在过程返回消息后创建过程并运行代码（C ++）

Question

I am trying to create a new process (which shouldn't block the current program) in C++, and to have the C++ listen for a message. When the message arrives, I want to run some more code.

I have this method which executes the command and results the result immediately, but I have no idea how to make one that runs code when the process returns a message:

    string exec(const char* cmd)
    {
        // popen for *nix
        FILE* pipe = _popen(cmd, "r");

        if (!pipe)
            return "";

        char buffer[128];
        string result = "";

        while(!feof(pipe)) {
            if(fgets(buffer, 128, pipe) != NULL)
                result += buffer;
        }

        // pclose for *nix
        _pclose(pipe);

        return result;
    }

Note: it may take a 1-3 seconds until the process returns the message -- and the process will continue execution after that so the above code is not sufficient as the executed program will never end.

Answer 1

您可以将执行该过程的代码放在单独的线程中，这样，主线程可以继续执行，而辅助线程可以启动该过程，并等待子进程的消息。

Answer 2

After you start the process, periodically check to see if it has sent you the message. If so, call the code you need to call.

Exactly how you do this depends on the structure of the rest of your code. For example, if you already have a table of periodic events, then it's very simple. If you have a 'select' loop type design, it's again very simple.

If you don't have any of those things already, write a 'CheckIfTheresAnythingToDo' function. Have a boolean flag that indicates whether you have a child process running. In 'CheckIfTheresAnythingToDo' return immediately if the flag is clear. If it's set, call select and see if the message has been received. If so, call the code you need to call. (I'd recommend not using standard I/O because it doesn't play well with the non-blocking semantics you need to do more than one thing at a time.)

Then just call CheckIfTheresAnythingToDo at key points in your code. If you have some kind of loop structure, call it once on each pass of the loop. If you have code that waits for things, have your wait function call it.

But this is the kind of thing you should already know how to do when you started your design. "I need to check something every once in a while" is a very typical requirement and if your overall design has no way to accommodate it, you should rethink that design.