[antlr-interest] lexer: embedded quotes assistance

[Edwards, Waverly]

I *think* I understand.  I will have to try it out, run it through the
debugger to ensure I understand how it works.  Its much more elegant
than my solution, which strangely enough doesn't even work.


W. 

-----Original Message-----
From: Diehl, Matthew J [mailto:matthew.j.diehl at intel.com] 
Sent: Thursday, August 23, 2007 12:33 PM
To: Edwards, Waverly; antlr
Subject: RE: [antlr-interest] lexer: embedded quotes assistance

> History...
> I am replicating an existing language that uses embedded quotes to 
> indicate a quote character.

I'll show you something I wrote that worked...but there's probably some other simpler ways using '.'
My quoted text can only use graphic characters (a-z 0-9 SpaceCharacters, and other graphic characters (\u00a1 - \u00ff).

StringLiteral
  :Quote GraphicCharacter* Quote
  ;
fragment
GraphicCharacter
  : UpperCaseLetter | LowerCaseLetter | Digit | SpecialCharacter | SpaceCharacter | (Quote Quote)
  ;

But since it matches double quotes, it will continue when it sees one

> In addition to embedded quotes
> the quoted material may span multiple lines by using the ¬ character 
> followed by a CR.  Now a rule for treating ¬CR as whitespace is not an 
> issue as I just create a rule

I'm sure if you include the CR rule above, it would work as well

Ok,

Matt