#include "utfconvert.h"

#include <stdint.h>

static inline uint16_t byteswap_ushort(uint16_t number)
{
#if defined(_MSC_VER) && _MSC_VER > 1310
    return _byteswap_ushort(number);
#elif defined(__GNUC__)
    return __builtin_bswap16(number);
#else
    return (number >> 8) | (number << 8);
#endif
}


////////////////////////////////////////
//     以下转换都是在小端序下进行     //
////////////////////////////////////////

// 从UTF16编码字符串构建，需要带BOM标记
std::string utf16_to_utf8(const std::u16string& u16str)
{
    if (u16str.empty()){ return std::string(); }
    //Byte Order Mark
    char16_t bom = u16str[0];
    switch (bom){
    case 0xFEFF:    //Little Endian
        return utf16le_to_utf8(u16str);
        break;
    case 0xFFFE:    //Big Endian
        return utf16be_to_utf8(u16str);
        break;
    default:
        return std::string();
    }
}


// 从UTF16 LE编码的字符串创建
std::string utf16le_to_utf8(const std::u16string& u16str)
{
    if (u16str.empty()){ return std::string(); }
    const char16_t* p = u16str.data();
    std::u16string::size_type len = u16str.length();
    if (p[0] == 0xFEFF){
        p += 1; //带有bom标记，后移
        len -= 1;
    }

    // 开始转换
    std::string u8str;
    u8str.reserve(len * 3);

    char16_t u16char;
    for (std::u16string::size_type i = 0; i < len; ++i){
        // 这里假设是在小端序下(大端序不适用)
        u16char = p[i];

        // 1字节表示部分
        if (u16char < 0x0080){
            // u16char <= 0x007f
            // U- 0000 0000 ~ 0000 07ff : 0xxx xxxx
            u8str.push_back((char)(u16char & 0x00FF));  // 取低8bit
            continue;
        }
        // 2 字节能表示部分
        if (u16char >= 0x0080 && u16char <= 0x07FF){
            // * U-00000080 - U-000007FF:  110xxxxx 10xxxxxx
            u8str.push_back((char)(((u16char >> 6) & 0x1F) | 0xC0));
            u8str.push_back((char)((u16char & 0x3F) | 0x80));
            continue;
        }
        // 代理项对部分(4字节表示)
        if (u16char >= 0xD800 && u16char <= 0xDBFF) {
            // * U-00010000 - U-001FFFFF: 1111 0xxx 10xxxxxx 10xxxxxx 10xxxxxx
            uint32_t highSur = u16char;
            uint32_t lowSur = p[++i];
            // 从代理项对到UNICODE代码点转换
            // 1、从高代理项减去0xD800，获取有效10bit
            // 2、从低代理项减去0xDC00，获取有效10bit
            // 3、加上0x10000，获取UNICODE代码点值
            uint32_t codePoint = highSur - 0xD800;
            codePoint <<= 10;
            codePoint |= lowSur - 0xDC00;
            codePoint += 0x10000;
            // 转为4字节UTF8编码表示
            u8str.push_back((char)((codePoint >> 18) | 0xF0));
            u8str.push_back((char)(((codePoint >> 12) & 0x3F) | 0x80));
            u8str.push_back((char)(((codePoint >> 06) & 0x3F) | 0x80));
            u8str.push_back((char)((codePoint & 0x3F) | 0x80));
            continue;
        }
        // 3 字节表示部分
        {
            // * U-0000E000 - U-0000FFFF:  1110xxxx 10xxxxxx 10xxxxxx
            u8str.push_back((char)(((u16char >> 12) & 0x0F) | 0xE0));
            u8str.push_back((char)(((u16char >> 6) & 0x3F) | 0x80));
            u8str.push_back((char)((u16char & 0x3F) | 0x80));
            continue;
        }
    }

    return u8str;
}


// 从UTF16BE编码字符串创建
std::string utf16be_to_utf8(const std::u16string& u16str)
{
    if (u16str.empty()){ return std::string(); }
    const char16_t* p = u16str.data();
    std::u16string::size_type len = u16str.length();
    if (p[0] == 0xFEFF){
        p += 1; //带有bom标记，后移
        len -= 1;
    }


    // 开始转换
    std::string u8str;
    u8str.reserve(len * 2);
    char16_t u16char;   //u16le 低字节存低位，高字节存高位
    for (std::u16string::size_type i = 0; i < len; ++i) {
        // 这里假设是在小端序下(大端序不适用)
        u16char = p[i];
        // 将大端序转为小端序
        u16char = byteswap_ushort(u16char);

