Fix extension id calculation to take into account Windows encoding.

tools/crx_id/crx_id.py

 #!/usr/bin/env python
 # Copyright (c) 2012 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
 """ Read a CRX file and write out the App ID and the Full Hash of the ID.
 See: http://code.google.com/chrome/extensions/crx.html
 and 'http://stackoverflow.com/questions/'
   + '1882981/google-chrome-alphanumeric-hashes-to-identify-extensions'
 for docs on the format.
@@ skipping 57 matching lines @@
   version = f.read(4)
   if not version[0] != EXPECTED_CRX_VERSION:
     raise Exception('Invalid version number: %s (expecting %s)' %
                     (version,
                      EXPECTED_CRX_VERSION))
   pub_key_len_bytes = HexToInt(f.read(4))
   sig_len_bytes = HexToInt(f.read(4))
   pub_key = f.read(pub_key_len_bytes)
   return pub_key
 
-def GetPublicKeyFromPath(filepath):
+def GetPublicKeyFromPath(filepath, is_win_path=False):
   # Normalize the path for windows to have capital drive letters.
   # We intentionally don't check if sys.platform == 'win32' and just
   # check if this looks like drive letter so that we can test this
   # even on posix systems.
   if (len(filepath) >= 2 and
       filepath[0].islower() and
       filepath[1] == ':'):
-      return filepath[0].upper() + filepath[1:]
+      filepath = filepath[0].upper() + filepath[1:]
+
+  # On Windows, filepaths are encoded using UTF-16, little endian byte order,
+  # using "wide characters" that are 16 bits in size. On POSIX systems, the
+  # encoding is generally UTF-8, which has the property of being equivalent to
+  # ASCII when only ASCII characters are in the path.
+  if is_win_path:
+    filepath = filepath.encode('utf-16le')
+
   return filepath
 
 def GetPublicKeyUnpacked(f, filepath):
   manifest = json.load(f)
   if 'key' not in manifest:
     # Use the path as the public key.
     # See Extension::GenerateIdForPath in extension.cc
     return GetPublicKeyFromPath(filepath)
   else:
     return base64.standard_b64decode(manifest['key'])
 
 def HasPublicKey(filename):
   if os.path.isdir(filename):
     with open(os.path.join(filename, 'manifest.json'), 'rb') as f:
       manifest = json.load(f)
       return 'key' in manifest
   return False
 
-def GetPublicKey(filename, from_file_path):
+def GetPublicKey(filename, from_file_path, is_win_path=False):
   if from_file_path:
-    return GetPublicKeyFromPath(filename)
+    return GetPublicKeyFromPath(
+        filename, is_win_path=is_win_path)
 
   pub_key = ''
   if os.path.isdir(filename):
     # Assume it's an unpacked extension
     f = open(os.path.join(filename, 'manifest.json'), 'rb')
     pub_key = GetPublicKeyUnpacked(f, filename)
     f.close()
   else:
     # Assume it's a packed extension.
     f = open(filename, 'rb')
     pub_key = GetPublicKeyPacked(f)
     f.close()
   return pub_key
 
-def GetCRXHash(filename, from_file_path=False):
-  pub_key = GetPublicKey(filename, from_file_path)
+def GetCRXHash(filename, from_file_path=False, is_win_path=False):
+  pub_key = GetPublicKey(filename, from_file_path, is_win_path=is_win_path)
   pub_key_hash = hashlib.sha256(pub_key).digest()
   return HexTo256(pub_key_hash)
 
-def GetCRXAppID(filename, from_file_path=False):
-  pub_key = GetPublicKey(filename, from_file_path)
+def GetCRXAppID(filename, from_file_path=False, is_win_path=False):
+  pub_key = GetPublicKey(filename, from_file_path, is_win_path=is_win_path)
   pub_key_hash = hashlib.sha256(pub_key).digest()
   # AppID is the MPDecimal of only the first 128 bits of the hash.
   return HexToMPDecimal(pub_key_hash[:128/8])
 
 def main(argv):
   if len(argv) != 2:
     usage(argv)
     return 1
   print 'Raw Bytes: %s' % GetCRXHash(sys.argv[1])
   print 'AppID: %s' % GetCRXAppID(sys.argv[1])
 
 
 if __name__ == '__main__':
   sys.exit(main(sys.argv))