[antlr-interest] automatic token type per literal

[Matt Benson]

Thanks, Bryan!

-Matt

--- Bryan Ewbank <ewbank at gmail.com> wrote:

> ANTLR does all that you ask.  Look for
> "testLiterals" in the
> documentation.  It requires that you change your
> tokens section to be
> 
>   tokens {
>      FOO="foo";
>      ...
>   }
> 
> Also look at how to import and export token sets;
> very handy stuff.
> 
> Hope this is enough to get you started.  I've gotta
> run...
> - Bryan Ewbank
> 
> On 12/1/05, Matt Benson <gudnabrsam at yahoo.com>
> wrote:
> > Let's say I have a grammar having stolen the Java
> > IDENT lexer rule.  In my parser, I have the
> following:
> >
> > tokens { FOO; }
> >
> > somerule
> >     :    f:"foo"^ { #f.setType(FOO); } BAR
> >     ;
> >
> > Have I missed a feature by means of which I can
> (once)
> > permanently set the association between ttype FOO
> and
> > the literal "foo"?  If not, is there a rationale
> for
> > the absence of such a feature?  If the answer to
> THAT
> > is no, might such a feature find its way into
> ANTLR3?
> > ;)  I tried overriding the ttype in the IDENT
> lexer
> > rule but the ambiguity issues were (at my level of
> > expertise) insurmountable.  :)  I'm a longtime
> lurker
> > but still fairly green wrt actually making ANTLR
> do
> > anything...
> >
> > TIA,
> > Matt
> >
> >
> >
> >
> > __________________________________
> > Yahoo! Mail - PC Magazine Editors' Choice 2005
> > http://mail.yahoo.com
> >
> 



		
__________________________________ 
Yahoo! Music Unlimited 
Access over 1 million songs. Try it free. 
http://music.yahoo.com/unlimited/