让jsoncpp库能够支持直接解析unicode编码的字符串

看了看，jsoncpp里面的代码，发现其实并不难，代码很有规矩、易懂，一下就能定位到关键的位置： json_tool.h中的codePointToUTF8函数。

贴上这个函数的原型来看看就能明白了：

/// Converts a unicode code-point to UTF-8. static inline JSONCPP_STRING codePointToUTF8(unsigned int cp) { JSONCPP_STRING result; // based on description from http://en.wikipedia.org/wiki/UTF-8 if (cp <= 0x7f) { result.resize(1); result[0] = static_cast<char>(cp); } else if (cp <= 0x7FF) { result.resize(2); result[1] = static_cast<char>(0x80 | (0x3f & cp)); result[0] = static_cast<char>(0xC0 | (0x1f & (cp >> 6))); } else if (cp <= 0xFFFF) { result.resize(3); result[2] = static_cast<char>(0x80 | (0x3f & cp)); result[1] = static_cast<char>(0x80 | (0x3f & (cp >> 6))); result[0] = static_cast<char>(0xE0 | (0xf & (cp >> 12))); } else if (cp <= 0x10FFFF) { result.resize(4); result[3] = static_cast<char>(0x80 | (0x3f & cp)); result[2] = static_cast<char>(0x80 | (0x3f & (cp >> 6))); result[1] = static_cast<char>(0x80 | (0x3f & (cp >> 12))); result[0] = static_cast<char>(0xF0 | (0x7 & (cp >> 18))); } return result; }

稍微改改就可以在json解析的时候直接支持unicode编码的字符串了：

/// Converts a unicode code-point to UTF-8. static inline JSONCPP_STRING codePointToUTF8(unsigned int cp) { JSONCPP_STRING result; // based on description from http://en.wikipedia.org/wiki/UTF-8 if (cp <= 0x7f) { result.resize(1); result[0] = static_cast<char>(cp); } else if (cp <= 0x7FF) { result.resize(2); result[1] = static_cast<char>(0x80 | (0x3f & cp)); result[0] = static_cast<char>(0xC0 | (0x1f & (cp >> 6))); } else if (cp <= 0xFFFF) { if (cp >= 0x4E00 && cp <= 0x9FA5 || (cp >= 0xF900 && cp <= 0xFA2D)) { wchar_t src[2] = { 0 }; char dest[5] = { 0 }; src[0] = static_cast<wchar_t>(cp); std::string curLocale = setlocale(LC_ALL, NULL); setlocale(LC_ALL, "chs"); wcstombs_s(NULL, dest, 5, src, 2); result = dest; setlocale(LC_ALL, curLocale.c_str()); } else { result.resize(3); result[2] = static_cast<char>(0x80 | (0x3f & cp)); result[1] = static_cast<char>(0x80 | (0x3f & (cp >> 6))); result[0] = static_cast<char>(0xE0 | (0xf & (cp >> 12))); } } else if (cp <= 0x10FFFF) { result.resize(4); result[3] = static_cast<char>(0x80 | (0x3f & cp)); result[2] = static_cast<char>(0x80 | (0x3f & (cp >> 6))); result[1] = static_cast<char>(0x80 | (0x3f & (cp >> 12))); result[0] = static_cast<char>(0xF0 | (0x7 & (cp >> 18))); } return result; }

换上去之后，编译，运行，一切正常，让人头疼的编码问题终于告一段落了。

本文参考至：http://blog.csdn.net/harrycris/article/details/7733386