BaseTools/Scripts/ConvertUni.py

   1 ## @file
   2 #  Check a patch for various format issues
   3 #
   4 #  Copyright (c) 2015, Intel Corporation. All rights reserved.<BR>
   5 #
   6 #  This program and the accompanying materials are licensed and made
   7 #  available under the terms and conditions of the BSD License which
   8 #  accompanies this distribution. The full text of the license may be
   9 #  found at http://opensource.org/licenses/bsd-license.php
  10 #
  11 #  THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS"
  12 #  BASIS, WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER
  13 #  EXPRESS OR IMPLIED.
  14 #
  15
  16 from __future__ import print_function
  17
  18 VersionNumber = '0.1'
  19 __copyright__ = "Copyright (c) 2015, Intel Corporation  All rights reserved."
  20
  21 import argparse
  22 import codecs
  23 import os
  24 import sys
  25
  26 class ConvertOneArg:
  27     """Converts utf-16 to utf-8 for one command line argument.
  28
  29        This could be a single file, or a directory.
  30     """
  31
  32     def __init__(self, utf8, source):
  33         self.utf8 = utf8
  34         self.source = source
  35
  36         self.ok = True
  37
  38         if not os.path.exists(source):
  39             self.ok = False
  40         elif os.path.isdir(source):
  41             for (root, dirs, files) in os.walk(source):
  42                 files = filter(lambda a: a.endswith('.uni'), files)
  43                 for filename in files:
  44                     path = os.path.join(root, filename)
  45                     self.ok &= self.convert_one_file(path)
  46                     if not self.ok:
  47                         break
  48
  49                 if not self.ok:
  50                     break
  51         else:
  52             self.ok &= self.convert_one_file(source)
  53
  54     def convert_one_file(self, source):
  55         if self.utf8:
  56             new_enc, old_enc = 'utf-8', 'utf-16'
  57         else:
  58             new_enc, old_enc = 'utf-16', 'utf-8'
  59         #
  60         # Read file
  61         #
  62         f = open(source, mode='rb')
  63         file_content = f.read()
  64         f.close()
  65
  66         #
  67         # Detect UTF-16 Byte Order Mark at beginning of file.
  68         #
  69         bom = (file_content.startswith(codecs.BOM_UTF16_BE) or
  70                file_content.startswith(codecs.BOM_UTF16_LE))
  71         if bom != self.utf8:
  72             print("%s: already %s" % (source, new_enc))
  73             return True
  74
  75         #
  76         # Decode old string data
  77         #
  78         str_content = file_content.decode(old_enc, 'ignore')
  79
  80         #
  81         # Encode new string data
  82         #
  83         new_content = str_content.encode(new_enc, 'ignore')
  84
  85         #
  86         # Write converted data back to file
  87         #
  88         f = open(source, mode='wb')
  89         f.write(new_content)
  90         f.close()
  91
  92         print(source + ": converted, size", len(file_content), '=>', len(new_content))
  93         return True
  94
  95
  96 class ConvertUniApp:
  97     """Converts .uni files between utf-16 and utf-8."""
  98
  99     def __init__(self):
 100         self.parse_options()
 101         sources = self.args.source
 102
 103         self.ok = True
 104         for patch in sources:
 105             self.process_one_arg(patch)
 106
 107         if self.ok:
 108             self.retval = 0
 109         else:
 110             self.retval = -1
 111
 112     def process_one_arg(self, arg):
 113         self.ok &= ConvertOneArg(self.utf8, arg).ok
 114
 115     def parse_options(self):
 116         parser = argparse.ArgumentParser(description=__copyright__)
 117         parser.add_argument('--version', action='version',
 118                             version='%(prog)s ' + VersionNumber)
 119         parser.add_argument('source', nargs='+',
 120                             help='[uni file | directory]')
 121         group = parser.add_mutually_exclusive_group()
 122         group.add_argument("--utf-8",
 123                            action="store_true",
 124                            help="Convert from utf-16 to utf-8 [default]")
 125         group.add_argument("--utf-16",
 126                            action="store_true",
 127                            help="Convert from utf-8 to utf-16")
 128         self.args = parser.parse_args()
 129         self.utf8 = not self.args.utf_16
 130
 131 if __name__ == "__main__":
 132     sys.exit(ConvertUniApp().retval)