From garyo at genarts.com  Fri Jan 10 14:16:16 2014
From: garyo at genarts.com (Gary Oberbrunner)
Date: Fri, 10 Jan 2014 08:16:16 -0500 (EST)
Subject: [C++-sig] How to handle argument error exceptions in boost.python?
In-Reply-To: <1623265129.4204769.1389359712843.JavaMail.root@genarts.com>
Message-ID: <1279986594.4204774.1389359776309.JavaMail.root@genarts.com>

[I posted this on StackOverflow yesterday, but got nothing.  Hoping you will have some insight!]


I'm writing a C++ python extension.  I have a boost::python extension function

static void EXTrender_effect(EffectGlobals_t *effect_handle, 
                    std::string preset,
            bp::object dest, int xdim, int ydim)
{ ... }
that I export as usual in my boost.python extension:

def("render_effect", EXTrender_effect);

When I call that from python(2.7), I get a C++ exception boost::python::error_already_set. Tracing that down in Visual Studio, I can see it's coming from a boost::python::objects::function::argument_error. So OK, I have an argument error; I'm probably calling it with the wrong args. What I'd like to do is print or throw something sensible in my python extension when this happens, so users of my extension will see the nice message that I know is lurking in PyErr_Fetch. (I can see the message getting built in the boost.python argument_error code.)

But I can't catch that error_already_set; boost.python does that internally around the funcall. And I can't check for python errors in my extension code, since my function never gets called (the argument error is detected by boost.python). What ends up happening is it just silently fails.

What can I hook up so I (or my users) can see the argument error messages? And ideally convert those into python exceptions?

This is all Win7, Python2.7.

-- 
. . . . . . . . . . . . . . . . . . . . . .
Gary Oberbrunner           garyo at genarts.com
VP Engineering             Tel: 617-492-2888
GenArts, Inc.              www.genarts.com 

From legordian at gmx.net  Fri Jan 10 14:45:05 2014
From: legordian at gmx.net (Charly Bicker)
Date: Fri, 10 Jan 2014 14:45:05 +0100 (CET)
Subject: [C++-sig] How to handle argument error exceptions in
 boost.python?
Message-ID: <trinity-dbb4ce3e-82cc-4bda-91c1-088156f38893-1389361505426@3capp-gmx-bs37>

An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20140110/d73bf25e/attachment.html>

From karl.bicker at gordiano.net  Fri Jan 10 14:36:48 2014
From: karl.bicker at gordiano.net (Karl Bicker)
Date: Fri, 10 Jan 2014 14:36:48 +0100 (CET)
Subject: [C++-sig] How to handle argument error exceptions in
 boost.python?
In-Reply-To: <1279986594.4204774.1389359776309.JavaMail.root@genarts.com>
References: <1279986594.4204774.1389359776309.JavaMail.root@genarts.com>
Message-ID: <992310155.9582115.1389361008183.JavaMail.zimbra@gordiano.net>

Dear Gary,

I am not sure if this is what you want, but what about writing a wrapper function around EXTrender_effect which takes boost::python::objects and then use boost::python::extract to check if you can convert the objects you got to what they should be (cf. https://wiki.python.org/moin/boost.python/extract)? Does this help?

Regards,
Charly

----- Original Message -----
From: "Gary Oberbrunner" <garyo at genarts.com>
To: "Development of Python/C++ integration" <cplusplus-sig at python.org>
Sent: Friday, January 10, 2014 2:16:16 PM
Subject: [C++-sig] How to handle argument error exceptions in boost.python?

[I posted this on StackOverflow yesterday, but got nothing.  Hoping you will have some insight!]


I'm writing a C++ python extension.  I have a boost::python extension function

static void EXTrender_effect(EffectGlobals_t *effect_handle, 
                    std::string preset,
            bp::object dest, int xdim, int ydim)
{ ... }
that I export as usual in my boost.python extension:

def("render_effect", EXTrender_effect);

When I call that from python(2.7), I get a C++ exception boost::python::error_already_set. Tracing that down in Visual Studio, I can see it's coming from a boost::python::objects::function::argument_error. So OK, I have an argument error; I'm probably calling it with the wrong args. What I'd like to do is print or throw something sensible in my python extension when this happens, so users of my extension will see the nice message that I know is lurking in PyErr_Fetch. (I can see the message getting built in the boost.python argument_error code.)

But I can't catch that error_already_set; boost.python does that internally around the funcall. And I can't check for python errors in my extension code, since my function never gets called (the argument error is detected by boost.python). What ends up happening is it just silently fails.

What can I hook up so I (or my users) can see the argument error messages? And ideally convert those into python exceptions?

This is all Win7, Python2.7.

-- 
. . . . . . . . . . . . . . . . . . . . . .
Gary Oberbrunner           garyo at genarts.com
VP Engineering             Tel: 617-492-2888
GenArts, Inc.              www.genarts.com 
_______________________________________________
Cplusplus-sig mailing list
Cplusplus-sig at python.org
https://mail.python.org/mailman/listinfo/cplusplus-sig

From garyo at genarts.com  Sat Jan 11 15:52:42 2014
From: garyo at genarts.com (Gary Oberbrunner)
Date: Sat, 11 Jan 2014 09:52:42 -0500 (EST)
Subject: [C++-sig] How to handle argument error exceptions in
 boost.python?
In-Reply-To: <992310155.9582115.1389361008183.JavaMail.zimbra@gordiano.net>
Message-ID: <2032874050.4226045.1389451962920.JavaMail.root@genarts.com>



----- Original Message -----
> From: "Karl Bicker" <karl.bicker at gordiano.net>
> To: "Development of Python/C++ integration" <cplusplus-sig at python.org>
> Sent: Friday, January 10, 2014 8:36:48 AM
> Subject: Re: [C++-sig] How to handle argument error exceptions in boost.python?
> 
> Dear Gary,
> 
> I am not sure if this is what you want, but what about writing a
> wrapper function around EXTrender_effect which takes
> boost::python::objects and then use boost::python::extract to check
> if you can convert the objects you got to what they should be (cf.
> https://wiki.python.org/moin/boost.python/extract)? Does this help?

I suppose it might -- essentially what you're talking about is avoiding boost::python's overloading logic and putting in my own.  Then I could handle arg errors any way I want, because my function would be guaranteed to be called no matter what.  Right?

It seems like a bad solution overall though, because it means most of the coolness of boost::python is gone.  All my functions would just have to take a list of kw args and parse them out themselves, rather than boost::python doing that.  Seems like there must be a better way!  I mean, there is all this logic in the boost::python call framework to build nice error messages about which overloads were tried, but that message never comes out anywhere.

-- 
Gary Oberbrunner

From giuseppe.corbelli at copanitalia.com  Tue Jan 14 09:04:14 2014
From: giuseppe.corbelli at copanitalia.com (Giuseppe Corbelli)
Date: Tue, 14 Jan 2014 09:04:14 +0100
Subject: [C++-sig] How to handle argument error exceptions in
	boost.python?
In-Reply-To: <1279986594.4204774.1389359776309.JavaMail.root@genarts.com>
References: <1279986594.4204774.1389359776309.JavaMail.root@genarts.com>
Message-ID: <52D4EF7E.6030604@copanitalia.com>

On 10/01/2014 14:16, Gary Oberbrunner wrote:
> [I posted this on StackOverflow yesterday, but got nothing.  Hoping you will have some insight!]
>
>
> I'm writing a C++ python extension.  I have a boost::python extension function
>
> static void EXTrender_effect(EffectGlobals_t *effect_handle,
>                      std::string preset,
>              bp::object dest, int xdim, int ydim)
> { ... }
> that I export as usual in my boost.python extension:
>
> def("render_effect", EXTrender_effect);
>
> When I call that from python(2.7), I get a C++ exception
> boost::python::error_already_set. Tracing that down in Visual Studio, I can
> see it's coming from a boost::python::objects::function::argument_error. So
> OK, I have an argument error; I'm probably calling it with the wrong args.
> What I'd like to do is print or throw something sensible in my python
> extension when this happens, so users of my extension will see the nice
> message that I know is lurking in PyErr_Fetch. (I can see the message getting
> built in the boost.python argument_error code.)

What is the exception you see from the python side? something like:

ArgumentError: Python argument types in
     function_name(function_args)
did not match C++ signature:
     cpp_function_name(cpp_args)

-- 
             Giuseppe Corbelli
WASP Software Engineer, Copan Italia S.p.A
Phone: +390303666318  Fax: +390302659932
E-mail: giuseppe.corbelli at copanitalia.com

From garyo at genarts.com  Tue Jan 14 12:45:07 2014
From: garyo at genarts.com (Gary Oberbrunner)
Date: Tue, 14 Jan 2014 06:45:07 -0500 (EST)
Subject: [C++-sig] How to handle argument error exceptions
	in	boost.python?
In-Reply-To: <52D4EF7E.6030604@copanitalia.com>
Message-ID: <1031519715.4270700.1389699907378.JavaMail.root@genarts.com>



----- Original Message -----
> From: "Giuseppe Corbelli" <giuseppe.corbelli at copanitalia.com>
...
> What is the exception you see from the python side? something like:
> 
> ArgumentError: Python argument types in
>      function_name(function_args)
> did not match C++ signature:
>      cpp_function_name(cpp_args)

That's exactly what I'm trying to figure out!  The c++ exception gets handled in handle_exception and it never gets turned into a python exception as far as I can tell.

