我在Boost.Spirit中遇到了一个奇怪的词法分析器行为：我有一个规则来匹配正确工作的UTF-8 BOM R"(\xef\xbb\xbf)"。但是，一旦我引入了某些额外的规则，例如R"(\"([^\n\"\\]|\\[\"nrt\\])*\")"，我就不会得到任何匹配。我错过了什么或这是一个错误，更重要的是我如何解决/解决这个问题？
以下是一些显示问题的简化代码：
#include <boost/spirit/home/lex.hpp>
#include <boost/spirit/home/lex/lexer_lexertl.hpp>
#include <string>
#include <iostream>
namespace lex = boost::spirit::lex;
typedef lex::lexertl::token< std::string::const_iterator > token;
typedef lex::lexertl::lexer< token > lexer;
enum token_id
{
byte_order_mark = boost::spirit::lex::min_token_id,
bad_token
};
class working_lexer : public lex::lexer< lexer >
{
public:
working_lexer()
{
self.add
( R"(\xef\xbb\xbf)", byte_order_mark )
;
}
};
class broken_lexer : public lex::lexer< lexer >
{
public:
broken_lexer()
{
self.add
( R"(\xef\xbb\xbf)", byte_order_mark )
// why does the above no longer match once I have one of these?
//( R"(\/\/[^\n]*|\/\*([^\*]|\*[^\/])*\*\/)", bad_token ) // C-style comments
( R"(\"([^\n\"\\]|\\[\"nrt\\])*\")", bad_token ) // ""-strings with \n, \r, \t and \\ escapes
;
}
};
int main( int, char** )
{
std::string code( "\xef\xbb\xbf" );
const auto end = code.cend();
std::cout << "Working lexer:" << std::endl;
auto it = code.cbegin();
bool success = lex::tokenize( it, end, working_lexer{} );
std::cout << "Success: " << std::boolalpha << success << std::endl;
std::cout << "Consumed everything: " << std::boolalpha << ( it == end ) << std::endl;
std::cout << "Broken lexer:" << std::endl;
it = code.cbegin();
success = lex::tokenize( it, end, broken_lexer{} );
std::cout << "Success: " << std::boolalpha << success << std::endl;
std::cout << "Consumed everything: " << std::boolalpha << ( it == end ) << std::endl;
return 0;
}