用 boost 精神解析 boost uuid

Question

我正在尝试使用boost::spirit::qi 编写一个boost::uuids::uuid 解析器，以便以一种很好的方式与其他qi 解析器一起使用它并拥有一个很好的统一解析器api。

我的第一个想法是编写一个自定义的qi::grammar，它将使用boost::conversion::try_lexical_convert<boost::uuids::uuid &, const std::string &>，但是这将存在将开始迭代器正确设置为使用位置的问题，因为boost::conversion::try_lexical_convert<boost::uuids::uuid &, const std::string &>不仅会匹配16个字符长的输入但也可以带括号或不带破折号。

我的第二种方法是使用boost::spirit::qi::rule（如果愿意，也可以使用从boost::spirit::qi::grammar::base_type 派生的语法CRTP），但随后出现编译错误，可能来自BOOST_FUSION_ADAPT_STRUCT 表达式：

    #include <iostream>
    #include <string>
    #include <cstdint>
    #include <boost/uuid/uuid.hpp>
    #include <boost/spirit/include/qi.hpp>


    BOOST_FUSION_ADAPT_STRUCT(
            boost::uuids::uuid,
            (uint8_t, data[0])
            (uint8_t, data[1])
            (uint8_t, data[2])
            (uint8_t, data[3])
            (uint8_t, data[4])
            (uint8_t, data[5])
            (uint8_t, data[6])
            (uint8_t, data[7])
            (uint8_t, data[8])
            (uint8_t, data[9])
            (uint8_t, data[10])
            (uint8_t, data[11])
            (uint8_t, data[12])
            (uint8_t, data[13])
            (uint8_t, data[14])
            (uint8_t, data[15])
    )

    template<typename Iterator>
    boost::spirit::qi::rule<Iterator, boost::uuids::uuid>
            uuid_internal_{
            boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                    //time-low
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                    >> -boost::spirit::qi::lit("-")
                    //time-mid
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                    >> -boost::spirit::qi::lit("-")
                    //time-high-and-version
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                    >> -boost::spirit::qi::lit("-")
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>() //clock-seq-and-reserved
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>() //clock-seq-low
                    >> -boost::spirit::qi::lit("-")
                    //node
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                    >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
    };

    template<typename Iterator>
    struct uuid_
            : ::boost::spirit::qi::grammar<Iterator, boost::uuids::uuid()>{
        uuid_() : uuid_::base_type(start) {

            start %= (boost::spirit::qi::lit("{") >> uuid_internal_ >> boost::spirit::qi::lit("}")) |
                     uuid_internal_ ;
        }

        boost::spirit::qi::rule<Iterator, boost::uuids::uuid()> start;

        boost::spirit::qi::rule<Iterator, boost::uuids::uuid()>
                uuid_internal_{
                boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                        //time-low
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                        >> -boost::spirit::qi::lit("-")
                        //time-mid
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                        >> -boost::spirit::qi::lit("-")
                        //time-high-and-version
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                        >> -boost::spirit::qi::lit("-")
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>() //clock-seq-and-reserved
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>() //clock-seq-low
                        >> -boost::spirit::qi::lit("-")
                        //node
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
                        >> boost::spirit::qi::uint_parser<uint8_t, 16, 1, 1>()
        };

    };

    int main() {
        std::string input;
        std::cin >> input;
        uuid_<std::string::const_iterator> uuid_{};
        boost::uuids::uuid uuid{};
        auto begin = input.begin(), end = input.end();

        const bool success = boost::spirit::qi::parse(begin, end, uuid_, uuid);
        if (!success || begin != end)
            throw std::runtime_error("Parsing failed");

        return 0;


    }

/opt/local/include/boost/spirit/home/support/container.hpp:292:15：错误：'boost::uuids::uuid'中没有名为'insert'的成员 c.insert(c.end(), val);

似乎是由boost::spirit::qi::detail::pass_through_container 生成的问题，但是我将BOOST_FUSION_ADAPT_ADT 与*(obj.begin()+n) 结合使用的方法也因多个错误而失败。

Answer 1

您可以使用内置的 qi::stream 指令来获得 90% 的路径：

uuid_ = qi::stream;
start = '{' >> uuid_ >> '}' | uuid_;

看Live On Coliru

#include <boost/spirit/include/qi.hpp>
#include <boost/uuid/uuid_io.hpp>
#include <iomanip>

namespace qi = boost::spirit::qi;

template <typename Iterator> struct uuid_type : ::qi::grammar<Iterator, boost::uuids::uuid()> {
    uuid_type() : uuid_type::base_type(start) {

        start = '{' >> uuid_ >> '}' | uuid_;
        uuid_ = qi::stream;
    }
  private:
    qi::rule<Iterator, boost::uuids::uuid()> start, uuid_;
};

int main() {
    uuid_type<std::string::const_iterator> uuid_{};

    for (std::string const input : {
            "2bc69ead-4aba-4a39-92c0-9565f4d464b4",
            "2BC69EAD-4ABA-4A39-92C0-9565F4D464B4",
            "{2bc69ead-4aba-4a39-92c0-9565f4d464b4}",
            "{2BC69EAD-4ABA-4A39-92C0-9565F4D464B4}",
            //"{2bc69ead--4aba--4a39----92c0--9565f4d464b4}",
            //"{2BC69EAD--4ABA--4A39----92C0--9565F4D464B4}",
            })
    {
        boost::uuids::uuid uuid{};

        std::cout << "==== Input " << std::quoted(input) << "\n";

        if (qi::parse(input.begin(), input.end(), uuid_ >> qi::eoi, uuid))
            std::cout << "Parsed " << uuid << "\n";
        else
            std::cout << "Parsing failed\n";
    }
}

