Question

如何修改此boost :: spirit代码以解析范围或列表的任何组合？

请注意，这个问题与我的previous question有些不同。在这种情况下，我没有像RANGE:和LIST:之类的特殊预标记来帮助我解析，所以我不确定我们是否需要在这里进行某种预测。

我确实需要将解析分开，因为这样做会帮助我将结果捕获到不同的数据结构中。

目标让所有四个测试用例通过

编辑答案似乎适用于更复杂的表达

// #define BOOST_SPIRIT_DEBUG

#include <boost/foreach.hpp>
#include <boost/spirit/include/qi.hpp>
#include <boost/fusion/include/tuple.hpp>
#include <boost/tuple/tuple.hpp>
#include <boost/fusion/include/boost_tuple.hpp>

#include <iostream>

namespace qi = boost::spirit::qi;

typedef std::vector<boost::tuple<std::string,std::vector<int>>>   MY_TYPE;

template <typename Iterator, typename Skipper>
struct my_grammar :
qi::grammar<Iterator, MY_TYPE(), Skipper>
{
    my_grammar() :
        my_grammar::base_type( entries )
    {
        entries %=
            *(
                +(qi::char_ - '-')
                >> qi::lit( "->" )
                >> ( comma | range )
            )
            ;

        range %= '{' >> qi::int_ >> ':' >> qi::int_ >> ':' >> qi::int_ >> '}';
        comma %= '{' >> (qi::int_ % ',') >> '}';

        BOOST_SPIRIT_DEBUG_NODE( range   );
        BOOST_SPIRIT_DEBUG_NODE( comma   );
        BOOST_SPIRIT_DEBUG_NODE( entries );
    }

    qi::rule<Iterator, std::vector<int>(), Skipper> 
        range, comma;

    qi::rule<Iterator, MY_TYPE(), Skipper> 
        entries;
};

// -----------------------------------------------------------------------------

static void TryParse( const std::string& input, const std::string& label )
{
    MY_TYPE entries;
    auto it(input.cbegin()), end( input.cend() );
    my_grammar<decltype(it), qi::space_type> entry_grammar;

    if (qi::phrase_parse(it, end, entry_grammar, qi::space, entries)
            && it == end)
    {
        std::cerr << label << " SUCCESS" << std::endl;
    }
    else
    {
        std::cerr << label << " FAIL" << std::endl;
    }
}

int
main( int argc, char* argv[] )
{
    std::string range_first = "foo -> {1:9:1}\nbar -> {1,2,3,4,5}\n";
    std::string comma_first = "foo -> {1,2,3,4,5}\nbar -> {1:9:1}\n";
    std::string comma_only  = "foo -> {1,2,3,4,5}\n";
    std::string range_only  = "foo -> {1:9:1}\n";

    TryParse( range_first, "RANGE FIRST" );
    TryParse( comma_first, "COMMA FIRST" );
    TryParse( range_only,  "RANGE ONLY"  );
    TryParse( comma_only,  "COMMA ONLY"  );
}

编辑新输出（仍然无法处理异构列表）

/tmp$ g++ -g -std=c++11 sandbox.cpp -o sandbox && ./sandbox 
COMMA FIRST FAIL
RANGE FIRST FAIL
COMMA ONLY SUCCESS
RANGE ONLY SUCCESS
/tmp$

编辑改进的代码（仍然无法处理异构列表）

// #define BOOST_SPIRIT_DEBUG

#include <boost/foreach.hpp>
#include <boost/spirit/include/qi.hpp>
#include <iostream>

namespace qi = boost::spirit::qi;

template <typename Iterator, typename Skipper>
struct my_grammar :
qi::grammar<Iterator, std::vector<int>(), Skipper>
{
    my_grammar() :
        my_grammar::base_type(entries)
    {
        entries %= 
            *('{' >> qi::int_
            >> ( range_tail | comma_tail ))
        ;

        range_tail %= ':' >> qi::int_ >> ':' >> qi::int_ >> '}';
        comma_tail %= *( ',' >> qi::int_ ) >> '}';

