tools/idl_parser/idl_lexer.py

   1 #!/usr/bin/env python
   2 # Copyright (c) 2013 The Chromium Authors. All rights reserved.
   3 # Use of this source code is governed by a BSD-style license that can be
   4 # found in the LICENSE file.
   5
   6 """ Lexer for PPAPI IDL
   7
   8 The lexer uses the PLY library to build a tokenizer which understands both
   9 WebIDL and Pepper tokens.
  10
  11 WebIDL, and WebIDL regular expressions can be found at:
  12    http://www.w3.org/TR/2012/CR-WebIDL-20120419/
  13 PLY can be found at:
  14    http://www.dabeaz.com/ply/
  15 """
  16
  17 import os.path
  18 import sys
  19
  20 #
  21 # Try to load the ply module, if not, then assume it is in the third_party
  22 # directory.
  23 #
  24 try:
  25   # Disable lint check which fails to find the ply module.
  26   # pylint: disable=F0401
  27   from ply import lex
  28 except ImportError:
  29   module_path, module_name = os.path.split(__file__)
  30   third_party = os.path.join(module_path, '..', '..', 'third_party')
  31   sys.path.append(third_party)
  32   # pylint: disable=F0401
  33   from ply import lex
  34
  35 #
  36 # IDL Lexer
  37 #
  38 class IDLLexer(object):
  39   # 'literals' is a value expected by lex which specifies a list of valid
  40   # literal tokens, meaning the token type and token value are identical.
  41   literals = r'"*.(){}[],;:=+-/~|&^?<>'
  42
  43   # 't_ignore' contains ignored characters (spaces and tabs)
  44   t_ignore = ' \t'
  45
  46   # 'tokens' is a value required by lex which specifies the complete list
  47   # of valid token types.
  48   tokens = [
  49     # Data types
  50       'float',
  51       'integer',
  52       'string',
  53
  54     # Symbol and keywords types
  55       'COMMENT',
  56       'identifier',
  57
  58     # MultiChar operators
  59       'ELLIPSIS',
  60   ]
  61
  62   # 'keywords' is a map of string to token type.  All tokens matching
  63   # KEYWORD_OR_SYMBOL are matched against keywords dictionary, to determine
  64   # if the token is actually a keyword.
  65   keywords = {
  66     'any' : 'ANY',
  67     'attribute' : 'ATTRIBUTE',
  68     'boolean' : 'BOOLEAN',
  69     'byte' : 'BYTE',
  70     'ByteString' : 'BYTESTRING',
  71     'callback' : 'CALLBACK',
  72     'const' : 'CONST',
  73     'creator' : 'CREATOR',
  74     'Date' : 'DATE',
  75     'deleter' : 'DELETER',
  76     'dictionary' : 'DICTIONARY',
  77     'DOMString' : 'DOMSTRING',
  78     'double' : 'DOUBLE',
  79     'enum'  : 'ENUM',
  80     'false' : 'FALSE',
  81     'float' : 'FLOAT',
  82     'exception' : 'EXCEPTION',
  83     'getter': 'GETTER',
  84     'implements' : 'IMPLEMENTS',
  85     'Infinity' : 'INFINITY',
  86     'inherit' : 'INHERIT',
  87     'interface' : 'INTERFACE',
  88     'legacycaller' : 'LEGACYCALLER',
  89     'long' : 'LONG',
  90     'Nan' : 'NAN',
  91     'null' : 'NULL',
  92     'object' : 'OBJECT',
  93     'octet' : 'OCTET',
  94     'optional' : 'OPTIONAL',
  95     'or' : 'OR',
  96     'partial'  : 'PARTIAL',
  97     'readonly' : 'READONLY',
  98     'RegExp' : 'REGEXP',
  99     'sequence' : 'SEQUENCE',
 100     'serializer' : 'SERIALIZER',
 101     'setter': 'SETTER',
 102     'short' : 'SHORT',
 103     'static' : 'STATIC',
 104     'stringifier' : 'STRINGIFIER',
 105     'typedef' : 'TYPEDEF',
 106     'true' : 'TRUE',
 107     'unsigned' : 'UNSIGNED',
 108     'unrestricted' : 'UNRESTRICTED',
 109     'void' : 'VOID'
 110   }
 111
 112   # Token definitions
 113   #
 114   # Lex assumes any value or function in the form of 't_<TYPE>' represents a
 115   # regular expression where a match will emit a token of type <TYPE>.  In the
 116   # case of a function, the function is called when a match is made. These
 117   # definitions come from WebIDL.
 118   #
 119   # These need to be methods for lexer construction, despite not using self.
 120   # pylint: disable=R0201
 121   def t_ELLIPSIS(self, t):
 122     r'\.\.\.'
 123     return t
 124
 125   # Regex needs to be in the docstring
 126   # pylint: disable=C0301
 127   def t_float(self, t):
 128     r'-?(([0-9]+\.[0-9]*|[0-9]*\.[0-9]+)([Ee][+-]?[0-9]+)?|[0-9]+[Ee][+-]?[0-9]+)'
 129     return t
 130
 131   def t_integer(self, t):
 132     r'-?([1-9][0-9]*|0[Xx][0-9A-Fa-f]+|0[0-7]*)'
 133     return t
 134
 135
 136   # A line ending '\n', we use this to increment the line number
 137   def t_LINE_END(self, t):
 138     r'\n+'
 139     self.AddLines(len(t.value))
 140
 141   # We do not process escapes in the IDL strings.  Strings are exclusively
 142   # used for attributes and enums, and not used as typical 'C' constants.
