Re: Usefulness of automatic parse tree generation

> How difficult is it to upgrade an SLR table generator to an LALR
> table generator?


Difficult. There are pitfalls such as the NQLALR (not quite LALR)
problem.


> It may be that the user of the library needs to provide rules for
> the transformation.


Yes, the user provides node names, at least.


> If so, what would these rules look like? I'd love to hear from
> anybody that has experience in doing this and can offer sound
> guidance! >


I did it this way in a commercial PG I wrote. The AST turns out to be
what you want.


// Operator precedence goes here


// Grammar:


      Goal -> Stmt... <eof>


      Stmt -> Exp ';'
                            -> Target '=' Exp ';' +> assign


      Target -> <identifier> +> target


      Exp -> Primary
                            -> Exp '+' Exp +> add
                            -> Exp '-' Exp +> sub
                            -> Exp '*' Exp +> mul
                            -> Exp '/' Exp +> div


      Primary -> <identifier> +> ident
                                -> '(' Exp ')'
                                -> '+' Primary
                                -> '-' Primary +> neg
                                -> '@' Primary +> addr


// End.


Paul B Mann
paulbmann@yahoo.com

Return to the comp.compilers page.
Search the comp.compilers archives again.