Update Lexer/Parser to support '#inline' and 'label'

ppapi/generators/idl_lexer.py

 #!/usr/bin/python
 #
 # Copyright (c) 2011 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
 """ Lexer for PPAPI IDL """
 
 #
 # IDL Lexer
@@ skipping 31 matching lines @@
 # IDL Lexer
 #
 class IDLLexer(object):
   # 'tokens' is a value required by lex which specifies the complete list
   # of valid token types.
   tokens = [
     # Symbol and keywords types
       'COMMENT',
       'DESCRIBE',
       'ENUM',
+      'LABEL',
       'SYMBOL',
+      'INLINE',
       'INTERFACE',
       'STRUCT',
       'TYPEDEF',
 
     # Data types
       'FLOAT',
       'OCT',
       'INT',
       'HEX',
       'STRING',
 
     # Operators
       'LSHIFT'
   ]
 
   # 'keywords' is a map of string to token type.  All SYMBOL tokens are
   # matched against keywords, to determine if the token is actually a keyword.
   keywords = {
+    'attribute' : 'ATTRIBUTE',
     'describe' : 'DESCRIBE',
     'enum'  : 'ENUM',
+    'label' : 'LABEL',
     'interface' : 'INTERFACE',
     'readonly' : 'READONLY',
     'struct' : 'STRUCT',
     'typedef' : 'TYPEDEF',
   }
 
   # 'literals' is a value expected by lex which specifies a list of valid
   # literal tokens, meaning the token type and token value are identical.
   literals = '"*.(){}[],;:=+-'
 
@@ skipping 24 matching lines @@
   def t_STRING(self, t):
     r'"[^"]*"'
     t.value = t.value[1:-1]
     self.AddLines(t.value.count('\n'))
     return t
 
   # A C or C++ style comment:  /* xxx */ or //
   def t_COMMENT(self, t):
     r'(/\*(.|\n)*?\*/)|(//.*)'
     self.AddLines(t.value.count('\n'))
+    return t
 
-    # C++ comments should keep the newline
-    if t.value[:2] == '//': t.value += '\n'
+  # Return a "preprocessor" inline block
+  def t_INLINE(self, t):
+    r'\#inline (.|\n)*\#endinl.*'
+    self.AddLines(t.value.count('\n'))
     return t
 
   # A symbol or keyword.
   def t_KEYWORD_SYMBOL(self, t):
     r'[A-Za-z][A-Za-z_0-9]*'
 
     #All non-keywords are assumed to be symbols
     t.type = self.keywords.get(t.value, 'SYMBOL')
     return t
 
   def t_ANY_error(self, t):
+    msg = "Unrecognized input"
     line = self.lexobj.lineno
+
+    # If that line has not been accounted for, then we must have hit
+    # EoF, so compute the beginning of the line that caused the problem.
+    if line >= len(self.index):
+      # Find the offset in the line of the first word causing the issue
+      word = t.value.split()[0]
+      offs = self.lines[line - 1].find(word)
+      # Add the computed line's starting position
+      self.index.append(self.lexobj.lexpos - offs)
+      msg = "Unexpected EoF reached after"
+
     pos = self.lexobj.lexpos - self.index[line]
     file = self.lexobj.filename
-    out = self.ErrorMessage(file, line, pos, "Unrecognized input")
+    out = self.ErrorMessage(file, line, pos, msg)
     sys.stderr.write(out + '\n')
 
   def AddLines(self, count):
     # Set the lexer position for the beginning of the next line.  In the case
     # of multiple lines, tokens can not exist on any of the lines except the
     # last one, so the recorded value for previous lines are unused.  We still
     # fill the array however, to make sure the line count is correct.
     self.lexobj.lineno += count
     for i in range(count):
       self.index.append(self.lexobj.lexpos)
@@ skipping 37 matching lines @@
   for filename in filenames:
     data = open(filename).read()
     lexer.SetData(filename, data)
     if verbose: sys.stdout.write('  Loaded %s...\n' % filename)
     while 1:
       t = lexer.lexobj.token()
       if t is None: break
       outlist.append(t)
   return outlist
 
+
+def TokensFromText(text):
+  lexer = IDLLexer()
+  lexer.SetData('unknown', text)
+  outlist = []
+  while 1:
+    t = lexer.lexobj.token()
+    if t is None: break
+    outlist.append(t.value)
+  return outlist
+
 #
 # TextToTokens
 #
 # From a block of text, generate a list of tokens
 #
 def TextToTokens(source):
   lexer = IDLLexer()
   outlist = []
   lexer.SetData('AUTO', source)
   while 1:
     t = lexer.lexobj.token()
     if t is None: break
     outlist.append(t.value)
   return outlist
 
 
 #
 # TestSame
 #
 # From a set of token values, generate a new source text by joining with a
 # single space.  The new source is then tokenized and compared against the
 # old set.
 #
-def TestSame(values):
-  src1 = ' '.join(values)
-  src2 = ' '.join(TextToTokens(src1))
+def TestSame(values1):
+  # Recreate the source from the tokens.  We use newline instead of whitespace
+  # since the '//' and #inline regex are line sensitive.
+  text = '\n'.join(values1)
+  values2 = TextToTokens(text)
+
+  count1 = len(values1)
+  count2 = len(values2)
+  if count1 != count2:
+    print "Size mismatch original %d vs %d\n" % (count1, count2)
+    if count1 > count2: count1 = count2
+
+  for i in range(count1):
+    if values1[i] != values2[i]:
+      print "%d >>%s<< >>%s<<" % (i, values1[i], values2[i])
 
   if GetOption('output'):
     sys.stdout.write('Generating original.txt and tokenized.txt\n')
     open('original.txt', 'w').write(src1)
     open('tokenized.txt', 'w').write(src2)
 
-  if src1 == src2:
+  if values1 == values2:
     sys.stdout.write('Same: Pass\n')
     return 0
 
+  print "****************\n%s\n%s***************\n" % (src1, src2)
   sys.stdout.write('Same: Failed\n')
   return -1
 
 
 #
 # TestExpect
 #
 # From a set of tokens pairs, verify the type field of the second matches
 # the value of the first, so that:
 # INT 123 FLOAT 1.1
@@ skipping 38 matching lines @@
         return -1
     return 0
 
   except lex.LexError as le:
     sys.stderr.write('%s\n' % str(le))
   return -1
 
 if __name__ == '__main__':
   sys.exit(Main(sys.argv[1:]))