parser test raises some minor improvements (remove unnecessary ValueNode arrays, etc

lib/coffee_script/grammar.y

@@ -169,10 +169,10 @@ rule
 
   # Expressions that can be treated as values.
   Value:
-    IDENTIFIER                        { result = ValueNode.new(val) }
-  | Array                             { result = ValueNode.new(val) }
-  | Object                            { result = ValueNode.new(val) }
-  | Parenthetical                     { result = ValueNode.new(val) }
+    IDENTIFIER                        { result = ValueNode.new(val[0]) }
+  | Array                             { result = ValueNode.new(val[0]) }
+  | Object                            { result = ValueNode.new(val[0]) }
+  | Parenthetical                     { result = ValueNode.new(val[0]) }
   | Value Accessor                    { result = val[0] << val[1] }
   | Invocation Accessor               { result = ValueNode.new(val[0], [val[1]]) }
   ;

lib/coffee_script/lexer.rb

@@ -157,8 +157,9 @@ module CoffeeScript
       index = 0
       loop do
         tok = @tokens[index -= 1]
-        return if !tok || tok[0] != :IDENTIFIER
+        return if !tok
         next if tok[0] == ','
+        return if tok[0] != :IDENTIFIER
         tok[0] = :PARAM
       end
     end

lib/coffee_script/nodes.rb

@@ -148,7 +148,7 @@ module CoffeeScript
 
   # A value, indexed or dotted into or vanilla.
   class ValueNode < Node
-    attr_reader :last
+    attr_reader :name, :properties, :last
 
     def initialize(name, properties=[])
       @name, @properties = name, properties
@@ -212,6 +212,8 @@ module CoffeeScript
     statement
     custom_return
 
+    attr_reader :variable, :value, :context
+
     def initialize(variable, value, context=nil)
       @variable, @value, @context = variable, value, context
     end