From rwgk at yahoo.com  Thu Oct  1 06:48:19 2009
From: rwgk at yahoo.com (Ralf W. Grosse-Kunstleve)
Date: Wed, 30 Sep 2009 21:48:19 -0700 (PDT)
Subject: [C++-sig] boost python to/from python type conversion
In-Reply-To: <4AC3D04D.805@fnal.gov>
References: <4AC3D04D.805@fnal.gov>
Message-ID: <440430.92877.qm@web111406.mail.gq1.yahoo.com>

lvalue_from_pytype only works for passing by value or const references/pointers, but that seems to match your case,
since you MPI_Comm is passed by value to your sayhello().

However, wrapping extern "C" isn't portable, AFAIK. You'll need a thin C++ wrapper for each of the C functions.

Did you consider writing a thin object-oriented wrapper around the raw MPI interfaces? Then it should be very straightforward to bind the wrapper with Boost.Python, and you'll have a more intuitive Python interface.
E.g. make a class that owns MPI_Comm, passed as boost::python::object, and then give it a .sayhello() method using the comm data member.




----- Original Message ----
From: James Amundson <amundson at fnal.gov>
To: cplusplus-sig at python.org
Sent: Wednesday, September 30, 2009 2:40:29 PM
Subject: [C++-sig] boost python to/from python type conversion

Hi,

I'm trying to write to/from type converters for my code, and I'm stuck. My case doesn't map well to any of those in the boost python documentation. I'm trying to pass objects of type MPI_Comm (really a C type!) back and forth with mpi4py (which uses cython internally.) The mpi4py author has provided a working example using manual wrapping. I am attaching it to the end of this message. On the C++ side I can extract a point to the MPI_Comm object from a Python object like so:

PyObject* py_obj;
// obviously, the value of py_obj has to be set somewhere...
MPI_Comm *comm_p = PyMPIComm_Get(py_obj);

I thought it would be simple to write a type converter to do that, but I don't understand the mechanics. This example doesn't seem to map onto the lvalue_from_pytype paradigm, or the scitbx container_conversion paradigm.

Any suggestions?

Thanks,
Jim Amundson

-----------------------------------
// working example with manual conversion

#include <mpi.h>
#include <iostream>

static void sayhello(MPI_Comm comm)
{
  if (comm == MPI_COMM_NULL) {
    std::cout << "You passed MPI_COMM_NULL !!!" << std::endl;
    return;
  }
  int size;
  MPI_Comm_size(comm, &size);
  int rank;
  MPI_Comm_rank(comm, &rank);
  int plen; char pname[MPI_MAX_PROCESSOR_NAME];
  MPI_Get_processor_name(pname, &plen);
  std::cout <<
    "Hello, World! " <<
    "I am process " << rank <<
    " of " << size <<
    " on  " << pname <<
    "." << std::endl;
}


#include <boost/python.hpp>
#include <mpi4py/mpi4py.h>
using namespace boost::python;

static void hw_sayhello(object py_comm)
{
  PyObject* py_obj = py_comm.ptr();
  MPI_Comm *comm_p = PyMPIComm_Get(py_obj);
  if (comm_p == NULL) throw_error_already_set();
  sayhello(*comm_p);
}

BOOST_PYTHON_MODULE(helloworld)
{
  if (import_mpi4py() < 0) return; /* Python 2.X */

  def("sayhello", hw_sayhello);
}



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


From roman.yakovenko at gmail.com  Thu Oct  1 07:04:57 2009
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 1 Oct 2009 07:04:57 +0200
Subject: [C++-sig] [Py++] Does pyplusplus respect explicit constructors?
In-Reply-To: <b056e7240909300918j64274066r921cab5bc4654eff@mail.gmail.com>
References: <b056e7240909300918j64274066r921cab5bc4654eff@mail.gmail.com>
Message-ID: <7465b6170909302204l4ee75231g96bc7db0da0cbb0a@mail.gmail.com>

On Wed, Sep 30, 2009 at 6:18 PM, Christopher Bruns <cmbruns at stanford.edu> wrote:
> Single-argument C++ constructors in pyplusplus seem to generate a
> boost::python::implicitly_convertible tag, even when the constructor
> is declared "explicit" in C++. ?That doesn't sound right to me.

I don't completely agree with you.

> Shouldn't a pair of types that are not implicitly convertible in C++,
> also not be implicitly convertible in python?

It depends on your use cases.

> I suspect I can manually adjust the implicitly_convertible tag
> generation using the "allow_implicit_conversion" flag in pyplusplus.
> Is there a way to tell pyplusplus to avoid generating an
> implicitly_convertible tag for all explicit constructors?

I don't think so. I attached gccxml generated file for your code. As
you can see both constructors ( Foo and Bar )  have "explitic=1". So
Py++ has no way to find out whether a constructor explicit or not.

If such functionality would present, I think that the default would be
the opposite.

HTH

-- 
Roman Yakovenko
C++ Python language binding
http://www.language-binding.net/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 1.tar.bz
Type: application/x-bzip
Size: 4553 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20091001/ce6cf54f/attachment.bin>

From austin.bingham at gmail.com  Thu Oct  1 08:22:13 2009
From: austin.bingham at gmail.com (Austin Bingham)
Date: Thu, 1 Oct 2009 08:22:13 +0200
Subject: [C++-sig] boost python to/from python type conversion
In-Reply-To: <4AC3D04D.805@fnal.gov>
References: <4AC3D04D.805@fnal.gov>
Message-ID: <bf1abd10909302322w52973f31mf9b7f643c682d97c@mail.gmail.com>

If it's any help, I wrote a small bit on the details of writing
converters. It might help clarify some of the mechanics which, I
agree, are a bit mysterious:

  http://misspent.wordpress.com/2009/09/27/how-to-write-boost-python-converters/

Austin

On Wed, Sep 30, 2009 at 11:40 PM, James Amundson <amundson at fnal.gov> wrote:
> Hi,
>
> I'm trying to write to/from type converters for my code, and I'm stuck. My
> case doesn't map well to any of those in the boost python documentation. I'm
> trying to pass objects of type MPI_Comm (really a C type!) back and forth
> with mpi4py (which uses cython internally.) The mpi4py author has provided a
> working example using manual wrapping. I am attaching it to the end of this
> message. On the C++ side I can extract a point to the MPI_Comm object from a
> Python object like so:
>
> PyObject* py_obj;
> // obviously, the value of py_obj has to be set somewhere...
> MPI_Comm *comm_p = PyMPIComm_Get(py_obj);
>
> I thought it would be simple to write a type converter to do that, but I
> don't understand the mechanics. This example doesn't seem to map onto the
> lvalue_from_pytype paradigm, or the scitbx container_conversion paradigm.
>
> Any suggestions?
>
> Thanks,
> Jim Amundson
>
> -----------------------------------
> // working example with manual conversion
>
> #include <mpi.h>
> #include <iostream>
>
> static void sayhello(MPI_Comm comm)
> {
> ?if (comm == MPI_COMM_NULL) {
> ? ?std::cout << "You passed MPI_COMM_NULL !!!" << std::endl;
> ? ?return;
> ?}
> ?int size;
> ?MPI_Comm_size(comm, &size);
> ?int rank;
> ?MPI_Comm_rank(comm, &rank);
> ?int plen; char pname[MPI_MAX_PROCESSOR_NAME];
> ?MPI_Get_processor_name(pname, &plen);
> ?std::cout <<
> ? ?"Hello, World! " <<
> ? ?"I am process " << rank <<
> ? ?" of " << size <<
> ? ?" on ?" << pname <<
> ? ?"." << std::endl;
> }
>
>
> #include <boost/python.hpp>
> #include <mpi4py/mpi4py.h>
> using namespace boost::python;
>
> static void hw_sayhello(object py_comm)
> {
> ?PyObject* py_obj = py_comm.ptr();
> ?MPI_Comm *comm_p = PyMPIComm_Get(py_obj);
> ?if (comm_p == NULL) throw_error_already_set();
> ?sayhello(*comm_p);
> }
>
> BOOST_PYTHON_MODULE(helloworld)
> {
> ?if (import_mpi4py() < 0) return; /* Python 2.X */
>
> ?def("sayhello", hw_sayhello);
> }
>
>
>
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
>

From cmbruns at stanford.edu  Thu Oct  1 17:46:13 2009
From: cmbruns at stanford.edu (Christopher Bruns)
Date: Thu, 1 Oct 2009 08:46:13 -0700
Subject: [C++-sig] [Py++] Does pyplusplus respect explicit constructors?
Message-ID: <b056e7240910010846l572e1f02o1c1181b2eb47be60@mail.gmail.com>

Roman Yakovenko wrote:
> On Wed, Sep 30, 2009 at 6:18 PM, Christopher Bruns <cmbruns at stanford.edu> wrote:
> > I suspect I can manually adjust the implicitly_convertible tag
> > generation using the "allow_implicit_conversion" flag in pyplusplus.
> > Is there a way to tell pyplusplus to avoid generating an
> > implicitly_convertible tag for all explicit constructors?
>
> I don't think so. I attached gccxml generated file for your code. As
> you can see both constructors ( Foo and Bar )  have "explitic=1". So
> Py++ has no way to find out whether a constructor explicit or not.

I disagree.  Py++ does have a way to find out.  The only time
"explicit" matters is when a constructor takes exactly one argument.
gccxml sets "explicit=1" in all cases when the number of arguments is
other than 1.  That is why you saw "explicit=1" in all of the
constructors in the previous example.  However, in the case where
there is exactly one argument, gccxml does the right thing.  Consider
the following case:

### test.h ####
struct Foo1 {};
struct Foo2 {};

struct Bar {
    explicit Bar(Foo1&); // no implicit conversion
    Bar(Foo2&); // allows implicit conversion
};
#############


### fragment of resulting xml file from gccxml ###
  <Constructor id="_10" name="Bar" explicit="1" context="_3"
access="public" mangled="_ZN3BarC1ER4Foo1 *INTERNAL* "
demangled="Bar::Bar(Foo1&amp;)" location="f0:5" file="f0" line="5"
extern="1">
    <Argument type="_22" location="f0:5" file="f0" line="5"/>
  </Constructor>
  <Constructor id="_11" name="Bar" context="_3" access="public"
mangled="_ZN3BarC1ER4Foo2 *INTERNAL* " demangled="Bar::Bar(Foo2&amp;)"
location="f0:6" file="f0" line="6" extern="1">
    <Argument type="_23" location="f0:6" file="f0" line="6"/>
  </Constructor>
#######################

Notice that gccxml sets "explicit=1" for the first Bar constructor,
but not for the second.

However, Py++ sets "allow_implicit_conversion" to True for both
constructors.  I think gccxml is doing the right thing here.  But I am
not so sure about Py++ behavior.  I would have assumed that the first
Bar constructor would have "allow_implicit_conversion" set to False.
I was wrong about being able to use "allow_implicit_conversion" for
this task.  Is there another way to get at the "explicit" xml tag from
the module_builder object?  Or do I need to parse the xml file or
source code separately to discover this information?

Thank you Roman for your diligent feedback.  I sincerely appreciate
the quick and thoughtful feedback you provide to everyone on this
mailing list.

Chris Bruns

From amundson at fnal.gov  Thu Oct  1 23:30:01 2009
From: amundson at fnal.gov (James Amundson)
Date: Thu, 01 Oct 2009 16:30:01 -0500
Subject: [C++-sig] boost python to/from python type conversion
In-Reply-To: <440430.92877.qm@web111406.mail.gq1.yahoo.com>
References: <4AC3D04D.805@fnal.gov>
	<440430.92877.qm@web111406.mail.gq1.yahoo.com>
Message-ID: <4AC51F59.40303@fnal.gov>

On 09/30/2009 11:48 PM, Ralf W. Grosse-Kunstleve wrote:
> lvalue_from_pytype only works for passing by value or const references/pointers, but that seems to match your case,
> since you MPI_Comm is passed by value to your sayhello().
>
> However, wrapping extern "C" isn't portable, AFAIK. You'll need a thin C++ wrapper for each of the C functions.
>    
Yes. I think you are right. I'm going to try that now.
> Did you consider writing a thin object-oriented wrapper around the raw MPI interfaces? Then it should be very straightforward to bind the wrapper with Boost.Python, and you'll have a more intuitive Python interface.
> E.g. make a class that owns MPI_Comm, passed as boost::python::object, and then give it a .sayhello() method using the comm data member.
>    

It's more complicated than that. mpi4py provides a nice object-oriented 
wrapper on the Python side. The mpi4py wrapper is based on the 
"standard" MPI C++ bindings. Unfortunately, older versions of MPI don't 
support the C++ bindings and MPI has decided that the bindings will be 
dropped in the future, so I really don't want to start using them now. I 
think I will do as you suggest and just write very thin wrappers around 
the C interfaces.

Thanks for your help.

--Jim Amundson


From amundson at fnal.gov  Thu Oct  1 23:30:51 2009
From: amundson at fnal.gov (James Amundson)
Date: Thu, 01 Oct 2009 16:30:51 -0500
Subject: [C++-sig] boost python to/from python type conversion
In-Reply-To: <bf1abd10909302322w52973f31mf9b7f643c682d97c@mail.gmail.com>
References: <4AC3D04D.805@fnal.gov>
	<bf1abd10909302322w52973f31mf9b7f643c682d97c@mail.gmail.com>
Message-ID: <4AC51F8B.4000907@fnal.gov>

On 10/01/2009 01:22 AM, Austin Bingham wrote:
> If it's any help, I wrote a small bit on the details of writing
> converters. It might help clarify some of the mechanics which, I
> agree, are a bit mysterious:
>
>    http://misspent.wordpress.com/2009/09/27/how-to-write-boost-python-converters/
>
>    
Wonderful. Your blog post is very helpful. I wish you had written it 
(and I had found it!) a long time ago.

Thanks,
Jim Amundson


From roman.yakovenko at gmail.com  Fri Oct  2 01:10:17 2009
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Fri, 2 Oct 2009 01:10:17 +0200
Subject: [C++-sig] [Py++] Does pyplusplus respect explicit constructors?
In-Reply-To: <b056e7240910010846l572e1f02o1c1181b2eb47be60@mail.gmail.com>
References: <b056e7240910010846l572e1f02o1c1181b2eb47be60@mail.gmail.com>
Message-ID: <7465b6170910011610r2781c0bdw4ec1e4c0fc5622d7@mail.gmail.com>

On Thu, Oct 1, 2009 at 5:46 PM, Christopher Bruns <cmbruns at stanford.edu> wrote:
> I disagree. ?Py++ does have a way to find out. ?The only time
> "explicit" matters is when a constructor takes exactly one argument.
> gccxml sets "explicit=1" in all cases when the number of arguments is
> other than 1. ?That is why you saw "explicit=1" in all of the
> constructors in the previous example. ?However, in the case where
> there is exactly one argument, gccxml does the right thing. ?Consider
> the following case:
>
> ### test.h ####
> struct Foo1 {};
> struct Foo2 {};
>
> struct Bar {
> ? ?explicit Bar(Foo1&); // no implicit conversion
> ? ?Bar(Foo2&); // allows implicit conversion
> };
> #############
>
>
> ### fragment of resulting xml file from gccxml ###
> ?<Constructor id="_10" name="Bar" explicit="1" context="_3"
> access="public" mangled="_ZN3BarC1ER4Foo1 *INTERNAL* "
> demangled="Bar::Bar(Foo1&amp;)" location="f0:5" file="f0" line="5"
> extern="1">
> ? ?<Argument type="_22" location="f0:5" file="f0" line="5"/>
> ?</Constructor>
> ?<Constructor id="_11" name="Bar" context="_3" access="public"
> mangled="_ZN3BarC1ER4Foo2 *INTERNAL* " demangled="Bar::Bar(Foo2&amp;)"
> location="f0:6" file="f0" line="6" extern="1">
> ? ?<Argument type="_23" location="f0:6" file="f0" line="6"/>
> ?</Constructor>
> #######################
>
> Notice that gccxml sets "explicit=1" for the first Bar constructor,
> but not for the second.

