Boost.Spirit.X3 how to prevent token from being parsed by previous rule?

Question

given this grammar :

const auto grammar_def = x3::lit("start")
                        > x3::lit("{")
                        > (*(char_("a-zA-Z0-9\".{}=_~")))
                        > x3::lit("}") > x3::lit(";");

the last x3::lit("}") is being consumed by (*(char_("a-zA-Z0-9\".{}=_~"))) since it contains "}"
is there way to prevent that ? something like giving x3::lit("}") higher priority during parse?

score 1 · Answer 1 · answered Apr 09 '20 at 09:05

1

The only way I found was to disallow the kleene star from matching a '}' if a ; came after.

const auto grammar_def = x3::lit("start")
                        > x3::lit("{")
                        > *(char_("a-zA-Z0-9\".{}=_~") >> !lit(';'))
                        > x3::lit("}") > x3::lit(";");

answered Apr 09 '20 at 09:05

christianO

167
11

it works for this case, but I'm looking for something more generic, e.g : assuming ";" is also in the expression . – Abdullah Alattar Apr 10 '20 at 11:16

score 1 · Accepted Answer · answered Apr 14 '20 at 14:34

You can always write a custom parser that works like kleene star but can backtrack and also takes a second parser that must match as well. Something like this should work.

        template <class Left, class Right>
        struct backtracking : binary_parser<Left, Right, backtracking<Left, Right>>
        {
            using base = binary_parser<Left, Right, backtracking<Left, Right>>;
            backtracking(Left const& left, Right const& right)
                :  base(left, right)
            {
            }

            template<typename It, typename Ctx, typename Other>
            bool parse(It& f, It l, Ctx const& ctx, Other const& other, unused_type) const
            {
                auto end_it = l;
                while (end_it != f) {
                    auto save = f;
                    if (this->left.parse(f, end_it, ctx, other, unused)) {
                        if (this->right.parse(f, l, ctx, other, unused))
                            return true;
                    }
                    f = save;
                    --end_it;
                }
                return false;
            }
        };

const auto grammar_def = x3::lit("start")
                        > x3::lit("{")
                        > backtracking(
                             *(char_("a-zA-Z0-9\".{}=_~")), lit("}") >> lit(";"));

Boost.Spirit.X3 how to prevent token from being parsed by previous rule?

2 Answers2