-- 
Gary Oberbrunner

From herron at ELLINGTON.com  Tue Jan 14 00:01:35 2014
From: herron at ELLINGTON.com (Liam Herron)
Date: Mon, 13 Jan 2014 23:01:35 +0000
Subject: [C++-sig] BoostPython: C++ function with "non-const double
 reference" argument issue. Python looks for float causing argument
 mismatch.
Message-ID: <96426055F003C542B3FB8A0928736DF90EEC37@ex0.ELLINGTON.com>

Here is my example:

-------------------------------------------------------------------------------------------
CPP Code:
-------------------------------------------------------------------------------------------
#include <boost/python.hpp>
using namespace boost::python;


void updateDoubleReference(double& x)
{
  x = x + 5.4321;
}


BOOST_PYTHON_MODULE(_testDR)
{
    // Add regular functions to the module.
    def("updateDoubleReference", &updateDoubleReference );
}
-------------------------------------------------------------------------------------------

-------------------------------------------------------------------------------------------
PY Code:
-------------------------------------------------------------------------------------------
#!/usr/bin/python
#
#

from _testDR import *

x = 3.0
updateDoubleReference(x)
--------------------------------------------------------------------------------------------

-------------------------------------------------------------------------------------------
ERROR OUTPUT
-------------------------------------------------------------------------------------------
python2.5 testDR.py
Traceback (most recent call last):
  File "testDR.py", line 8, in <module>
    updateDoubleReference(x)
Boost.Python.ArgumentError: Python argument types in
    _testDR.updateDoubleReference(float)
did not match C++ signature:
    updateDoubleReference(double {lvalue})
-------------------------------------------------------------------------------------------


How do people typically handle this issue?   Wrap the function with 'float' versions of the arguments?  I tried to create a 'NativeDouble' struct that had a double member attribute but when exposed to python, it simply interprets this as a float and it still has a mismatch.

Any help would be greatly appreciated.

Thanks,
Liam








=============================================================================================
Email transmissions can not be guaranteed to be secure or error-free, as information 
could be intercepted, corrupted, lost, destroyed, arrive late or incomplete, or contain 
viruses.  The sender therefore does not accept liability for any errors or omissions in 
the contents of this message which arise as a result of email transmission.  In addition,
the information contained in this email message is intended only for use of the 
individual or entity named above.  If the reader of this message is not the intended
recipient, or the employee or agent responsible to deliver it to the intended recipient,
you are hereby notified that any dissemination, distribution,or copying of this communication,
disclosure of the parties to it, or any action taken or omitted to be taken in reliance on it,
is strictly prohibited, and may be unlawful.  If you are not the intended recipient please
delete this email message. 
==============================================================================================
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20140113/12d08aa6/attachment.html>

From talljimbo at gmail.com  Tue Jan 14 15:08:21 2014
From: talljimbo at gmail.com (Jim Bosch)
Date: Tue, 14 Jan 2014 09:08:21 -0500
Subject: [C++-sig] How to handle argument error exceptions in
	boost.python?
In-Reply-To: <1031519715.4270700.1389699907378.JavaMail.root@genarts.com>
References: <52D4EF7E.6030604@copanitalia.com>
 <1031519715.4270700.1389699907378.JavaMail.root@genarts.com>
Message-ID: <CAN+0q26o7n+FZegfb=gy+UTjNqskP_aBHzLYVuiDjPFi+-m2AQ@mail.gmail.com>

On Tue, Jan 14, 2014 at 6:45 AM, Gary Oberbrunner <garyo at genarts.com> wrote:
>
>
> ----- Original Message -----
>> From: "Giuseppe Corbelli" <giuseppe.corbelli at copanitalia.com>
> ...
>> What is the exception you see from the python side? something like:
>>
>> ArgumentError: Python argument types in
>>      function_name(function_args)
>> did not match C++ signature:
>>      cpp_function_name(cpp_args)
>
> That's exactly what I'm trying to figure out!  The c++ exception gets handled in handle_exception and it never gets turned into a python exception as far as I can tell.
>

My memory on this is a bit fuzzy, but I think it does get translated
to a custom Boost.Python exception (it's either that or a built-in
Python exception, which would make your task much more difficult, if
not impossible).  Unfortunately there's no clear-cut way to import
that exception.  You could try intentionally raising one in a Python
try/except block when your module is imported, though, so you could
store the exception type and reuse it later in subsequent try/except
blocks.


Jim

From talljimbo at gmail.com  Tue Jan 14 15:13:34 2014
From: talljimbo at gmail.com (Jim Bosch)
Date: Tue, 14 Jan 2014 09:13:34 -0500
Subject: [C++-sig] BoostPython: C++ function with "non-const double
 reference" argument issue. Python looks for float causing argument
 mismatch.
In-Reply-To: <96426055F003C542B3FB8A0928736DF90EEC37@ex0.ELLINGTON.com>
References: <96426055F003C542B3FB8A0928736DF90EEC37@ex0.ELLINGTON.com>
Message-ID: <CAN+0q27hHx-z90OUOSHo5Y1EUbEwk5z9eQ6W1tQgCyApn7EHnw@mail.gmail.com>

The problem here isn't a mismatch between double and float, it's that
Python's float object is immutable, and hence there's no way to
translate it into a C++ lvalue.  You simply can't wrap a C++ function
with a "double &" (or "int &" or "string &") argument to Python with
the same signature, regardless of whether you're using Boost.Python or
something else, because Python won't allow the argument you pass to be
changed.

You have two options:

 - Add an extra struct (similar to your NativeDouble idea) that is a
wrapped class that only contains a double.  But you'll need explicit
get/set methods, not just a public data member, though you can add a
Python property to make the get/set look like a data member.

 - Transform the function signature, by wrapping a new function that
takes a non-reference double and returns the new one, delegating it's
work to the original function.

Jim

From garyo at genarts.com  Tue Jan 14 15:24:44 2014
From: garyo at genarts.com (Gary Oberbrunner)
Date: Tue, 14 Jan 2014 09:24:44 -0500 (EST)
Subject: [C++-sig] How to handle argument error exceptions
	in	boost.python?
In-Reply-To: <CAN+0q26o7n+FZegfb=gy+UTjNqskP_aBHzLYVuiDjPFi+-m2AQ@mail.gmail.com>
Message-ID: <1062900681.4273373.1389709483992.JavaMail.root@genarts.com>



----- Original Message -----
> From: "Jim Bosch" <talljimbo at gmail.com>
...
> My memory on this is a bit fuzzy, but I think it does get translated
> to a custom Boost.Python exception (it's either that or a built-in
> Python exception, which would make your task much more difficult, if
> not impossible).  Unfortunately there's no clear-cut way to import
> that exception.  You could try intentionally raising one in a Python
> try/except block when your module is imported, though, so you could
> store the exception type and reuse it later in subsequent try/except
> blocks.

Thanks Jim; if that's happening, something on the python side must be eating that exception, because I don't see it there.  Still, this is a complex piece of code so it's entirely possible it is being swallowed on the python side.  I'll continue to track it down based on that theory.

-- 
Gary Oberbrunner

From giuseppe.corbelli at copanitalia.com  Tue Jan 14 16:42:39 2014
From: giuseppe.corbelli at copanitalia.com (Giuseppe Corbelli)
Date: Tue, 14 Jan 2014 16:42:39 +0100
Subject: [C++-sig] How to handle argument error exceptions in
	boost.python?
In-Reply-To: <1031519715.4270700.1389699907378.JavaMail.root@genarts.com>
References: <1031519715.4270700.1389699907378.JavaMail.root@genarts.com>
Message-ID: <52D55AEF.404@copanitalia.com>

On 14/01/2014 12:45, Gary Oberbrunner wrote:
> ...
>> What is the exception you see from the python side? something like:
>>
>> ArgumentError: Python argument types in
>>       function_name(function_args)
>> did not match C++ signature:
>>       cpp_function_name(cpp_args)
>
> That's exactly what I'm trying to figure out! The c++ exception gets
> handled  in handle_exception and it never gets turned into a python exception as far as
> I can tell.

So, what you see in the python side? And what boost version?
I've always seen an ArgumentError being raised when passing wrong argument 
types to the C++ code object and ValueError (I think) when the C++ code object 
raises an exception without a specific C++ to python exception translator 
installed.

-- 
             Giuseppe Corbelli
WASP Software Engineer, Copan Italia S.p.A
Phone: +390303666318  Fax: +390302659932
E-mail: giuseppe.corbelli at copanitalia.com

From petr_aleksandrov at mail.ru  Thu Jan 16 14:36:27 2014
From: petr_aleksandrov at mail.ru (=?UTF-8?B?0JDQu9C10LrRgdCw0L3QtNGA0L7QsiDQn9C10YLRgA==?=)
Date: Thu, 16 Jan 2014 17:36:27 +0400
Subject: [C++-sig] Using Numpy C API with boost python numeric
Message-ID: <52D7E05B.5060608@mail.ru>

Is it safe to use direct Numpy C API in Boost Python module together 
with Boost Python numeric::array (also using numpy)?
The function "import_array" must be called before using 
PyArray_SimpeNew. Will it cause problems if the function "import_array" 
will be called twice in one module?

From talljimbo at gmail.com  Fri Jan 17 14:49:34 2014
From: talljimbo at gmail.com (Jim Bosch)
Date: Fri, 17 Jan 2014 08:49:34 -0500
Subject: [C++-sig] Using Numpy C API with boost python numeric
In-Reply-To: <52D7E05B.5060608@mail.ru>
References: <52D7E05B.5060608@mail.ru>
Message-ID: <CAN+0q27GtNh+tQLqsMHKh1jCsowuGKoj5wj3diNBxncvdEiGHA@mail.gmail.com>

I'm pretty sure the numeric::array stuff doesn't actually use the
NumPy C API, and hence it doesn't call import_array.  It just relies
on the Python interface to NumPy objects, which is why it doesn't
support things like accessing the raw data pointers.  So there
shouldn't be any problem with using the NumPy C API with it.

