[antlr-interest] Extracting meaning out of Antlr's exceptions

[rmsanjivp <rmsanjivp at yahoo.com>]

Hi!

I am trying to make out a meaningful error message from the 
exception thrown by the Antlr. I am catching the exception in the 
grammar rule and trying to make out the meaning. Consider the 
following:

expr: m1:mult_expr ( (p:PLUS | m:MINUS) m2:mult_expr)*;
    exception
     catch[TokenStreamException ex]
     {
       // will this section catch unrecognized symbol from the lexer
     }
     catch[RecognitionException ex]
     {
       // I want to make a decision if the operator is missing or
       // operand is missing. Will the following code help??
       if( p == NULL && m == NULL)
         printf("Operator Missing\n")
       else if( (p != NULL || m != NULL) && #m2 == NULL)
         printf("OPerand Missing\n");
     }



Similar kind of exception handling for mult_expr rule.

I want to know how to make what is the real problem. I can have an 
input that has unrecognized symbol by lexer. How do I catch that in 
the exception handler in the grammar? I have gone through the list 
of ecception thrown on different problems. I found that 
RecognitionException handles both unrecognized input from the lexer 
and unexpected token from the input. How do I differentiate between 
them? Is there a way to catch these exception and generate really 
meaningfule messages??

Is the way I catch exception in the rule fine?

Kindly advise.

Thanks,
Sanjiv



 

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