[antlr-interest] "unexpected token" error

Monty Zukowski monty at codetransform.com
Tue Mar 8 09:35:48 PST 2005


togol machillan wrote:
> Hi,
> 
> The whitespace and the newline rules are as follows.
> 
> NEWLINE 	: ('\r''\n')=> '\r''\n' //DOS
> 		| '\r' //MAC
> 		| '\n' //UNIX
> 		{ newline(); 
> 		}
> 		;
> 
> WS : (' '|'\t'|'\f') { _ttype = ANTLR_USE_NAMESPACE(antlr)Token::SKIP; } ;
> 
> Regards,
> 
> Mach
> ----- Original Message -----
> From: "Monty Zukowski" <monty at codetransform.com>
> To: antlr-interest at antlr.org
> Subject: Re: [antlr-interest] "unexpected token" error
> Date: Tue, 08 Mar 2005 10:03:28 -0500
> 
> 
>>togol machillan wrote:
>>
>>>Hi,
>>>
>>>Another newbie question. I hope somebody can help me out here.
>>>
>>>I have the following lexer rules (omitting some which are clear by name)
>>>
>>>VALUE  options {testLiterals = true; }
>>>       : CAPITAL_LETTER (CAPITAL_LETTER|DIGIT|','|'-')*
>>>       ;
>>>
>>>FIGURE: ('-')? ((DIGIT)+ '.' (DIGIT)* | '.'(DIGIT)+ );
>>>
>>>ASSIGNMENT: VALUE '=' (VALUE|FIGURE)
>>>          ;
>>>
>>>I want the ASSIGNMENT rule to match a token like MOLE-FRAC=1.0 or 
>>>STOIC=IN, for example. But when I include this rule, the program 
>>>throws the unexpected token: error at the point where first space 
>>>character occurs in the input file. Is this due to some problem 
>>>with the greedy nature of the the VALUE rule? Thanks in advance.
>>>
>>>Regards,
>>>
>>>Togol Mach
>>
>>What does your whitespace rule look like?
>>
>>Monty
> 
> 

It's still not enough information to diagnose.  Best thing is the 
minimal grammar & input to reproduce, plus the actual error message.

Monty


More information about the antlr-interest mailing list