C++怎么处理字符串编码 C++字符串编码转换的方法介绍

#include <iostream>
#include <unicode/ustream.h>
#include <unicode/ucnv.h>
#include <string>

std::string convertEncoding(const std::string& str, const char* fromEncoding, const char* toEncoding) {
  UErrorCode status = U_ZERO_ERROR;
  UConverter* fromConverter = ucnv_open(fromEncoding, &status);
  if (U_FAILURE(status)) {
    std::cerr << "Failed to open fromConverter: " << u_errorName(status) << std::endl;
    return "";
  }

  UConverter* toConverter = ucnv_open(toEncoding, &status);
  if (U_FAILURE(status)) {
    std::cerr << "Failed to open toConverter: " << u_errorName(status) << std::endl;
    ucnv_close(fromConverter);
    return "";
  }

  std::string result;
  const char* source = str.c_str();
  const char* sourceEnd = source + str.length();
  char buffer[4096]; // 缓冲区大小，根据实际情况调整
  char* target = buffer;
  char* targetEnd = buffer + sizeof(buffer);

  ucnv_convertEx(toConverter, fromConverter, &target, targetEnd, &source, sourceEnd, NULL, NULL, &status);

  if (U_FAILURE(status) && status != U_BUFFER_OVERFLOW_ERROR) {
    std::cerr << "Conversion failed: " << u_errorName(status) << std::endl;
    ucnv_close(fromConverter);
    ucnv_close(toConverter);
    return "";
  }

  result.assign(buffer, target - buffer);

  ucnv_close(fromConverter);
  ucnv_close(toConverter);

  return result;
}

int main() {
  std::string utf8String = "你好，世界！";
  std::string gbkString = convertEncoding(utf8String, "UTF-8", "GBK");

  if (!gbkString.empty()) {
    std::cout << "UTF-8: " << utf8String << std::endl;
    std::cout << "GBK: " << gbkString << std::endl;
  }

  return 0;
}

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

std::string convertEncoding(const std::string& str, UINT fromCodePage, UINT toCodePage) {
  if (str.empty()) {
    return "";
  }

  int wLen = MultiByteToWideChar(fromCodePage, 0, str.c_str(), -1, NULL, 0);
  if (wLen == 0) {
    std::cerr << "MultiByteToWideChar failed: " << GetLastError() << std::endl;
    return "";
  }

  wchar_t* wBuffer = new wchar_t[wLen];
  MultiByteToWideChar(fromCodePage, 0, str.c_str(), -1, wBuffer, wLen);

  int len = WideCharToMultiByte(toCodePage, 0, wBuffer, -1, NULL, 0, NULL, NULL);
  if (len == 0) {
    std::cerr << "WideCharToMultiByte failed: " << GetLastError() << std::endl;
    delete[] wBuffer;
    return "";
  }

  char* buffer = new char[len];
  WideCharToMultiByte(toCodePage, 0, wBuffer, -1, buffer, len, NULL, NULL);

  std::string result(buffer);

  delete[] wBuffer;
  delete[] buffer;

  return result;
}

int main() {
  std::string utf8String = "你好，世界！";
  std::string gbkString = convertEncoding(utf8String, CP_UTF8, 936); // 936是GBK的代码页

  if (!gbkString.empty()) {
    std::cout << "UTF-8: " << utf8String << std::endl;
    std::cout << "GBK: " << gbkString << std::endl;
  }

  return 0;
}

#include <iostream>

int main() {
  char16_t utf16Char = u'你';
  char32_t utf32Char = U'好';

  std::u16string utf16String = u"你好";
  std::u32string utf32String = U"世界";

  std::cout << "UTF-16 char: " << utf16Char << std::endl;
  std::cout << "UTF-32 char: " << utf32Char << std::endl;
  std::cout << "UTF-16 string: " << utf16String.c_str() << std::endl;
  std::cout << "UTF-32 string: " << utf32String.c_str() << std::endl; // 注意：cout直接输出u32string可能会有问题，需要转换

  return 0;
}