I didn't know that, so pygccxml didn't provided access to the information.

> However, Py++ sets "allow_implicit_conversion" to True for both
> constructors. ?I think gccxml is doing the right thing here. ?But I am
> not so sure about Py++ behavior.

Well, these days I would implement "allow_implicit_conversion" to be a
mix of the user desire and "explicit" attribute. But, this change will
introduce backward compatibility problem, which I would like to avoid.

> I would have assumed that the first
> Bar constructor would have "allow_implicit_conversion" set to False.
> I was wrong about being able to use "allow_implicit_conversion" for
> this task. ?Is there another way to get at the "explicit" xml tag from
> the module_builder object? ?Or do I need to parse the xml file or
> source code separately to discover this information?

No. I updated pygccxml package
http://pygccxml.svn.sourceforge.net/pygccxml/?rev=1761&view=rev

All you need is to write the following code:

mb = module_builder_t(...)
mb.constructors( lambda c: c.explicit == True
).allow_implicit_conversion = False

That's all

> Thank you Roman for your diligent feedback. ?I sincerely appreciate
> the quick and thoughtful feedback you provide to everyone on this
> mailing list.

Wow :-). Thank you for such words!

-- 
Roman Yakovenko
C++ Python language binding
http://www.language-binding.net/

From micdestefano at gmail.com  Fri Oct  2 16:06:57 2009
From: micdestefano at gmail.com (Michele De Stefano)
Date: Fri, 2 Oct 2009 16:06:57 +0200
Subject: [C++-sig] Conversion from Python to C++
Message-ID: <c410adc90910020706n696da461q65bfee4eadda7330@mail.gmail.com>

Hello to everybody.

I was wondering if there is a general way to convert a C++ class
wrapped in Python into a C++ class not wrapped in Python.
I've not fully understood the documentation on this point.

For example, assume we have two C++ classes, A and B. A is exported to
Python with class_<>, while B is not exported to Python.
Assume also that, in C++, A is convertible to B (i.e. I can construct
a B object from an existing A object).

Assume I have a C++ function (f) which takes a B object as input and I
want to expose this function, but I don't want to expose the B class.

Is there a general way to create a converter from A to B in order to
be able to call f(A) from Python ?

Thank you in advance.

-- 
Michele De Stefano

From seefeld at sympatico.ca  Fri Oct  2 16:22:36 2009
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Fri, 02 Oct 2009 10:22:36 -0400
Subject: [C++-sig] Conversion from Python to C++
In-Reply-To: <c410adc90910020706n696da461q65bfee4eadda7330@mail.gmail.com>
References: <c410adc90910020706n696da461q65bfee4eadda7330@mail.gmail.com>
Message-ID: <4AC60CAC.4050907@sympatico.ca>

On 10/02/2009 10:06 AM, Michele De Stefano wrote:
> For example, assume we have two C++ classes, A and B. A is exported to
> Python with class_<>, while B is not exported to Python.
> Assume also that, in C++, A is convertible to B (i.e. I can construct
> a B object from an existing A object).
>
> Assume I have a C++ function (f) which takes a B object as input and I
> want to expose this function, but I don't want to expose the B class.
>
> Is there a general way to create a converter from A to B in order to
> be able to call f(A) from Python ?
>    

Depending on how many such functions you have, you may either simply 
write a wrapper function with the right signature (i.e. accepting an 'A' 
and (implicitly) converting that to a 'B' when calling the actual 
function), or you may write a custom from_python converter.

Regards,
         Stefan

-- 

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


From micdestefano at gmail.com  Fri Oct  2 16:35:50 2009
From: micdestefano at gmail.com (Michele De Stefano)
Date: Fri, 2 Oct 2009 16:35:50 +0200
Subject: [C++-sig] Conversion from Python to C++
In-Reply-To: <4AC60CAC.4050907@sympatico.ca>
References: <c410adc90910020706n696da461q65bfee4eadda7330@mail.gmail.com>
	<4AC60CAC.4050907@sympatico.ca>
Message-ID: <c410adc90910020735n2fa2d1c9h1ebd5fe2d781aef@mail.gmail.com>

Ok.

I'd like to write a custom from_python converter, but the point is
that I've not understood how it works.

The explanation in the FAQs it's too obscure for me.

2009/10/2 Stefan Seefeld <seefeld at sympatico.ca>:
> On 10/02/2009 10:06 AM, Michele De Stefano wrote:
>>
>> For example, assume we have two C++ classes, A and B. A is exported to
>> Python with class_<>, while B is not exported to Python.
>> Assume also that, in C++, A is convertible to B (i.e. I can construct
>> a B object from an existing A object).
>>
>> Assume I have a C++ function (f) which takes a B object as input and I
>> want to expose this function, but I don't want to expose the B class.
>>
>> Is there a general way to create a converter from A to B in order to
>> be able to call f(A) from Python ?
>>
>
> Depending on how many such functions you have, you may either simply write a
> wrapper function with the right signature (i.e. accepting an 'A' and
> (implicitly) converting that to a 'B' when calling the actual function), or
> you may write a custom from_python converter.
>
> Regards,
> ? ? ? ?Stefan
>
> --
>
> ? ? ?...ich hab' noch einen Koffer in Berlin...
>
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
>



-- 
Michele De Stefano
http://www.linkedin.com/in/micdestefano
http://xoomer.virgilio.it/michele_de_stefano

From nat at lindenlab.com  Fri Oct  2 17:48:08 2009
From: nat at lindenlab.com (Nat Goodspeed)
Date: Fri, 02 Oct 2009 11:48:08 -0400
Subject: [C++-sig] Conversion from Python to C++
In-Reply-To: <c410adc90910020735n2fa2d1c9h1ebd5fe2d781aef@mail.gmail.com>
References: <c410adc90910020706n696da461q65bfee4eadda7330@mail.gmail.com>	<4AC60CAC.4050907@sympatico.ca>
	<c410adc90910020735n2fa2d1c9h1ebd5fe2d781aef@mail.gmail.com>
Message-ID: <4AC620B8.6030904@lindenlab.com>

Michele De Stefano wrote:

> I'd like to write a custom from_python converter, but the point is
> that I've not understood how it works.

Did you see Austin Bingham's recent post on another thread?

> If it's any help, I wrote a small bit on the details of writing
> converters. It might help clarify some of the mechanics which, I
> agree, are a bit mysterious:
> 
>   http://misspent.wordpress.com/2009/09/27/how-to-write-boost-python-converters/

From micdestefano at gmail.com  Fri Oct  2 18:02:22 2009
From: micdestefano at gmail.com (Michele De Stefano)
Date: Fri, 2 Oct 2009 18:02:22 +0200
Subject: [C++-sig] Conversion from Python to C++
In-Reply-To: <4AC620B8.6030904@lindenlab.com>
References: <c410adc90910020706n696da461q65bfee4eadda7330@mail.gmail.com>
	<4AC60CAC.4050907@sympatico.ca>
	<c410adc90910020735n2fa2d1c9h1ebd5fe2d781aef@mail.gmail.com>
	<4AC620B8.6030904@lindenlab.com>
Message-ID: <c410adc90910020902q6b8882d1v9227df98a6723f96@mail.gmail.com>

Honestly, no.

Thank you for pointing me to that. I will look at it and if I have
problems again, I will post again.

Thank you very much.

2009/10/2 Nat Goodspeed <nat at lindenlab.com>:
> Michele De Stefano wrote:
>
>> I'd like to write a custom from_python converter, but the point is
>> that I've not understood how it works.
>
> Did you see Austin Bingham's recent post on another thread?
>
>> If it's any help, I wrote a small bit on the details of writing
>> converters. It might help clarify some of the mechanics which, I
>> agree, are a bit mysterious:
>>
>>
>> ?http://misspent.wordpress.com/2009/09/27/how-to-write-boost-python-converters/
>
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
>



-- 
Michele De Stefano
http://www.linkedin.com/in/micdestefano
http://xoomer.virgilio.it/michele_de_stefano

From rfritz333 at gmail.com  Mon Oct  5 02:36:34 2009
From: rfritz333 at gmail.com (Randolph Fritz)
Date: Mon, 5 Oct 2009 00:36:34 +0000 (UTC)
Subject: [C++-sig] Setting up a boost.python workflow on a Mac
Message-ID: <slrnhcifsk.8an.rfritz333@Lobo.local>

I have just successfully--I hope--done this & I want to get my notes
down before I forget the problems I encountered.  These occurred
with Boost 1.39 and Mac OS 10.5.8.

1. Bjam did not automatically detect the Mac OS environment and
   compilation consistently failed.  So, I added the following line to
   /etc/site-config.jam: 
     using darwin ;

2. I built boost.python according to the getting started guide.  This
   left me with four files in stage/bin:
     libboost_python-xgcc40-mt-1_39.a
     libboost_python-xgcc40-mt-1_39.dylib
     libboost_python-xgcc40-mt.a
     libboost_python-xgcc40-mt.dylib

3. The install command recommended by bjam was "bjam install
   --prefix=<place>".  I attempted to use the command, and found it
   consistently tried to place files in /lib.  A manual move to my
   target library directory fixed that problem.

4. It turned out that the default build script automatically looked
   for debug libraries, which weren't automatically built.  I tried
   "bjam debug" and it placed the following four files in stage/bin:
     libboost_python-xgcc40-mt-d-1_39.a
     libboost_python-xgcc40-mt-d-1_39.dylib
     libboost_python-xgcc40-mt-d.a
     libboost_python-xgcc40-mt-d.dylib
   These I also placed manually.

5. I then discovered that the sample Jamroot provided did not support
   multi-threaded (*-mt-*) libraries.  I added the following symbolic
   links to my target directory:
     libboost_python-xgcc40-1.39.dylib -> 
       libboost_python-xgcc40-mt-1_39.dylib
     libboost_python-xgcc40-d-1_39.dylib -> 
       libboost_python-xgcc40-mt-d-1_39.dylib

6. Voila!  My test application built and apparently ran.

-- 
Randolph Fritz
  design machine group, architecture department, university of washington
rfritz at u.washington.edu -or- rfritz333 at gmail.com


From blake at hailmail.net  Mon Oct  5 03:11:36 2009
From: blake at hailmail.net (Dane Springmeyer)
Date: Sun, 4 Oct 2009 18:11:36 -0700
Subject: [C++-sig] Setting up a boost.python workflow on a Mac
In-Reply-To: <slrnhcifsk.8an.rfritz333@Lobo.local>
References: <slrnhcifsk.8an.rfritz333@Lobo.local>
Message-ID: <F84AA595-F603-4207-8D28-F178484ECDAB@hailmail.net>

Hello Randolph,

On Oct 4, 2009, at 5:36 PM, Randolph Fritz wrote:

> I have just successfully--I hope--done this & I want to get my notes
> down before I forget the problems I encountered.  These occurred
> with Boost 1.39 and Mac OS 10.5.8.
>
> 1. Bjam did not automatically detect the Mac OS environment and
>   compilation consistently failed.  So, I added the following line to
>   /etc/site-config.jam:
>     using darwin ;
>

You can pass as a command line option to bjam:

./bjam toolset=darwin

(see also: https://trac.mapnik.org/wiki/MacInstallation#Step1:RouteBC)


Dane

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20091004/2c6763f0/attachment.htm>

From gustavoborges at rjxcapital.com  Thu Oct  8 21:44:26 2009
From: gustavoborges at rjxcapital.com (Gustavo Adolfo Borges)
Date: Thu, 8 Oct 2009 16:44:26 -0300
Subject: [C++-sig] call_method crash
Message-ID: <948705c00910081244i469e8d3ehfccc951036a7b75a@mail.gmail.com>

Since Python Win32 lib does not implement DDE Advise Loop and i need it for
a project, i guessed that Boost Python would be a way to integrate the ddeml
lib with my python code. I guessed i could just call call_method inside the
DDE Callback function, but when i do that python crashes. I can't(or couldnt
find how to) add the DDE Callback function to the Python Module and i tried
through a 2 stages call(DdeCallback calls a Python Module function that
call_method) without success. Everything else works great... I can connect
to the DDE Server, the advise loop works, my setPyCallback/getPyCallback
works... I just can't get any python to work on the C++ DDECallback
function.

Here is part of the code
------------------

PyObject* myPythonCallback;
void CallPy(TCHAR SomeValue){
    boost::python::call<void>(myPythonCallback,SomeValue);
}
HDDEDATA CALLBACK DdeCallback(UINT wType,UINT uFmt,HCONV hconv,HSZ hsz1,HSZ
hsz2,HDDEDATA hData,DWORD dwData1,DWORD dwData2){
..........
    switch (wType) {
            case XTYP_ADVDATA:
                        .................. bunch of code ..........
                        CallPy(mybuffer);
                        return((HDDEDATA)DDE_FACK);
}
int ddeConnect();

void ddeAdvise(TCHAR* topic);

void setCallback(PyObject* p){
    boost::python::call<void>(p,"3");// test calling function and WORKS
    myPythonCallback = p;
}

BOOST_PYTHON_MODULE(extending)
{
    using namespace boost::python;
    def("ddeConnect", ddeConnect);
    def("ddeAdvise", ddeAdvise);
    def("ddeSetCallback", setCallback);
    def("CallPy", CallPy);
}

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

Python
------------
import extending
import wx
import justAFrame

def mCB (valor):
    print "mCB",str(valor)
    return float(str(valor))

class MyFrame(justAFrame.MyFrame):
    def initAfter (self):
        justAFrame.MyFrame.initAfter(self)
        extending.ddeConnect()
        extending.ddeSetCallback(mCB)
        extending.CallPy("1") #WORKS
        extending.ddeAdvise("0012PETR4;1")

app = wx.PySimpleApp()
frame=MyFrame(None,title="DDE Test")
app.MainLoop()
--------------


How can i make this DdeCallback to interact with a python function defined
by the user ? Is there a better way to do this ? I'm open to suggestions.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20091008/0dd2769e/attachment.htm>

From rogeeff at gmail.com  Fri Oct  9 08:25:26 2009
From: rogeeff at gmail.com (Gennadiy Rozental)
Date: Fri, 9 Oct 2009 06:25:26 +0000 (UTC)
Subject: [C++-sig] "pass all args in list" syntax
Message-ID: <loom.20091009T082053-738@post.gmane.org>

Hi,

In native Python  I can do something like this:

    def goo(a,b,c):
        return a+b+c

    ll = [1,2,3]
    goo( *ll )

And it will pass all the arguments properly. I need similar functionality in C++:

Given 

bp::object func;
bp::list args;

I'd like to invoke the function like this func( *args );

Is there syntax in Boost.Python to achieve this? Or I need to resort to ugly
switch statement?

Gennadiy


From seefeld at sympatico.ca  Fri Oct  9 13:58:40 2009
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Fri, 09 Oct 2009 07:58:40 -0400
Subject: [C++-sig] "pass all args in list" syntax
In-Reply-To: <loom.20091009T082053-738@post.gmane.org>
References: <loom.20091009T082053-738@post.gmane.org>
Message-ID: <4ACF2570.8030205@sympatico.ca>

On 10/09/2009 02:25 AM, Gennadiy Rozental wrote:
> Hi,
>
> In native Python  I can do something like this:
>
>      def goo(a,b,c):
>          return a+b+c
>
>      ll = [1,2,3]
>      goo( *ll )
>
> And it will pass all the arguments properly. I need similar functionality in C++:
>
> Given
>
> bp::object func;
> bp::list args;
>
> I'd like to invoke the function like this func( *args );
>
> Is there syntax in Boost.Python to achieve this? Or I need to resort to ugly
> switch statement?
>    

I added support for the above quite a while ago (and I'm assured that 
this time it even made it into the release branch in time ;-) ).
Thus, you can construct a list and keyword, then pass to a python 
callable as '*args' and '**kwds' respectively.

