[antlr-interest] Specifying limit on number of digits in the Grammar in ANTLR

[Alexandre Porcelli]

This problem is well described in the new Ter book (page 292, in my
last beta version).

In fact there is 3 ways to do it:

1) Simple context-free grammar

ID: DIGIT DIGIT DIGIT DIGIT DIGIT
| DIGIT DIGIT DIGIT DIGIT
| DIGIT DIGIT DIGIT
| DIGIT DIGIT
| DIGIT
;

2) Verify the size using actions

ID: ( d+=DIGIT)+ {if ( $d.size() > 5 ) <<error action>> }?
;


3) Using validating semantic predicate

ID: ( d+=DIGIT)+ {$d.size() <= 5}?
;


What is the best solution? It always depends (on my point of view) but
if sometimes you will need to change your grammar to accept 7 instead
of 5 digits… I would choose the third one.

Alexandre Porcelli


On 4/26/07, Bharath R <rbharath24 at yahoo.com> wrote:
> Hi,
>
>       I'm a newbie to ANTLR.
>         I am trying to write a simple parser, that accepts number whose
> number of digits vary from 1 to 5.
>
>     Below is the ABNF format
>  ID: 1*5(DIGIT)
>  DIGIT: '0'..'9'
>
>  i.e the ID can be any number varying from a 1 digit number to a max of
> 5digit number. Each digit varying between 0 and 9.
>  How do I achieve this in ANTLR? I have been trying to write an equivalent
> EBNF format but without success.
>  Thanks for you help.
>
> Regards,
> Bharath
>
>
>  ________________________________
> Ahhh...imagining that irresistible "new car" smell?
>  Check out new cars at Yahoo! Autos.
>
>