        // 1字节表示部分
        if (u16char < 0x0080) {
            // u16char <= 0x007f
            // U- 0000 0000 ~ 0000 07ff : 0xxx xxxx
            u8str.push_back((char)(u16char & 0x00FF));
            continue;
        }
        // 2 字节能表示部分
        if (u16char >= 0x0080 && u16char <= 0x07FF) {
            // * U-00000080 - U-000007FF:  110xxxxx 10xxxxxx
            u8str.push_back((char)(((u16char >> 6) & 0x1F) | 0xC0));
            u8str.push_back((char)((u16char & 0x3F) | 0x80));
            continue;
        }
        // 代理项对部分(4字节表示)
        if (u16char >= 0xD800 && u16char <= 0xDBFF) {
            // * U-00010000 - U-001FFFFF: 1111 0xxx 10xxxxxx 10xxxxxx 10xxxxxx
            uint32_t highSur = u16char;
            uint32_t lowSur = byteswap_ushort(p[++i]);
            // 从代理项对到UNICODE代码点转换
            // 1、从高代理项减去0xD800，获取有效10bit
            // 2、从低代理项减去0xDC00，获取有效10bit
            // 3、加上0x10000，获取UNICODE代码点值
            uint32_t codePoint = highSur - 0xD800;
            codePoint <<= 10;
            codePoint |= lowSur - 0xDC00;
            codePoint += 0x10000;
            // 转为4字节UTF8编码表示
            u8str.push_back((char)((codePoint >> 18) | 0xF0));
            u8str.push_back((char)(((codePoint >> 12) & 0x3F) | 0x80));
            u8str.push_back((char)(((codePoint >> 06) & 0x3F) | 0x80));
            u8str.push_back((char)((codePoint & 0x3F) | 0x80));
            continue;
        }
        // 3 字节表示部分
        {
            // * U-0000E000 - U-0000FFFF:  1110xxxx 10xxxxxx 10xxxxxx
            u8str.push_back((char)(((u16char >> 12) & 0x0F) | 0xE0));
            u8str.push_back((char)(((u16char >> 6) & 0x3F) | 0x80));
            u8str.push_back((char)((u16char & 0x3F) | 0x80));
            continue;
        }
    }
    return u8str;
}






// 获取转换为UTF-16 LE编码的字符串
std::u16string utf8_to_utf16le(const std::string& u8str, bool addbom, bool* ok)
{
    std::u16string u16str;
    u16str.reserve(u8str.size());
    if (addbom) {
        u16str.push_back(0xFEFF);   //bom (字节表示为 FF FE)
    }
    std::string::size_type len = u8str.length();

    const unsigned char* p = (unsigned char*)(u8str.data());
    // 判断是否具有BOM(判断长度小于3字节的情况)
    if (len > 3 && p[0] == 0xEF && p[1] == 0xBB && p[2] == 0xBF){
        p += 3;
        len -= 3;
    }

