[antlr-interest] Re: High level semantic analysis

Matt Benson gudnabrsam at yahoo.com
Wed May 21 07:24:55 PDT 2003


Or one person could just write such an AST and make it
publicly available as an extension to Antlr...


--- "Tiller, Michael (M.M.)" <mtiller at ford.com> wrote:
> > From: mzukowski at yci.com [mailto:mzukowski at yci.com]
> > Subject: RE: [antlr-interest] Re: High level
> semantic analysis
> > 
> > TNode.java--has support for double linking (so a
> node can 
> > find its parent)
> > which is currently unused.  To use it you would
> first build 
> > the complete
> > tree and then call doubleLink().  TNode Also has a
> slot for 
> > the defNode, the
> > idea being to have a tree pass which annotates
> every use of a 
> > variable with
> > the definition to make it easy to find the
> definition without 
> > having to keep
> > track of the current scope.  It also has a
> Hashtable of 
> > attributes so you
> > could put anything you want in for the node.  That
> is how you 
> > decorate trees
> > in ANTLR.
> 
> For ANTLR 3, it would be nice (for people like me)
> if a "rich" AST node type was included that
> implemented references to parent nodes and some kind
> of hash or map structure for annotating the AST with
> additional information.  I'm sure I can figure out
> how to write my own, but it seems like this would be
> a really common need and having a built-in type
> would just spare everyone the exercise of
> implementing such a node.
> 
> --
> Mike
> 
>  
> 
> Your use of Yahoo! Groups is subject to
> http://docs.yahoo.com/info/terms/ 
> 
> 


__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ 




More information about the antlr-interest mailing list