Re: [Grammatica-users] Question about tokens

[Per Cederberg]

Well, generally what you want to do is impossible in a context-
insensitive grammar (like the ones Grammatica handles). However,
you can implement the context-sensitive parts of the grammar in
source code form inside your Analyzer subclass. If the grammar
looks like this:

  Reference = aLetter+ ;

You'd then write something like:

  protected Node exitReference(Production node)
      throws ParseException {

      if (node.getChildCount() > 3) {
          throw new ParseException(ParseException.ANALYSIS_ERROR,
                                   "reference code too long",
                                   node.getStartLine(),
                                   node.getStartColumn());
      }
      ... check that token children are identical ...
      return node;
  }

So, this takes away some of the nice code/grammar separation
that Grammatica attempts to uphold, but it also greatly reduces
the complexity of the grammar language itself.

Cheers,

/Per

On thu, 2005-03-17 at 10:20 -0700, Anant Mistry wrote:
> 
> Hopefully, this question is not going to be a stupid as the last one I
> posted :)
> 
> In my grammar definition I want to represent a part of a doc number
> that could be a sequence of the same letter (only A-Z),  but not more
> than three.
> 
> e.g.  A or AA or AAA or B or BB or BBB
> 
> So I have a token defined for every letter and then built a grammar
> like
> 
> a = "A"
> b = "B"
> c = "C"
> etc.
> 
> reference = a [a[a]] | b[b[b]] | c[c[c]] | .... etc.
> 
> There are other places I want to use the token 
> 
> aLetter = <<[A-Z]>>
> 
> Obviously, the two conflict. Is there a way I can define the
> requirement for multiples of the same letter just using a production
> of the aLetter token?
> 
> Thanks
> 
> Anant
> 
> -- 
> Anant Mistry
> Director of Infrastructure Technology
> Information Handling Services
> 15 Inverness Way East
> Englewood
> CO 80112
> USA
>  
> Confidentiality Notice: The information in this e-mail may be
> confidential
> and / or privileged. This e-mail is intended to be reviewed by only the
> individual or organization named in the e-mail address. If you are not
> the
> intended recipient, you are hereby notified that any review,
> dissemination
> or copying of this e-mail and attachments, if any, or the information
> contained herein, is strictly prohibited.
> _______________________________________________
> Grammatica-users mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/grammatica-users