That said, I think it's always good practice to follow the NumPy C API
docs for multiple source files and define a unique preprocessor symbol
to import the NumPy API with, i.e.:

http://docs.scipy.org/doc/numpy/reference/c-api.array.html#miscellaneous

You may also be interested in this project:

https://github.com/ndarray/Boost.NumPy



Jim

From garyo at genarts.com  Tue Jan 21 00:14:46 2014
From: garyo at genarts.com (Gary Oberbrunner)
Date: Mon, 20 Jan 2014 18:14:46 -0500 (EST)
Subject: [C++-sig] create a python object from scratch in boost.python?
In-Reply-To: <367439895.4443291.1390259684599.JavaMail.root@genarts.com>
Message-ID: <903128184.4443292.1390259686007.JavaMail.root@genarts.com>

I'd like to create a python object with some attributes in C++, and return it to python.  Something like this:

    Plugin_Info *pinfo = get_plugin_info_i(i);
    bp::object plugin;  // this creates None; how can I make an empty class?
    plugin.attr("name") = std::string(pinfo->plugin_name);
    plugin.attr("category") = std::string(pinfo->category);
    plugin.attr("n_effects") = (int)pinfo->n_effects;
    plugin.attr("n_inputs") = (int)pinfo->n_inputs;
    plugin.attr("custom_p") = (int)pinfo->custom_p;
    effect_list.append(plugin);

But this fails on the third line, because the "plugin" object is None, and None has no attributes.  Instead of None, I'd like to create a simple empty python class I can add attributes to.  How can I do that?  (Yes, I could use a dict instead but the python side would like to see attributes with values rather than a dict.)

-- 
. . . . . . . . . . . . . . . . . . . . . .
Gary Oberbrunner           garyo at genarts.com
VP Engineering             Tel: 617-492-2888
GenArts, Inc.              www.genarts.com 


From ndbecker2 at gmail.com  Tue Jan 21 02:39:55 2014
From: ndbecker2 at gmail.com (Neal Becker)
Date: Mon, 20 Jan 2014 20:39:55 -0500
Subject: [C++-sig] create a python object from scratch in boost.python?
References: <367439895.4443291.1390259684599.JavaMail.root@genarts.com>
 <903128184.4443292.1390259686007.JavaMail.root@genarts.com>
Message-ID: <lbkj50$kg7$1@ger.gmane.org>

Gary Oberbrunner wrote:

> I'd like to create a python object with some attributes in C++, and return it
> to python.  Something like this:
> 
>     Plugin_Info *pinfo = get_plugin_info_i(i);
>     bp::object plugin;  // this creates None; how can I make an empty class?
>     plugin.attr("name") = std::string(pinfo->plugin_name);
>     plugin.attr("category") = std::string(pinfo->category);
>     plugin.attr("n_effects") = (int)pinfo->n_effects;
>     plugin.attr("n_inputs") = (int)pinfo->n_inputs;
>     plugin.attr("custom_p") = (int)pinfo->custom_p;
>     effect_list.append(plugin);
> 
> But this fails on the third line, because the "plugin" object is None, and
> None has no attributes.  Instead of None, I'd like to create a simple empty
> python class I can add attributes to.  How can I do that?  (Yes, I could use a
> dict instead but the python side would like to see attributes with values
> rather than a dict.)
> 

I think if you've exposed your c++ class, then there is a conversion from c++ 
object to python object.  So

bp::object plugin (Plugin_Info());

I believe would work.


From garyo at genarts.com  Tue Jan 21 14:29:48 2014
From: garyo at genarts.com (Gary Oberbrunner)
Date: Tue, 21 Jan 2014 08:29:48 -0500 (EST)
Subject: [C++-sig] create a python object from scratch in boost.python?
In-Reply-To: <lbkj50$kg7$1@ger.gmane.org>
Message-ID: <2086157964.4451583.1390310988666.JavaMail.root@genarts.com>



----- Original Message -----
> From: "Neal Becker" <ndbecker2 at gmail.com>
> To: cplusplus-sig at python.org
> Sent: Monday, January 20, 2014 8:39:55 PM
> Subject: Re: [C++-sig] create a python object from scratch in boost.python?
> 
> Gary Oberbrunner wrote:
> 
> > I'd like to create a python object with some attributes in C++, and
> > return it
> > to python. ...  Instead of None, I'd like to create a
> > simple empty
> > python class I can add attributes to....
> 
> I think if you've exposed your c++ class, then there is a conversion
> from c++ object to python object.  So
> 
> bp::object plugin (Plugin_Info());

Thanks -- I can indeed do that (or a proxy C++ class in any case).  I was wondering though if it's possible to create an anonymous python class in BP, without a "starter" C++ class as it were.  But I suppose that's mostly an academic question right now.  

-- 
Gary Oberbrunner

From nikolay.mladenov at gmail.com  Tue Jan 21 14:55:16 2014
From: nikolay.mladenov at gmail.com (Nikolay Mladenov)
Date: Tue, 21 Jan 2014 08:55:16 -0500
Subject: [C++-sig] create a python object from scratch in boost.python?
In-Reply-To: <2086157964.4451583.1390310988666.JavaMail.root@genarts.com>
References: <lbkj50$kg7$1@ger.gmane.org>
 <2086157964.4451583.1390310988666.JavaMail.root@genarts.com>
Message-ID: <CACbdn-+hHOP7s6T+Jrw6gu_OHDqrGXoDEi_yB-jWr=tZb_yAQQ@mail.gmail.com>

My be construct bp::object from the result of _PyObject_New?


On Tue, Jan 21, 2014 at 8:29 AM, Gary Oberbrunner <garyo at genarts.com> wrote:

>
>
> ----- Original Message -----
> > From: "Neal Becker" <ndbecker2 at gmail.com>
> > To: cplusplus-sig at python.org
> > Sent: Monday, January 20, 2014 8:39:55 PM
> > Subject: Re: [C++-sig] create a python object from scratch in
> boost.python?
> >
> > Gary Oberbrunner wrote:
> >
> > > I'd like to create a python object with some attributes in C++, and
> > > return it
> > > to python. ...  Instead of None, I'd like to create a
> > > simple empty
> > > python class I can add attributes to....
> >
> > I think if you've exposed your c++ class, then there is a conversion
> > from c++ object to python object.  So
> >
> > bp::object plugin (Plugin_Info());
>
> Thanks -- I can indeed do that (or a proxy C++ class in any case).  I was
> wondering though if it's possible to create an anonymous python class in
> BP, without a "starter" C++ class as it were.  But I suppose that's mostly
> an academic question right now.
>
> --
> Gary Oberbrunner
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> https://mail.python.org/mailman/listinfo/cplusplus-sig
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20140121/89c9499d/attachment.html>

From garyo at genarts.com  Tue Jan 21 16:37:01 2014
From: garyo at genarts.com (Gary Oberbrunner)
Date: Tue, 21 Jan 2014 10:37:01 -0500 (EST)
Subject: [C++-sig] create a python object from scratch in boost.python?
In-Reply-To: <CACbdn-+hHOP7s6T+Jrw6gu_OHDqrGXoDEi_yB-jWr=tZb_yAQQ@mail.gmail.com>
Message-ID: <2079615492.4456011.1390318621152.JavaMail.root@genarts.com>

...
> 
> > > I think if you've exposed your c++ class, then there is a conversion
> > > from c++ object to python object. So
> > > bp::object plugin (Plugin_Info());
> 

> > Thanks -- I can indeed do that (or a proxy C++ class in any case).
> > I
> > was wondering though if it's possible to create an anonymous python
> > class in BP, without a "starter" C++ class as it were. But I
> > suppose that's mostly an academic question right now.

Well, it turns out I really do want to do this without a predefined C++ class, or something like that.

I want to have the C++ code be easily extensible, so I don't have to repeat myself too much when adding new attributes.  I tried making the proxy C++ class and exposing it, but it's very verbose and hard to maintain.

I switched to using a dict, and what I have now is almost there:

static bp::list EXTget_effects()
{
  bp::list effect_list;
  int n_effects = get_n_sapphire_effects();
  for (int i = 0; i < n_effects; i++) {
    Plugin_Info *pinfo = get_plugin_info_i(i);
    bp::dict ppinfo;
    ppinfo["name"] = (std::string)pinfo->plugin_name;
    ppinfo["category"] = (std::string)pinfo->category;
    ppinfo["n_effects"] = (int)pinfo->n_effects;
    ppinfo["n_inputs"] = (int)pinfo->n_inputs;
    ppinfo["custom_p"] = (int)pinfo->custom_p;
    effect_list.append(ppinfo);
  }
  return effect_list;
}

So now it returns a python list of python dicts.  All I want now is to override __getattr__ on each ppinfo dict so it returns the dict value as the attribute value, so in python I can reference effect_list[i].name instead of effect_list[i]['name'].  Or any alternative way to get the same effect.  Is that possible?

