How to handle ambiguous token in grammar rule in ANTLR?

Question

I have the following grammar for parsing person name after normalizing it.

exp : fullName EOF;
fullName : title? f=name m=name? l=name;

title: TITLE;
name : NAME;

TITLE : 'mr'| 'mrs' | 'ms';
NAME : ('a'..'z')+;

WHITESPACE : ('\t' | ' ' | '\r' | '\n'| '\u0020' | '\u000C' )+ -> skip ;

When I parse a name like "mr john me smith" it works correctly but when one of the title tokens appears as a name like "mr john mr smith", I got the following error

line 1:8 extraneous input 'mr' expecting NAME
line 1:16 missing NAME at '<EOF>'
(exp (fullName (title mr) (name john) (name mr smith) (name <missing NAME>)) <EOF>)

Is there a way to use the token according to it position in the rule only and neglect it if it appeared in another location?

The lexer operates independently of the parser, so mr will always be parsed as a TITLE. See @SaraVF 's solution — Adrian Leonhard

vanilla vanilla · Accepted Answer · 2015-04-10T20:37:04

As long as the lexer will not be able to neglect it, the parser rule should be changed to

name : NAME | TITLE;

Modifying the lexer rule will not fix the problem and will generate another error.

How to handle ambiguous token in grammar rule in ANTLR?

2 Answers