C++实现uft8和gbk编码字符串互相转换

最新推荐文章于 2024-07-31 17:46:22 发布

柒僖

最新推荐文章于 2024-07-31 17:46:22 发布

阅读量988

点赞数 3

文章标签： c++ 开发语言

本文链接：https://blog.csdn.net/weixin_40429361/article/details/125785775

版权

在windows下用<windows.h>头文件里的函数进行多字节和宽字符转换，linux下采用<iconv.h>头文件里的函数进行编码互相解析。

#include <iostream>
#include <stdlib.h>
#include <string.h>
#include <string>

#ifdef _WIN32
#include <Windows.h>
#else
#include <iconv.h>
#endif

#ifdef _WIN32
std::string Utf8ToGbk(const char* src_str)
{
	std::string result;
	wchar_t* strSrc;
	char* szRes;
	int len = MultiByteToWideChar(CP_UTF8, 0, src_str, -1, NULL, 0);
	strSrc = new wchar_t[len + 1];
	MultiByteToWideChar(CP_UTF8, 0, src_str, -1, strSrc, len);

	len = WideCharToMultiByte(CP_ACP, 0, strSrc, -1, NULL, 0, NULL, NULL);
	szRes = new char[len + 1];
	WideCharToMultiByte(CP_ACP, 0, strSrc, -1, szRes, len, NULL, NULL);
	result = szRes;
	if(strSrc) 
		delete[]strSrc;
	if(szRes)
		delete[]szRes;
	return result;
}


std::string GbkToUtf8(const char* src_str)
{
	std::string result;
	wchar_t* strSrc;
	char* szRes;
	int len = MultiByteToWideChar(CP_ACP, 0, src_str, -1, NULL, 0);
	strSrc = new wchar_t[len + 1];
	MultiByteToWideChar(CP_ACP, 0, src_str, -1, strSrc, len);

	len = WideCharToMultiByte(CP_UTF8, 0, strSrc, -1, NULL, 0, NULL, NULL);
	szRes = new char[len + 1];
	WideCharToMultiByte(CP_UTF8, 0, strSrc, -1, szRes, len, NULL, NULL);
	result = szRes;
	if (strSrc)
		delete[]strSrc;
	if (szRes)
		delete[]szRes;
	return result;
}
#else
int code_convert(char *from_charset, char *to_charset, char *inbuf, size_t inlen, char *outbuf, size_t outlen)
{
	iconv_t cd;
	int rc;
	char **pin = &inbuf;
	char **pout = &outbuf;

	cd = iconv_open(to_charset, from_charset);
	if (cd == 0) return -1;
	memset(outbuf, 0, outlen);
	if (iconv(cd, pin, &inlen, pout, &outlen) == -1) return -1;
	iconv_close(cd);
	return 0;
}
int GbkToUtf8(char *inbuf, size_t inlen, char *outbuf, size_t outlen)
{
	return code_convert("gb2312", "utf-8", inbuf, inlen, outbuf, outlen);
}
int Utf8ToGbk(char *inbuf, size_t inlen, char *outbuf, size_t outlen)
{
	return code_convert("utf-8", "gb2312", inbuf, inlen, outbuf, outlen);
}
#endif

int main()
{
	std::string teststr = "测试字符串";
	std::cout << "原始字符串：" << teststr.c_str() << std::endl;
#ifdef _WIN32
	std::cout << Utf8ToGbk(GbkToUtf8(teststr.c_str()).c_str()) << std::endl;
#else
	char result_g[1024];
	GbkToUtf8((char*)teststr.c_str(), strlen(teststr.c_str()), result_g, 1024);
	char result_u[1024];
	Utf8ToGbk(result_g, strlen(result_g), result_u, 1024);
	std::cout << result_g << std::endl;
	std::cout << result_u << std::endl;
#endif
	return 0;
}