 143   def t_string(self, t):
 144     r'"[^"]*"'
 145     t.value = t.value[1:-1]
 146     self.AddLines(t.value.count('\n'))
 147     return t
 148
 149   # A C or C++ style comment:  /* xxx */ or //
 150   def t_COMMENT(self, t):
 151     r'(/\*(.|\n)*?\*/)|(//.*(\n[ \t]*//.*)*)'
 152     self.AddLines(t.value.count('\n'))
 153     return t
 154
 155   # A symbol or keyword.
 156   def t_KEYWORD_OR_SYMBOL(self, t):
 157     r'_?[A-Za-z][A-Za-z_0-9]*'
 158
 159     # All non-keywords are assumed to be symbols
 160     t.type = self.keywords.get(t.value, 'identifier')
 161
 162     # We strip leading underscores so that you can specify symbols with the same
 163     # value as a keywords (E.g. a dictionary named 'interface').
 164     if t.value[0] == '_':
 165       t.value = t.value[1:]
 166     return t
 167
 168   def t_ANY_error(self, t):
 169     msg = 'Unrecognized input'
 170     line = self.Lexer().lineno
 171
 172     # If that line has not been accounted for, then we must have hit
 173     # EoF, so compute the beginning of the line that caused the problem.
 174     if line >= len(self.index):
 175       # Find the offset in the line of the first word causing the issue
 176       word = t.value.split()[0]
 177       offs = self.lines[line - 1].find(word)
 178       # Add the computed line's starting position
 179       self.index.append(self.Lexer().lexpos - offs)
 180       msg = 'Unexpected EoF reached after'
 181
 182     pos = self.Lexer().lexpos - self.index[line]
 183     out = self.ErrorMessage(line, pos, msg)
 184     sys.stderr.write(out + '\n')
 185     self._lex_errors += 1
 186
 187
 188   def AddLines(self, count):
 189     # Set the lexer position for the beginning of the next line.  In the case
 190     # of multiple lines, tokens can not exist on any of the lines except the
 191     # last one, so the recorded value for previous lines are unused.  We still
 192     # fill the array however, to make sure the line count is correct.
 193     self.Lexer().lineno += count
 194     for _ in range(count):
 195       self.index.append(self.Lexer().lexpos)
 196
 197   def FileLineMsg(self, line, msg):
 198     # Generate a message containing the file and line number of a token.
 199     filename = self.Lexer().filename
 200     if filename:
 201       return "%s(%d) : %s" % (filename, line + 1, msg)
 202     return "<BuiltIn> : %s" % msg
 203
 204   def SourceLine(self, line, pos):
 205     # Create a source line marker
 206     caret = ' ' * pos + '^'
 207     # We decrement the line number since the array is 0 based while the
 208     # line numbers are 1 based.
 209     return "%s\n%s" % (self.lines[line - 1], caret)
 210
 211   def ErrorMessage(self, line, pos, msg):
 212     return "\n%s\n%s" % (
 213         self.FileLineMsg(line, msg),
 214         self.SourceLine(line, pos))
 215
 216 #
 217 # Tokenizer
 218 #
 219 # The token function returns the next token provided by IDLLexer for matching
 220 # against the leaf paterns.
 221 #
 222   def token(self):
 223     tok = self.Lexer().token()
 224     if tok:
 225       self.last = tok
 226     return tok
 227
 228
 229   def GetTokens(self):
 230     outlist = []
 231     while True:
 232       t = self.Lexer().token()
 233       if not t:
 234         break
 235       outlist.append(t)
 236     return outlist
 237
 238   def Tokenize(self, data, filename='__no_file__'):
 239     lexer = self.Lexer()
 240     lexer.lineno = 1
 241     lexer.filename = filename
 242     lexer.input(data)
 243     self.lines = data.split('\n')
 244
 245   def KnownTokens(self):
 246     return self.tokens
 247
 248   def Lexer(self):
 249     if not self._lexobj:
 250       self._lexobj = lex.lex(object=self, lextab=None, optimize=0)
 251     return self._lexobj
 252
 253   def _AddToken(self, token):
 254     if token in self.tokens:
 255       raise RuntimeError('Same token: ' + token)
 256     self.tokens.append(token)
 257
 258   def _AddTokens(self, tokens):
 259     for token in tokens:
 260       self._AddToken(token)
 261
 262   def _AddKeywords(self, keywords):
 263     for key in keywords:
 264       value = key.upper()
 265       self._AddToken(value)
 266       self.keywords[key] = value
 267
 268   def _DelKeywords(self, keywords):
 269     for key in keywords:
 270       self.tokens.remove(key.upper())
 271       del self.keywords[key]
 272
 273   def __init__(self):
 274     self.index = [0]
 275     self._lex_errors = 0
 276     self.linex = []
 277     self.filename = None
 278     self.keywords = {}
 279     self.tokens = []
 280     self._AddTokens(IDLLexer.tokens)
 281     self._AddKeywords(IDLLexer.keywords)
 282     self._lexobj = None
 283     self.last = None
 284     self.lines = None
 285
 286 # If run by itself, attempt to build the lexer
 287 if __name__ == '__main__':
 288   lexer_object = IDLLexer()