Antlr grammar: declaring the grammatical context

I've been looking at:

http://www.antlr.org/wiki/display/~admin/2008/04/11/Rewrite+rules

There is the section on grammatical context that says:

[classDef var]: // context is "field"
  type ID -> blort

[... method var]: // context is "local"
  type ID -> blech

which as I understand is a way for you to look at the past non-terminals that derived the current non-terminal. I've tried to use this in my grammar but I don't know the correct Antlr syntax for doing so. I've been trying to do something like this in my tree grammar:

enclosed returns [String s]
 :  
        [andCond enclosed*]:
 ^(ENCLOSED andCond)
 {$s = $andCond.s; }

 |

 [orCond enclosed*]:
 ^(ENCLOSED orCond) 
 {$s = $orCond.s; }

 |

 ^(ENCLOSED enclosed) // return whatever is enclosed
 {$s = $enclosed.s; }

 |

 ^(ENCLOSED condition)  // last resort, have to keep parenthesis
 {$s = "(" + $condition.s + ")"; } 
 ;

where enclosed is the nonterminal that matches AST nodes that represents an expression that's nested inside parenthesis. What I try to accomplish is the removal of the useless parentheses in the string produced.

For instance, if user puts in (((a==b))) I expect to get the string a==b after I run it through the parser and the tree walker.

ansaurus

tags:

views:

answers:

Antlr grammar: declaring the grammatical context

related questions