From fedry.kemilau at gmail.com  Fri Jul  3 10:01:25 2009
From: fedry.kemilau at gmail.com (Fedry Kemilau)
Date: Fri, 3 Jul 2009 16:01:25 +0800
Subject: [Python.NET] Scripting with a managed object using python.net engine
Message-ID: <2758edf90907030101h22d05ffdx1c7a9c4afb318899@mail.gmail.com>

Dear all,

Currently, I am facing an issue to do scripting with a managed object in
python engine.
I managed to import the module and set an attribute value to the managed
object.

The issue is in the next step where I am trying to use the module in a
statement/expression via PythonEngine.RunString or RunSimpleString method.
As far as I know:

   - RunString() is not working because it always return null value. I tried
   running a simple expression, e.g., "1 + 1", and it returned null value. I
   believe the fix
here<http://osdir.com/ml/python.dotnet/2007-03/msg00009.html> could
   solve this bug (if it really is...)
   - RunSimpleString() executes expression from the __main__ module
   implicitly. I have not found any possible way yet to include the module I've
   imported in the beginning into __main__ module.

So...

I made a little modification on the code by creating a new class that
represents the Python Module object.
This is how I use the modification in my code.

>             PythonEngine.Initialize();



            // create __main__ and my own module

            PyModule __main__ = new
> PyModule(PythonEngine.ImportModule("__main__"));

            PyModule birdwrapper = new
> PyModule(PythonEngine.ImportModule("birdwrapper"));


>             // set the value of the attribute to a managed object

            birdwrapper.SetAttr("pet", PyObject.FromManagedObject(new
> Eagle("Foo")));


>             // include my own module in the __main__ module

            __main__.AddObject("birdwrapper", birdwrapper);


>             // run the script from this method because it is called from
> the __main__ module

            PythonEngine.RunSimpleString("birdwrapper.Fly()");


>             PythonEngine.Shutdown();


 ++ in runtime.cs:

>         internal static bool PyModule_Check(IntPtr ob)

          {

                return PyObject_TYPE(ob) == Runtime.PyModuleType;

          }



        [DllImport(Runtime.dll, CallingConvention = CallingConvention.Cdecl,

          ExactSpelling = true, CharSet = CharSet.Ansi)]

          internal unsafe static extern int

          PyModule_AddObject(IntPtr module, string name, IntPtr value);


 pymodule.cs:

>     public class PyModule : PyObject

    {


>         public PyModule(IntPtr ptr) : base(ptr) { }


>         public PyModule(PyObject o) : base()

        {

            if (!IsModuleType(o))

            {

                throw new ArgumentException("object is not a module");

            }

            Runtime.Incref(o.obj);

            obj = o.obj;

        }


>         public PyModule(string name) : base()

        {

            obj = Runtime.PyModule_New(name);

            if (obj == IntPtr.Zero)

            {

                throw new PythonException();

            }

        }


>         public static bool IsModuleType(PyObject value)

        {

            return Runtime.PyModule_Check(value.obj);

        }


>         public void AddObject(string name, PyObject value)

        {

            int op = Runtime.PyModule_AddObject(obj, name, value.obj);

            if (op != 0)

                throw new PythonException();

        }


>     }



I hope the code could illustrate what I am basically trying to achieve with
python.net.
I wonder if there is a possible way to do this without modifying the code.

Could anyone please advise?

--
Best regards,
Fedry Kemilau
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pythondotnet/attachments/20090703/00d64df4/attachment.htm>