cpython/Tools/unicode/genwincodec.py
Benjamin Peterson 75ad1fc089 Merged revisions 78806 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk

........
  r78806 | benjamin.peterson | 2010-03-08 16:15:11 -0600 (Mon, 08 Mar 2010) | 1 line

  set svn:eol-style on various files
........
2010-03-08 22:17:58 +00:00

62 lines
1.7 KiB
Python

"""This script generates a Python codec module from a Windows Code Page.
It uses the function MultiByteToWideChar to generate a decoding table.
"""
import ctypes
from ctypes import wintypes
from gencodec import codegen
import unicodedata
def genwinmap(codepage):
MultiByteToWideChar = ctypes.windll.kernel32.MultiByteToWideChar
MultiByteToWideChar.argtypes = [wintypes.UINT, wintypes.DWORD,
wintypes.LPCSTR, ctypes.c_int,
wintypes.LPWSTR, ctypes.c_int]
MultiByteToWideChar.restype = ctypes.c_int
enc2uni = {}
for i in list(range(32)) + [127]:
enc2uni[i] = (i, 'CONTROL CHARACTER')
for i in range(256):
buf = ctypes.create_unicode_buffer(2)
ret = MultiByteToWideChar(
codepage, 0,
bytes([i]), 1,
buf, 2)
assert ret == 1, "invalid code page"
assert buf[1] == '\x00'
try:
name = unicodedata.name(buf[0])
except ValueError:
try:
name = enc2uni[i][1]
except KeyError:
name = ''
enc2uni[i] = (ord(buf[0]), name)
return enc2uni
def genwincodec(codepage):
import platform
map = genwinmap(codepage)
encodingname = 'cp%d' % codepage
code = codegen("", map, encodingname)
# Replace first lines with our own docstring
code = '''\
"""Python Character Mapping Codec %s generated on Windows:
%s with the command:
python Tools/unicode/genwincodec.py %s
"""#"
''' % (encodingname, ' '.join(platform.win32_ver()), codepage
) + code.split('"""#"', 1)[1]
print(code)
if __name__ == '__main__':
import sys
genwincodec(int(sys.argv[1]))