Regards,
         Stefan

-- 

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


From rogeeff at gmail.com  Fri Oct  9 22:19:59 2009
From: rogeeff at gmail.com (Gennadiy Rozental)
Date: Fri, 9 Oct 2009 20:19:59 +0000 (UTC)
Subject: [C++-sig] "pass all args in list" syntax
References: <loom.20091009T082053-738@post.gmane.org>
	<4ACF2570.8030205@sympatico.ca>
Message-ID: <loom.20091009T221949-946@post.gmane.org>

Stefan Seefeld <seefeld <at> sympatico.ca> writes:
> I added support for the above quite a while ago (and I'm assured that 
> this time it even made it into the release branch in time  ).
> Thus, you can construct a list and keyword, then pass to a python 
> callable as '*args' and '**kwds' respectively.

Right.

So no luck with 1.33.1 guess ;)

Will it get into 1.41?

Gennadiy


From seefeld at sympatico.ca  Fri Oct  9 23:09:31 2009
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Fri, 09 Oct 2009 17:09:31 -0400
Subject: [C++-sig] "pass all args in list" syntax
In-Reply-To: <loom.20091009T221949-946@post.gmane.org>
References: <loom.20091009T082053-738@post.gmane.org>	<4ACF2570.8030205@sympatico.ca>
	<loom.20091009T221949-946@post.gmane.org>
Message-ID: <4ACFA68B.4070004@sympatico.ca>

On 10/09/2009 04:19 PM, Gennadiy Rozental wrote:
> Stefan Seefeld<seefeld<at>  sympatico.ca>  writes:
>    
>> I added support for the above quite a while ago (and I'm assured that
>> this time it even made it into the release branch in time  ).
>> Thus, you can construct a list and keyword, then pass to a python
>> callable as '*args' and '**kwds' respectively.
>>      
> Right.
>
> So no luck with 1.33.1 guess ;)
>
> Will it get into 1.41?
>    

With 'this time' I was referring to the 1.40 release. So yes, it will be 
in 1.41, too. :-)

     Stefan

-- 

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


From yooi at msn.com  Sat Oct 10 06:00:52 2009
From: yooi at msn.com (Cheng Lee)
Date: Sat, 10 Oct 2009 04:00:52 +0000
Subject: [C++-sig] [Urgent Question] Import Qt builded Lib into python with
 Boost.python
Message-ID: <BLU126-W21AE6D04583DDBEC9268F8D5CA0@phx.gbl>


Hello All,

 

Our project made some of the functions with Qt and also encapsulate them with dll library, 

the imported module, we made it following the boost.python syntax, was built successfully, 

and imported into python 2.6.2 with no errors, but our issues are, the functions donot work, no response

 

as our investigation, it supposes to be Qt signal/slot donot work in the python module

 

I made a simple example below:

Would someone help me check what codes i need add to make the signal/slot work in the python, Thanks a lot

 

Boost 1.40

Python 2.6.2

++++++++++++++++++++++++++++++++++++++++++++++++++.h

#ifndef SIGALE_H
#define SIGALE_H
#include<QObject>

 

class SignalMine : public QObject
{
    Q_OBJECT
public:
    SignalMine(QObject* parent = 0);
    
    void submit(void){emit trigger();}
signals:
    void trigger(void);
    
protected slots:
    void printOut(void);
    
};

#endif

++++++++++++++++++++++++++++++++++++++++++++++++++.cpp

#include<iostream.h>
#include"example.h"

 

SignalMine::SignalMine(QObject* parent):QObject(parent)
{
    connect(this, SIGNAL(trigger()), this, SLOT(printOut()));    
}

void SignalMine::printOut()
{
      cout << "Get signal!" << endl;  
}

 

//! for boost.python

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

BOOST_PYTHON_MODULE(SIGN)
{
    class_<QObject, boost::noncopyable>("QObject");
    class_<SignalMine, boost::noncopyable, bases<QObject>>("SignalMineller")
        .def("submit", &SignalMine::submit);
}

 

 
 		 	   		  
_________________________________________________________________
Keep your friends updated?even when you?re not signed in.
http://www.microsoft.com/middleeast/windows/windowslive/see-it-in-action/social-network-basics.aspx?ocid=PID23461::T:WLMTAGL:ON:WL:en-xm:SI_SB_5:092010
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20091010/3af8642c/attachment.htm>

From lists_ravi at lavabit.com  Sat Oct 10 17:58:46 2009
From: lists_ravi at lavabit.com (Ravi)
Date: Sat, 10 Oct 2009 11:58:46 -0400
Subject: [C++-sig] [python] Function objects in place of member functions
Message-ID: <200910101158.46996.lists_ravi@lavabit.com>

Hello,
  If a free function 'func' has X* as its first argument, then, boost.python 
allows it to be bound to a member function on the python side, i.e., the 
following is legal:
  void func( X* x, arg1_t arg ) { ... }
  class_<X>( "X" ).def( "func", &func );

In order to use a function object in place of a free function, one must 
specialize/overload
  boost::python::detail::get_signature
which, for some reason, does not account for function objects. Here's a very 
simple example that works:

----------------------------------------------
#include <boost/mpl/vector.hpp>

struct X { int y; };

// Function object
struct Z { int operator()( X *x, int z ) { return z + x->y; } };

namespace boost { namespace python { namespace detail {
mpl::vector<int, X*, int> get_signature( Z&, X* )
  {return mpl::vector<int, X*, int>();}
}}}

#include <boost/python/class.hpp>
#include <boost/python/module.hpp>

BOOST_PYTHON_MODULE( mft ) {
  boost::python::class_<X>( "X" )
    .def( "z", Z() ).def_readwrite( "y", &X::y ); }
-------------------------------------------------

However, note that the overload of get_signature precedes the inclusion of the 
boost.python headers, which is extremely inconvenient. However, if the headers 
are moved to their proper location as in the following,

----------------------------------------------
#include <boost/mpl/vector.hpp>
#include <boost/python/class.hpp>
#include <boost/python/module.hpp>

struct X { int y; };

// Function object
struct Z { int operator()( X *x, int z ) { return z + x->y; } };

namespace boost { namespace python { namespace detail {
boost::mpl::vector<int, X*, int> get_signature( Z&, X* )
  {return boost::mpl::vector<int, X*, int>();}
}}}

BOOST_PYTHON_MODULE( mft ) {
  boost::python::class_<X>( "X" )
    .def( "z", Z() ).def_readwrite( "y", &X::y ); }
-------------------------------------------------

the compilation fails with the following error message (gcc 4.4.1 with boost 
1.37 or 1.39):

$ g++ -Wall -shared -o mft.so memft.cc -lboost_python-mt -
I/usr/include/python2.6 -fPIC
In file included from memft.cc:1:
/usr/include/boost/python/class.hpp: In member function ?void 
boost::python::class_<T, X1, X2, X3>::def_impl(T*, const char*, Fn, const 
Helper&, ...) [with T = X, Fn = Z, Helper = 
boost::python::detail::def_helper<const char*, 
boost::python::detail::not_specified, boost::python::detail::not_specified, 
boost::python::detail::not_specified>, W = X, X1 = 
boost::python::detail::not_specified, X2 = 
boost::python::detail::not_specified, X3 = 
boost::python::detail::not_specified]?:
/usr/include/boost/python/class.hpp:235:   instantiated from 
?boost::python::class_<T, X1, X2, X3>& boost::python::class_<T, X1, X2, 
X3>::def(const char*, F) [with F = Z, W = X, X1 = 
boost::python::detail::not_specified, X2 = 
boost::python::detail::not_specified, X3 = 
boost::python::detail::not_specified]?
memft.cc:16:   instantiated from here
/usr/include/boost/python/class.hpp:536: error: no matching function for call 
to ?get_signature(Z&, X*)?

Why is the overloaded get_signature not picked up when it is declared *after* 
the inclusion of the headers?

The reason for using function objects is to change the argument types of a 
member function of X as seen by python; for example, the functor would have an 
argument type 'double' and would then internally convert the received 'double' 
argument from the python side into some custom type prior to passing it to the 
member function of X. In other words, the functor would act as a C++ 
equivalent of a python decorator.

Regards,
Ravi


From troy at resophonic.com  Mon Oct 12 01:44:29 2009
From: troy at resophonic.com (troy d. straszheim)
Date: Sun, 11 Oct 2009 19:44:29 -0400
Subject: [C++-sig] [python] Function objects in place of member functions
In-Reply-To: <200910101158.46996.lists_ravi@lavabit.com>
References: <200910101158.46996.lists_ravi@lavabit.com>
Message-ID: <4AD26DDD.5020402@resophonic.com>

Ravi wrote:
[snip]
> 
> In order to use a function object in place of a free function, one must 
> specialize/overload
>   boost::python::detail::get_signature
> which, for some reason, does not account for function objects. Here's a very 
> simple example that works:
> 
[snip]
> 
> However, note that the overload of get_signature precedes the inclusion of the 
> boost.python headers, which is extremely inconvenient. However, if the headers 
> are moved to their proper location as in the following,
> 
[snip]
> 
> Why is the overloaded get_signature not picked up when it is declared *after* 
> the inclusion of the headers?
> 

I'm not sure why it isn't picked up.  I've been working in this area, 
replacing most of detail/caller.hpp and detail/invoke.hpp with 
boost.fusion, seen here:

http://gitorious.org/~straszheim/boost/straszheim/blobs/python/boost/python/detail/caller.hpp

In the process, I overhauled get_signature to use boost::function_types, 
and to be a metafunction, not a function:

http://gitorious.org/~straszheim/boost/straszheim/blobs/python/boost/python/signature.hpp

The overall effect is a lot less preprocessor stuff to look around.

I have function objects (given a specialization of get_signature), and 
boost::function (only up to 3 arguments ATM) working,
here's the test (which passes):

/////////////////////////////////////

#include <boost/mpl/vector.hpp>
#include <boost/function.hpp>
#include <boost/python/class.hpp>
#include <boost/python/module.hpp>

namespace mpl = boost::mpl;

struct X { int y; };

struct FnObject
{
   typedef int result_type;

   int operator()(X *x, int z) const
   {
     return z + x->y;
   }
};

int f1(X* x) { return x->y ; }
int f2(X* x, int i) { return x->y * i; }
int f3(X* x, int i, int j) { return x->y * i + j; }

namespace boost {
   namespace python {
     namespace detail {

       template<>
       struct get_signature<FnObject, X>
       {
	typedef mpl::vector<int, X*, int> type;
       };

     }
   }
}

using namespace boost::python;

BOOST_PYTHON_MODULE( function_objects_ext )
{
   FnObject fobj;
   boost::function<int(X*, int)> bf0(fobj);

   boost::function<int(X*)> bf1(f1);
   boost::function<int(X*, int)> bf2(f2);
   boost::function<int(X*, int, double)> bf3(f3);

   boost::python::class_<X>( "X" )
     .def( "fobj", fobj)
     .def( "bf0",  bf0)
     .def( "bf1",  bf1)
     .def( "bf2",  bf2)
     .def( "bf3",  bf3)
     .def_readwrite( "y", &X::y )
     ;
}

/////////////////////////////

 >>> from function_objects_ext import *
 >>> x = X()
 >>> x.y = 13
 >>> x.fobj(12)
25
 >>> x.bf1()
13
 >>> x.bf2(2)
26
 >>> x.bf3(2, -27)
-1

/////////////////////////////

I'm fairly new to the internals of boost.python, and only just now got 
this working...  Do you see problems with this, specifically the 
conversion of get_signature from function to metafunction?

-t



From lists_ravi at lavabit.com  Mon Oct 12 02:32:59 2009
From: lists_ravi at lavabit.com (Ravi)
Date: Sun, 11 Oct 2009 20:32:59 -0400
Subject: [C++-sig] [python] Function objects in place of member functions
In-Reply-To: <4AD26DDD.5020402@resophonic.com>
References: <200910101158.46996.lists_ravi@lavabit.com>
	<4AD26DDD.5020402@resophonic.com>
Message-ID: <200910112032.59896.lists_ravi@lavabit.com>

On Sunday 11 October 2009 19:44:29 troy d. straszheim wrote:
> > Why is the overloaded get_signature not picked up when it is declared
> > after  the inclusion of the headers?
> 
> I'm not sure why it isn't picked up.

Does that mean that you can reproduce the problem I pointed out?

> I've been working in this area, 
> replacing most of detail/caller.hpp and detail/invoke.hpp with 
> boost.fusion, seen here:
> 
> http://gitorious.org/~straszheim/boost/straszheim/blobs/python/boost/python
> /detail/caller.hpp
> 
> In the process, I overhauled get_signature to use boost::function_types, 
> and to be a metafunction, not a function:
> 
> http://gitorious.org/~straszheim/boost/straszheim/blobs/python/boost/python
> /signature.hpp

IMHO, this is the right way to do it. This avoids relying on the compiler to 
optimize out all the ugly tag-dispatching. Of course, Dave A & Ralf WGK did 
not have function types when they wrote this originally.

[snip]

>    boost::function<int(X*, int)> bf0(fobj);

Why do you need to use boost::function here? Shouldn't the type be deduced 
automatically?

> I'm fairly new to the internals of boost.python, and only just now got 
> this working...  Do you see problems with this, specifically the 
> conversion of get_signature from function to metafunction?

I don't see any problems with the conversion of get_signature to a 
metafunction. Do compile times get any longer?

Regards,
Ravi


From rfritz333 at gmail.com  Mon Oct 12 06:14:46 2009
From: rfritz333 at gmail.com (Randolph Fritz)
Date: Mon, 12 Oct 2009 04:14:46 +0000 (UTC)
Subject: [C++-sig] Boost.python: extracting an array
Message-ID: <slrnhd5b9m.qjp.rfritz333@D-128-95-103-236.dhcp4.washington.edu>

Is there any way to extract a vector of floats from a python array?

In Python, can I write something like:
  from array import array
  from bpclass import foo

  vec = array('f', (1,2,3))
  foo ([vec])

And, in C++:
  void bpclass::foo (list v) {
    ... extract< {magic} >(v[0]) ...

Or would it be better to replace the array with a list?
-- 
Randolph Fritz
  design machine group, architecture department, university of washington
rfritz at u.washington.edu -or- rfritz333 at gmail.com


From hwang.roman at gmail.com  Mon Oct 12 08:59:58 2009
From: hwang.roman at gmail.com (Roman Hwang)
Date: Mon, 12 Oct 2009 15:59:58 +0900
Subject: [C++-sig] Returning pointer to abstract class from pure virtual
	class method
Message-ID: <499f6e2b0910112359m1c2f9fe7j7fd8fb65066af95c@mail.gmail.com>

Hello.

I'm new to Python, but really want to use it with my C++ code.

I faced a problem while returning pointer to abstract class from pure
virtual class method.
I tried to search the web, but found nothing.

struct ICoreWrapper : ICore, wrapper<ICore>
{
	virtual IRegistry* CreateRegistry()
		{ return this->get_override("CreateRegistry")(); }

};

struct IRegistryWrapper : IRegistry, wrapper<IRegistry>
{
	virtual void DoSomestuff()
                        { return this->get_override("DoSomestuff")(); }
};

ICore* CreateCoreW()
{
	return new CCore;
}

BOOST_PYTHON_MODULE_INIT(CorePy)
{
	class_<IRegistryWrapper, boost::noncopyable>("IRegistry")
		.def("DoSomestuff", pure_virtual(&IRegistry::DoSomestuff))
		;

	class_<ICoreWrapper, boost::noncopyable>("ICore")
		.def("CreateRegistry", pure_virtual(&ICore::CreateRegistry),
return_value_policy<reference_existing_object>())
		;

	def("CreateCore", CreateCoreW,
return_value_policy<reference_existing_object>());
}

The Python code looks like this:
>>> import CorePy
>>> c = CorePy.CreateCore()
>>> b = c.CreateRegistry()

I get Debug Error

Run-Time Check Failure #0 - The value of ESP was not properly saved
across a function call.
This is usually a result of calling a function declared with one
calling convention with a function pointer declared
with a different calling convention.