        BOOST_SPIRIT_DEBUG_NODE(entries   );
        BOOST_SPIRIT_DEBUG_NODE(comma_tail);
        BOOST_SPIRIT_DEBUG_NODE(range_tail);
    }

    qi::rule<Iterator, std::vector<int>(), Skipper> entries, comma_tail, range_tail;
};

// -----------------------------------------------------------------------------

static void TryParse( const std::string& input, const std::string& label )
{
    std::vector<int> entries;
    auto it(input.cbegin()), end( input.cend() );
    my_grammar<decltype(it), qi::blank_type> entry_grammar;

    if (qi::phrase_parse(it, end, entry_grammar, qi::blank, entries)
            && it == end)
    {
        std::cerr << label << " SUCCESS" << std::endl;
    }
    else
    {
        std::cerr << label << " FAIL" << std::endl;
    }
}

int
main( int argc, char* argv[] )
{
    std::string range_first = "{3:5:7}\n{1,2,3,4,5}";
    std::string comma_first = "{1,2,3,4,5}\n{3:5:7}";
    std::string comma_only  = "{1,2,3,4,5}";
    std::string range_only  = "{3:5:7}";

    TryParse( comma_first, "COMMA FIRST" );
    TryParse( range_first, "RANGE FIRST" );
    TryParse( comma_only,  "COMMA ONLY"  );
    TryParse( range_only,  "RANGE ONLY"  );
}

原始代码

编译并运行

/tmp$ g++ -g -std=c++11 sandbox.cpp -o sandbox && ./sandbox 
COMMA FIRST FAIL
RANGE FIRST FAIL
COMMA ONLY SUCCESS
RANGE ONLY FAIL
/tmp$

sandbox.cpp

// #define BOOST_SPIRIT_DEBUG

#include <boost/foreach.hpp>
#include <boost/spirit/include/qi.hpp>
#include <iostream>


namespace qi = boost::spirit::qi;
typedef int Entry;

template <typename Iterator, typename Skipper>
struct my_grammar :
qi::grammar<Iterator, std::vector<Entry>(), Skipper>
{
    my_grammar() :
        my_grammar::base_type(entries)
    {
        entries %= 
            *( 
                '{' >>
                    ( comma_list | range_list )
                >> '}'
            )
        ;

        comma_list %= qi::int_ % ',';
        range_list %= qi::int_ >> ':' >> qi::int_ >> ':' >> qi::int_;

        BOOST_SPIRIT_DEBUG_NODE(entries);
        BOOST_SPIRIT_DEBUG_NODE(comma_list);
        BOOST_SPIRIT_DEBUG_NODE(range_list);
    }

    qi::rule<Iterator, std::vector<Entry>(), Skipper> entries, comma_list, range_list;
};

// -----------------------------------------------------------------------------
static void TryParse( const std::string& input, const std::string& label )
{
    std::vector<Entry> entries;
    auto it(input.cbegin()), end( input.cend() );
    my_grammar<decltype(it), qi::blank_type> entry_grammar;

    if (qi::phrase_parse(it, end, entry_grammar, qi::blank, entries)
            && it == end)
    {
        std::cerr << label << " SUCCESS" << std::endl;
    }
    else
    {
        std::cerr << label << " FAIL" << std::endl;
    }
}

int
main( int argc, char* argv[] )
{
    std::string range_first = "{3:5:7}\n{1,2,3,4,5}";
    std::string comma_first = "{1,2,3,4,5}\n{3:5:7}";
    std::string comma_only  = "{1,2,3,4,5}";
    std::string range_only  = "{3:5:7}";

    TryParse( range_first, "COMMA FIRST" );
    TryParse( comma_first, "RANGE FIRST" );
    TryParse( comma_only,  "COMMA ONLY"  );
    TryParse( range_only,  "RANGE ONLY"  );
}

Answer 1

在公共代码中使用前导号码。在该数字后面有一个范围尾部或一个列表尾部。列表尾部以a开头，范围尾部以：。

开头

可能将}折叠到尾部，所以列表尾部是（，数字）*}，范围尾部是：number：number}，这使得空列表更容易解析。

如何在没有预标记的情况下解析异构列表列表？

1 个答案: