What is the equivalent of Peg.js labels ?

[loloof64]

Hi ! What is the equivalent of peg.js labels ?

Let's say for example I want to convert the following Peg.js rule

additive
  = left:multiplicative "+" right:additive { return left + right; }
  / multiplicative

This rule uses two labels in the grammar : left and right. Then it uses them in the parser (return left+right;).

What is the best way to do it in PetitParser ?

I would like to use such a feature because in many rules I've got optional parts, and so it would make my task easier.

Also I'm using the same structure as in the examples (prolog, json)

• a grammar file
• a parser file
• later I plan to set up an evaluator file

Regards

[renggli]

PetitParser doesn't use code generation (like peg.js). This would enable things like labels, but also comes at the cost of introducing another language and making it impossible to arbitrarily mix Dart and parser code. This might change hopefully soon when Dart gets support for static metaprogramming.

In the meantime a sequence of parsers generates a list of parse results that you can access by index, i.e.

Parser additive() => (ref0(multiplicative) & char("+") & ref0(additive)).map((values) => values[0] + values[2])
     / ref(multiplicative);

If this is a common pattern, I recommend to extract it to a helper method, i.e.

Parser term(Parser left, Parser op, Parser right, int Function(int left, int right) callback) {
  return (left & op & right).map((values) => callback(values[0], values[2])) / ref(left);
}

Parser additive() => term(ref0(multiplicative), char("+"), ref0(additive), (left, right) => values[0] + values[2]);

Also note that for expressions there is an expression-builder that does exactly this out of the box:

...
builder.group()
   ..left(char('+').trim(), (a, op, b) => a + b)
...

[loloof64]

Thank you very much for your answer.

In fact, in the grammar I'm writing, I've written the following rule :

import 'package:petitparser/petitparser.dart';

class PgnGrammarDefinition extends GrammarDefinition {

    Parser game() =>
      ref0(tags).optional() & ref0(comments).optional() & ref0(pgn);

}

And in the parser file, I've written something like this

import 'package:petitparser/petitparser.dart';
import 'grammar.dart';

class PgnParserDefinition extends PgnGrammarDefinition {
  var _messages = [];

   @override
   Parser game() => super.game().map((each) {
      var message = _messages; _messages = [];
      return {
        'tags': each[0],
        'gameComment': each[1],
        'moves': each[2],
        'messages': message,
     };
  });
}

So is it warranted that in the overriden game() function:

• each[0] matches the ref0(tags) even if not providen as it is optionnal ?
• each[1] matches the ref0(comments) even if not providen ?
• each[2] matches the ref0(pgn) ?
  So that there is no need to make use condtionnal, which for sure I'll code this in an ugly way ?

Regards

[renggli]

Yes, that's correct. The optional parsers get a fixed entry in the resulting list. By default the resulting absent value will be null, but there is also the option to specify any other default value, see the documentation.

[loloof64]

Thank you very much 😃