Please tell me what I do wrong.
Or whether this code is supported or not.

Thanks

Roman Hwang.

From diepen at astron.nl  Mon Oct 12 09:38:42 2009
From: diepen at astron.nl (Ger van Diepen)
Date: Mon, 12 Oct 2009 09:38:42 +0200
Subject: [C++-sig] Compile error when using exception_translation.hpp
Message-ID: <4AD2F921.6086.00A9.0@astron.nl>

When compiling the following code using gcc-4.3.2 and boost-1.37, I get
errors as listed at the end. It also fails for boost-1.40 and gcc-4.3.4.

It is solved when patching translate_exception.hpp by adding
# include <boost/type_traits/add_reference.hpp>

Is it a Boost-Python problem or do I do something wrong?
Of course, I can include that file in my own code, but that does not
seem the correct solution to me.

Thanks,
Ger van Diepen


  #include <boost/translate_exception.hpp>
  #include <stdexcept>
  void translate_stdexcp (const std::exception& e)
  {
    PyErr_SetString(PyExc_RuntimeError, e.what());
  }

  //# Note that the most general exception must be registered first.
  void register_convert_excp()
  {
    boost::python::register_exception_translator<std::exception>
      (&translate_stdexcp);
  }


/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:34:
error: expected nested-name-specifier before ?add_reference?
/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:34:
error: expected ?;? before ?<? token
/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:
In member function ?bool
boost::python::detail::translate_exception<ExceptionType,
Translate>::operator()(const
boost::python::detail::exception_handler&,
const boost::function0<void>&, typename
boost::call_traits<Translate>::param_type) const?:
/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:56:
error: expected type-specifier before ?exception_cref?
/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:56:
error: expected `)' before ?e?
/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:56:
error: expected `{' before ?e?
/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:56:
error: ?e? was not declared in this scope
/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:56:
error: expected `;' before ?)? token


From troy at resophonic.com  Mon Oct 12 14:47:22 2009
From: troy at resophonic.com (troy d. straszheim)
Date: Mon, 12 Oct 2009 08:47:22 -0400
Subject: [C++-sig] [python] Function objects in place of member functions
In-Reply-To: <200910112032.59896.lists_ravi@lavabit.com>
References: <200910101158.46996.lists_ravi@lavabit.com>	<4AD26DDD.5020402@resophonic.com>
	<200910112032.59896.lists_ravi@lavabit.com>
Message-ID: <4AD3255A.9090505@resophonic.com>

Ravi wrote:
> On Sunday 11 October 2009 19:44:29 troy d. straszheim wrote:
>>> Why is the overloaded get_signature not picked up when it is declared
>>> after  the inclusion of the headers?
>> I'm not sure why it isn't picked up.
> 
> Does that mean that you can reproduce the problem I pointed out?
> 

Yes I can.

>> I've been working in this area, 
>> replacing most of detail/caller.hpp and detail/invoke.hpp with 
>> boost.fusion, seen here:
>>
>> http://gitorious.org/~straszheim/boost/straszheim/blobs/python/boost/python
>> /detail/caller.hpp
>>
>> In the process, I overhauled get_signature to use boost::function_types, 
>> and to be a metafunction, not a function:
>>
>> http://gitorious.org/~straszheim/boost/straszheim/blobs/python/boost/python
>> /signature.hpp
> 
> IMHO, this is the right way to do it. This avoids relying on the compiler to 
> optimize out all the ugly tag-dispatching. Of course, Dave A & Ralf WGK did 
> not have function types when they wrote this originally.
> 

Yes, the code was all dated ~2002.  There are more corners of the code 
that could use modernization as well.

> [snip]
> 
>>    boost::function<int(X*, int)> bf0(fobj);
> 
> Why do you need to use boost::function here? Shouldn't the type be deduced 
> automatically?
> 

Note that the .def() of the various boost::function objects work without 
requiring the user to specialize get_signature, (since this modified 
get_signature metafunction knows about boost::function objects).  Here's 
a different (working) example:

#include <boost/function.hpp>
#include <boost/lexical_cast.hpp>
#include <boost/python.hpp>
#include <boost/python/class.hpp>
#include <boost/python/module.hpp>

struct X { int y; };

struct add_to
{
   template <typename T>
   T operator()(X* x, T t) const
   {
     return t + x->y;
   }

   std::string operator()(X* x, std::string s)
   {
     return s + boost::lexical_cast<std::string>(x->y);
   }

};

using namespace boost::python;

BOOST_PYTHON_MODULE( function_objects_ext )
{
   add_to add_to_inst;
   boost::function<int(X*, int)> f_int(add_to_inst);
   boost::function<std::string(X*, std::string)> f_string(add_to_inst);

   boost::python::class_<X>( "X" )
     .def_readwrite( "y", &X::y )
     .def( "add_to", f_int)
     .def( "add_to", f_string)
     ;
}

 >>> from function_objects_ext import *
 >>> x = X()
 >>> x.y = 13
 >>> x.add_to(1)
14
 >>> x.add_to("foo")
'foo13'

Some musings:

I think interoperability with boost::function is a Good Thing, but in 
the example above (and the previous), all it does is give def() a way to 
tell what the signature of the function object is at compile time, and 
it does so at a runtime cost.   A specialization of get_signature works, 
but strikes me as kinda ugly (it should at least be moved out of 
'detail' if it is a user customization point).  How about something like

   class_<T>("T")
     .def("f1", f1, sig<void(T*, double)>)
     .def<void(T*, double)>("fi", f1)

Which looks pretty straightforward when dealing with function objects, 
but when dealing with overloads, things get tricky:

   void f2(T*, int);
   void f2(T*, double);

   class_<T>("T")
     .def("f2", f2)                     //  #1 error
     .def("f2", (void(*)(T*, int))f2)   //  #2 ok but ugly
     .def<void(T*, int)>("f2", f2)      //  #3 works but tricky
     ;

#1 is an error, because f2's type is undetermined.

#2 is how you can fix #1 in current practice (or with an alias).

#3 could work (I've tried, it works) if def, when passed a function 
type, only accepts pointers to functions of that type, so that the 
overload is resolved.  But how to use this same form to also specify the 
types of arbitrary function objects... haven't thought about that yet.

>> I'm fairly new to the internals of boost.python, and only just now got 
>> this working...  Do you see problems with this, specifically the 
>> conversion of get_signature from function to metafunction?
> 
> I don't see any problems with the conversion of get_signature to a 
> metafunction. Do compile times get any longer?

I haven't checked this yet.  I have other concerns re typechecking and 
automatic conversions...  I'll go back and look more closely.

-t



From p.schellart at gmail.com  Mon Oct 12 17:36:20 2009
From: p.schellart at gmail.com (Pim Schellart)
Date: Mon, 12 Oct 2009 17:36:20 +0200
Subject: [C++-sig] Numpy ndarray as argument or return value using boost
	python
Message-ID: <7d3dada90910120836q6ba2b545sede71b6aef0543fe@mail.gmail.com>

Hi Everyone,

I also posted this to the boost users list.
I tried to find a good example for this but after spending two days on
google and one day getting pyublas to compile I decided to call for
help.
I want to give a numpy ndarray as an argument to a c++ method,
calculate something and return the output as an ndarray to Python.
I would prefer not to use another library (such as PyUblas) for this
so I do not have to add another dependency to my program.
Since the latest information on boost python is quite outdated (uses
numeric) I would like to know the following.

1. Is boost python still maintained or should I switch to another tool.
2. Is numpy supported, and if so,
3. can someone give me a basic example of how to use it.

Kind regards,

Pim Schellart

From seefeld at sympatico.ca  Mon Oct 12 19:57:21 2009
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Mon, 12 Oct 2009 13:57:21 -0400
Subject: [C++-sig] Numpy ndarray as argument or return value using boost
 python
In-Reply-To: <7d3dada90910120836q6ba2b545sede71b6aef0543fe@mail.gmail.com>
References: <7d3dada90910120836q6ba2b545sede71b6aef0543fe@mail.gmail.com>
Message-ID: <4AD36E01.4080303@sympatico.ca>

On 10/12/2009 11:36 AM, Pim Schellart wrote:
> Hi Everyone,
>
> I also posted this to the boost users list.
> I tried to find a good example for this but after spending two days on
> google and one day getting pyublas to compile I decided to call for
> help.
> I want to give a numpy ndarray as an argument to a c++ method,
> calculate something and return the output as an ndarray to Python.
> I would prefer not to use another library (such as PyUblas) for this
> so I do not have to add another dependency to my program.
> Since the latest information on boost python is quite outdated (uses
> numeric) I would like to know the following.
>
> 1. Is boost python still maintained or should I switch to another tool.
>    

Boost.Python is definitely maintained.

> 2. Is numpy supported, and if so,
>    

That is a good question, to which I don't have an answer. In fact, I 
submitted an issue about the docs 
(https://svn.boost.org/trac/boost/ticket/3340), as a first step to 
getting an up-to-date picture.
Meanwhile, I have written my own wrapper for PyArray, as the existing 
one is by far not enough for my needs. (In particular, I need direct 
memory access.)

I would be happy to share this code and move it into boost.python, if in 
fact everyone agrees that what exists should be replaced.

David, Ralf, do you have an opinion on this ?


> 3. can someone give me a basic example of how to use it.
>    


Regards,
         Stefan

-- 

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


From AFoglia at princeton.com  Mon Oct 12 19:54:43 2009
From: AFoglia at princeton.com (Anthony Foglia)
Date: Mon, 12 Oct 2009 13:54:43 -0400
Subject: [C++-sig] Numpy ndarray as argument or return value using boost
 python
In-Reply-To: <7d3dada90910120836q6ba2b545sede71b6aef0543fe@mail.gmail.com>
References: <7d3dada90910120836q6ba2b545sede71b6aef0543fe@mail.gmail.com>
Message-ID: <4AD36D63.1050207@princeton.com>

Pim Schellart wrote:
> I also posted this to the boost users list.

Most of the Boost Python discussion goes on here, so this is probably 
the better group.

> I want to give a numpy ndarray as an argument to a c++ method,
> calculate something and return the output as an ndarray to Python.
> I would prefer not to use another library (such as PyUblas) for this
> so I do not have to add another dependency to my program.

I did this myself last year, and here's what I can tell you.

> 1. Is boost python still maintained or should I switch to another tool.

Yes it is maintained.

> 2. Is numpy supported, and if so,

Yes and no.  Boost.Python has a header for dealing with Numeric arrays 
that works just as well with NumPy. 
<http://www.boost.org/doc/libs/1_40_0/libs/python/doc/v2/numeric.html>. 
  But it works by passing all methods through the Python interpreter, so 
you will see no increase in speed compared to using a more low-level 
NumPy API.

There is num_util header created by a third-party that claims to do a 
lot of the wrapping. 
<http://www.eos.ubc.ca/research/clouds/software/pythonlibs/num_util/num_util_release2/> 
  I ended up not using it, but instead going getting using NumPy's C API 
directly and keeping the pointers wrapped in boost::python::handle 
objects as much as possible.

> 3. can someone give me a basic example of how to use it.

The num_util page has an example of their code.

-- 
Anthony Foglia
Princeton Consultants
(609) 987-8787 x233


From cmbruns at stanford.edu  Mon Oct 12 22:51:32 2009
From: cmbruns at stanford.edu (Christopher Bruns)
Date: Mon, 12 Oct 2009 13:51:32 -0700
Subject: [C++-sig] [Py++] How to wrap default arg as enum with complicated
	scope?
Message-ID: <b056e7240910121351u1d3d1956ufd87f50e34d38fc2@mail.gmail.com>

Below is a simplified example of a tricky wrapping problem I have
encountered.  The example remains somewhat complicated, but further
simplification results in a file that causes no trouble.
I get a compile error when I try to build the boost python code
generated by the following case:

#### test_enum.h ####
struct System {
    struct ProjectOptions {
        enum Option {
            All = 1
        };
        ProjectOptions(Option);
    };
    void project(ProjectOptions = ProjectOptions::All);
};
#### end test_enum.h ####

The trouble comes from the default value ('ProjectOptions::All') for
the project() method.  Notice also the constructor for ProjectOptions
from Options.  That is required for reproducing the trouble here.

The resulting boost.python code is as follows:

#### wrap_enum2.py ####
from pyplusplus import module_builder

mb = module_builder.module_builder_t(["test_enum2.h"]
    , gccxml_path = "gccxml.exe")
mb.build_code_creator( module_name='test' )
mb.write_module( 'test_wrap_enum2.cpp' )
#### end wrap_enum2.py ####

#### test_wrap_enum2.cpp ####
// This file has been generated by Py++.

#include "boost/python.hpp"

#include "test_enum2.h"

namespace bp = boost::python;

BOOST_PYTHON_MODULE(test){
    { //::System
        typedef bp::class_< System > System_exposer_t;
        System_exposer_t System_exposer = System_exposer_t( "System" );
        bp::scope System_scope( System_exposer );
        { //::System::ProjectOptions
            typedef bp::class_< System::ProjectOptions >
ProjectOptions_exposer_t;
            ProjectOptions_exposer_t ProjectOptions_exposer =
ProjectOptions_exposer_t( "ProjectOptions", bp::init<
System::ProjectOptions::Option >(( bp::arg("arg0") )) );
            bp::scope ProjectOptions_scope( ProjectOptions_exposer );
            bp::enum_< System::ProjectOptions::Option>("Option")
                .value("All", System::ProjectOptions::All)
                .export_values()
                ;
            bp::implicitly_convertible<
System::ProjectOptions::Option, System::ProjectOptions >();
        }
        { //::System::project

            typedef void ( ::System::*project_function_type )(
::System::ProjectOptions ) ;

            System_exposer.def(
                "project"
                , project_function_type( &::System::project )
                , ( bp::arg("arg0")=All ) );
        }
    }
}
#### test_wrap_enum2.cpp ####

The symbol "All" in the line ", ( bp::arg("arg0")=All ) );" is out of
scope here.  If I change the line from

  , ( bp::arg("arg0")=All ) ); // compile failure

to

  , ( bp::arg("arg0")=System::ProjectOptions::All ) ); // works

it compiles fine.

Is there any way to instruct pyplusplus to generate the default
argument value for the project() method with the scope
"System::ProjectOptions::All"?

Any enlightenment is much appreciated.

--Chris Bruns

From roman.yakovenko at gmail.com  Tue Oct 13 00:39:04 2009
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Tue, 13 Oct 2009 00:39:04 +0200
Subject: [C++-sig] [Py++] How to wrap default arg as enum with
	complicated scope?
In-Reply-To: <b056e7240910121351u1d3d1956ufd87f50e34d38fc2@mail.gmail.com>
References: <b056e7240910121351u1d3d1956ufd87f50e34d38fc2@mail.gmail.com>
Message-ID: <7465b6170910121539i4b16cc8am5ab1d38b42fc474@mail.gmail.com>

On Mon, Oct 12, 2009 at 10:51 PM, Christopher Bruns
<cmbruns at stanford.edu> wrote:
> Below is a simplified example of a tricky wrapping problem I have
> encountered. ?The example remains somewhat complicated, but further
> simplification results in a file that causes no trouble.
> I get a compile error when I try to build the boost python code
> generated by the following case:
>
> #### test_enum.h ####
> struct System {
> ? ?struct ProjectOptions {
> ? ? ? ?enum Option {
> ? ? ? ? ? ?All = 1
> ? ? ? ?};
> ? ? ? ?ProjectOptions(Option);
> ? ?};
> ? ?void project(ProjectOptions = ProjectOptions::All);
> };
> #### end test_enum.h ####
>
> The trouble comes from the default value ('ProjectOptions::All') for
> the project() method. ?Notice also the constructor for ProjectOptions
> from Options. ?That is required for reproducing the trouble here.
>
> The resulting boost.python code is as follows:
> ...
> #### test_wrap_enum2.cpp ####
>
> The symbol "All" in the line ", ( bp::arg("arg0")=All ) );" is out of
> scope here. ?If I change the line from
>
> ?, ( bp::arg("arg0")=All ) ); // compile failure
>
> to
>
> ?, ( bp::arg("arg0")=System::ProjectOptions::All ) ); // works
>
> it compiles fine.
>
> Is there any way to instruct pyplusplus to generate the default
> argument value for the project() method with the scope
> "System::ProjectOptions::All"?

