third_party/WebKit/Source/build/scripts/make_css_tokenizer_codepoints.py

   1 #!/usr/bin/env python
   2
   3 # Copyright 2014 The Chromium Authors. All rights reserved.
   4 # Use of this source code is governed by a BSD-style license that can be
   5 # found in the LICENSE file.
   6
   7 import in_generator
   8 import sys
   9 import os
  10
  11 module_basename = os.path.basename(__file__)
  12 module_pyname = os.path.splitext(module_basename)[0] + '.py'
  13
  14 CPP_TEMPLATE = """
  15 // Copyright 2014 The Chromium Authors. All rights reserved.
  16 // Use of this source code is governed by a BSD-style license that can be
  17 // found in the LICENSE file.
  18
  19 // Auto-generated by {module_pyname}
  20
  21 const CSSTokenizer::CodePoint CSSTokenizer::codePoints[{array_size}] = {{
  22 {token_lines}
  23 }};
  24 const unsigned codePointsNumber = {array_size};
  25 """
  26
  27
  28 def token_type(i):
  29     codepoints = {'(': 'leftParenthesis',
  30                   ')': 'rightParenthesis',
  31                   '[': 'leftBracket',
  32                   ']': 'rightBracket',
  33                   '{': 'leftBrace',
  34                   '}': 'rightBrace',
  35                   '+': 'plusOrFullStop',
  36                   '.': 'plusOrFullStop',
  37                   '-': 'hyphenMinus',
  38                   '*': 'asterisk',
  39                   '<': 'lessThan',
  40                   ',': 'comma',
  41                   '/': 'solidus',
  42                   '\\': 'reverseSolidus',
  43                   ':': 'colon',
  44                   ';': 'semiColon',
  45                   '#': 'hash',
  46                   '^': 'circumflexAccent',
  47                   '$': 'dollarSign',
  48                   '|': 'verticalLine',
  49                   '~': 'tilde',
  50                   '@': 'commercialAt',
  51                   'u': 'letterU',
  52                   'U': 'letterU',
  53                   }
  54     c = chr(i)
  55     if c in codepoints:
  56         return codepoints[c]
  57     whitespace = '\n\r\t\f '
  58     quotes = '"\''
  59     if c in whitespace:
  60         return 'whiteSpace'
  61     if c.isdigit():
  62         return 'asciiDigit'
  63     if c.isalpha() or c == '_':
  64         return 'nameStart'
  65     if c in quotes:
  66         return 'stringStart'
  67     if i == 0:
  68         return 'endOfFile'
  69
  70
  71 class MakeCSSTokenizerCodePointsWriter(in_generator.Writer):
  72     def __init__(self, in_file_path):
  73         super(MakeCSSTokenizerCodePointsWriter, self).__init__(in_file_path)
  74
  75         self._outputs = {
  76             ('CSSTokenizerCodepoints.cpp'): self.generate,
  77         }
  78
  79     def generate(self):
  80         array_size = 128  # SCHAR_MAX + 1
  81         token_lines = ['    &CSSTokenizer::%s,' % token_type(i)
  82                         if token_type(i) else '    0,'
  83                         for i in range(array_size)]
  84         return CPP_TEMPLATE.format(array_size=array_size, token_lines='\n'.join(token_lines), module_pyname=module_pyname)
  85
  86 if __name__ == '__main__':
  87     in_generator.Maker(MakeCSSTokenizerCodePointsWriter).main(sys.argv)