[antlr-interest] How to intervene automatic keywords generati on

[mzukowski at yci.com]

Since you want to keep "error" out of the literals table, then treat it as
an ID:

on_error_stmt
   : "on" ID "goto" label
   ;

If you want to make sure that it is always "error" you can use a validating
semantic predicate:

on_error_stmt
   : "on" { "error".equals(LA(1).getText()) }? ID "goto" label
   ;

Monty
www.codetransform.com

> -----Original Message-----
> From: Fan Yang [mailto:yhhf_dy at yahoo.com]
> Sent: Wednesday, October 16, 2002 2:17 PM
> To: antlr-interest at yahoogroups.com
> Subject: [antlr-interest] How to intervene automatic keywords 
> generation
> 
> 
> Hi,
> 
> According to the manual of ANTLR about Keywords and literals: any 
> double-quoted string used in a parser is automatically entered into 
> the literals table of the associated lexer. For instance,
> 
> on_error_stmt 
>    : "on" "error" "goto" label)
>    ;
> 
> So "on", "error" and "goto" will be automatically putted into the 
> literals table. But in my target language "error" is not a keyword. 
> so "error" can be an "identifier" in a statement as following pseudo 
> code line shows. 
> 
>   printf("%s", e.error)
> 
> In fact, this line will cause a problem. because the parser will look 
> for the identifier -"error" in the literals table first, then 
> treat "error" as keyword not an general identifier. 
> 
> IDENT
>   options {testLiterals=true;}
>   :	('a'..'z'|'_') ('a'..'z'|'_'|'0'..'9')*	
>   ;
> 
> My question is that may I determine which one ,double-quoted strings, 
> can be entered into literals table instead of automatically 
> generatation. Or there is an alternative way to deal with this kind 
> of problem
> 
> Thanks,
> 
> Fan
> 
> 
> 
>  
> 
> Your use of Yahoo! Groups is subject to 
http://docs.yahoo.com/info/terms/ 


 

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