[antlr-interest] Getting started with Antlr: Resolving a conflicting lexer rule

[Carl Crowder]

I believe using lookahead would work - options { k=2; } 

Carl

On 3 Dec 2008, 5:52 PM, "Howard Lewis Ship" <hlship at gmail.com> wrote:

I'm working on an improved property expression language for Tapestry
5.1. My first step was to recreate what T5.0 accepted in the past
(using an ad-hoc parser based on regular expressions).

lexer grammar PELexer;


@header
{
package propexp;
}



fragment LETTER
       :       ('a'..'z'|'A'..'Z');
fragment DIGIT
       :       '0'..'9';
fragment SIGN
       :       ('+'|'-')?;
LPAREN  :       '(';
RPAREN  :       ')';

LBRACKET
       :        '[';
RBRACKET:       ']';

LBRACE  :       '{';
RBRACE  :       '}';

IDENTIFIER
       :       LETTER (LETTER | DIGIT | '_')+;
DEREF   :        '.';
SAFEDEREF
       :       '?.';

RANGEOP :       '.' (options { greedy = true; } : '.');

INT     :       SIGN DIGIT+;
FLOAT   :       SIGN DIGIT?  '.' DIGIT+;


NULL    :       'null';
TRUE    :       'true';
FALSE   :       'false';
THIS    :       'this';

WS      :       (' '|'\t'|'\n'|'\r')+ { skip(); };


My problem is a conflict between FLOAT and RANGEOP in my parser grammar:

parser grammar PEParser;


options
{
 tokenVocab=PELexer;
 output=AST;
 ASTLabelType=CommonTree;
}

tokens
{
 INVOKE;
}


@header
{
package propexp;
}




start   :       expression^ EOF!;

expression
       :       term DEREF expression -> ^(DEREF term expression)
       |       term SAFEDEREF expression -> ^(SAFEDEREF term expression)
       |       term;

term    :       reservedliteral
       |       range
       |       INT
       |       FLOAT
       |       IDENTIFIER
       |       id=IDENTIFIER LPAREN RPAREN -> ^(INVOKE $id)
       ;

reservedliteral: NULL | TRUE | FALSE | THIS;

range   :       from=INT RANGEOP to=INT  -> ^(RANGEOP $from $to);


(Obviosly, I've been playing around with this, trying different
options, such as greedy).

Everything works fine except for ranges:

    input: 1..10
line 1:2 required (...)+ loop did not match anything at character '.'
      AST: .10


In other words, I can't get it to see that a RANGE (i.e. "1..10")
should take precedence over a FLOAT (i.e., "1.234") ... it gets as far
as the "1." in a FLOAT, then gags on the extra ".".

I'm at a loss as to fixing this; I've been working through the book
and the FAQ and haven't stumbled over a solution yet.

Thanks in advance for any help!


--
Howard M. Lewis Ship

Creator Apache Tapestry and Apache HiveMind

List: http://www.antlr.org/mailman/listinfo/antlr-interest
Unsubscribe: 
http://www.antlr.org/mailman/options/antlr-interest/your-email-address

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.antlr.org/pipermail/antlr-interest/attachments/20081203/610c7de8/attachment.html