-- 
Gary Oberbrunner

From stefan at seefeld.name  Tue Jan 21 16:45:52 2014
From: stefan at seefeld.name (Stefan Seefeld)
Date: Tue, 21 Jan 2014 10:45:52 -0500
Subject: [C++-sig] create a python object from scratch in boost.python?
In-Reply-To: <2079615492.4456011.1390318621152.JavaMail.root@genarts.com>
References: <2079615492.4456011.1390318621152.JavaMail.root@genarts.com>
Message-ID: <52DE9630.3010907@seefeld.name>

On 01/21/2014 10:37 AM, Gary Oberbrunner wrote:
> So now it returns a python list of python dicts. All I want now is to
> override __getattr__ on each ppinfo dict so it returns the dict value
> as the attribute value, so in python I can reference
> effect_list[i].name instead of effect_list[i]['name']. Or any
> alternative way to get the same effect. Is that possible? 
There are many ways to do that elegantly in Python. For example:

class Effect:
  def __init__(self, mydict):
    self.__dict__.update(mydict)

If you want to combine that with existing classes / APIs, you could also
consider merging the new attributes in via metaclasses.

        Stefan

-- 

      ...ich hab' noch einen Koffer in Berlin...


From garyo at genarts.com  Tue Jan 21 17:03:48 2014
From: garyo at genarts.com (Gary Oberbrunner)
Date: Tue, 21 Jan 2014 11:03:48 -0500 (EST)
Subject: [C++-sig] create a python object from scratch in boost.python?
In-Reply-To: <52DE9630.3010907@seefeld.name>
Message-ID: <138792326.4456593.1390320228445.JavaMail.root@genarts.com>



----- Original Message -----
> From: "Stefan Seefeld" <stefan at seefeld.name>
> To: "Development of Python/C++ integration" <cplusplus-sig at python.org>
> Sent: Tuesday, January 21, 2014 10:45:52 AM
> Subject: Re: [C++-sig] create a python object from scratch in boost.python?
> 
> On 01/21/2014 10:37 AM, Gary Oberbrunner wrote:
> > So now it returns a python list of python dicts. All I want now is
> > to
> > override __getattr__ on each ppinfo dict so it returns the dict
> > value
> > as the attribute value, so in python I can reference
> > effect_list[i].name instead of effect_list[i]['name']. Or any
> > alternative way to get the same effect. Is that possible?
> There are many ways to do that elegantly in Python. For example:
> 
> class Effect:
>   def __init__(self, mydict):
>     self.__dict__.update(mydict)

I like the idea of using self.__dict__.  If that's accessible from C++, then I suppose I could have an empty C++ class, export that as a simple container, and dynamically add attributes to it by manipulating its __dict__.  Does that seem possible?

-- 
Gary Oberbrunner

From stefan at seefeld.name  Tue Jan 21 17:35:20 2014
From: stefan at seefeld.name (Stefan Seefeld)
Date: Tue, 21 Jan 2014 11:35:20 -0500
Subject: [C++-sig] create a python object from scratch in boost.python?
In-Reply-To: <138792326.4456593.1390320228445.JavaMail.root@genarts.com>
References: <138792326.4456593.1390320228445.JavaMail.root@genarts.com>
Message-ID: <52DEA1C8.9040705@seefeld.name>

On 01/21/2014 11:03 AM, Gary Oberbrunner wrote:
>
> ----- Original Message -----
>> From: "Stefan Seefeld" <stefan at seefeld.name>
>> To: "Development of Python/C++ integration" <cplusplus-sig at python.org>
>> Sent: Tuesday, January 21, 2014 10:45:52 AM
>> Subject: Re: [C++-sig] create a python object from scratch in boost.python?
>>
>> On 01/21/2014 10:37 AM, Gary Oberbrunner wrote:
>>> So now it returns a python list of python dicts. All I want now is
>>> to
>>> override __getattr__ on each ppinfo dict so it returns the dict
>>> value
>>> as the attribute value, so in python I can reference
>>> effect_list[i].name instead of effect_list[i]['name']. Or any
>>> alternative way to get the same effect. Is that possible?
>> There are many ways to do that elegantly in Python. For example:
>>
>> class Effect:
>>   def __init__(self, mydict):
>>     self.__dict__.update(mydict)
> I like the idea of using self.__dict__.  If that's accessible from C++, then I suppose I could have an empty C++ class, export that as a simple container, and dynamically add attributes to it by manipulating its __dict__.  Does that seem possible?
>

Why do you want to export an empty C++ class ? Do you have an API that
would take that as argument or return type ? If not, I see no point in
doing that, as all that does is provide implicit type conversion for you.
It seems what you really want is a way for your C++ code to manipulate
pure Python objects, without any automatic type conversion.
So the cleanest way to do that would be to use the above definition of
the "Effect" class (in Python), import that into your C++ runtime (via
bpl::import()), and then instantiate an "Effect" object with your
pre-defined dict object.

  bpl::object effect_type = bpl::import("Effect"); // Define Effect type
in C++ runtime
  bpl::object effect = effect_type(ppinfo);  // Instantiate Effect
object with pre-defined attribute dictionary
  ...


    Stefan

-- 

      ...ich hab' noch einen Koffer in Berlin...


From garyo at genarts.com  Tue Jan 21 19:52:06 2014
From: garyo at genarts.com (Gary Oberbrunner)
Date: Tue, 21 Jan 2014 13:52:06 -0500 (EST)
Subject: [C++-sig] create a python object from scratch in boost.python?
In-Reply-To: <52DEA1C8.9040705@seefeld.name>
Message-ID: <1752024852.4460451.1390330326308.JavaMail.root@genarts.com>



----- Original Message -----
> From: "Stefan Seefeld" <stefan at seefeld.name>
...
> It seems what you really want is a way for your C++ code to
> manipulate pure Python objects, without any automatic type conversion.

Yes!

> So the cleanest way to do that would be to use the above definition of
> the "Effect" class (in Python), import that into your C++ runtime
> (via bpl::import()), and then instantiate an "Effect" object with your
> pre-defined dict object.

I'm trying to do this in a boost python extension, so it has to be all in C++.  If I have to wrap some python around it, it'll become a bit too heavyweight.  

I'd love to do what you suggest above, if it can be done all in C++.  I think all I need is a way to create, in C++, an empty Python class object with a __dict__.  Then, as you say, I can add attributes to it (at least I think I should be able to - but I'm not sure how far the boost.python world goes).

-- 
Gary Oberbrunner

From stefan at seefeld.name  Tue Jan 21 23:07:54 2014
From: stefan at seefeld.name (Stefan Seefeld)
Date: Tue, 21 Jan 2014 17:07:54 -0500
Subject: [C++-sig] create a python object from scratch in boost.python?
In-Reply-To: <1752024852.4460451.1390330326308.JavaMail.root@genarts.com>
References: <1752024852.4460451.1390330326308.JavaMail.root@genarts.com>
Message-ID: <52DEEFBA.9090601@seefeld.name>

On 01/21/2014 01:52 PM, Gary Oberbrunner wrote:
>
> ----- Original Message -----
>> From: "Stefan Seefeld" <stefan at seefeld.name>
> ...
>> It seems what you really want is a way for your C++ code to
>> manipulate pure Python objects, without any automatic type conversion.
> Yes!
>
>> So the cleanest way to do that would be to use the above definition of
>> the "Effect" class (in Python), import that into your C++ runtime
>> (via bpl::import()), and then instantiate an "Effect" object with your
>> pre-defined dict object.
> I'm trying to do this in a boost python extension, so it has to be all in C++.  If I have to wrap some python around it, it'll become a bit too heavyweight.  

Have a look at the attached code; I don't think that counts as
heavy-weight. In particular, trying to do the same without  embedding a
little "script" would be just more cumbersome, if it worked at all.


> I'd love to do what you suggest above, if it can be done all in C++.  I think all I need is a way to create, in C++, an empty Python class object with a __dict__.  Then, as you say, I can add attributes to it (at least I think I should be able to - but I'm not sure how far the boost.python world goes).

Compile the attached as a stand-alone app. I think it demonstrates what
you want, so you should be able to easily adjust it to your use-case.

        Stefan


-- 

      ...ich hab' noch einen Koffer in Berlin...

-------------- next part --------------
A non-text attachment was scrubbed...
Name: bpl2.cc
Type: text/x-c++src
Size: 889 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20140121/39611e43/attachment.cc>

From garyo at genarts.com  Tue Jan 21 23:24:15 2014
From: garyo at genarts.com (Gary Oberbrunner)
Date: Tue, 21 Jan 2014 17:24:15 -0500 (EST)
Subject: [C++-sig] create a python object from scratch in boost.python?
In-Reply-To: <52DEEFBA.9090601@seefeld.name>
Message-ID: <831082552.4463977.1390343055177.JavaMail.root@genarts.com>



----- Original Message -----
> From: "Stefan Seefeld" <stefan at seefeld.name>

> Have a look at the attached code; I don't think that counts as
> heavy-weight. In particular, trying to do the same without  embedding
> a
> little "script" would be just more cumbersome, if it worked at all.

Whoa, that is cool!  I had no idea boost python could do this.  Thanks, Stefan!

(It's like "python-ception": python calls C++ which calls python...)

-- 
Gary Oberbrunner

From stefan at seefeld.name  Tue Jan 21 23:34:00 2014
From: stefan at seefeld.name (Stefan Seefeld)
Date: Tue, 21 Jan 2014 17:34:00 -0500
Subject: [C++-sig] create a python object from scratch in boost.python?
In-Reply-To: <831082552.4463977.1390343055177.JavaMail.root@genarts.com>
References: <831082552.4463977.1390343055177.JavaMail.root@genarts.com>
Message-ID: <52DEF5D8.6050808@seefeld.name>

On 01/21/2014 05:24 PM, Gary Oberbrunner wrote:
>
> ----- Original Message -----
>> From: "Stefan Seefeld" <stefan at seefeld.name>
>> Have a look at the attached code; I don't think that counts as
>> heavy-weight. In particular, trying to do the same without  embedding
>> a
>> little "script" would be just more cumbersome, if it worked at all.
> Whoa, that is cool!  I had no idea boost python could do this.  Thanks, Stefan!
>
> (It's like "python-ception": python calls C++ which calls python...)

