Problem with embedding python in c++.

[Benke]

Hello, I'm quite new to Python and embedding python in c++. I'm trying to write a function that i can use to call a python function. It should take 3 arguments, the name of the python file, the python function name and a std:vector with arguments for the python function. This is my code:

int main()
{
    string arg1 = "runme"; //name of python file
    string arg2 = "sub"; // name of python function
 
    Py_Initialize();
 
    vector<string> arguments;
    arguments.push_back("100");
    arguments.push_back("10");
 
    callPython(arg1,arg2,arguments);
 
    Py_Finalize(); // crash here
 
    return 1;
}
 

void callPython(string py_file, string py_function, vector<string> args)
{
    int i;
    PyObject *pName, *pModule, *pDict, *pFunc, *pArgs, *pValue;
 
    // Build the name object
    pName = PyString_FromString( py_file.c_str() );
 
    // Load the module object
    pModule = PyImport_Import(pName);
 
    // pDict is a borrowed reference 
    pDict = PyModule_GetDict(pModule);
 
    // pFunc is also a borrowed reference 
    pFunc = PyDict_GetItemString(pDict, py_function.c_str() );
 
    if (PyCallable_Check(pFunc)) 
    {
        pValue = PyObject_CallObject(pFunc, NULL);//måste returna något tex 1
 
        // Prepare the argument list for the call
        if( args.size() > 0 )
        {
                pArgs = PyTuple_New(args.size());
                for (i = 0; i < args.size(); i++)
                {
                    pValue = PyInt_FromLong(atoi(args[i].c_str()));
                    if (!pValue)
                    {                        
                        PyErr_Print();
                    }
                    PyTuple_SetItem(pArgs, i, pValue);    
                }
 
                pValue = PyObject_CallObject(pFunc, pArgs);
 
                if (pArgs != NULL)
                {
                    Py_DECREF(pArgs);
                }
        } else
        {
                pValue = PyObject_CallObject(pFunc, NULL);
        }
 
        if (pValue != NULL) 
        {
            Py_DECREF(pValue);
        }
        else 
        {
            PyErr_Print();
        }
    } else 
    {
        PyErr_Print();
    }
 
    // Clean up
    Py_DECREF(pModule);
    Py_DECREF(pName);
}
 

And the python file (runme.py):

def sub(a,b):
    print a-b
    return 0
 

Now to the problem. It does print out the proper result (90 in this case) so it calls the python function. But I get an error and it crashes when at Py_Finalize() in main.

I get the following error:

Exception exceptions.TypeError: 'sub() takes exactly 2 arguments (0 given)' in 'garbage collection' ignored
Fatal Python error: unexpected exception during garbage collection
 

What have i done wrong? I did get it to work with adding PyErr_Clear(); before Py_Finalize() but that doesnt seem like the right thing to do?

I'm using Python 2.5.1, Windows XP pro and VS2005.

Thanks

[novotny3]

Have you tried called PyErr_Print() before Py_Finalize() to ensure that there are no other errors?