打印

==== Input "2bc69ead-4aba-4a39-92c0-9565f4d464b4"
Parsed 2bc69ead-4aba-4a39-92c0-9565f4d464b4
==== Input "2BC69EAD-4ABA-4A39-92C0-9565F4D464B4"
Parsed 2bc69ead-4aba-4a39-92c0-9565f4d464b4
==== Input "{2bc69ead-4aba-4a39-92c0-9565f4d464b4}"
Parsed 2bc69ead-4aba-4a39-92c0-9565f4d464b4
==== Input "{2BC69EAD-4ABA-4A39-92C0-9565F4D464B4}"
Parsed 2bc69ead-4aba-4a39-92c0-9565f4d464b4

剩下的 10%

根据modified Pareto principle，剩下的 10% 是困难的部分。

我什至不确定你是否想要这个，但+qi::lit("-") 暗示评论的测试用例也应该被接受（？！）：

        //"{2bc69ead--4aba--4a39----92c0--9565f4d464b4}",
        //"{2BC69EAD--4ABA--4A39----92C0--9565F4D464B4}",

如果那是 /really/ 你想要的，我确实建议使用词法转换实现进行两阶段解析操作：

好的，因为现在已经过了一个多小时，这意味着它更像是“10% 的功能需要 900% 的努力”——我希望你真的想要它 :)

Live On Coliru

#include <boost/spirit/include/qi.hpp>
#include <boost/lexical_cast.hpp>
#include <boost/uuid/uuid_io.hpp>
#include <iomanip>

using Uuid = boost::uuids::uuid;

namespace boost::spirit::traits {
    template <> struct is_container<Uuid> : mpl::false_ {};

    template <> struct assign_to_attribute_from_value<Uuid, std::string> {
        static void call(std::string const& s, Uuid& v) { v = lexical_cast<Uuid>(s); }
    };
}

namespace qi = boost::spirit::qi;

template <typename Iterator> struct uuid_type : qi::grammar<Iterator, Uuid()> {
    uuid_type() : uuid_type::base_type(start) {
        using namespace qi;

        auto sep_  = copy(+lit('-') >> qi::attr('-'));
        auto hex2_ = copy(xdigit >> xdigit >> xdigit >> xdigit);
        auto hex4_ = copy(hex2_ >> hex2_);
        auto hex6_ = copy(hex4_ >> hex2_);
        auto fmt_  = copy(
            hex4_ >> sep_ >> hex2_ >> sep_ >> hex2_ >> sep_ >> hex2_ >> sep_ >> hex6_
        );

        start = as_string['{' >> fmt_ >> '}' | fmt_];
    }
  private:
    qi::rule<Iterator, Uuid()> start;
};

int main() {
    uuid_type<std::string::const_iterator> uuid_{};

    for (std::string const input : {
            "2bc69ead-4aba-4a39-92c0-9565f4d464b4",
            "2BC69EAD-4ABA-4A39-92C0-9565F4D464B4",
            "{2bc69ead-4aba-4a39-92c0-9565f4d464b4}",
            "{2BC69EAD-4ABA-4A39-92C0-9565F4D464B4}",
            "{2bc69ead--4aba--4a39----92c0--9565f4d464b4}",
            "{2BC69EAD--4ABA--4A39----92C0--9565F4D464B4}",
            })
    {
        Uuid uuid{};

        std::cout << "==== Input " << std::quoted(input) << "\n";

        auto f = input.begin(), l = input.end();
        if (qi::parse(f, l, uuid_ >> qi::eoi, uuid))
            std::cout << "Parsed " << uuid << "\n";
        else
            std::cout << "Parsing failed\n";
    }
}

打印

==== Input "2bc69ead-4aba-4a39-92c0-9565f4d464b4"
Parsed 2bc69ead-4aba-4a39-92c0-9565f4d464b4
==== Input "2BC69EAD-4ABA-4A39-92C0-9565F4D464B4"
Parsed 2bc69ead-4aba-4a39-92c0-9565f4d464b4
==== Input "{2bc69ead-4aba-4a39-92c0-9565f4d464b4}"
Parsed 2bc69ead-4aba-4a39-92c0-9565f4d464b4
==== Input "{2BC69EAD-4ABA-4A39-92C0-9565F4D464B4}"
Parsed 2bc69ead-4aba-4a39-92c0-9565f4d464b4
==== Input "{2bc69ead--4aba--4a39----92c0--9565f4d464b4}"
Parsed 2bc69ead-4aba-4a39-92c0-9565f4d464b4
==== Input "{2BC69EAD--4ABA--4A39----92C0--9565F4D464B4}"
Parsed 2bc69ead-4aba-4a39-92c0-9565f4d464b4