Yeah, the boundary between the two languages really becomes blurry,
precisely because the two really share the same runtime, so passing
objects across the boundary becomes so seamless you can almost forget
it's there. The beauty of hybrid programming... :-)

    Stefan

-- 

      ...ich hab' noch einen Koffer in Berlin...


From arm at rfh-koeln.de  Tue Jan 28 17:33:18 2014
From: arm at rfh-koeln.de (=?iso-8859-1?Q?Peter_Armbr=FCster?=)
Date: Tue, 28 Jan 2014 17:33:18 +0100
Subject: [C++-sig] PhysX3.x Wrapping Problems
Message-ID: <039301cf1c46$a6c0b7d0$f4422770$@rfh-koeln.de>

Dear Boost.python, C++ and Py++ Experts,

 

this is my first post in your excellent Forum.

 

I have chosen as the wrapper software the boost.python/Py++ Package to write
a physx 3.x Wrapper, because I think it is the best.

Many many thanks to David Abrahams and Roman Yakovenko for the great Tools.

 

The wrapping of old versions physx 2.x had made no problems.

And now I come across a  very big problem at the 3.x Versions.

All wrappings of classes works very well.

The problem only exists in the PhysX free functions, which are responsible
for the initialization of PhysX.

 

I try to explain with a reduced Example.

The initialization of PhysX starts with create of foundation :

 

0 static PxFoundation *mFoundation = NULL;

1 static PxDefaultErrorCallback gDefaultErrorCallback; 
2 static PxDefaultAllocator gDefaultAllocatorCallback; 
3 mFoundation = PxCreateFoundation(PX_PHYSICS_VERSION,
gDefaultAllocatorCallback, gDefaultErrorCallback); 4
4 if(!mFoundation) 

5   fatalError("PxCreateFoundation failed!"); 

 

The wrapped python program generate a runtime error at line 3: RuntimeError:
Access violation - no RTTI data!

Its impossible to get mFoundation.

 

 

 

Notes: 

- physx comes as SDK:  only Headers, libs and dlls.

- PxFoundation is a Abstract class:

 

class PX_FOUNDATION_API PxFoundation
{
public:
   virtual   void release() = 0;
   virtual PxErrorCallback& getErrorCallback() const = 0;
   virtual void setErrorLevel(PxErrorCode::Enum mask =
PxErrorCode::eMASK_ALL) = 0;
   virtual PxErrorCode::Enum getErrorLevel() const = 0;
   virtual PxBroadcastingAllocator& getAllocator() const = 0;
   virtual PxAllocatorCallback& getAllocatorCallback() const = 0;
   virtual bool getReportAllocationNames() const = 0;
   virtual void setReportAllocationNames(bool value) = 0;

   virtual ~PxFoundation() {}
};

 

- The wrapping of PxFoundation makes no problem.

 

 

To make sure that I made no mistake with the  initialization of PhysX, I
tried a different solution.

A small function InitPhysX create the foundation, and  GetFoundation get
this.

 

C++ Code:


using namespace physx;

PxU32 PHYSICS_VERSION = PX_PHYSICS_VERSION;  
static PxFoundation *mFoundation = NULL;
static PxDefaultErrorCallback gDefaultErrorCallback;
static PxDefaultAllocator gDefaultAllocatorCallback;


PxFoundation* GetFoundation(){
   return mFoundation;
}

void InitPhysX() {
   mFoundation = PxCreateFoundation(
      PX_PHYSICS_VERSION,
      gDefaultAllocatorCallback,
      gDefaultErrorCallback);

   if(!mFoundation)
      printf("ERROR: PxCreateFoundation failed!\n");

   else
      printf("PxCreateFoundation created!\n");
}

 

The wrapped python program: 

 

1 import _physx324_d as PX

2 PX.InitPhysX()

3 foundation = PX.GetFoundation()

 

I get on line 2 the message: "PxCreateFoundation created!" 
=> OK!

 

On Line 3 the same error RuntimeError: Access violation - no RTTI data!

 

Now i check the wrappings of PxCreateFoundation in
_physx324_d_free_functions.pypp.cpp

 

void register_free_functions(){

    . . .

 

    { //::GetFoundation
    
        typedef ::physx::PxFoundation * ( *GetFoundation_function_type )(
);
        
        bp::def( 
            "GetFoundation"
            , GetFoundation_function_type( &::GetFoundation )
            , bp::return_value_policy< bp::reference_existing_object >() );
    
    }

 

    . . . 

 

 

I think: return_value_policy< bp::reference_existing_object > thats right!?

 

Then a debug session with PTVS, start with a break point in InitPhysX 

and inspect the mFoundation after the successful generation.

Its a vptr with 9 Entries => thats right, PxFoundation has 9 Methods.

That's all what I can see, because I have no source code and no pdb
(symbols).

 

I place the next Breakpoint in

PxFoundation* GetFoundation(){
   return mFoundation;
}

 

to search the Error location.

 

Then followed a trace trip through the code, for me is a journey through an
unreadable jungles of wrapper C + + code.

Unfortunately, I am not an C++Expert.

The Error is not to find for me.

 

 

The same problem occurs with PxCreatePhysics.

 

My work enviroment:

?  Windows 7 x64

?  Python 2.7

?  physxVersion 3.24  (i try: 3.01, 3.23, 3.3, same problem) 

?  visualStudioVersion 11 (i try: VS 9 and  VS 10,  same problem)  

?  PythonTools for Visual Studio (PTVS)

?  Boost 1.55 (i try: 1.45,  same problem) 

?  gccxml 0.9

?  pyplusplus 1.0 svn 1.5.1

?  pygccxml 1.0 svn 1.5.1

 

 

 

What is my mistake?

I would be very grateful if someone could an expert help.

Thanks for your response.

 

    Peter Armold

 

 

 

 

 

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20140128/0b2f2743/attachment-0001.html>

From niki at vintech.bg  Wed Jan 29 09:19:59 2014
From: niki at vintech.bg (niki)
Date: Wed, 29 Jan 2014 10:19:59 +0200
Subject: [C++-sig] PhysX3.x Wrapping Problems
In-Reply-To: <039301cf1c46$a6c0b7d0$f4422770$@rfh-koeln.de>
References: <039301cf1c46$a6c0b7d0$f4422770$@rfh-koeln.de>
Message-ID: <52E8B9AF.4050602@vintech.bg>

On 28.01.2014 18:33, Peter Armbr?ster wrote:
> Dear Boost.python, C++ and Py++ Experts,
>
> this is my first post in your excellent Forum.
>
> I have chosen as the wrapper software the boost.python/Py++ Packageto
> write a physx 3.x Wrapper, because I think it is the best.
>
> Many many thanks to David Abrahams and Roman Yakovenko for the great Tools.
>
> The wrapping of old versions physx 2.x had made no problems.
>
> And now I come across a  very big problem at the 3.x Versions.
>
> All wrappings of classes works very well.
>
> The problem only exists in the PhysX free functions, which are
> responsible for the initialization of PhysX.
>
> I try to explain with a reduced Example.
>
> The initialization of PhysX starts with create of foundation :
>
> *0 static*PxFoundation*mFoundation=NULL;
>
> *1 static*PxDefaultErrorCallbackgDefaultErrorCallback;
> *2 static*PxDefaultAllocatorgDefaultAllocatorCallback;
> 3
> mFoundation=PxCreateFoundation(PX_PHYSICS_VERSION,gDefaultAllocatorCallback,gDefaultErrorCallback);4
> *4 if*(!mFoundation)
>
> 5   fatalError("PxCreateFoundation failed!");
>
> The wrapped python program generate a runtime error at line
> 3:RuntimeError: Access violation - no RTTI data!

Just wild guess: add compiler option to enable RTTI

HTH
Niki

From giuseppe.corbelli at copanitalia.com  Wed Jan 29 11:03:00 2014
From: giuseppe.corbelli at copanitalia.com (Giuseppe Corbelli)
Date: Wed, 29 Jan 2014 11:03:00 +0100
Subject: [C++-sig] PhysX3.x Wrapping Problems
In-Reply-To: <039301cf1c46$a6c0b7d0$f4422770$@rfh-koeln.de>
References: <039301cf1c46$a6c0b7d0$f4422770$@rfh-koeln.de>
Message-ID: <52E8D1D4.7060003@copanitalia.com>

On 28/01/2014 17:33, Peter Armbr?ster wrote:
> Dear Boost.python, C++ and Py++ Experts,
>
> this is my first post in your excellent Forum.
>
> I have chosen as the wrapper software the boost.python/Py++ Packageto write a
> physx 3.x Wrapper, because I think it is the best.
>
> Many many thanks to David Abrahams and Roman Yakovenko for the great Tools.
>
> The wrapping of old versions physx 2.x had made no problems.
>
> And now I come across a  very big problem at the 3.x Versions.
>
> All wrappings of classes works very well.
>
> The problem only exists in the PhysX free functions, which are responsible for
> the initialization of PhysX.
>
> I try to explain with a reduced Example.
>
> The initialization of PhysX starts with create of foundation :
>
> *0 static*PxFoundation*mFoundation=NULL;
>
> *1 static*PxDefaultErrorCallbackgDefaultErrorCallback;
> *2 static*PxDefaultAllocatorgDefaultAllocatorCallback;
> 3
> mFoundation=PxCreateFoundation(PX_PHYSICS_VERSION,gDefaultAllocatorCallback,gDefaultErrorCallback);4
> *4 if*(!mFoundation)
>
> 5   fatalError("PxCreateFoundation failed!");
>
> The wrapped python program generate a runtime error at line 3:RuntimeError:
> Access violation - no RTTI data!
>
> Its impossible to get mFoundation.
>
> *Notes: *
>
> - physx comes as SDK:  only Headers, libs and dlls.

