[antlr-interest] Someting about range (to change)

[Fabien Hermenier]

Hi Bart

Your fix works fine. So the problem was in reality due to the fact that 
LEFTY and DEC_NUMBER can start with the same character.

Thanks,
Fabien
Le 29/06/11 05:32, Bart Kiers a écrit :
> Hi Fabien,
>
> Note that the fact that `..` (let's call it DOT_DOT) does not appear 
> in a lexer rule, there _is_ a token being created because you defined 
> DOT_DOT inside a parser rule.
>
> The lexer has a problem with input like:
>
>   123..
>
> The lexer tries to construct a LEFTY from it instead of a DEC_NUMBER 
> followed by a DOT_DOT.
>
> Better let LEFTY _not_ possibly start with a Digit:
>
> LEFTY : Letter ('.' | Letter | Digit | '-')* '[';
>
> Bart.
>
>
> On Wed, Jun 29, 2011 at 7:53 AM, Fabien Hermenier 
> <hermenierfabien at gmail.com <mailto:hermenierfabien at gmail.com>> wrote:
>
>     Hi Bart
>
>     Here is. It's just a change in LEFTY:
>     ---
>
>     fragment Digit :'0'..'9';
>     DEC_NUMBER: '1'..'9' Digit*;
>
>     fragment Letter    :'a'..'z'|'A'..'Z';
>
>     WS    :    ('\n'|'\r'|'\t'|' ') {$channel=HIDDEN;};
>     LEFTY: (Letter|Digit) ('.'|Letter|Digit|'-')* '[';
>     RIGHTY: ']' (('-' (Letter|Digit))|Letter|Digit)* ;
>
>     number :DEC_NUMBER;
>
>     range: LEFTY number '..' number RIGHTY;
>     ---
>
>     Le 28/06/11 23:51, Bart Kiers a écrit :
>>     Hi Fabien,
>>
>>     Can you post the grammar that produces the error?
>>
>>     Bart.
>>
>>
>>     On Wed, Jun 29, 2011 at 5:57 AM, Fabien Hermenier
>>     <hermenierfabien at gmail.com <mailto:hermenierfabien at gmail.com>> wrote:
>>
>>         Hi
>>
>>         I am sorry, but this is another problem related to sequence
>>         of integer.
>>         I don't understand my error(s), despite severals
>>         pass on the wiki page related to this use case.
>>
>>         I want to parse a string with a "sequence" parameter into it,
>>         such as
>>         "toto-[2 ..3]-toto".
>>         Here is a short version of the grammar that works perfectly:
>>         ---
>>         grammar tryout;
>>
>>         options {
>>             k=3;
>>         }
>>         fragment Digit :'0'..'9';
>>         DEC_NUMBER: '1'..'9' Digit*;
>>
>>         fragment Letter    :'a'..'z'|'A'..'Z';
>>
>>         WS    :    ('\n'|'\r'|'\t'|' ') {$channel=HIDDEN;};
>>         LEFTY: (Letter|Digit) (Letter|Digit|'-')* '[';
>>         RIGHTY: ']' (('-' (Letter|Digit))|Letter|Digit)* ;
>>
>>
>>         number :DEC_NUMBER; //HEX_NUMBER, OCT_NUMBER, ... are
>>         following but
>>         removed for this example.
>>
>>         range: LEFTY number '..' number RIGHTY;
>>         ---
>>
>>         This grammar accepts "toto-[1..3]-toto" or "toto-[1 .. 3]-toto".
>>         Now, I want to be able to accept the LEFTY token or the
>>         RIGHTY token
>>         even if they contains a '.' inside (not at the beginning). So
>>         I've
>>         modified LEFTY as following:
>>         LEFTY: (Letter|Digit) ('.'|Letter|Digit|'-')* '[';
>>
>>         Now, ANTLR does no longer accept "toto-[1..3]-toto". It
>>         requires at
>>         least one space between the first number and the range.
>>         I have read the wiki page related to range, integer, and so
>>         one. But in
>>         my case, I don't see where my grammar is ambiguous as no token
>>         can start with a '.' . So it seems there is a concept I don't
>>         get. Can
>>         anyone try to help me ?
>>
>>         Thanks in advance,
>>         Fabien.
>>
>>         List: http://www.antlr.org/mailman/listinfo/antlr-interest
>>         Unsubscribe:
>>         http://www.antlr.org/mailman/options/antlr-interest/your-email-address
>>
>>
>
>
>     -- 
>     postdoctoral researcher at Flux
>     School of Computing, University of Utah
>     http://sites.google.com/site/hermenierfabien/home
>
>


-- 
postdoctoral researcher at Flux
School of Computing, University of Utah
http://sites.google.com/site/hermenierfabien/home