[antlr-interest] how to put lexer rules syntactic check

[Ivo Kondapaneni]

Hi everyone,

during creation of lexer for our language I have run across one trouble. My
lexer rules looks as follows:

Special_PERIOD    :    '.';
Special_DOUBLE_PERIOD    :    '..';

//
fragment DIGIT
    :     '0'..'9';
//
fragment FIXEDPOINTNUMBER
    :    INTEGERNUMBER Special_PERIOD INTEGERNUMBER;

fragment INTEGERNUMBER
    :     DIGIT ('_'? DIGIT)*;
//
Literal_INTEGER
    :    INTEGERNUMBER;

Literal_REAL
    :    FIXEDPOINTNUMBER ( 'e' ('+'|'-')? INTEGERNUMBER )?;

and lexer generated from these throws MismatchedRange exception during
parsing of input like:

"1234.." or "1234. something",

because it automatically choose to start with Literal_REAL instead of
Literal_INTEGER and it fails when it comes to the second period (or anything
which differs from a digit). I have tried to put there an syntactic
predicate to real and integer rules, but that obviously didn't worked,
because there is always only one possibility how to expand.

To summarize, I would like to parse string "1234.." to "Literal_INTEGER
Special_DOUBLE_PERIOD" tokens

So, have anyone an idea, how to solve it in lexer definition?

Thanks in advance.

-- 
Ivo Kondapaneni
programátor
Energocentrum Plus, s.r.o.

Mobile: +420 739 074 964

http://energocentrum.cz, http://rcware.eu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.antlr.org/pipermail/antlr-interest/attachments/20090724/1cf71711/attachment.html