Before reading on I'd investigate the "no RTTI data" message.
Make sure you're using the /GR compiler option.

-- 
             Giuseppe Corbelli
WASP Software Engineer, Copan Italia S.p.A
Phone: +390303666318  Fax: +390302659932
E-mail: giuseppe.corbelli at copanitalia.com

From arm at rfh-koeln.de  Wed Jan 29 12:26:18 2014
From: arm at rfh-koeln.de (=?iso-8859-1?Q?Peter_Armbr=FCster?=)
Date: Wed, 29 Jan 2014 12:26:18 +0100
Subject: [C++-sig] Cplusplus-sig Digest, Vol 64, Issue 11
In-Reply-To: <mailman.41.1390993202.22219.cplusplus-sig@python.org>
References: <mailman.41.1390993202.22219.cplusplus-sig@python.org>
Message-ID: <03d701cf1ce4$ee69dc30$cb3d9490$@rfh-koeln.de>

Hi Niki and Giuseppe,

thanks for your answers, but once I had already considered .

My compiler flags are:  /DWIN32 /DDEBUG /GR  /RTCsu /DPX_DEBUG  

I wish you a nice day.

Peter




-----Urspr?ngliche Nachricht-----
Von: Cplusplus-sig
[mailto:cplusplus-sig-bounces+arm=rfh-koeln.de at python.org] Im Auftrag von
cplusplus-sig-request at python.org
Gesendet: Mittwoch, 29. Januar 2014 12:00
An: cplusplus-sig at python.org
Betreff: Cplusplus-sig Digest, Vol 64, Issue 11

Send Cplusplus-sig mailing list submissions to
	cplusplus-sig at python.org

To subscribe or unsubscribe via the World Wide Web, visit
	https://mail.python.org/mailman/listinfo/cplusplus-sig
or, via email, send a message with subject or body 'help' to
	cplusplus-sig-request at python.org

You can reach the person managing the list at
	cplusplus-sig-owner at python.org

When replying, please edit your Subject line so it is more specific than
"Re: Contents of Cplusplus-sig digest..."


Today's Topics:

   1. Re: PhysX3.x Wrapping Problems (niki)
   2. Re: PhysX3.x Wrapping Problems (Giuseppe Corbelli)


----------------------------------------------------------------------

Message: 1
Date: Wed, 29 Jan 2014 10:19:59 +0200
From: niki <niki at vintech.bg>
To: Development of Python/C++ integration <cplusplus-sig at python.org>
Subject: Re: [C++-sig] PhysX3.x Wrapping Problems
Message-ID: <52E8B9AF.4050602 at vintech.bg>
Content-Type: text/plain; charset=UTF-8; format=flowed

On 28.01.2014 18:33, Peter Armbr?ster wrote:
> Dear Boost.python, C++ and Py++ Experts,
>
> this is my first post in your excellent Forum.
>
> I have chosen as the wrapper software the boost.python/Py++ Packageto 
> write a physx 3.x Wrapper, because I think it is the best.
>
> Many many thanks to David Abrahams and Roman Yakovenko for the great
Tools.
>
> The wrapping of old versions physx 2.x had made no problems.
>
> And now I come across a  very big problem at the 3.x Versions.
>
> All wrappings of classes works very well.
>
> The problem only exists in the PhysX free functions, which are 
> responsible for the initialization of PhysX.
>
> I try to explain with a reduced Example.
>
> The initialization of PhysX starts with create of foundation :
>
> *0 static*PxFoundation*mFoundation=NULL;
>
> *1 static*PxDefaultErrorCallbackgDefaultErrorCallback;
> *2 static*PxDefaultAllocatorgDefaultAllocatorCallback;
> 3
> mFoundation=PxCreateFoundation(PX_PHYSICS_VERSION,gDefaultAllocatorCal
> lback,gDefaultErrorCallback);4
> *4 if*(!mFoundation)
>
> 5   fatalError("PxCreateFoundation failed!");
>
> The wrapped python program generate a runtime error at line
> 3:RuntimeError: Access violation - no RTTI data!

Just wild guess: add compiler option to enable RTTI

HTH
Niki


------------------------------

Message: 2
Date: Wed, 29 Jan 2014 11:03:00 +0100
From: Giuseppe Corbelli <giuseppe.corbelli at copanitalia.com>
To: Development of Python/C++ integration <cplusplus-sig at python.org>
Subject: Re: [C++-sig] PhysX3.x Wrapping Problems
Message-ID: <52E8D1D4.7060003 at copanitalia.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed

On 28/01/2014 17:33, Peter Armbr?ster wrote:
> Dear Boost.python, C++ and Py++ Experts,
>
> this is my first post in your excellent Forum.
>
> I have chosen as the wrapper software the boost.python/Py++ Packageto 
> write a physx 3.x Wrapper, because I think it is the best.
>
> Many many thanks to David Abrahams and Roman Yakovenko for the great
Tools.
>
> The wrapping of old versions physx 2.x had made no problems.
>
> And now I come across a  very big problem at the 3.x Versions.
>
> All wrappings of classes works very well.
>
> The problem only exists in the PhysX free functions, which are 
> responsible for the initialization of PhysX.
>
> I try to explain with a reduced Example.
>
> The initialization of PhysX starts with create of foundation :
>
> *0 static*PxFoundation*mFoundation=NULL;
>
> *1 static*PxDefaultErrorCallbackgDefaultErrorCallback;
> *2 static*PxDefaultAllocatorgDefaultAllocatorCallback;
> 3
> mFoundation=PxCreateFoundation(PX_PHYSICS_VERSION,gDefaultAllocatorCal
> lback,gDefaultErrorCallback);4
> *4 if*(!mFoundation)
>
> 5   fatalError("PxCreateFoundation failed!");
>
> The wrapped python program generate a runtime error at line
3:RuntimeError:
> Access violation - no RTTI data!
>
> Its impossible to get mFoundation.
>
> *Notes: *
>
> - physx comes as SDK:  only Headers, libs and dlls.

Before reading on I'd investigate the "no RTTI data" message.
Make sure you're using the /GR compiler option.

-- 
             Giuseppe Corbelli
WASP Software Engineer, Copan Italia S.p.A
Phone: +390303666318  Fax: +390302659932
E-mail: giuseppe.corbelli at copanitalia.com


------------------------------

Subject: Digest Footer

_______________________________________________
Cplusplus-sig mailing list
Cplusplus-sig at python.org
https://mail.python.org/mailman/listinfo/cplusplus-sig

------------------------------

End of Cplusplus-sig Digest, Vol 64, Issue 11
*********************************************


From nikolay.mladenov at gmail.com  Wed Jan 29 16:21:50 2014
From: nikolay.mladenov at gmail.com (Nikolay Mladenov)
Date: Wed, 29 Jan 2014 10:21:50 -0500
Subject: [C++-sig] PhysX3.x Wrapping Problems
In-Reply-To: <039301cf1c46$a6c0b7d0$f4422770$@rfh-koeln.de>
References: <039301cf1c46$a6c0b7d0$f4422770$@rfh-koeln.de>
Message-ID: <CACbdn-J9sPWzS4cVqB9BPnJ5A=ecz__c2qnbFoQzRtj=F0O2mw@mail.gmail.com>

Recently I had such symptom - it went away when I turned "Whole Program
Optimization" off


On Tue, Jan 28, 2014 at 11:33 AM, Peter Armbr?ster <arm at rfh-koeln.de> wrote:

