are closed correctly."""