Need to use multiple parser rules?

[rempAut]

I have an identifier described by:
identifier ::= (letter | (’’ (letter | digit))) {[’’] (letter | digit)}

• can begin with letter or underscore followed by letter or digit
• continues with letter or digit, and allows sequences of only single underscore in the identifier

This grammar is working:

grammar TestRule

entry Model:
    (ident=Identifier);

Identifier:
    value= (Identifier_underscore_prefix | Letter) ((Underscore)? (Digit | Letter))*;

Identifier_underscore_prefix:
    value= Underscore (Digit | Letter);

hidden terminal WS: /\s+/;
terminal Digit: /[0-9]/;
terminal Letter: /[a-zA-Z]/;
terminal Underscore: /[_]/;
terminal ID: /[_a-zA-Z][\w_]*/;

hidden terminal ML_COMMENT: /\/\*[\s\S]*?\*\//;
hidden terminal SL_COMMENT: /\/\/[^\n\r]*/;

But when I try to insert the Identifier_underscore_prefix expression into the Identifier rule directly, I got an error. Is this supposed to happen? Is there some limit to nesting the rules expressions?

Rule does not work example :

value= ((Underscore (Digit | Letter))| Letter) ((Underscore)? (Digit | Letter))*;

[msujew]

Hey @rempAut,

even though we use EBNF syntax for most of our grammar DSL, the assignment syntax is a bit more limited, and doesn't allow for arbitrary amount of tokens in there. In your case, it looks as though your Identifier rule is actually supposed to be a data type rule. This change would address your problem:

entry Model:
    (ident=Identifier);

Identifier returns string:
    (Underscore | Letter) (Underscore | Digit | Letter)*;

I've simplified the Identifier rule, but it still parses the same input. Please note that your ID terminal at the bottom of your grammar is already able to parse the input that would be parsed by the Identifier rule.

[rempAut]

Thanks, it works when I use the data type rule.

However your rule and the ID terminal is not the same as my Identifier, because it parses
this_
th__is
etc...

But its not supposed to parse it.

[msujew]

Ah, I see, makes sense 👍