> Dear Boost.python, C++ and Py++ Experts,
>
>
>
> this is my first post in your excellent Forum.
>
>
>
> I have chosen as the wrapper software the boost.python/Py++ Package to
> write a physx 3.x Wrapper, because I think it is the best.
>
> Many many thanks to David Abrahams and Roman Yakovenko for the great Tools.
>
>
>
> The wrapping of old versions physx 2.x had made no problems.
>
> And now I come across a  very big problem at the 3.x Versions.
>
> All wrappings of classes works very well.
>
> The problem only exists in the PhysX free functions, which are responsible
> for the initialization of PhysX.
>
>
>
> I try to explain with a reduced Example.
>
> The initialization of PhysX starts with create of foundation :
>
>
>
> *0 static* PxFoundation *mFoundation = NULL;
>
> *1 static* PxDefaultErrorCallback gDefaultErrorCallback;
> *2 static* PxDefaultAllocator gDefaultAllocatorCallback;
> 3 mFoundation = PxCreateFoundation(PX_PHYSICS_VERSION,
> gDefaultAllocatorCallback, gDefaultErrorCallback); 4
> *4 if*(!mFoundation)
>
> 5   fatalError("PxCreateFoundation failed!");
>
>
>
> The wrapped python program generate a runtime error at line 3:RuntimeError: Access violation - no RTTI data!
>
> Its impossible to get mFoundation.
>
>
>
>
>
>
>
> *Notes: *
>
> - physx comes as SDK:  only Headers, libs and dlls.
>
> - PxFoundation is a Abstract class:
>
>
>
> *class* PX_FOUNDATION_API PxFoundation
> {
> *public*:
>    *virtual*   *void* release() = 0;
>    *virtual* PxErrorCallback& getErrorCallback() *const* = 0;
>    *virtual* *void* setErrorLevel(PxErrorCode::Enum mask =
> PxErrorCode::eMASK_ALL) = 0;
>    *virtual* PxErrorCode::Enum getErrorLevel() *const* = 0;
>    *virtual* PxBroadcastingAllocator& getAllocator() *const* = 0;
>    *virtual* PxAllocatorCallback& getAllocatorCallback() *const* = 0;
>    *virtual* *bool* getReportAllocationNames() *const* = 0;
>    *virtual* *void* setReportAllocationNames(*bool* value) = 0;
>
>    *virtual* ~PxFoundation() {}
> };
>
>
>
> - The wrapping of PxFoundation makes no problem.
>
>
>
>
>
> To make sure that I made no mistake with the  initialization of PhysX, I
> tried a different solution.
>
> A small function InitPhysX create the foundation, and  GetFoundation get
> this.
>
>
>
> C++ Code:
>
>
> *using* *namespace* physx;
>
> PxU32 PHYSICS_VERSION = PX_PHYSICS_VERSION;
> *static* PxFoundation *mFoundation = NULL;
> *static* PxDefaultErrorCallback gDefaultErrorCallback;
> *static* PxDefaultAllocator gDefaultAllocatorCallback;
>
>
> PxFoundation* GetFoundation(){
>    *return* mFoundation;
> }
>
> *void* InitPhysX() {
>    mFoundation = PxCreateFoundation(
>       PX_PHYSICS_VERSION,
>       gDefaultAllocatorCallback,
>       gDefaultErrorCallback);
>
>    *if*(!mFoundation)
>       printf("ERROR: PxCreateFoundation failed!\n");
>
>
> *else*      printf("PxCreateFoundation created!\n");
> }
>
>
>
> The wrapped python program:
>
>
>
> 1 import _physx324_d as PX
>
> 2 PX.InitPhysX()
>
> 3 foundation = PX.GetFoundation()
>
>
>
> I get on line 2 the message: "PxCreateFoundation created!"
> => OK!
>
>
>
> On Line 3 the same error RuntimeError: Access violation - no RTTI data!
>
>
>
> Now i check the wrappings of PxCreateFoundation in
> _physx324_d_free_functions.pypp.cpp
>
>
>
> void register_free_functions(){
>
>     . . .
>
>
>
>     { //::GetFoundation
>
>         typedef ::physx::PxFoundation * ( *GetFoundation_function_type )(
> );
>
>         bp::def(
>             "GetFoundation"
>             , GetFoundation_function_type( &::GetFoundation )
>             , bp::return_value_policy< bp::reference_existing_object >() );
>
>     }
>
>
>
>     . . .
>
>
>
>
>
> I think: return_value_policy< bp::reference_existing_object > thats
> right!?
>
>
>
> Then a debug session with PTVS, start with a break point in InitPhysX
>
> and inspect the mFoundation after the successful generation.
>
> Its a vptr with 9 Entries => thats right, PxFoundation has 9 Methods.
>
> That's all what I can see, because I have no source code and no pdb
> (symbols).
>
>
>
> I place the next Breakpoint in
>
> PxFoundation* GetFoundation(){
>    *return* mFoundation;
> }
>
>
>
> to search the Error location.
>
>
>
> Then followed a trace trip through the code, for me is a journey through
> an unreadable jungles of wrapper C + + code.
>
> Unfortunately, I am not an C++Expert.
>
> The Error is not to find for me.
>
>
>
>
>
> The same problem occurs with PxCreatePhysics.
>
>
>
> My work enviroment:
>
> ?  Windows 7 x64
>
> ?  Python 2.7
>
> ?  physxVersion 3.24  (i try: 3.01, 3.23, 3.3, same problem)
>
> ?  visualStudioVersion 11 (i try: VS 9 and  VS 10,  same problem)
>
> ?  PythonTools for Visual Studio (PTVS)
>
> ?  Boost 1.55 (i try: 1.45,  same problem)
>
> ?  gccxml 0.9
>
> ?  pyplusplus 1.0 svn 1.5.1
>
> ?  pygccxml 1.0 svn 1.5.1
>
>
>
>
>
>
>
> What is my mistake?
>
> I would be very grateful if someone could an expert help.
>
> Thanks for your response.
>
>
>
>     Peter Armold
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> https://mail.python.org/mailman/listinfo/cplusplus-sig
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20140129/35d6f148/attachment-0001.html>

From arm at rfh-koeln.de  Thu Jan 30 16:28:59 2014
From: arm at rfh-koeln.de (=?iso-8859-1?Q?Peter_Armbr=FCster?=)
Date: Thu, 30 Jan 2014 16:28:59 +0100
Subject: [C++-sig] PhysX3.x Wrapping Problems
Message-ID: <03dc01cf1dd0$01ac35e0$0504a1a0$@rfh-koeln.de>

Hi Nicolay,

thanks for your Tip.
I tested this, but same Error.
I think that is the biggest sh... Error in my live.

Maybe I would better become a musician. :-)


Best Regards
   Peter


-----Urspr?ngliche Nachricht-----
Von: Cplusplus-sig
[mailto:cplusplus-sig-bounces+arm=rfh-koeln.de at python.org] Im Auftrag von
cplusplus-sig-request at python.org
Gesendet: Mittwoch, 29. Januar 2014 16:22
An: cplusplus-sig at python.org
Betreff: Cplusplus-sig Digest, Vol 64, Issue 12

Send Cplusplus-sig mailing list submissions to
	cplusplus-sig at python.org

To subscribe or unsubscribe via the World Wide Web, visit
	https://mail.python.org/mailman/listinfo/cplusplus-sig
or, via email, send a message with subject or body 'help' to
	cplusplus-sig-request at python.org

You can reach the person managing the list at
	cplusplus-sig-owner at python.org

When replying, please edit your Subject line so it is more specific than
"Re: Contents of Cplusplus-sig digest..."


Today's Topics:

   1. Re: Cplusplus-sig Digest, Vol 64, Issue 11 (Peter Armbr?ster)
   2. Re: PhysX3.x Wrapping Problems (Nikolay Mladenov)


----------------------------------------------------------------------

Message: 1
Date: Wed, 29 Jan 2014 12:26:18 +0100
From: Peter Armbr?ster <arm at rfh-koeln.de>
To: <cplusplus-sig at python.org>
Subject: Re: [C++-sig] Cplusplus-sig Digest, Vol 64, Issue 11
Message-ID: <03d701cf1ce4$ee69dc30$cb3d9490$@rfh-koeln.de>
Content-Type: text/plain;	charset="iso-8859-1"

Hi Niki and Giuseppe,

thanks for your answers, but once I had already considered .

My compiler flags are:  /DWIN32 /DDEBUG /GR  /RTCsu /DPX_DEBUG  

I wish you a nice day.

Peter




-----Urspr?ngliche Nachricht-----
Von: Cplusplus-sig
[mailto:cplusplus-sig-bounces+arm=rfh-koeln.de at python.org] Im Auftrag von
cplusplus-sig-request at python.org
Gesendet: Mittwoch, 29. Januar 2014 12:00
An: cplusplus-sig at python.org
Betreff: Cplusplus-sig Digest, Vol 64, Issue 11

Send Cplusplus-sig mailing list submissions to
	cplusplus-sig at python.org

To subscribe or unsubscribe via the World Wide Web, visit
	https://mail.python.org/mailman/listinfo/cplusplus-sig
or, via email, send a message with subject or body 'help' to
	cplusplus-sig-request at python.org

You can reach the person managing the list at
	cplusplus-sig-owner at python.org

When replying, please edit your Subject line so it is more specific than
"Re: Contents of Cplusplus-sig digest..."


Today's Topics:

   1. Re: PhysX3.x Wrapping Problems (niki)
   2. Re: PhysX3.x Wrapping Problems (Giuseppe Corbelli)


----------------------------------------------------------------------

Message: 1
Date: Wed, 29 Jan 2014 10:19:59 +0200
From: niki <niki at vintech.bg>
To: Development of Python/C++ integration <cplusplus-sig at python.org>
Subject: Re: [C++-sig] PhysX3.x Wrapping Problems
Message-ID: <52E8B9AF.4050602 at vintech.bg>
Content-Type: text/plain; charset=UTF-8; format=flowed

On 28.01.2014 18:33, Peter Armbr?ster wrote:
> Dear Boost.python, C++ and Py++ Experts,
>
> this is my first post in your excellent Forum.
>
> I have chosen as the wrapper software the boost.python/Py++ Packageto 
> write a physx 3.x Wrapper, because I think it is the best.
>
> Many many thanks to David Abrahams and Roman Yakovenko for the great
Tools.
>
> The wrapping of old versions physx 2.x had made no problems.
>
> And now I come across a  very big problem at the 3.x Versions.
>
> All wrappings of classes works very well.
>
> The problem only exists in the PhysX free functions, which are 
> responsible for the initialization of PhysX.
>
> I try to explain with a reduced Example.
>
> The initialization of PhysX starts with create of foundation :
>
> *0 static*PxFoundation*mFoundation=NULL;
>
> *1 static*PxDefaultErrorCallbackgDefaultErrorCallback;
> *2 static*PxDefaultAllocatorgDefaultAllocatorCallback;
> 3
> mFoundation=PxCreateFoundation(PX_PHYSICS_VERSION,gDefaultAllocatorCal
> lback,gDefaultErrorCallback);4
> *4 if*(!mFoundation)
>
> 5   fatalError("PxCreateFoundation failed!");
>
> The wrapped python program generate a runtime error at line
> 3:RuntimeError: Access violation - no RTTI data!