Yes:

mb = module_builder_t(...)
project = mb.mem_fun( 'project' )
project.arguments[0].default_value = "System::ProjectOptions::All"

> Any enlightenment is much appreciated.

Basically this is GCCXML limitations. The default argument could be
full-blown C++ expression. GCCXML doesn't handles\dumps them. It do
writes something for default arguments and Py++(pygccxml) try to
decode what it is.

May be the following link could help you:
http://language-binding.net/pygccxml/upgrade_issues.html
http://language-binding.net/pyplusplus/documentation/functions/functions.html

Also try google, this problem was discussed on this, pygccxml and
gccxml mailing lists.

HTH

-- 
Roman Yakovenko
C++ Python language binding
http://www.language-binding.net/

From lists_ravi at lavabit.com  Tue Oct 13 03:04:46 2009
From: lists_ravi at lavabit.com (Ravi)
Date: Mon, 12 Oct 2009 21:04:46 -0400
Subject: [C++-sig] [python] Function objects in place of member functions
In-Reply-To: <4AD3255A.9090505@resophonic.com>
References: <200910101158.46996.lists_ravi@lavabit.com>
	<200910112032.59896.lists_ravi@lavabit.com>
	<4AD3255A.9090505@resophonic.com>
Message-ID: <200910122104.46505.lists_ravi@lavabit.com>

On Monday 12 October 2009 08:47:22 troy d. straszheim wrote:
> >>    boost::function<int(X*, int)> bf0(fobj);
> >
> > 
> > Why do you need to use boost::function here? Shouldn't the type be
> > deduced  automatically?
> > 
> 
> Note that the .def() of the various boost::function objects work without 
> requiring the user to specialize get_signature, (since this modified 
> get_signature metafunction knows about boost::function objects).

[snip]

> I think interoperability with boost::function is a Good Thing, but in 
> the example above (and the previous), all it does is give def() a way to 
> tell what the signature of the function object is at compile time, and 
> it does so at a runtime cost.

At least for my case, that extra virtual function call is a killer.

> A specialization of get_signature works, 
> but strikes me as kinda ugly (it should at least be moved out of 
> 'detail' if it is a user customization point).  How about something like
> 
>    class_<T>("T")
>      .def("f1", f1, sig<void(T*, double)>)
>      .def<void(T*, double)>("fi", f1)

I'd rather have something along the lines of 
      .def< mpl::vector<void,T*,double> >("f1",f1)
by using function_types to compose the currect function signature. 
Alternatively, one could have the sig<> template above be a def_visitor and 
take the appropriate template arguments without the need to modify class_::def 
at all.

> >> I'm fairly new to the internals of boost.python, and only just now got 
> >> this working...  Do you see problems with this, specifically the 
> >> conversion of get_signature from function to metafunction?
> >
> > 
> > I don't see any problems with the conversion of get_signature to a 
> > metafunction. Do compile times get any longer?
> 
> I haven't checked this yet.  I have other concerns re typechecking and 
> automatic conversions...

Care to elaborate?

Regards,
Ravi



From lists_ravi at lavabit.com  Tue Oct 13 03:11:44 2009
From: lists_ravi at lavabit.com (Ravi)
Date: Mon, 12 Oct 2009 21:11:44 -0400
Subject: [C++-sig] Numpy ndarray as argument or return value using boost
	python
In-Reply-To: <7d3dada90910120836q6ba2b545sede71b6aef0543fe@mail.gmail.com>
References: <7d3dada90910120836q6ba2b545sede71b6aef0543fe@mail.gmail.com>
Message-ID: <200910122111.44171.lists_ravi@lavabit.com>

On Monday 12 October 2009 11:36:20 Pim Schellart wrote:
> 1. Is boost python still maintained or should I switch to another tool.
> 2. Is numpy supported, and if so,
> 3. can someone give me a basic example of how to use it.

Please see 
 http://mail.python.org/pipermail/cplusplus-sig/2008-October/013825.html
which supports numpy arrays with and without copying, along with pass by 
reference and pass by value.

Regards,
Ravi


From p.schellart at gmail.com  Tue Oct 13 12:18:32 2009
From: p.schellart at gmail.com (Pim Schellart)
Date: Tue, 13 Oct 2009 12:18:32 +0200
Subject: [C++-sig] Cplusplus-sig Digest, Vol 13, Issue 11
In-Reply-To: <mailman.25.1255428002.28310.cplusplus-sig@python.org>
References: <mailman.25.1255428002.28310.cplusplus-sig@python.org>
Message-ID: <7d3dada90910130318t79361207sa7110a416c0f841e@mail.gmail.com>

Dear Ravi,

thanks for the link. I'll be sure to try that.
Any chance this will be included (and documented) in boost python?
Numpy is needed for just about any scientific software package using
Python and it would be a real plus if boost Python had an easy to use
interface to C++ for it.

Kind regards,

Pim Schellart

2009/10/13  <cplusplus-sig-request at python.org>:
> Send Cplusplus-sig mailing list submissions to
> ? ? ? ?cplusplus-sig at python.org
>
> To subscribe or unsubscribe via the World Wide Web, visit
> ? ? ? ?http://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: [Py++] How to wrap default arg as enum with ? complicated
> ? ? ?scope? (Roman Yakovenko)
> ? 2. Re: [python] Function objects in place of member functions (Ravi)
> ? 3. Re: Numpy ndarray as argument or return value using boost
> ? ? ?python (Ravi)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Tue, 13 Oct 2009 00:39:04 +0200
> From: Roman Yakovenko <roman.yakovenko at gmail.com>
> To: "Development of Python/C++ integration" <cplusplus-sig at python.org>
> Subject: Re: [C++-sig] [Py++] How to wrap default arg as enum with
> ? ? ? ?complicated scope?
> Message-ID:
> ? ? ? ?<7465b6170910121539i4b16cc8am5ab1d38b42fc474 at mail.gmail.com>
> Content-Type: text/plain; charset=ISO-8859-1
>
> On Mon, Oct 12, 2009 at 10:51 PM, Christopher Bruns
> <cmbruns at stanford.edu> wrote:
>> Below is a simplified example of a tricky wrapping problem I have
>> encountered. ?The example remains somewhat complicated, but further
>> simplification results in a file that causes no trouble.
>> I get a compile error when I try to build the boost python code
>> generated by the following case:
>>
>> #### test_enum.h ####
>> struct System {
>> ? ?struct ProjectOptions {
>> ? ? ? ?enum Option {
>> ? ? ? ? ? ?All = 1
>> ? ? ? ?};
>> ? ? ? ?ProjectOptions(Option);
>> ? ?};
>> ? ?void project(ProjectOptions = ProjectOptions::All);
>> };
>> #### end test_enum.h ####
>>
>> The trouble comes from the default value ('ProjectOptions::All') for
>> the project() method. ?Notice also the constructor for ProjectOptions
>> from Options. ?That is required for reproducing the trouble here.
>>
>> The resulting boost.python code is as follows:
>> ...
>> #### test_wrap_enum2.cpp ####
>>
>> The symbol "All" in the line ", ( bp::arg("arg0")=All ) );" is out of
>> scope here. ?If I change the line from
>>
>> ?, ( bp::arg("arg0")=All ) ); // compile failure
>>
>> to
>>
>> ?, ( bp::arg("arg0")=System::ProjectOptions::All ) ); // works
>>
>> it compiles fine.
>>
>> Is there any way to instruct pyplusplus to generate the default
>> argument value for the project() method with the scope
>> "System::ProjectOptions::All"?
>
> Yes:
>
> mb = module_builder_t(...)
> project = mb.mem_fun( 'project' )
> project.arguments[0].default_value = "System::ProjectOptions::All"
>
>> Any enlightenment is much appreciated.
>
> Basically this is GCCXML limitations. The default argument could be
> full-blown C++ expression. GCCXML doesn't handles\dumps them. It do
> writes something for default arguments and Py++(pygccxml) try to
> decode what it is.
>
> May be the following link could help you:
> http://language-binding.net/pygccxml/upgrade_issues.html
> http://language-binding.net/pyplusplus/documentation/functions/functions.html
>
> Also try google, this problem was discussed on this, pygccxml and
> gccxml mailing lists.
>
> HTH
>
> --
> Roman Yakovenko
> C++ Python language binding
> http://www.language-binding.net/
>
>
> ------------------------------
>
> Message: 2
> Date: Mon, 12 Oct 2009 21:04:46 -0400
> From: Ravi <lists_ravi at lavabit.com>
> To: "Development of Python/C++ integration" <cplusplus-sig at python.org>
> Subject: Re: [C++-sig] [python] Function objects in place of member
> ? ? ? ?functions
> Message-ID: <200910122104.46505.lists_ravi at lavabit.com>
> Content-Type: Text/Plain; ?charset="iso-8859-1"
>
> On Monday 12 October 2009 08:47:22 troy d. straszheim wrote:
>> >> ? ?boost::function<int(X*, int)> bf0(fobj);
>> >
>> >
>> > Why do you need to use boost::function here? Shouldn't the type be
>> > deduced ?automatically?
>> >
>>
>> Note that the .def() of the various boost::function objects work without
>> requiring the user to specialize get_signature, (since this modified
>> get_signature metafunction knows about boost::function objects).
>
> [snip]
>
>> I think interoperability with boost::function is a Good Thing, but in
>> the example above (and the previous), all it does is give def() a way to
>> tell what the signature of the function object is at compile time, and
>> it does so at a runtime cost.
>
> At least for my case, that extra virtual function call is a killer.
>
>> A specialization of get_signature works,
>> but strikes me as kinda ugly (it should at least be moved out of
>> 'detail' if it is a user customization point). ?How about something like
>>
>> ? ?class_<T>("T")
>> ? ? ?.def("f1", f1, sig<void(T*, double)>)
>> ? ? ?.def<void(T*, double)>("fi", f1)
>
> I'd rather have something along the lines of
> ? ? ?.def< mpl::vector<void,T*,double> >("f1",f1)
> by using function_types to compose the currect function signature.
> Alternatively, one could have the sig<> template above be a def_visitor and
> take the appropriate template arguments without the need to modify class_::def
> at all.
>
>> >> I'm fairly new to the internals of boost.python, and only just now got
>> >> this working... ?Do you see problems with this, specifically the
>> >> conversion of get_signature from function to metafunction?
>> >
>> >
>> > I don't see any problems with the conversion of get_signature to a
>> > metafunction. Do compile times get any longer?
>>
>> I haven't checked this yet. ?I have other concerns re typechecking and
>> automatic conversions...
>
> Care to elaborate?
>
> Regards,
> Ravi
>
>
>
>
> ------------------------------
>
> Message: 3
> Date: Mon, 12 Oct 2009 21:11:44 -0400
> From: Ravi <lists_ravi at lavabit.com>
> To: "Development of Python/C++ integration" <cplusplus-sig at python.org>
> Subject: Re: [C++-sig] Numpy ndarray as argument or return value using
> ? ? ? ?boost ? python
> Message-ID: <200910122111.44171.lists_ravi at lavabit.com>
> Content-Type: Text/Plain; ?charset="iso-8859-1"
>
> On Monday 12 October 2009 11:36:20 Pim Schellart wrote:
>> 1. Is boost python still maintained or should I switch to another tool.
>> 2. Is numpy supported, and if so,
>> 3. can someone give me a basic example of how to use it.
>
> Please see
> ?http://mail.python.org/pipermail/cplusplus-sig/2008-October/013825.html
> which supports numpy arrays with and without copying, along with pass by
> reference and pass by value.
>
> Regards,
> Ravi
>
>
>
> ------------------------------
>
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
>
> End of Cplusplus-sig Digest, Vol 13, Issue 11
> *********************************************
>

From loeb at fkp.tu-darmstadt.de  Tue Oct 13 15:44:42 2009
From: loeb at fkp.tu-darmstadt.de (=?ISO-8859-1?Q?Daniel_L=F6b?=)
Date: Tue, 13 Oct 2009 15:44:42 +0200
Subject: [C++-sig] object constructor + integer == fail
Message-ID: <4AD4844A.7070302@fkp.tu-darmstadt.de>

Hi!

I seem to be unable to call the constructor of "object" with an integer 
parameter.
Working example:

#include <boost/python.hpp>
#include <iostream>

using namespace boost::python;
using std::cout;
using std::endl;

int main()
{
	object toast(3);
	cout << extract<int>(toast) << endl;
}

It compiles just fine, but dies with a segmentation fault on execution 
of the first line within main.

Backtrace:

0x00007ffff7ac9ca3 in PyInt_FromLong () from /usr/lib/libpython2.5.so.1.0
#1  0x0000000000401860 in arg_to_python (this=0x7fffffffeb30, 
x=@0x7fffffffebdc) at 
/usr/include/boost/python/converter/builtin_converters.hpp:113
#2  0x0000000000401888 in 
boost::python::api::object_initializer_impl<false, false>::get<int> 
(x=@0x7fffffffebdc)
     at /usr/include/boost/python/object_core.hpp:374
#3  0x00000000004018cb in 
boost::python::api::object_base_initializer<int> (x=@0x7fffffffebdc) at 
/usr/include/boost/python/object_core.hpp:296
#4  0x00000000004018e7 in object<int> (this=0x7fffffffebd0, 
x=@0x7fffffffebdc) at /usr/include/boost/python/object_core.hpp:315
#5  0x0000000000401196 in main () at bp_bug.cpp:10

Doing the same as above with a string instead of an integer works fine.
I'm running debian with libboost-python1.38.0 .

Greetings,
Daniel

From troy at resophonic.com  Tue Oct 13 16:12:32 2009
From: troy at resophonic.com (troy d. straszheim)
Date: Tue, 13 Oct 2009 10:12:32 -0400
Subject: [C++-sig] object constructor + integer == fail
In-Reply-To: <4AD4844A.7070302@fkp.tu-darmstadt.de>
References: <4AD4844A.7070302@fkp.tu-darmstadt.de>
Message-ID: <4AD48AD0.5060104@resophonic.com>

Daniel L?b wrote:
> Hi!
> 
> I seem to be unable to call the constructor of "object" with an integer 
> parameter.
> Working example:
> 
> #include <boost/python.hpp>
> #include <iostream>
> 
> using namespace boost::python;
> using std::cout;
> using std::endl;
> 
> int main()
> {
>     object toast(3);
>     cout << extract<int>(toast) << endl;
> }
> 
> It compiles just fine, but dies with a segmentation fault on execution 
> of the first line within main.
> 

Does calling Py_Initialize() help?

