Numeric C extension: bug or error ?

[Philippe Grosse]

Hi,

I don't understand this strange behaviour:

I compile this code :

#include <Python.h>
#include"Numeri c/arrayobject.h"

static PyObject *
return_vector(P yObject *self, PyObject *args)
{
PyObject *input1;
PyArrayObject *vector;

if (!PyArg_ParseTu ple(args, "O", &input1))
return NULL;

vector = (PyArrayObject
*)PyArray_Conti guousFromObject (input1,PyArray _DOUBLE, 1, 1);

if (vector == NULL)
return NULL;

return PyArray_Return( vector);

}

/* registration table */
static struct PyMethodDef testMethods[] = {
{"return_vector ", return_vector, 1}, /* method name, C
funcptr, always-tuple */
{NULL, NULL} /* end of table marker */
};

/* module initializer */
void inittest() /* called on first import */
{ /* name matters if loaded
dynamically */
(void) Py_InitModule(" test",testMetho ds); /* mod name, table ptr */
import_array(); /* indispensable pour utiliser les arrays */
}

Very simple: this module takes a Numeric array (vector) as argument and
send this array back to python...

If I compile it under macOSX, the result in python is:
[color=blue][color=green][color=darkred]
>>> import test
>>> from Numeric import *
>>> v=array([1.0,2.0,3.0,4.0],Float)
>>> v[/color][/color][/color]
array([ 1., 2., 3., 4.])[color=blue][color=green][color=darkred]
>>> test.return_vec tor(v)[/color][/color][/color]
array([ 1., 2., 3., 4.])

but in linux the result is:
[color=blue][color=green][color=darkred]
>>> import test
>>> from Numeric import *
>>> v=array([1.0,2.0,3.0,4.0],Float)
>>> v[/color][/color][/color]
array([ 1., 2., 3., 4.])[color=blue][color=green][color=darkred]
>>> test.return_vec tor(v)[/color][/color][/color]
array([ 1.+0.j, 2.+0.j, 3.+0.j, 4.+0.j])

Strange !!! the result is a complex array.
The C vector is really complex because if I try to use it in the C module,
the result is...incomplete .
For example, if I try to multiply each item of the vector array by a
scalar ( pi for example) in a loop, I get this result in python:

[color=blue][color=green][color=darkred]
>>> import test
>>> from Numeric import *
>>> v=array([1.0,2.0,3.0,4.0],Float)
>>> test.return_vec tor(v)[/color][/color][/color]
array([ 3.14159265+0.j, 0.+0.j, 6.28318531+0.j, 0.+0.j])

It's the result of pi* 1.0, 0.0, 2.0, 0.0 ( the four first elements of the
internal representation of the complex array ([ 1.+0.j, 2.+0.j, 3.+0.j,
4.+0.j]).

Anyone any idea where I'm going wrong?
Thank you for any help.

Philippe

[David M. Cooke]

Re: Numeric C extension: bug or error ?

At some point, veronique.gross e@wanadoo.fr (Philippe Grosse) wrote:[color=blue]
> Hi,
>
> I don't understand this strange behaviour:
>
> I compile this code :[/color]
[code clipped; it looks good]
....[color=blue]
> /* registration table */
> static struct PyMethodDef testMethods[] = {
> {"return_vector ", return_vector, 1}, /* method name, C
> funcptr, always-tuple */
> {NULL, NULL} /* end of table marker */
> };[/color]

Instead of the magic constant "1", use METH_VARARGS.

....[color=blue]
> Very simple: this module takes a Numeric array (vector) as argument and
> send this array back to python...
>
> If I compile it under macOSX, the result in python is:
>[color=green][color=darkred]
>>>> import test
>>>> from Numeric import *
>>>> v=array([1.0,2.0,3.0,4.0],Float)
>>>> v[/color][/color]
> array([ 1., 2., 3., 4.])[color=green][color=darkred]
>>>> test.return_vec tor(v)[/color][/color]
> array([ 1., 2., 3., 4.])
>
> but in linux the result is:
>[color=green][color=darkred]
>>>> import test
>>>> from Numeric import *
>>>> v=array([1.0,2.0,3.0,4.0],Float)
>>>> v[/color][/color]
> array([ 1., 2., 3., 4.])[color=green][color=darkred]
>>>> test.return_vec tor(v)[/color][/color]
> array([ 1.+0.j, 2.+0.j, 3.+0.j, 4.+0.j])[/color]

This isn't what I get; I get the same result under linux as you get
from Mac OS X. Have you checked that the copies of the code are the
same?

--
|>|\/|<
/--------------------------------------------------------------------------\
|David M. Cooke
|cookedm(at)phy sics(dot)mcmast er(dot)ca