[antlr-interest] Specifying literals in the Lexer

[Remi Koutcherawy]

Hi
> I really do want to specify one or two of my keywords in the lexer
> instead of the parser. This is because the keywords contain '[', ']'
> characters that are valid characters in my IDENTIFIER expression.

I had the almost same problem parsing Fortran to identify ".EQ." as a token and not
a wrong float number.

I had to declare it as a token AND as a rule in the lexer :

class LexF90 extends Lexer;
  k = 3;
  testLiterals=true;
  caseSensitive=false;
  caseSensitiveLiterals=false;
}
tokens {
EQ   = ".eq.";
}
EQ   : ".eq.";
NUM  : ('0'..'9')+ ('.' ('0'..'9')*)? (EXPONENT)?
     | '.' ('0'..'9')+ (EXPONENT)? ;
protected EXPONENT:('e') ('+'|'-')? ('0'..'9')+;

Rémi


 

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