Just wild guess: add compiler option to enable RTTI

HTH
Niki


------------------------------

Message: 2
Date: Wed, 29 Jan 2014 11:03:00 +0100
From: Giuseppe Corbelli <giuseppe.corbelli at copanitalia.com>
To: Development of Python/C++ integration <cplusplus-sig at python.org>
Subject: Re: [C++-sig] PhysX3.x Wrapping Problems
Message-ID: <52E8D1D4.7060003 at copanitalia.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed

On 28/01/2014 17:33, Peter Armbr?ster wrote:
> Dear Boost.python, C++ and Py++ Experts,
>
> this is my first post in your excellent Forum.
>
> I have chosen as the wrapper software the boost.python/Py++ Packageto 
> write a physx 3.x Wrapper, because I think it is the best.
>
> Many many thanks to David Abrahams and Roman Yakovenko for the great
Tools.
>
> The wrapping of old versions physx 2.x had made no problems.
>
> And now I come across a  very big problem at the 3.x Versions.
>
> All wrappings of classes works very well.
>
> The problem only exists in the PhysX free functions, which are 
> responsible for the initialization of PhysX.
>
> I try to explain with a reduced Example.
>
> The initialization of PhysX starts with create of foundation :
>
> *0 static*PxFoundation*mFoundation=NULL;
>
> *1 static*PxDefaultErrorCallbackgDefaultErrorCallback;
> *2 static*PxDefaultAllocatorgDefaultAllocatorCallback;
> 3
> mFoundation=PxCreateFoundation(PX_PHYSICS_VERSION,gDefaultAllocatorCal
> lback,gDefaultErrorCallback);4
> *4 if*(!mFoundation)
>
> 5   fatalError("PxCreateFoundation failed!");
>
> The wrapped python program generate a runtime error at line
3:RuntimeError:
> Access violation - no RTTI data!
>
> Its impossible to get mFoundation.
>
> *Notes: *
>
> - physx comes as SDK:  only Headers, libs and dlls.

Before reading on I'd investigate the "no RTTI data" message.
Make sure you're using the /GR compiler option.

-- 
             Giuseppe Corbelli
WASP Software Engineer, Copan Italia S.p.A
Phone: +390303666318  Fax: +390302659932
E-mail: giuseppe.corbelli at copanitalia.com


------------------------------

Subject: Digest Footer

_______________________________________________
Cplusplus-sig mailing list
Cplusplus-sig at python.org
https://mail.python.org/mailman/listinfo/cplusplus-sig

------------------------------

End of Cplusplus-sig Digest, Vol 64, Issue 11
*********************************************



------------------------------

Message: 2
Date: Wed, 29 Jan 2014 10:21:50 -0500
From: Nikolay Mladenov <nikolay.mladenov at gmail.com>
To: "Development of Python/C++ integration" <cplusplus-sig at python.org>
Subject: Re: [C++-sig] PhysX3.x Wrapping Problems
Message-ID:
	<CACbdn-J9sPWzS4cVqB9BPnJ5A=ecz__c2qnbFoQzRtj=F0O2mw at mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"

Recently I had such symptom - it went away when I turned "Whole Program
Optimization" off


On Tue, Jan 28, 2014 at 11:33 AM, Peter Armbr?ster <arm at rfh-koeln.de> wrote:

> Dear Boost.python, C++ and Py++ Experts,
>
>
>
> this is my first post in your excellent Forum.
>
>
>
> I have chosen as the wrapper software the boost.python/Py++ Package to 
> write a physx 3.x Wrapper, because I think it is the best.
>
> Many many thanks to David Abrahams and Roman Yakovenko for the great
Tools.
>
>
>
> The wrapping of old versions physx 2.x had made no problems.
>
> And now I come across a  very big problem at the 3.x Versions.
>
> All wrappings of classes works very well.
>
> The problem only exists in the PhysX free functions, which are 
> responsible for the initialization of PhysX.
>
>
>
> I try to explain with a reduced Example.
>
> The initialization of PhysX starts with create of foundation :
>
>
>
> *0 static* PxFoundation *mFoundation = NULL;
>
> *1 static* PxDefaultErrorCallback gDefaultErrorCallback;
> *2 static* PxDefaultAllocator gDefaultAllocatorCallback;
> 3 mFoundation = PxCreateFoundation(PX_PHYSICS_VERSION,
> gDefaultAllocatorCallback, gDefaultErrorCallback); 4
> *4 if*(!mFoundation)
>
> 5   fatalError("PxCreateFoundation failed!");
>
>
>
> The wrapped python program generate a runtime error at line
3:RuntimeError: Access violation - no RTTI data!
>
> Its impossible to get mFoundation.
>
>
>
>
>
>
>
> *Notes: *
>
> - physx comes as SDK:  only Headers, libs and dlls.
>
> - PxFoundation is a Abstract class:
>
>
>
> *class* PX_FOUNDATION_API PxFoundation {
> *public*:
>    *virtual*   *void* release() = 0;
>    *virtual* PxErrorCallback& getErrorCallback() *const* = 0;
>    *virtual* *void* setErrorLevel(PxErrorCode::Enum mask =
> PxErrorCode::eMASK_ALL) = 0;
>    *virtual* PxErrorCode::Enum getErrorLevel() *const* = 0;
>    *virtual* PxBroadcastingAllocator& getAllocator() *const* = 0;
>    *virtual* PxAllocatorCallback& getAllocatorCallback() *const* = 0;
>    *virtual* *bool* getReportAllocationNames() *const* = 0;
>    *virtual* *void* setReportAllocationNames(*bool* value) = 0;
>
>    *virtual* ~PxFoundation() {}
> };
>
>
>
> - The wrapping of PxFoundation makes no problem.
>
>
>
>
>
> To make sure that I made no mistake with the  initialization of PhysX, 
> I tried a different solution.
>
> A small function InitPhysX create the foundation, and  GetFoundation 
> get this.
>
>
>
> C++ Code:
>
>
> *using* *namespace* physx;
>
> PxU32 PHYSICS_VERSION = PX_PHYSICS_VERSION;
> *static* PxFoundation *mFoundation = NULL;
> *static* PxDefaultErrorCallback gDefaultErrorCallback;
> *static* PxDefaultAllocator gDefaultAllocatorCallback;
>
>
> PxFoundation* GetFoundation(){
>    *return* mFoundation;
> }
>
> *void* InitPhysX() {
>    mFoundation = PxCreateFoundation(
>       PX_PHYSICS_VERSION,
>       gDefaultAllocatorCallback,
>       gDefaultErrorCallback);
>
>    *if*(!mFoundation)
>       printf("ERROR: PxCreateFoundation failed!\n");
>
>
> *else*      printf("PxCreateFoundation created!\n");
> }
>
>
>
> The wrapped python program:
>
>
>
> 1 import _physx324_d as PX
>
> 2 PX.InitPhysX()
>
> 3 foundation = PX.GetFoundation()
>
>
>
> I get on line 2 the message: "PxCreateFoundation created!"
> => OK!
>
>
>
> On Line 3 the same error RuntimeError: Access violation - no RTTI data!
>
>
>
> Now i check the wrappings of PxCreateFoundation in 
> _physx324_d_free_functions.pypp.cpp
>
>
>
> void register_free_functions(){
>
>     . . .
>
>
>
>     { //::GetFoundation
>
>         typedef ::physx::PxFoundation * ( *GetFoundation_function_type 
> )( );
>
>         bp::def(
>             "GetFoundation"
>             , GetFoundation_function_type( &::GetFoundation )
>             , bp::return_value_policy< bp::reference_existing_object 
> >() );
>
>     }
>
>
>
>     . . .
>
>
>
>
>
> I think: return_value_policy< bp::reference_existing_object > thats 
> right!?
>
>
>
> Then a debug session with PTVS, start with a break point in InitPhysX
>
> and inspect the mFoundation after the successful generation.
>
> Its a vptr with 9 Entries => thats right, PxFoundation has 9 Methods.
>
> That's all what I can see, because I have no source code and no pdb 
> (symbols).
>
>
>
> I place the next Breakpoint in
>
> PxFoundation* GetFoundation(){
>    *return* mFoundation;
> }
>
>
>
> to search the Error location.
>
>
>
> Then followed a trace trip through the code, for me is a journey 
> through an unreadable jungles of wrapper C + + code.
>
> Unfortunately, I am not an C++Expert.
>
> The Error is not to find for me.
>
>
>
>
>
> The same problem occurs with PxCreatePhysics.
>
>
>
> My work enviroment:
>
> ?  Windows 7 x64
>
> ?  Python 2.7
>
> ?  physxVersion 3.24  (i try: 3.01, 3.23, 3.3, same problem)
>
> ?  visualStudioVersion 11 (i try: VS 9 and  VS 10,  same problem)
>
> ?  PythonTools for Visual Studio (PTVS)
>
> ?  Boost 1.55 (i try: 1.45,  same problem)
>
> ?  gccxml 0.9
>
> ?  pyplusplus 1.0 svn 1.5.1
>
> ?  pygccxml 1.0 svn 1.5.1
>
>
>
>
>
>
>
> What is my mistake?
>
> I would be very grateful if someone could an expert help.
>
> Thanks for your response.
>
>
>
>     Peter Armold
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> https://mail.python.org/mailman/listinfo/cplusplus-sig
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.python.org/pipermail/cplusplus-sig/attachments/20140129/35d6f14
8/attachment.html>

------------------------------

Subject: Digest Footer

_______________________________________________
Cplusplus-sig mailing list
Cplusplus-sig at python.org
https://mail.python.org/mailman/listinfo/cplusplus-sig

------------------------------

End of Cplusplus-sig Digest, Vol 64, Issue 12
*********************************************