读取 UTF-8 文件，使用 SendInput 将内容传递给其他应用程序

Question

I have a utf-8 formatted file that contains a character ' ů ', when I read it in c++ using ifstream and cout it, the output is ' ů '. I can solve it by adding ' SetConsoleOutputCP( 65001 ); ', afterwards Im sending the char using ' SendInput(); ' to another window. But because the SetConsoleOutput() doesnt affect SendInput() the application recieves the corrupted character ' ů '. The question is how to SendInput() accented character?

        ifstream file(path);
        string str;
        while (getline (file, str)) 
        {
            cout << endl << str <<endl;
            for (char &c : str) {
                INPUT ip;
                ip.type = INPUT_KEYBOARD;
                ip.ki.wScan = 0;
                ip.ki.time = 0;
                ip.ki.dwExtraInfo = 0;
                ip.ki.dwFlags = KEYEVENTF_UNICODE;
                if (isupper(c)) {
                    ip.ki.wVk = VK_LSHIFT;
                    SendInput(1, &ip, sizeof(INPUT));
                } else {
                    ip.ki.dwFlags = KEYEVENTF_KEYUP;
                    ip.ki.wVk = VK_LSHIFT;
                    SendInput(1, &ip, sizeof(INPUT));
                    ip.ki.dwFlags = KEYEVENTF_UNICODE;
                } 
                ip.ki.wVk = VkKeyScanA(c);
                SendInput(1, &ip, sizeof(INPUT));
                Sleep(100);
            }
            INPUT enter;
            enter.type = INPUT_KEYBOARD;
            enter.ki.wScan = 0;
            enter.ki.time = 0;
            enter.ki.dwExtraInfo = 0;
            enter.ki.dwFlags = 0;
            enter.ki.wVk = VK_RETURN;
            SendInput(1, &enter, sizeof(INPUT));
            cout << "enter pressed";
        }

Answer 1

1. UTF-8 and Windows are not good friends. Almost everything you are going to do with Windows requires UTF-16. You need to convert your UTF-8 string to UTF-16. On Windows, that would be "wstring", but you should use the right conversion routine. Alas the C++ standard cannot be bothered to give you a working one one right now, so Windows-specific MultiByteToWideChar seems like a good option.

2.  ip.ki.dwFlags = KEYEVENTF_UNICODE; if (isupper(c)) { ip.ki.wVk = VK_LSHIFT;

   It doesn't work this way.

   If the dwFlags member specifies KEYEVENTF_UNICODE, wVk must be 0.

   Unicode characters cannot be combined with Shift or other modifiers. They are not key codes, they are characters. A and a are different characters, there is no need to send shift presses to differentiate between them. You also don't need VkKeyScanA , because you want to send Unicode characters, not scan codes.

Answer 2

You need to convert the UTF-8 data to UTF-16 first, such as with MultiByteToWideChar() or equivalent, and then you can pass the individual UTF-16 codeunits to SendInput() using the KEYEVENTF_UNICODE flag. See my answer to Sending Two or more chars using SendInput for an example of sending UTF-16 strings with SendInput() .