Add support for using expressions as enum values.

mojo/public/bindings/parse/mojo_parser.py

Index: mojo/public/bindings/parse/mojo_parser.py
diff --git a/mojo/public/bindings/parse/mojo_parser.py b/mojo/public/bindings/parse/mojo_parser.py
index 3e8e991b51f447b9ae8972538881944ddfdbc310..80d0a121684dbd985ba3ee6f23ae1de5490830d3 100755
--- a/mojo/public/bindings/parse/mojo_parser.py
+++ b/mojo/public/bindings/parse/mojo_parser.py
@@ -9,7 +9,6 @@
 import sys
 import os.path
 
-
 # Try to load the ply module, if not, then assume it is in the third_party
 # directory.
 try:
@@ -26,6 +25,8 @@ except ImportError:
   from ply import lex
   from ply import yacc
 
+from .mojo_lexer import Lexer
+
 
 def ListFromConcat(*items):
   """Generate list by concatenating inputs"""
@@ -41,114 +42,13 @@ def ListFromConcat(*items):
   return itemsout
 
 
-class Lexer(object):
-
-  # This field is required by lex to specify the complete list of valid tokens.
-  tokens = (
-    'NAME',
-    'NUMBER',
-
-    'ORDINAL',
-
-    'HANDLE',
-    'DATAPIPECONSUMER',
-    'DATAPIPEPRODUCER',
-    'MESSAGEPIPE',
-
-    'MODULE',
-    'STRUCT',
-    'INTERFACE',
-    'ENUM',
-    'VOID',
-
-    'LCURLY',
-    'RCURLY',
-    'LPAREN',
-    'RPAREN',
-    'LANGLE',
-    'RANGLE',
-    'LBRACKET',
-    'RBRACKET',
-    'COMMA',
-    'SEMICOLON',
-    'EQUALS',
-  )
-
-  t_LCURLY     = r'{'
-  t_RCURLY     = r'}'
-  t_LPAREN     = r'\('
-  t_RPAREN     = r'\)'
-  t_LANGLE     = r'<'
-  t_RANGLE     = r'>'
-  t_LBRACKET   = r'\['
-  t_RBRACKET   = r'\]'
-  t_COMMA      = r','
-  t_SEMICOLON  = r';'
-  t_EQUALS     = r'='
-  t_NAME       = r'[a-zA-Z_][a-zA-Z0-9_]*'
-  t_NUMBER     = r'\d+'
-  t_ORDINAL    = r'@[0-9]*'
-
-  def t_HANDLE(self, t):
-    r'handle'
-    return t
-
-  def t_DATAPIPECONSUMER(self, t):
-    r'data_pipe_consumer'
-    return t
-
-  def t_DATAPIPEPRODUCER(self, t):
-    r'data_pipe_producer'
-    return t
-
-  def t_MESSAGEPIPE(self, t):
-    r'message_pipe'
-    return t
-
-  def t_MODULE(self, t):
-    r'module'
-    return t
-
-  def t_STRUCT(self, t):
-    r'struct'
-    return t
-
-  def t_INTERFACE(self, t):
-    r'interface'
-    return t
-
-  def t_ENUM(self, t):
-    r'enum'
-    return t
-
-  def t_VOID(self, t):
-    r'void'
-    return t
-
-  # Ignore C and C++ style comments
-  def t_COMMENT(self, t):
-    r'(/\*(.|\n)*?\*/)|(//.*(\n[ \t]*//.*)*)'
-    pass
-
-  # Ignored characters
-  t_ignore = " \t"
-
-  def t_newline(self, t):
-    r'\n+'
-    t.lexer.lineno += t.value.count("\n")
-
-  def t_error(self, t):
-    print("Illegal character '%s'" % t.value[0])
-    t.lexer.skip(1)
-
-
 class Parser(object):
 
   def __init__(self, lexer):
     self.tokens = lexer.tokens
 
   def p_module(self, p):
-    """module : MODULE NAME LCURLY definitions RCURLY"""
+    """module : MODULE NAME LBRACE definitions RBRACE"""
     p[0] = ('MODULE', p[2], p[4])
 
   def p_definitions(self, p):
@@ -179,12 +79,12 @@ class Parser(object):
       p[0] = ListFromConcat(p[1], p[3])
 
   def p_attribute(self, p):
-    """attribute : NAME EQUALS NUMBER
+    """attribute : NAME EQUALS expression
                  | NAME EQUALS NAME"""
     p[0] = ('ATTRIBUTE', p[1], p[3])
 
   def p_struct(self, p):
-    """struct : attribute_section STRUCT NAME LCURLY struct_body RCURLY SEMICOLON"""
+    """struct : attribute_section STRUCT NAME LBRACE struct_body RBRACE SEMI"""
     p[0] = ('STRUCT', p[3], p[1], p[5])
 
   def p_struct_body(self, p):
@@ -195,11 +95,11 @@ class Parser(object):
       p[0] = ListFromConcat(p[1], p[2])
 
