The following will nicely do this:

class MyLexer extends Lexer;

options {
	charVocabulary = '3'..'377';
	k=3;
}

SYMBOL 
  : "te" 
  | "test" 
  ;

Note the k=3 - it is necessary so that the lexer can look ahead more than the length of "te" and check whether the next character is an 's' or (in this case only a) 't'.

This will correctly parse

tetetetetesttete

into

te/te/te/te/test/te/te

However, with more complex grammars it may not be so easy.

A fundamentally different approach is to use "keywords" with a keyword lookup table. See the ANTLR documentation for doing this (you would have something like

IDENTIFIER
  : ('a'..'z')+ 
  ;

and then look up the string in a table.

Regards


Harald M.