我有一個 C++ 類，它將 Python 模塊作為文件導入，并在返回接收到的數據之前連續調用它的函數：ProgramLiveProcess::ProgramLiveProcess() {    //Start Python to run the Program Python code    Py_Initialize();    //Import the Program live processing module    program_live_processing_module = PyImport_ImportModule(MODULE_NAME);    //Get the objects for the functions to call    live_process = PyObject_GetAttrString(program_live_processing_module, "Program_LiveProcess");}//Creates Python list from intensities arrayPyObject* ProgramLiveProcess::get_intensities_list(unsigned int intensities[INTENSITIES_DATA_SIZE]) {    PyObjec* tr = PyList_New(0);    for (int i = 0; i < INTENSITIES_DATA_SIZE; i++) {        PyObject* ta = PyLong_FromLong(intensities[i]);        PyList_Append(tr, ta);    }    return tr;}//TODO: Make this actually work//Frees objects in intensities Python list and the list itselfvoid ProgramLiveProcess::free_intensities_list(PyObject* i_list) {    //std::cout << std::to_string(i_list->ob_refcnt) << std::endl;    for (int i = 0; i < INTENSITIES_DATA_SIZE; i++) {        PyObject* ta = PyList_GET_ITEM(i_list, i);        Py_DECREF(ta);    }    Py_DECREF(i_list);}processed_data* ProgramLiveProcess::send_to_program_live_process(unsigned int intensities[INTENSITIES_DATA_SIZE], bool use_sensor_1) {    //Call the Program_LiveProcess function    PyObject* intensities_py = get_intensities_list(intensities);    PyObject* calculate_args = PyTuple_Pack(2, intensities_py, use_sensor_1 ? Py_True : Py_False); //True or false depending on if using sensor 1 or 2    PyObject* processed_data_tuple = PyObject_CallObject(live_process, calculate_args);    Py_DECREF(calculate_args);    free_intensities_list(intensities_py);    //Get the data from the function    PyObject* s0p = PyTuple_GetItem(processed_data_tuple, 0);    PyObject* s0t = PyTuple_GetItem(processed_data_tuple, 1);}雖然程序接收數據很好，但在不同條件下運行時，我在內存泄漏和崩潰方面遇到了不一致的情況。雖然我無法提供堆棧跟蹤，但我想知道我在創建或取消引用 Python 對象方面是否做錯了什么。通常，程序僅在我注釋掉 free_intensities_list 調用時運行。