   def p_field(self, p):
-    """field : typename NAME ordinal SEMICOLON"""
+    """field : typename NAME ordinal SEMI"""
     p[0] = ('FIELD', p[1], p[2], p[3])
 
   def p_interface(self, p):
-    """interface : attribute_section INTERFACE NAME LCURLY interface_body RCURLY SEMICOLON"""
+    """interface : attribute_section INTERFACE NAME LBRACE interface_body RBRACE SEMI"""
     p[0] = ('INTERFACE', p[3], p[1], p[5])
 
   def p_interface_body(self, p):
@@ -210,7 +110,7 @@ class Parser(object):
       p[0] = ListFromConcat(p[1], p[2])
 
   def p_method(self, p):
-    """method : VOID NAME LPAREN parameters RPAREN ordinal SEMICOLON"""
+    """method : VOID NAME LPAREN parameters RPAREN ordinal SEMI"""
     p[0] = ('METHOD', p[2], p[4], p[6])
 
   def p_parameters(self, p):
@@ -240,13 +140,13 @@ class Parser(object):
     p[0] = p[1]
 
   def p_specializedhandle(self, p):
-    """specializedhandle : HANDLE LANGLE specializedhandlename RANGLE"""
+    """specializedhandle : HANDLE LT specializedhandlename GT"""
     p[0] = "handle<" + p[3] + ">"
 
   def p_specializedhandlename(self, p):
-    """specializedhandlename : DATAPIPECONSUMER
-                             | DATAPIPEPRODUCER
-                             | MESSAGEPIPE"""
+    """specializedhandlename : DATA_PIPE_CONSUMER
+                             | DATA_PIPE_PRODUCER
+                             | MESSAGE_PIPE"""
     p[0] = p[1]
 
   def p_array(self, p):
@@ -260,7 +160,7 @@ class Parser(object):
       p[0] = p[1]
 
   def p_enum(self, p):
-    """enum : ENUM NAME LCURLY enum_fields RCURLY SEMICOLON"""
+    """enum : ENUM NAME LBRACE enum_fields RBRACE SEMI"""
     p[0] = ('ENUM', p[2], p[4])
 
   def p_enum_fields(self, p):
@@ -274,12 +174,84 @@ class Parser(object):
 
   def p_enum_field(self, p):
     """enum_field : NAME
-                  | NAME EQUALS NUMBER"""
+                  | NAME EQUALS expression"""
     if len(p) == 2:
       p[0] = ('ENUM_FIELD', p[1], None)
     else:
       p[0] = ('ENUM_FIELD', p[1], p[3])
 
+  ### Expressions ###
+
+  def p_expression(self, p):
+    """expression : conditional_expression"""
+    p[0] = p[1]
+
+  def p_conditional_expression(self, p):
+    """conditional_expression : binary_expression
+                              | binary_expression CONDOP expression COLON conditional_expression"""
+    # Just pass the arguments through. I don't think it's possible to preserve
+    # the spaces of the original, so just put a single space between them.
+    p[0] = ' '.join(p[1:])
+
+  # PLY lets us specify precedence of operators, but since we don't actually

[darin (slow to review), 2014/01/10 00:36:34]

Perhaps we should evaluate these expressions here?

See for example: http://www.dabeaz.com/ply/example.html

Is there a reason to prefer evaluating expressions downstream in the code
generator?

+  # evaluate them, we don't need that here.
+  def p_binary_expression(self, p):
+    """binary_expression : unary_expression
+                         | binary_expression binary_operator binary_expression"""
+    p[0] = ' '.join(p[1:])
+
+  def p_binary_operator(self, p):
+    """binary_operator : TIMES
+                       | DIVIDE
+                       | MOD
+                       | PLUS
+                       | MINUS
+                       | RSHIFT
+                       | LSHIFT
+                       | LT
+                       | LE
+                       | GE
+                       | GT
+                       | EQ
+                       | NE
+                       | AND
+                       | OR
+                       | XOR
+                       | LAND
+                       | LOR"""
+    p[0] = p[1]
+
+  def p_unary_expression(self, p):
+    """unary_expression : primary_expression
+                        | unary_operator expression"""
+    p[0] = ''.join(p[1:])
+
+  def p_unary_operator(self, p):
+    """unary_operator : TIMES
+                      | PLUS
+                      | MINUS
+                      | NOT
+                      | LNOT"""
+    p[0] = p[1]
+
+  def p_primary_expression(self, p):
+    """primary_expression : constant
+                          | NAME
+                          | LPAREN expression RPAREN"""
+    p[0] = ''.join(p[1:])
+
+  def p_constant(self, p):
+    """constant : INT_CONST_DEC
+                | INT_CONST_OCT
+                | INT_CONST_HEX
+                | FLOAT_CONST
+                | HEX_FLOAT_CONST
+                | CHAR_CONST
+                | WCHAR_CONST
+                | STRING_LITERAL
+                | WSTRING_LITERAL"""
+    p[0] = ''.join(p[1:])
+
   def p_error(self, e):
     print('error: %s'%e)