-t



From loeb at fkp.tu-darmstadt.de  Tue Oct 13 16:19:02 2009
From: loeb at fkp.tu-darmstadt.de (=?ISO-8859-1?Q?Daniel_L=F6b?=)
Date: Tue, 13 Oct 2009 16:19:02 +0200
Subject: [C++-sig] object constructor + integer == fail
In-Reply-To: <4AD48AD0.5060104@resophonic.com>
References: <4AD4844A.7070302@fkp.tu-darmstadt.de>
	<4AD48AD0.5060104@resophonic.com>
Message-ID: <4AD48C56.6010203@fkp.tu-darmstadt.de>

troy d. straszheim wrote:
> Daniel L?b wrote:
>> Hi!
>>
>> I seem to be unable to call the constructor of "object" with an 
>> integer parameter.
>> Working example:
>>
>> #include <boost/python.hpp>
>> #include <iostream>
>>
>> using namespace boost::python;
>> using std::cout;
>> using std::endl;
>>
>> int main()
>> {
>>     object toast(3);
>>     cout << extract<int>(toast) << endl;
>> }
>>
>> It compiles just fine, but dies with a segmentation fault on execution 
>> of the first line within main.
>>
> 
> Does calling Py_Initialize() help?
> 
Yes, that did it.

Thank you!

Best regards,
Daniel

From rwgk at yahoo.com  Tue Oct 13 18:51:29 2009
From: rwgk at yahoo.com (Ralf W. Grosse-Kunstleve)
Date: Tue, 13 Oct 2009 09:51:29 -0700 (PDT)
Subject: [C++-sig] Compile error when using exception_translation.hpp
In-Reply-To: <4AD2F921.6086.00A9.0@astron.nl>
References: <4AD2F921.6086.00A9.0@astron.nl>
Message-ID: <797789.92141.qm@web111402.mail.gq1.yahoo.com>

translate_exception.hpp is in the "detail" directory.
The convention is that this isn't a public interface.
#include <boost/python/exception_translator.hpp>
fixes your problem (at least under Fedora 8).
Ralf



----- Original Message ----
From: Ger van Diepen <diepen at astron.nl>
To: boost-users at lists.boost.org; cplusplus-sig at python.org
Sent: Mon, October 12, 2009 12:38:42 AM
Subject: [C++-sig] Compile error when using exception_translation.hpp

When compiling the following code using gcc-4.3.2 and boost-1.37, I get
errors as listed at the end. It also fails for boost-1.40 and gcc-4.3.4.

It is solved when patching translate_exception.hpp by adding
# include <boost/type_traits/add_reference.hpp>

Is it a Boost-Python problem or do I do something wrong?
Of course, I can include that file in my own code, but that does not
seem the correct solution to me.

Thanks,
Ger van Diepen


  #include <boost/translate_exception.hpp>
  #include <stdexcept>
  void translate_stdexcp (const std::exception& e)
  {
    PyErr_SetString(PyExc_RuntimeError, e.what());
  }

  //# Note that the most general exception must be registered first.
  void register_convert_excp()
  {
    boost::python::register_exception_translator<std::exception>
      (&translate_stdexcp);
  }


/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:34:
error: expected nested-name-specifier before ?add_reference?
/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:34:
error: expected ?;? before ?<? token
/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:
In member function ?bool
boost::python::detail::translate_exception<ExceptionType,
Translate>::operator()(const
boost::python::detail::exception_handler&,
const boost::function0<void>&, typename
boost::call_traits<Translate>::param_type) const?:
/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:56:
error: expected type-specifier before ?exception_cref?
/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:56:
error: expected `)' before ?e?
/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:56:
error: expected `{' before ?e?
/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:56:
error: ?e? was not declared in this scope
/home/jds/Work/svn-usg/release/include/boost/python/detail/translate_exception.hpp:56:
error: expected `;' before ?)? token

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

From troy at resophonic.com  Thu Oct 15 03:59:42 2009
From: troy at resophonic.com (troy d. straszheim)
Date: Wed, 14 Oct 2009 21:59:42 -0400
Subject: [C++-sig] [python] python + phoenix
In-Reply-To: <200910122104.46505.lists_ravi@lavabit.com>
References: <200910101158.46996.lists_ravi@lavabit.com>	<200910112032.59896.lists_ravi@lavabit.com>	<4AD3255A.9090505@resophonic.com>
	<200910122104.46505.lists_ravi@lavabit.com>
Message-ID: <4AD6820E.9040302@resophonic.com>

Ravi wrote:
 >
 > I'd rather have something along the lines of
 >       .def< mpl::vector<void,T*,double> >("f1",f1)

I have something working.  There is still a bunch of stuff to iron out
yet.  I went with

   def("name", as<Signature>(callable));

where Signature is the signature with which callable will be called,
e.g.

   struct times_seven
   {
     template <typename T> struct result;

     template <typename Self, typename T>
     struct result<Self(T&)>  // annoying: reference
     {
       typedef T type;
     };

     template <typename T>
     T operator()(const T& t)
     {
       return t * 7;
     }
   };

   BOOST_PYTHON_MODULE( function_objects3_ext )
   {
     def("times_seven", as<int(int)>(times_seven()));
   }

the T& in the struct result is something I haven't yet sorted out, has
something to do with how I am constructing fusion sequences from the
converted python objects before calling fusion::invoke.  Anyhow, it
works:

   >>> times_seven(7)
   49

I thought that one might try to remove the necessity of having the
return type in the Signature and/or implementing the result_of
protocol, but maybe it is nice that they don't have to match:

   def("times_seven", as<double(int)>(times_seven()));

   >>> times_seven(7)
   49.0

boost::function still works, and doesn't require as<>:

   boost::function<int(int, int)> myplus = std::plus<int>();
   def("myplus", myplus);

and old-school function objects:

   def("otherplus", std::plus<int>())

I was surprised to find that it all plays nice with boost::phoenix:

   def("plus", as<int(int, int)>(arg1 + arg2));

   def("throw_if_gt_5",
       as<void(int)>(if_(arg1 > 5)
		    [ throw_(std::runtime_error("too big!!")) ]
		    .else_
		    [ std::cout << val("okay")                ]));

And boost::phoenix plays nice with shared_ptr, which plays nice with
boost::python:

   struct X { int x; };   typedef boost::shared_ptr<X> XPtr;

   class_<X, XPtr>("X")
     .def_readwrite("x", &X::x)
     .def("add7", as<void(X*)>  (arg1->*&X::x      += 7))
     .def("add8", as<void(XPtr)>(arg1->*&X::x      += 8))
     .def("add9", as<void(X&)>  (bind(&X::x, arg1) += 9))
   ;

the other little thing is overload resolution, for which I propose a
little helper function "unload" (underload? deload? resolve?  I dunno):

   void foo(int);
   void foo(double);

   def("foo", unload<void(int)>(foo));

Which is safer than

   def("foo", (void(*)(int)) foo);

and just as safe as, and less verbose than

   void(*fooint)(int) = foo;
   def("foo", fooint);

So that's where I'm at...  what do you make of that interface?

-t


From micdestefano at gmail.com  Mon Oct 19 17:34:35 2009
From: micdestefano at gmail.com (Michele De Stefano)
Date: Mon, 19 Oct 2009 17:34:35 +0200
Subject: [C++-sig] Help with converters ...
Message-ID: <c410adc90910190834j553865bbof1410be6c4a85aac@mail.gmail.com>

Hello to everyone.

I'm stuck with this problem. I'm trying to write a "from Python
converter", from a tuple(tuple,tuple) type to C++.
Clarifying more, the Python type is a tuple, containing 2 tuples.

Now, when I try to implement the "convertible" static member function
of the converter class, I write this (I've extracted only the
interesting part):

static void* convertible(PyObject* obj_ptr) {
		
	using namespace boost::python;

	handle<>	hndl(borrowed(obj_ptr));
	object	in(hndl);
		
	if (!PyTuple_Check(in.ptr())) return NULL;
		
	extract<const tuple&>
             get_tup0(in[0]),
             get_tup1(in[1]);
		
        bool
	     t1 = get_tup0.check(),
	     t2 = get_tup1.check();

...... the rest of the code is not necessary ....


The problem is that even if I pass a tuple containing two tuples from
Python, t1 and t2 become "false".
Surely I'm not interfacing properly with Python, but I don't understand why.

Can someone help me to understand ?

Thank you in advance.
Michele


-- 
Michele De Stefano
http://www.linkedin.com/in/micdestefano
http://xoomer.virgilio.it/michele_de_stefano

From hans_meine at gmx.net  Mon Oct 19 20:44:30 2009
From: hans_meine at gmx.net (Hans Meine)
Date: Mon, 19 Oct 2009 20:44:30 +0200
Subject: [C++-sig] Help with converters ...
In-Reply-To: <c410adc90910190834j553865bbof1410be6c4a85aac@mail.gmail.com>
References: <c410adc90910190834j553865bbof1410be6c4a85aac@mail.gmail.com>
Message-ID: <200910192044.39774.hans_meine@gmx.net>

Hi Michele!

On Montag 19 Oktober 2009, Michele De Stefano wrote:
> Now, when I try to implement the "convertible" static member function
> of the converter class, I write this (I've extracted only the
> interesting part):
>
> [...]
> 	extract<const tuple&>
>              get_tup0(in[0]),
>              get_tup1(in[1]);

This looks like the problem to me; you're effectively trying to get a 
*reference* from python, i.e. this can only work if there is already an 
*existing* boost::python::tuple object hidden in (in[0]) and (in[1]).
(This will typically only work with BPL-exported C++ objects wrapped into 
their python shells.)

I'd try simply removing the const &: extract<tuple>..

-- 
Ciao, /  /                                                    .o.
     /--/                                                     ..o
    /  / ANS                                                  ooo
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20091019/6cf35908/attachment.pgp>

From micdestefano at gmail.com  Tue Oct 20 09:15:21 2009
From: micdestefano at gmail.com (Michele De Stefano)
Date: Tue, 20 Oct 2009 09:15:21 +0200
Subject: [C++-sig] Help with converters ...
In-Reply-To: <200910192044.39774.hans_meine@gmx.net>
References: <c410adc90910190834j553865bbof1410be6c4a85aac@mail.gmail.com>
	<200910192044.39774.hans_meine@gmx.net>
Message-ID: <c410adc90910200015v7a7d5a62l5ec1f9fbaa830152@mail.gmail.com>

Hans,

thank you very much. Removing the substituting "const tuple&" with
"tuple" works.

Just to let me understand more your answer, what does it mean "BPL-exported" ?
I'm not so comfortable with acronyms ...


2009/10/19 Hans Meine <hans_meine at gmx.net>:
> Hi Michele!
>
> On Montag 19 Oktober 2009, Michele De Stefano wrote:
>> Now, when I try to implement the "convertible" static member function
>> of the converter class, I write this (I've extracted only the
>> interesting part):
>>
>> [...]
>> ? ? ? extract<const tuple&>
>> ? ? ? ? ? ? ?get_tup0(in[0]),
>> ? ? ? ? ? ? ?get_tup1(in[1]);
>
> This looks like the problem to me; you're effectively trying to get a
> *reference* from python, i.e. this can only work if there is already an
> *existing* boost::python::tuple object hidden in (in[0]) and (in[1]).
> (This will typically only work with BPL-exported C++ objects wrapped into
> their python shells.)
>
> I'd try simply removing the const &: extract<tuple>..
>
> --
> Ciao, / ?/ ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?.o.
> ? ? /--/ ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ..o
> ? ?/ ?/ ANS ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?ooo
>
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
>



-- 
Michele De Stefano
http://www.linkedin.com/in/micdestefano
http://xoomer.virgilio.it/michele_de_stefano

From nitroamos at gmail.com  Thu Oct 22 21:13:21 2009
From: nitroamos at gmail.com (Amos Anderson)
Date: Thu, 22 Oct 2009 12:13:21 -0700
Subject: [C++-sig] How to install extensions using bjam(?)
Message-ID: <9e910970910221213j5b3d5a73xcc4d34fddc095ca4@mail.gmail.com>

Hello --

I'm working on a project which is mixed C++/Python, and we use Boost
functionality in our C++, and we've set up bjam to compile our project
(on OSX). Everything seems to be working quite well.

To use our libraries from python scripts right now, we have to run this first:

#!/bin/bash
trunk=$HOME/project/trunk/
export DYLD_LIBRARY_PATH=$DYLD_LIBRARY_PATH:$trunk/lib
export PYTHONPATH=$trunk:$trunk/lib
python $@


which is quite inconvenient. Now obviously I can skip this step by
pasting those two export statements into my .profile, however, at some
point we're going to start distributing this software and it would be
nice if we didn't require the end user to do something like that.

I can't believe that either of these 2 steps are how it's supposed to
work... so how should I be doing it?  Ideally, I want to set something
up so that the end user only has to run 1 (one) command to build &
install. Here are some things I've thought about:

1) Having bjam edit the user's .profile to set up these paths,
possibly by calling a script I write. This doesn't seem like the right
approach to me.

2) It looks like python's distutils is supposed to be able to help,
but it seems like a waste of time to figure out how to get distutils
to compile my code when I already have bjam pretty much figured out.
First, I assume that distutils can put my libraries on an automatic
DYLD_LIBRARY_PATH path that python knows about, because if it can't
then there's no point. But assuming that's not a problem, I can't
figure out how to tell distutils about Extensions that it shouldn't
try to build...

3) bjam includes a way to install libraries into a folder I can name,
but does it have a way to set environment variables? or automatically
figure out the right places to put all the python files? If so, I
can't find it...

4) ?


thanks!

Amos.

From nat at lindenlab.com  Thu Oct 22 21:48:09 2009
From: nat at lindenlab.com (Nat Goodspeed)
Date: Thu, 22 Oct 2009 15:48:09 -0400
Subject: [C++-sig] How to install extensions using bjam(?)
In-Reply-To: <9e910970910221213j5b3d5a73xcc4d34fddc095ca4@mail.gmail.com>
References: <9e910970910221213j5b3d5a73xcc4d34fddc095ca4@mail.gmail.com>
Message-ID: <4AE0B6F9.2040705@lindenlab.com>

Amos Anderson wrote:

> I'm working on a project which is mixed C++/Python, and we use Boost
> functionality in our C++, and we've set up bjam to compile our project
> (on OSX). Everything seems to be working quite well.
> 
> 2) It looks like python's distutils is supposed to be able to help,
> but it seems like a waste of time to figure out how to get distutils
> to compile my code when I already have bjam pretty much figured out.
> First, I assume that distutils can put my libraries on an automatic
> DYLD_LIBRARY_PATH path that python knows about, because if it can't
> then there's no point. But assuming that's not a problem, I can't
> figure out how to tell distutils about Extensions that it shouldn't
> try to build...

In a previous project I've had success with py2app:
http://svn.pythonmac.org/py2app/py2app/trunk/doc/index.html

We had a number of extensions we built ourselves (not with distutils or 
py2app itself) and it happily packaged them into a working app bundle.

From micdestefano at gmail.com  Fri Oct 23 10:38:30 2009
From: micdestefano at gmail.com (Michele De Stefano)
Date: Fri, 23 Oct 2009 10:38:30 +0200
Subject: [C++-sig] Python converters for raw pointers: request for help
Message-ID: <c410adc90910230138k77675a29n6e2fe435d2c2b21b@mail.gmail.com>

Hello to everyone.

I'm trying to write "from Python" converters for PETSc objects of the
petsc4py package.
Even if no-one of you uses petsc4py, I think you can help me, if I
explain some basics.
Please, be patient.

I want to make converters for the PETSc Mat type.

Mat is defined as below:

typedef struct _p_Mat*           Mat;

where _p_Mat is a structure whose details are not interesting for my problem.

Now, I managed to make a functioning "to Python converter".

My problem is the "from Python converter".

