C++中实现GBK、KSC5601和JIS0208到Unicode的转换，可以利用Unicode转换库，如iconv库，或者手动编写转换表。

原创于 2024-07-10 11:24:08 发布 · 571 阅读

1 ·

本内容遵循CC 4.0 BY-SA版权协议

GEO检测

标签

#c++ #开发语言

要在C++中实现GBK、KSC5601和JIS0208到Unicode的转换，可以利用Unicode转换库，如iconv库，或者手动编写转换表。以下是使用iconv库的示例代码：

首先，确保在你的系统上安装了iconv库。

GBK to Unicode

#include <iostream>
#include <iconv.h>
#include <cstring>

void convertGBKToUnicode(const char* input, char* output, size_t outputSize) {
    iconv_t cd = iconv_open("UTF-8", "GBK");
    if (cd == (iconv_t)-1) {
        std::cerr << "iconv_open failed" << std::endl;
        return;
    }

    size_t inBytesLeft = strlen(input);
    size_t outBytesLeft = outputSize;
    char* inBuf = const_cast<char*>(input);
    char* outBuf = output;

    if (iconv(cd, &inBuf, &inBytesLeft, &outBuf, &outBytesLeft) == (size_t)-1) {
        std::cerr << "iconv conversion failed" << std::endl;
    }

    iconv_close(cd);
}

int main() {
    const char* gbkStr = "GBK字符串"; // GBK encoded string
    char unicodeStr[256] = {0};
    convertGBKToUnicode(gbkStr, unicodeStr, sizeof(unicodeStr));
    std::cout << "Unicode: " << unicodeStr << std::endl;
    return 0;
}

KSC5601 to Unicode

#include <iostream>
#include <iconv.h>
#include <cstring>

void convertKSC5601ToUnicode(const char* input, char* output, size_t outputSize) {
    iconv_t cd = iconv_open("UTF-8", "CP949");
    if (cd == (iconv_t)-1) {
        std::cerr << "iconv_open failed" << std::endl;
        return;
    }

    size_t inBytesLeft = strlen(input);
    size_t outBytesLeft = outputSize;
    char* inBuf = const_cast<char*>(input);
    char* outBuf = output;

    if (iconv(cd, &inBuf, &inBytesLeft, &outBuf, &outBytesLeft) == (size_t)-1) {
        std::cerr << "iconv conversion failed" << std::endl;
    }

    iconv_close(cd);
}

int main() {
    const char* ksc5601Str = "KSC5601 문자열"; // KSC5601 encoded string
    char unicodeStr[256] = {0};
    convertKSC5601ToUnicode(ksc5601Str, unicodeStr, sizeof(unicodeStr));
    std::cout << "Unicode: " << unicodeStr << std::endl;
    return 0;
}

JIS0208 to Unicode

#include <iostream>
#include <iconv.h>
#include <cstring>

void convertJIS0208ToUnicode(const char* input, char* output, size_t outputSize) {
    iconv_t cd = iconv_open("UTF-8", "EUC-JP");
    if (cd == (iconv_t)-1) {
        std::cerr << "iconv_open failed" << std::endl;
        return;
    }

    size_t inBytesLeft = strlen(input);
    size_t outBytesLeft = outputSize;
    char* inBuf = const_cast<char*>(input);
    char* outBuf = output;

    if (iconv(cd, &inBuf, &inBytesLeft, &outBuf, &outBytesLeft) == (size_t)-1) {
        std::cerr << "iconv conversion failed" << std::endl;
    }

    iconv_close(cd);
}

int main() {
    const char* jis0208Str = "JIS0208文字列"; // JIS0208 encoded string
    char unicodeStr[256] = {0};
    convertJIS0208ToUnicode(jis0208Str, unicodeStr, sizeof(unicodeStr));
    std::cout << "Unicode: " << unicodeStr << std::endl;
    return 0;
}