if json file contain Internationalization chars , get exception

[tan00]

	json j;
	j["name"] = "中文";

	std::cout << j.dump() << std::endl;
	std::ofstream ofs;
	ofs.open("test.json");
	ofs << j;
	ofs.close();

	std::ifstream ifs("test.json");
	if (!ifs.is_open())
	{
		std::cout << "open error" << std::endl;
		return -1;
	}
	json j2;
	j2 << ifs;
	std::cout << j2.dump() << std::endl;

exception at j2 << ifs;

[nlohmann]

Please make sure that your code is compiled with UTF-8, assuming you use MSVC.

[OvermindDL1]

j["name"] = "中文";

For note, this is not entirely valid in source code. The C++ spec only ensures that the ascii range is viable source code, anything beyond that is up to the compiler (and though most compilers will support it, you have to add the proper argument to most compilers to tell it to).

In general, keep to the ascii range in C++ source, keep UTF-8 and so forth out in other files that load up.

[tan00]

Tested under Linux, everything is fine with utf-8.

Under win10 MSVS, my system uses GBK encoding.
Whether the source code is utf8 or not, the resulting json file is the system encoding.

Converting the json file to utf8 format by hand , reading does not throw an exception, but the out data is encoded with utf8. It cannot be displayed correctly in the terminal.
need a conversion

std::string Utf8ToAnsi(const char *pSrc, int nLen)
{
	if (nLen == 0)
	{
		return "";
	}

	//计算需要的Unicode字符长度
	int nCount = MultiByteToWideChar(CP_UTF8, 0, pSrc, nLen, NULL, 0);
	std::vector<WCHAR> vecTmp;
	vecTmp.resize(nCount);

	//将UTF-8转为Unicode
	MultiByteToWideChar(CP_UTF8, 0, pSrc, nLen, &vecTmp[0], nCount);

	//计算需要的Ansi字符长度
	nCount = WideCharToMultiByte(CP_ACP, 0, &vecTmp[0], vecTmp.size(), NULL, 0, 0, 0);;
	std::vector<char> vecAnsi;
	vecAnsi.resize(nCount);

	//将Unicode转为Ansi
	WideCharToMultiByte(CP_ACP, 0, &vecTmp[0], vecTmp.size(), &vecAnsi[0], nCount, 0, 0);

	//反馈结果
	std::string strRet;
	strRet.append(&vecAnsi[0], nCount);
	char *pChar = &vecAnsi[0];
	return strRet;
}

[OvermindDL1]

Tested under Linux, everything is fine with utf-8.

Nothing to do with the OS, everything to do with the compiler and the options for it.
Even then the only reliable source code is source code kept within the standard ASCII range as that is the only one guaranteed by the Standard.

As for the terminal, sounds like a new terminal is needed, windows does not have great methods for detecting terminal capabilities unlike linux, but even then that is all user-code stuff, not json. ^.^

[weili0677]

Tested under Linux, everything is fine with utf-8.

Under win10 MSVS, my system uses GBK encoding. Whether the source code is utf8 or not, the resulting json file is the system encoding.

Converting the json file to utf8 format by hand , reading does not throw an exception, but the out data is encoded with utf8. It cannot be displayed correctly in the terminal. need a conversion

std::string Utf8ToAnsi(const char *pSrc, int nLen)
{
	if (nLen == 0)
	{
		return "";
	}

	//计算需要的Unicode字符长度
	int nCount = MultiByteToWideChar(CP_UTF8, 0, pSrc, nLen, NULL, 0);
	std::vector<WCHAR> vecTmp;
	vecTmp.resize(nCount);

	//将UTF-8转为Unicode
	MultiByteToWideChar(CP_UTF8, 0, pSrc, nLen, &vecTmp[0], nCount);

	//计算需要的Ansi字符长度
	nCount = WideCharToMultiByte(CP_ACP, 0, &vecTmp[0], vecTmp.size(), NULL, 0, 0, 0);;
	std::vector<char> vecAnsi;
	vecAnsi.resize(nCount);

	//将Unicode转为Ansi
	WideCharToMultiByte(CP_ACP, 0, &vecTmp[0], vecTmp.size(), &vecAnsi[0], nCount, 0, 0);

	//反馈结果
	std::string strRet;
	strRet.append(&vecAnsi[0], nCount);
	char *pChar = &vecAnsi[0];
	return strRet;
}

这个临时解决了我的问题^^