I made a test function

void print_Mat(Mat m) {
    ... // The implementation is not interesting
};

And I tried to make a "from Python converter" with the code below
(please, look at it ... it's not too long and it's very simple):


struct PETSc_Mat_from_Python {
	
	PETSc_Mat_from_Python() {

		static bool		do_registration(true);
		
		if (do_registration) {
			
			boost::python::converter::registry::push_back(
					&convertible,&construct,boost::python::type_id<Mat>());
			
			do_registration = false;
		}
	}
	
	static void* convertible(PyObject* obj_ptr) {
		
                // In the line below, PyPetscMat_Get takes a PyObject*
as input and returns a Mat
                // (if it succedes) or a PETSC_NULL (if it fails).
This is a function of the petsc4py C API.
                // The ownership of the returned Mat is not
transferred (I think)

		if (PyPetscMat_Get(obj_ptr) == PETSC_NULL) return NULL;
		
		return obj_ptr;
	}
	
	static void construct(PyObject* obj_ptr,
			boost::python::converter::rvalue_from_python_stage1_data* data) {

		void* storage = (
				(boost::python::converter::
				rvalue_from_python_storage<Mat>*)data)->storage.bytes;
		
		data->convertible = storage;
		
                // NOTE: The instruction below does not transfer
ownership, but this is not my problem, for now
		new (storage) Mat(PyPetscMat_Get(obj_ptr));
	}
};


Then, obviously ...

BOOST_PYTHON_MODULE(petsc_ext) {

        ... some not important lines ...

        PETSc_Mat_from_Python(); // Registers the converter
	
	def("print_Mat",print_Mat);
}


The code compiles without errors.
The problem is that when calling "print_Mat" from the Python shell,
the interpreter gives the error:

print_Mat(m)  # where m was previously and correctly built from Python
and it is a Python object containing a Mat

Traceback (most recent call last):
  File "<stdin>", line 1, in ?
Boost.Python.ArgumentError: Python argument types in
    petsc_ext.print_Mat(petsc4py.PETSc.Mat)
did not match C++ signature:
    print_Mat(_p_Mat*)


I tried also to modify "void print_Mat(Mat m)" into "void
print_Mat(const Mat m)" or "void print_Mat(const Mat& m)", but the
result is the same.

I tried also to debug the code, but the debugger does not enter into
the "convertible" method, so I guess there is something wrong that I
can't see.

So may be I am missing something for the process of writing a "from
Python converter" for pointer arguments.

Can anyone help me, please ?


-- 
Michele De Stefano
http://www.linkedin.com/in/micdestefano

From rfritz333 at gmail.com  Mon Oct 26 01:26:24 2009
From: rfritz333 at gmail.com (Randolph Fritz)
Date: Mon, 26 Oct 2009 00:26:24 +0000 (UTC)
Subject: [C++-sig] Python converters for raw pointers: request for help
References: <c410adc90910230138k77675a29n6e2fe435d2c2b21b@mail.gmail.com>
Message-ID: <slrnhe9r5g.clo.rfritz333@D-128-95-103-160.dhcp4.washington.edu>

Do you want to return an opaque type, would you like to expose some
of the mat's methods in Python, or would you perhaps like to convert a
numpy Matrix to a PETSc mat?  These all take different approaches.
Could you perhaps give an example of some of the Python code you would
like to write?
-- 
Randolph Fritz
  design machine group, architecture department, university of washington
rfritz at u.washington.edu -or- rfritz333 at gmail.com


From rfritz333 at gmail.com  Mon Oct 26 01:32:15 2009
From: rfritz333 at gmail.com (Randolph Fritz)
Date: Mon, 26 Oct 2009 00:32:15 +0000 (UTC)
Subject: [C++-sig] Python converters for raw pointers: request for help
References: <c410adc90910230138k77675a29n6e2fe435d2c2b21b@mail.gmail.com>
Message-ID: <slrnhe9rgf.cq9.rfritz333@D-128-95-103-160.dhcp4.washington.edu>

Additional thought.  Take a look at:
  <http://wiki.python.org/moin/boost.python/PointersAndSmartPointers#OSGexample>
& see if it helps clarify some of the problems.

-- 
Randolph Fritz
  design machine group, architecture department, university of washington
rfritz at u.washington.edu -or- rfritz333 at gmail.com


From micdestefano at gmail.com  Mon Oct 26 09:29:50 2009
From: micdestefano at gmail.com (Michele De Stefano)
Date: Mon, 26 Oct 2009 09:29:50 +0100
Subject: [C++-sig] Python converters for raw pointers: request for help
In-Reply-To: <slrnhe9rgf.cq9.rfritz333@D-128-95-103-160.dhcp4.washington.edu>
References: <c410adc90910230138k77675a29n6e2fe435d2c2b21b@mail.gmail.com>
	<slrnhe9rgf.cq9.rfritz333@D-128-95-103-160.dhcp4.washington.edu>
Message-ID: <c410adc90910260129q4ebfc9b6g684b9970851300e0@mail.gmail.com>

Ok. Let me clarify (I looked at the link you suggested, but I've not
found the answers I was looking for).

There exist a "petsc4py" package, which can be extended using its C
API. This package, already exposes PETSc types (and also the Mat
type). petsc4py is not programmed using Boost Python.

Now, assume I already have a C++ function that takes a Mat parameter;
to make it simple, let's take

void print_Mat(Mat m) {
   ... // The implementation is not interesting
};

I'd like to expose this function to Python, with a seamless
integration mechanism, using Boost Python.

Unfortunately, if I use the petsc4py API, I have to expose a wrapper
for "print_Mat", in this way:

void print_Mat_Wrap(const object& mat_obj) {
    Mat m( PyPetscMat_Get(mat_obj.ptr()) );
    ... // The rest is not important
}

I don't know anything of the petsc4py implementation, but I know that
it already exposes PETSc Mat to Python. I know also that, using the
petsc4py C API, the only way to retrieve a Mat from a
PyObject* is through the PyPetscMat_Get function, whose signature is this one:

Mat (*PyPetscMat_Get)(PyObject *);

I know for sure that Mat objects are not allocated with the C++ new
operator (I don't know if this is a useful information for you).

What I'd really like to do, would be to write a from-Python-converter
from a PyObject* (containing a petsc4py exposed Mat) into a Mat, in
order to be able to directly expose my "void print_Mat(Mat m)" without
writing any wrapper ... but I didn't manage to properly make it ...

Is this possible ?


2009/10/26 Randolph Fritz <rfritz333 at gmail.com>:
> Additional thought. ?Take a look at:
> ?<http://wiki.python.org/moin/boost.python/PointersAndSmartPointers#OSGexample>
> & see if it helps clarify some of the problems.
>
> --
> Randolph Fritz
> ?design machine group, architecture department, university of washington
> rfritz at u.washington.edu -or- rfritz333 at gmail.com
>
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
>



-- 
Michele De Stefano
http://www.linkedin.com/in/micdestefano
http://xoomer.virgilio.it/michele_de_stefano

From micdestefano at gmail.com  Mon Oct 26 09:38:41 2009
From: micdestefano at gmail.com (Michele De Stefano)
Date: Mon, 26 Oct 2009 09:38:41 +0100
Subject: [C++-sig] Python converters for raw pointers: request for help
In-Reply-To: <c410adc90910260129q4ebfc9b6g684b9970851300e0@mail.gmail.com>
References: <c410adc90910230138k77675a29n6e2fe435d2c2b21b@mail.gmail.com>
	<slrnhe9rgf.cq9.rfritz333@D-128-95-103-160.dhcp4.washington.edu>
	<c410adc90910260129q4ebfc9b6g684b9970851300e0@mail.gmail.com>
Message-ID: <c410adc90910260138o53c05697j8ce45cb4fe62c506@mail.gmail.com>

I was forgetting to give an example of the Python code I would like to write:

>>> from petsc4py import PETSc

... now some uninteresting code for building a PETSc.Mat python object ...
... the result will be a PETSc.Mat m object ....
... once created, if I write "m" and press RETURN, this is the output ...

>>> m
<petsc4py.PETSc.Mat object at 0xb7bea548>

... so, basically, I already have a PETSc Mat in Python, because it is
exposed by the petsc4py package ...

... Now I want to pass this Python object to my "print_Mat" function ...
... What I'd like to write is ...

>>> form my_extension import *
>>> print_Mat(m)

I hope now it's all clear.

Thank you in advance.
Michele

2009/10/26 Michele De Stefano <micdestefano at gmail.com>:
> Ok. Let me clarify (I looked at the link you suggested, but I've not
> found the answers I was looking for).
>
> There exist a "petsc4py" package, which can be extended using its C
> API. This package, already exposes PETSc types (and also the Mat
> type). petsc4py is not programmed using Boost Python.
>
> Now, assume I already have a C++ function that takes a Mat parameter;
> to make it simple, let's take
>
> void print_Mat(Mat m) {
> ? ... // The implementation is not interesting
> };
>
> I'd like to expose this function to Python, with a seamless
> integration mechanism, using Boost Python.
>
> Unfortunately, if I use the petsc4py API, I have to expose a wrapper
> for "print_Mat", in this way:
>
> void print_Mat_Wrap(const object& mat_obj) {
> ? ?Mat m( PyPetscMat_Get(mat_obj.ptr()) );
> ? ?... // The rest is not important
> }
>
> I don't know anything of the petsc4py implementation, but I know that
> it already exposes PETSc Mat to Python. I know also that, using the
> petsc4py C API, the only way to retrieve a Mat from a
> PyObject* is through the PyPetscMat_Get function, whose signature is this one:
>
> Mat (*PyPetscMat_Get)(PyObject *);
>
> I know for sure that Mat objects are not allocated with the C++ new
> operator (I don't know if this is a useful information for you).
>
> What I'd really like to do, would be to write a from-Python-converter
> from a PyObject* (containing a petsc4py exposed Mat) into a Mat, in
> order to be able to directly expose my "void print_Mat(Mat m)" without
> writing any wrapper ... but I didn't manage to properly make it ...
>
> Is this possible ?
>
>
> 2009/10/26 Randolph Fritz <rfritz333 at gmail.com>:
>> Additional thought. ?Take a look at:
>> ?<http://wiki.python.org/moin/boost.python/PointersAndSmartPointers#OSGexample>
>> & see if it helps clarify some of the problems.
>>
>> --
>> Randolph Fritz
>> ?design machine group, architecture department, university of washington
>> rfritz at u.washington.edu -or- rfritz333 at gmail.com
>>
>> _______________________________________________
>> Cplusplus-sig mailing list
>> Cplusplus-sig at python.org
>> http://mail.python.org/mailman/listinfo/cplusplus-sig
>>
>
>
>
> --
> Michele De Stefano
> http://www.linkedin.com/in/micdestefano
> http://xoomer.virgilio.it/michele_de_stefano
>



-- 
Michele De Stefano
http://www.linkedin.com/in/micdestefano
http://xoomer.virgilio.it/michele_de_stefano

From AFoglia at princeton.com  Mon Oct 26 15:08:16 2009
From: AFoglia at princeton.com (Anthony Foglia)
Date: Mon, 26 Oct 2009 10:08:16 -0400
Subject: [C++-sig] How to install extensions using bjam(?)
In-Reply-To: <9e910970910221213j5b3d5a73xcc4d34fddc095ca4@mail.gmail.com>
References: <9e910970910221213j5b3d5a73xcc4d34fddc095ca4@mail.gmail.com>
Message-ID: <4AE5AD50.6010009@princeton.com>

Amos Anderson wrote:
> I'm working on a project which is mixed C++/Python, and we use Boost
> functionality in our C++, and we've set up bjam to compile our project
> (on OSX). Everything seems to be working quite well.
> 
> To use our libraries from python scripts right now, we have to run this first:
> 
> #!/bin/bash
> trunk=$HOME/project/trunk/
> export DYLD_LIBRARY_PATH=$DYLD_LIBRARY_PATH:$trunk/lib
> export PYTHONPATH=$trunk:$trunk/lib
> python $@
> 
> 
> which is quite inconvenient. Now obviously I can skip this step by
> pasting those two export statements into my .profile, however, at some
> point we're going to start distributing this software and it would be
> nice if we didn't require the end user to do something like that.
> 
> I can't believe that either of these 2 steps are how it's supposed to
> work... so how should I be doing it?  Ideally, I want to set something
> up so that the end user only has to run 1 (one) command to build &
> install. Here are some things I've thought about:

[...]

> 3) bjam includes a way to install libraries into a folder I can name,
> but does it have a way to set environment variables? or automatically
> figure out the right places to put all the python files? If so, I
> can't find it...

	Almost certainly not.  bjam would be running in a child process and 
won't be able to change the environment of the parent (shell) process. 
Even if it did, any changes wouldn't persist into newly launched shells.

	I don't think there you can get the "right place" to put the python 
files, especially since different users will have different "right 
places".  I think your best bet is to follow the usual convention of 
using a default location of "/usr/lib/python<version>/site-packages", 
but allow users to override it from the command line.  I'm not 100% sure 
of the syntax, but something like

bjam -s prefix=/usr/local/lib/python2.5

should set the variable prefix to /usr/local/lib/python2.5 inside the 
Jamfile.

Now you just need a way to get the default python version to set the 
default value.  bjam's python module has a .version-countdown variable 
that might help.

The boost-build mailing list can probably give you better help.

-- 
Anthony Foglia
Princeton Consultants
(609) 987-8787 x233


From micdestefano at gmail.com  Mon Oct 26 17:28:57 2009
From: micdestefano at gmail.com (Michele De Stefano)
Date: Mon, 26 Oct 2009 17:28:57 +0100
Subject: [C++-sig] How to call reshape((M,N),order='F') ?
Message-ID: <c410adc90910260928j388fe761tf331912d282f484d@mail.gmail.com>

Hello to everyone.

I have this problem.

Assume I'm building a to-Python converter for a 3x2 matrix type,
FORTRAN storage (these are not the real dimensions of my problem, but
I give these numbers only to simplify the discussion).

Into the "convert" C++ method, first, I build a list with all the
matrix elements (lout);
then, I instantiate a boost::python::numeric::array in this way (note:
I had previously used
set_module_and_type("numpy","ndarray"), for using the NumPy's array):

boost::python::numeric::array   v(lout)

Finally, I'd like to call "reshape" on this array, in this way

v.attr("reshape")(make_tuple(3,2),"F");

But it does not work. It builds, but, on runtime, I obtain the error:

Traceback (most recent call last):
  File "<stdin>", line 1, in ?
TypeError: an integer is required

I tried to build and re-shape a matrix entirely from Python and I
discovered that, when using reshape,
you are forced to give the second parameter as

order = 'C' or 'F'

Now the question is: is it possible to specify a named parameter from
the C++ code ?
If the answer is "yes", how ?

Thank you in advance,
Michele

-- 
Michele De Stefano
http://www.linkedin.com/in/micdestefano

From eilif at gmx.de  Mon Oct 26 18:01:49 2009
From: eilif at gmx.de (Eilif Mueller)
Date: Mon, 26 Oct 2009 18:01:49 +0100
Subject: [C++-sig] function with >15 args yields get_signature error
Message-ID: <20091026170149.143750@gmx.net>

Hi,

Wrapping a function f with 16 arguments:

int f(int x1, int x2, int x3, int x4, int x5,
      int x6, int x7, int x8, int x9, int x10,
      int x11, int x12, int x13, int x14, int x15,
      int x16) {

  return x1;

}

BOOST_PYTHON_MODULE(test)
{

  def("f",f);

}


yields

/usr/include/boost/python/make_function.hpp: In function ?boost::python::api::object boost::python::make_function(F) [with F = int (*)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int)]?:
/usr/include/boost/python/def.hpp:82:   instantiated from ?boost::python::api::object boost::python::detail::make_function1(T, ...) [with T = int (*)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int)]?
/usr/include/boost/python/def.hpp:91:   instantiated from ?void boost::python::def(const char*, Fn) [with Fn = int (*)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int)]?
test_module.cpp:20:   instantiated from here
/usr/include/boost/python/make_function.hpp:104: error: no matching function for call to ?get_signature(int (*&)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int))?
make: *** [stream.o] Error 1


whereas all is fine if that last arg "int x16" is removed.  All of gcc-4.2, gcc-4.3 and gcc-4.4 seem to exhibit the same behaviour.  Same effect for libboost-python1.38-dev on Ubuntu karmic and libboost-python1.35-dev on Ubuntu jaunty.

