2000-03-11 07:17:24 +08:00
|
|
|
""" Python 'utf-16' Codec
|
|
|
|
|
|
|
|
|
|
|
|
Written by Marc-Andre Lemburg (mal@lemburg.com).
|
|
|
|
|
|
|
|
(c) Copyright CNRI, All Rights Reserved. NO WARRANTY.
|
|
|
|
|
|
|
|
"""
|
2001-06-20 04:07:51 +08:00
|
|
|
import codecs, sys
|
2000-03-11 07:17:24 +08:00
|
|
|
|
|
|
|
### Codec APIs
|
|
|
|
|
2004-09-08 04:24:22 +08:00
|
|
|
encode = codecs.utf_16_encode
|
2000-03-11 07:17:24 +08:00
|
|
|
|
2004-09-08 04:24:22 +08:00
|
|
|
def decode(input, errors='strict'):
|
|
|
|
return codecs.utf_16_decode(input, errors, True)
|
2000-03-11 07:17:24 +08:00
|
|
|
|
2004-09-08 04:24:22 +08:00
|
|
|
class StreamWriter(codecs.StreamWriter):
|
2001-06-20 04:07:51 +08:00
|
|
|
def __init__(self, stream, errors='strict'):
|
2004-09-08 04:24:22 +08:00
|
|
|
self.bom_written = False
|
2001-06-20 04:07:51 +08:00
|
|
|
codecs.StreamWriter.__init__(self, stream, errors)
|
|
|
|
|
2004-09-08 04:24:22 +08:00
|
|
|
def encode(self, input, errors='strict'):
|
|
|
|
self.bom_written = True
|
|
|
|
result = codecs.utf_16_encode(input, errors)
|
|
|
|
if sys.byteorder == 'little':
|
|
|
|
self.encode = codecs.utf_16_le_encode
|
|
|
|
else:
|
|
|
|
self.encode = codecs.utf_16_be_encode
|
2001-06-20 04:07:51 +08:00
|
|
|
return result
|
2002-08-09 04:19:19 +08:00
|
|
|
|
2004-09-08 04:24:22 +08:00
|
|
|
class StreamReader(codecs.StreamReader):
|
|
|
|
|
|
|
|
def decode(self, input, errors='strict'):
|
|
|
|
(object, consumed, byteorder) = \
|
|
|
|
codecs.utf_16_ex_decode(input, errors, 0, False)
|
|
|
|
if byteorder == -1:
|
|
|
|
self.decode = codecs.utf_16_le_decode
|
|
|
|
elif byteorder == 1:
|
|
|
|
self.decode = codecs.utf_16_be_decode
|
|
|
|
elif consumed>=2:
|
|
|
|
raise UnicodeError,"UTF-16 stream does not start with BOM"
|
|
|
|
return (object, consumed)
|
2002-04-05 20:12:00 +08:00
|
|
|
|
2000-03-11 07:17:24 +08:00
|
|
|
### encodings module API
|
|
|
|
|
|
|
|
def getregentry():
|
|
|
|
|
2004-09-08 04:24:22 +08:00
|
|
|
return (encode,decode,StreamReader,StreamWriter)
|