[antlr-interest] my parser loops

[Jerome Pequery]

Hi everyone,

I have written a java parser using antlr 2.7.2, having for particularity 
to integrate comments into the AST while parsing. And i have a problem 
using class contents and what is between { and }:

> // This is the body of a class.  You can have fields and extra semicolons,
> // That's about it (until you see what a field is...)
> classBlock
>     : LCURLY!
>         ((j:commentDefinition)?
>             (    (field[#j])=>field[#j]
>             |     (semiOrSemiComment)
>                 {
>                     astFactory.addASTChild(currentAST, j_AST);
>                     break;
>                 }
>             )
>         )*
>     RCURLY!
>       {#classBlock = #([OBJBLOCK, "OBJBLOCK"], #classBlock);}
>     ;


I parse an optional comment (commentdefinition) and, if there is, parse 
a field (attribute or operation), passing it its association comment as 
parameter.

But now, if the comment close the class content, as in this example :

> public class GameStatusLabel extends JLabel implements GameListener {
>     /**
>      * commentaire
>      */
>     public void updateStatus(String message)
>     {
>         frame.addWindowListener(new WindowAdapter()
>         {
>         });
>     }
>
>     //public Dimension getMaximumSize() {
>     //Dimension d = getPreferredSize();
>     //d.width = Short.MAX_VALUE;
>     //return d;
>     //}
> }

There is no associated field for my comment. So the parser give me a 
syntax error on }.

Now, if I use sytactic predicates to check for this case and simply drop 
theses comments, so my rule is modified.

> classBlock
>     : LCURLY!
>         ((j:commentDefinition)?
>             (    (field[#j])=>field[#j]
>             |     (semiOrSemiComment)
>                 {
>                     astFactory.addASTChild(currentAST, j_AST);
>                     break;
>                 }
>             | (RCURLY)=> {
>                     astFactory.addASTChild(currentAST, j_AST);
>                     break;
>             }
>             )
>         )*
>     RCURLY!
>       {#classBlock = #([OBJBLOCK, "OBJBLOCK"], #classBlock);}
>     ;

The syntax error disappear, and my code is parsed. But anonymous classes 
parsing are now looping. In my upper java sample, the generated parser 
loop indefinitly with LA(1)=RCURLY and LA(2)=RPAREN.

How can I correct my grammar to support both comments  at end of a class 
and anonymous classes ?

Thanks in advance,

Jerome Pequery