    bool is_ok = true;
    // 开始转换
    for (std::string::size_type i = 0; i < len; ++i) {
        uint32_t ch = p[i]; // 取出UTF8序列首字节
        if ((ch & 0x80) == 0) {
            // 最高位为0，只有1字节表示UNICODE代码点
            u16str.push_back((char16_t)ch);
            continue;
        }
        switch (ch & 0xF0)
        {
        case 0xF0: // 4 字节字符, 0x10000 到 0x10FFFF
        {
            uint32_t c2 = p[++i];
            uint32_t c3 = p[++i];
            uint32_t c4 = p[++i];
            // 计算UNICODE代码点值(第一个字节取低3bit，其余取6bit)
            uint32_t codePoint = ((ch & 0x07U) << 18) | ((c2 & 0x3FU) << 12) | ((c3 & 0x3FU) << 6) | (c4 & 0x3FU);
            if (codePoint >= 0x10000)
            {
                // 在UTF-16中 U+10000 到 U+10FFFF 用两个16bit单元表示, 代理项对.
                // 1、将代码点减去0x10000(得到长度为20bit的值)
                // 2、high 代理项 是将那20bit中的高10bit加上0xD800(110110 00 00000000)
                // 3、low  代理项 是将那20bit中的低10bit加上0xDC00(110111 00 00000000)
                codePoint -= 0x10000;
                u16str.push_back((char16_t)((codePoint >> 10) | 0xD800U));
                u16str.push_back((char16_t)((codePoint & 0x03FFU) | 0xDC00U));
            }
            else
            {
                // 在UTF-16中 U+0000 到 U+D7FF 以及 U+E000 到 U+FFFF 与Unicode代码点值相同.
                // U+D800 到 U+DFFF 是无效字符, 为了简单起见，这里假设它不存在(如果有则不编码)
                u16str.push_back((char16_t)codePoint);
            }
        }
        break;
        case 0xE0: // 3 字节字符, 0x800 到 0xFFFF
        {
            uint32_t c2 = p[++i];
            uint32_t c3 = p[++i];
            // 计算UNICODE代码点值(第一个字节取低4bit，其余取6bit)
            uint32_t codePoint = ((ch & 0x0FU) << 12) | ((c2 & 0x3FU) << 6) | (c3 & 0x3FU);
            u16str.push_back((char16_t)codePoint);
        }
        break;
        case 0xD0: // 2 字节字符, 0x80 到 0x7FF
        case 0xC0:
        {
            uint32_t c2 = p[++i];
            // 计算UNICODE代码点值(第一个字节取低5bit，其余取6bit)
            uint32_t codePoint = ((ch & 0x1FU) << 12) | ((c2 & 0x3FU) << 6);
            u16str.push_back((char16_t)codePoint);
        }
        break;
        default:    // 单字节部分(前面已经处理，所以不应该进来)
            is_ok = false;
            break;
        }
    }
    if (ok != NULL) { *ok = is_ok; }

    return u16str;
}


// 获取转换为UTF-16 BE的字符串
std::u16string utf8_to_utf16be(const std::string& u8str, bool addbom, bool* ok)
{
    // 先获取utf16le编码字符串
    std::u16string u16str = utf8_to_utf16le(u8str, addbom, ok);
    // 将小端序转换为大端序
    for (size_t i = 0; i < u16str.size(); ++i) {
        u16str[i] = byteswap_ushort(u16str[i]);
    }
    return u16str;
}

// 从UTF16编码的字符串创建
size_t utf16_to_char(const std::u16string& u16str, char *buffer) {
    size_t length = 2;
    if (u16str.empty()){
        *buffer++ = '\0';
        *buffer++ = '\0';
        return length;
    }
    const char16_t* p = u16str.data();
    std::u16string::size_type len = u16str.length();
    if (p[0] == 0xFEFF){
        p += 1; //带有bom标记，后移
        len -= 1;
    }
    // 开始转换
    char16_t u16char;   //u16le 低字节存低位，高字节存高位
    for (std::u16string::size_type i = 0; i < len; ++i) {
        u16char = p[i];
        *buffer++ = (u16char & 0xFF00) >> 8;
        *buffer++ = (u16char & 0xFF);
        length += 2;
    }
    *buffer++ = '\0';
    *buffer++ = '\0';
    return length;
}

size_t utf8_to_char(const std::string& u8str, char* buffer) {
    size_t length = 1;
    if (u8str.empty()){
        *buffer++ = '\0';
        return length;
    }
    const char* p = u8str.c_str();
    std::string::size_type len = u8str.length();
    // 开始转换
    for (std::string::size_type i = 0; i < len; ++i) {
        *buffer++ = p[i];
        length ++;
    }
    *buffer++ = '\0';
    return length;
}

size_t hex_to_char(const char *hex, char *buffer) {
    size_t i,n = 0;
    for(i = 0; hex[i]; i += 2) {
        if(hex[i] >= 'A' && hex[i] <= 'F')
            buffer[n] = hex[i] - 'A' + 10;
        else if(hex[i] >= 'a' && hex[i] <= 'f')
            buffer[n] = hex[i] - 'a' + 10;
        else buffer[n] = hex[i] - '0';
        if(hex[i + 1] >= 'A' && hex[i + 1] <= 'F')
            buffer[n] = (buffer[n] << 4) | (hex[i + 1] - 'A' + 10);
        else if(hex[i + 1] >= 'a' && hex[i + 1] <= 'f')
            buffer[n] = (buffer[n] << 4) | (hex[i + 1] - 'a' + 10);
        else buffer[n] = (buffer[n] << 4) | (hex[i + 1] - '0');
        ++n;
    }
    buffer[n++] = '\0';
    return n;
}