[Versa] Updated ebnf with no conflicts

[Jeremy Kloth]

On Monday 26 September 2005 10:13 am, Michael Olson wrote:
> On Sep 23, 2005, at 3:11 PM, Jeremy Kloth wrote:
> > One question I've been meaning to ask, why is there a need to make all
> > single
> > character tokens a named token.  In the grammar section, just use
> > <symbol>'c'</symbol> where c is the single character.  It would make
> > the
> > resulting EBNF much easier to grok.
>
> I didn't know I could do that is the short answer.  I'll work on
> updating it today.
>
> Does it work for multiple characters?  ie
>
> <symbol>'<='</symbol>
>

Not yet, but I really want to navigate towards eliminating the need for a 
separate lexer section, that is, all patterns are defined inline.  To follow 
bison's lead, that would be a symbol of "[a-z]+" would automatically be 
translated into an anonymous (well, not named) token behind the scenes.  
However its not there yet.

-- 
Jeremy Kloth
Fourthought, Inc.
http://fourthought.com/
http://4suite.org/