[antlr-interest] unexpected char exception

[Loring Craymer]

Replace ABC by "abc" in startRule, and delete the ABC rule.

Any quoted string in the parser grammar (like "abc") is assigned a unique
token type.  However, there must be a rule that recognizes that string
(first message) and it must have the testLiterals = true (default case)
option set.  Then ANTLR automatically looks up these special case tokens in
a hash table to determine token type.  So IDENTIFIER matches "abc" and marks
the returned token as being of type LITERAL_abc (I think that I have the
prefix right).

--Loring

> -----Original Message-----
> From: antlr-interest-bounces at antlr.org [mailto:antlr-interest-
> bounces at antlr.org] On Behalf Of jean.morissette666 at videotron.ca
> Sent: Saturday, February 26, 2005 1:44 PM
> To: antlr-interest at antlr.org
> Subject: Re: [antlr-interest] unexpected char exception
> 
> Thank for your reply.
> 
> But now I have another kind of error.  In my grammar, rule ABC is a
> special
> case of IDENTIFIER and Antlr complain about it:
> warning:lexical nondeterminism between rules ABC and IDENTIFIER
> 
> Also, I have read this
> http://www.doc.ic.ac.uk/lab/secondyear/Antlr/options.html#_bb8
> but sadly I don't understand the purpose of the testLiterals option.
> 
> Any help would be appreciated.
> 
> 
> class P extends Parser;
> 
> startRule
>     :   ABC
>     ;
> 
> class L extends Lexer;
> 
> ABC
> 	:	"abc"
> 	;
> 
> IDENTIFIER
> 	:	( 'a'..'z' | 'A'..'Z' ) ( 'a'..'z' | 'A'..'Z' | '0'..'9' )*
> 	;
> 
> 
> Le 26 Février 2005 16:17, Loring Craymer a écrit :
> > You need at least a lexer rule which recognizes text; "abc" is a special
> > case for that rule.  Also look at the testLiterals option.