I need that 16th arg and more ... about 32 args, I think.

Thanks for any help you can offer.

cheers,

Eilif





 
-- 
Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5 -
sicherer, schneller und einfacher! http://portal.gmx.net/de/go/atbrowser

-- 
Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5 -
sicherer, schneller und einfacher! http://portal.gmx.net/de/go/atbrowser

From troy at resophonic.com  Mon Oct 26 18:08:42 2009
From: troy at resophonic.com (troy d. straszheim)
Date: Mon, 26 Oct 2009 13:08:42 -0400
Subject: [C++-sig] function with >15 args yields get_signature error
In-Reply-To: <20091026170149.143750@gmx.net>
References: <20091026170149.143750@gmx.net>
Message-ID: <4AE5D79A.6040505@resophonic.com>

Eilif Mueller wrote:
> Hi,
> 
> Wrapping a function f with 16 arguments:
> 
> int f(int x1, int x2, int x3, int x4, int x5,
>       int x6, int x7, int x8, int x9, int x10,
>       int x11, int x12, int x13, int x14, int x15,
>       int x16) {
> 
>   return x1;
> 
> }
> 
> BOOST_PYTHON_MODULE(test)
> {
> 
>   def("f",f);
> 
> }
> 
> 
> yields
> 
> /usr/include/boost/python/make_function.hpp: In function ?boost::python::api::object boost::python::make_function(F) [with F = int (*)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int)]?:
> /usr/include/boost/python/def.hpp:82:   instantiated from ?boost::python::api::object boost::python::detail::make_function1(T, ...) [with T = int (*)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int)]?
> /usr/include/boost/python/def.hpp:91:   instantiated from ?void boost::python::def(const char*, Fn) [with Fn = int (*)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int)]?
> test_module.cpp:20:   instantiated from here
> /usr/include/boost/python/make_function.hpp:104: error: no matching function for call to ?get_signature(int (*&)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int))?
> make: *** [stream.o] Error 1
> 
> 
> whereas all is fine if that last arg "int x16" is removed.  All of gcc-4.2, gcc-4.3 and gcc-4.4 seem to exhibit the same behaviour.  Same effect for libboost-python1.38-dev on Ubuntu karmic and libboost-python1.35-dev on Ubuntu jaunty.
> 
> I need that 16th arg and more ... about 32 args, I think.
> 
> Thanks for any help you can offer.
> 

This came up recently.  You're going to have various problems (not all 
boost.python problems) getting arity that large.  The workaround is to 
introduce an intermediate function that takes a boost::python::tuple
and forward to the zillion-arguments one, example here:
 
http://gitorious.org/~straszheim/boost/straszheim/blobs/python/libs/python/test/zillion_arguments_workaround.cpp

-t


From renatox at gmail.com  Mon Oct 26 18:21:22 2009
From: renatox at gmail.com (Renato Araujo)
Date: Mon, 26 Oct 2009 14:21:22 -0300
Subject: [C++-sig] function with >15 args yields get_signature error
In-Reply-To: <20091026170149.143750@gmx.net>
References: <20091026170149.143750@gmx.net>
Message-ID: <95291a80910261021i7b1eaad3i64f1fbc9c180c69d@mail.gmail.com>

Hi,

Try use this flag in your compilation, -DBOOST_PYTHON_MAX_ARITY=XX


The default value is 15, the I think thi can solve your problem.

BR
Renato Araujo Oliveira Filho




On Mon, Oct 26, 2009 at 2:01 PM, Eilif Mueller <eilif at gmx.de> wrote:
> Hi,
>
> Wrapping a function f with 16 arguments:
>
> int f(int x1, int x2, int x3, int x4, int x5,
> ? ? ?int x6, int x7, int x8, int x9, int x10,
> ? ? ?int x11, int x12, int x13, int x14, int x15,
> ? ? ?int x16) {
>
> ?return x1;
>
> }
>
> BOOST_PYTHON_MODULE(test)
> {
>
> ?def("f",f);
>
> }
>
>
> yields
>
> /usr/include/boost/python/make_function.hpp: In function ?boost::python::api::object boost::python::make_function(F) [with F = int (*)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int)]?:
> /usr/include/boost/python/def.hpp:82: ? instantiated from ?boost::python::api::object boost::python::detail::make_function1(T, ...) [with T = int (*)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int)]?
> /usr/include/boost/python/def.hpp:91: ? instantiated from ?void boost::python::def(const char*, Fn) [with Fn = int (*)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int)]?
> test_module.cpp:20: ? instantiated from here
> /usr/include/boost/python/make_function.hpp:104: error: no matching function for call to ?get_signature(int (*&)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int))?
> make: *** [stream.o] Error 1
>
>
> whereas all is fine if that last arg "int x16" is removed. ?All of gcc-4.2, gcc-4.3 and gcc-4.4 seem to exhibit the same behaviour. ?Same effect for libboost-python1.38-dev on Ubuntu karmic and libboost-python1.35-dev on Ubuntu jaunty.
>
> I need that 16th arg and more ... about 32 args, I think.
>
> Thanks for any help you can offer.
>
> cheers,
>
> Eilif
>
>
>
>
>
>
> --
> Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5 -
> sicherer, schneller und einfacher! http://portal.gmx.net/de/go/atbrowser
>
> --
> Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5 -
> sicherer, schneller und einfacher! http://portal.gmx.net/de/go/atbrowser
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig

From wladwig at wdtinc.com  Mon Oct 26 18:20:14 2009
From: wladwig at wdtinc.com (William Ladwig)
Date: Mon, 26 Oct 2009 12:20:14 -0500
Subject: [C++-sig] function with >15 args yields get_signature error
In-Reply-To: <20091026170149.143750@gmx.net>
References: <20091026170149.143750@gmx.net>
Message-ID: <765CBD9053EA2B438625895F30D1856F02DC395454@storm.wdtinc.com>

There is a config macro BOOST_PYTHON_MAX_ARITY which may work for you.  I'm not sure how high you can increase this number though.  Refer to:

http://www.boost.org/doc/libs/1_40_0/libs/python/doc/v2/configuration.html



-----Original Message-----
From: cplusplus-sig-bounces+wladwig=wdtinc.com at python.org [mailto:cplusplus-sig-bounces+wladwig=wdtinc.com at python.org] On Behalf Of Eilif Mueller
Sent: Monday, October 26, 2009 12:02 PM
To: cplusplus-sig at python.org
Subject: [C++-sig] function with >15 args yields get_signature error

Hi,

Wrapping a function f with 16 arguments:

int f(int x1, int x2, int x3, int x4, int x5,
      int x6, int x7, int x8, int x9, int x10,
      int x11, int x12, int x13, int x14, int x15,
      int x16) {

  return x1;

}

BOOST_PYTHON_MODULE(test)
{

  def("f",f);

}


yields

/usr/include/boost/python/make_function.hpp: In function 'boost::python::api::object boost::python::make_function(F) [with F = int (*)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int)]':
/usr/include/boost/python/def.hpp:82:   instantiated from 'boost::python::api::object boost::python::detail::make_function1(T, ...) [with T = int (*)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int)]'
/usr/include/boost/python/def.hpp:91:   instantiated from 'void boost::python::def(const char*, Fn) [with Fn = int (*)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int)]'
test_module.cpp:20:   instantiated from here
/usr/include/boost/python/make_function.hpp:104: error: no matching function for call to 'get_signature(int (*&)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int, int))'
make: *** [stream.o] Error 1


whereas all is fine if that last arg "int x16" is removed.  All of gcc-4.2, gcc-4.3 and gcc-4.4 seem to exhibit the same behaviour.  Same effect for libboost-python1.38-dev on Ubuntu karmic and libboost-python1.35-dev on Ubuntu jaunty.

I need that 16th arg and more ... about 32 args, I think.

Thanks for any help you can offer.

cheers,

Eilif





 
-- 
Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5 -
sicherer, schneller und einfacher! http://portal.gmx.net/de/go/atbrowser

-- 
Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5 -
sicherer, schneller und einfacher! http://portal.gmx.net/de/go/atbrowser
_______________________________________________
Cplusplus-sig mailing list
Cplusplus-sig at python.org
http://mail.python.org/mailman/listinfo/cplusplus-sig

From nitroamos at gmail.com  Mon Oct 26 19:11:37 2009
From: nitroamos at gmail.com (Amos Anderson)
Date: Mon, 26 Oct 2009 11:11:37 -0700
Subject: [C++-sig] How to install extensions using bjam(?)
In-Reply-To: <4AE5AD50.6010009@princeton.com>
References: <9e910970910221213j5b3d5a73xcc4d34fddc095ca4@mail.gmail.com>
	<4AE5AD50.6010009@princeton.com>
Message-ID: <9e910970910261111qe312a9ai1d1f16be387477f6@mail.gmail.com>

On Mon, Oct 26, 2009 at 7:08 AM, Anthony Foglia <AFoglia at princeton.com> wrote:
> Amos Anderson wrote:
>>
>> I'm working on a project which is mixed C++/Python, and we use Boost
>> functionality in our C++, and we've set up bjam to compile our project
>> (on OSX). Everything seems to be working quite well.
>>
>> To use our libraries from python scripts right now, we have to run this
>> first:
>>
>> #!/bin/bash
>> trunk=$HOME/project/trunk/
>> export DYLD_LIBRARY_PATH=$DYLD_LIBRARY_PATH:$trunk/lib
>> export PYTHONPATH=$trunk:$trunk/lib
>> python $@
>>
>>
>> which is quite inconvenient. Now obviously I can skip this step by
>> pasting those two export statements into my .profile, however, at some
>> point we're going to start distributing this software and it would be
>> nice if we didn't require the end user to do something like that.
>>
>> I can't believe that either of these 2 steps are how it's supposed to
>> work... so how should I be doing it? ?Ideally, I want to set something
>> up so that the end user only has to run 1 (one) command to build &
>> install. Here are some things I've thought about:
>
> [...]
>
>> 3) bjam includes a way to install libraries into a folder I can name,
>> but does it have a way to set environment variables? or automatically
>> figure out the right places to put all the python files? If so, I
>> can't find it...
>
> ? ? ? ?Almost certainly not. ?bjam would be running in a child process and
> won't be able to change the environment of the parent (shell) process. Even
> if it did, any changes wouldn't persist into newly launched shells.


Well, what I had in mind was getting bjam to implement my solution #1.
So bjam would scan my .profile to see if it had ever modified the file
previously, and if so modify previously set variables, otherwise add
new environmental variables, etc.


>
> ? ? ? ?I don't think there you can get the "right place" to put the python
> files, especially since different users will have different "right places".
> ?I think your best bet is to follow the usual convention of using a default
> location of "/usr/lib/python<version>/site-packages", but allow users to
> override it from the command line. ?I'm not 100% sure of the syntax, but
> something like
>
> bjam -s prefix=/usr/local/lib/python2.5
>
> should set the variable prefix to /usr/local/lib/python2.5 inside the
> Jamfile.
>
> Now you just need a way to get the default python version to set the default
> value. ?bjam's python module has a .version-countdown variable that might
> help.


Based on your idea, I did some searching, and I found that I can do this:

python -c "from distutils.sysconfig import get_python_lib; print
get_python_lib()"
/Library/Python/2.6/site-packages

to automatically find the right location. And, it looks like:
http://developer.apple.com/mac/library/documentation/Darwin/Reference/ManPages/man1/dyld.1.html
says I can put the dylib's in one of these places:
$(HOME)/lib:/usr/local/lib:/lib:/usr/lib

and the linker will automatically find them. I guess one of those is
good enough for me, but I'd imagine that python would add a few new
places to look? I tried to put my dylib files into site-packages, but
that didn't work.

It looks like this is making some progress. It looks like at the very
least I can get bjam to put the dylib into $(HOME)/lib directly, and
edit the python scripts to add the right python paths.


Amos.


>
> The boost-build mailing list can probably give you better help.
>

I've tried a few times to sign up to that list... but it doesn't seem to work.


thanks!

Amos.



> --
> Anthony Foglia
> Princeton Consultants
> (609) 987-8787 x233
>
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
>



-- 
~<>~<>~<>~<>~<>~<>~<>~<>~
nitroamos at gmail.com
amosa at wag.caltech.edu (secure)

+1-626-399-8958 (cell)
+1-626-794-1971 (home)

From AFoglia at princeton.com  Mon Oct 26 19:52:47 2009
From: AFoglia at princeton.com (Anthony Foglia)
Date: Mon, 26 Oct 2009 14:52:47 -0400
Subject: [C++-sig] How to install extensions using bjam(?)
In-Reply-To: <9e910970910261111qe312a9ai1d1f16be387477f6@mail.gmail.com>
References: <9e910970910221213j5b3d5a73xcc4d34fddc095ca4@mail.gmail.com>	<4AE5AD50.6010009@princeton.com>
	<9e910970910261111qe312a9ai1d1f16be387477f6@mail.gmail.com>
Message-ID: <4AE5EFFF.1050505@princeton.com>

Amos Anderson wrote:
> Well, what I had in mind was getting bjam to implement my solution #1.
> So bjam would scan my .profile to see if it had ever modified the file
> previously, and if so modify previously set variables, otherwise add
> new environmental variables, etc.

But what if the user isn't using bash, but tcsh?  Or if they have a 
.bash_profile?  Or if they source their .bashrc via their .profile and 
set their PYTHONPATH in there?  There's a lot that could go wrong.

The second approach sounds much more robust and is the more conventional 
approach.  Hopefully you can get it to do what you want.

-- 
Anthony Foglia
Princeton Consultants
(609) 987-8787 x233


From rfritz333 at gmail.com  Mon Oct 26 20:10:55 2009
From: rfritz333 at gmail.com (Randolph Fritz)
Date: Mon, 26 Oct 2009 19:10:55 +0000 (UTC)
Subject: [C++-sig] Python converters for raw pointers: request for help
References: <c410adc90910230138k77675a29n6e2fe435d2c2b21b@mail.gmail.com>
	<slrnhe9rgf.cq9.rfritz333@D-128-95-103-160.dhcp4.washington.edu>
	<c410adc90910260129q4ebfc9b6g684b9970851300e0@mail.gmail.com>
	<c410adc90910260138o53c05697j8ce45cb4fe62c506@mail.gmail.com>
Message-ID: <slrnhebt1v.gpl.rfritz333@160.86.161.25.in-addr.arpa>

I think what you want is here:
  <http://www.boost.org/doc/libs/1_39_0/libs/python/doc/tutorial/doc/html/python/object.html#python.extracting_c___objects>

and here:
  <http://www.boost.org/doc/libs/1_40_0/libs/python/doc/v2/extract.html>

Take a look at that--see if it helps.

-- 
Randolph Fritz
  design machine group, architecture department, university of washington
rfritz at u.washington.edu -or- rfritz333 at gmail.com


From eilif at gmx.de  Mon Oct 26 21:59:13 2009
From: eilif at gmx.de (Eilif Mueller)
Date: Mon, 26 Oct 2009 21:59:13 +0100
Subject: [C++-sig] function with >15 args yields get_signature error
In-Reply-To: <765CBD9053EA2B438625895F30D1856F02DC395454@storm.wdtinc.com>
References: <20091026170149.143750@gmx.net>
	<765CBD9053EA2B438625895F30D1856F02DC395454@storm.wdtinc.com>
Message-ID: <20091026205913.294100@gmx.net>

Hi,

Thanks!  -DBOOST_PYTHON_MAX_ARITY=34 did indeed help ... for simple functions it did the trick, and it did increase my limit for member functions, but only to 25.  For over 25 args for a member functions of an exposed class I get the following compiler error:

/usr/include/boost/python/detail/invoke.hpp:81: error: must use ?.*? or ?->*? to call pointer-to-member function in ?f (...)?


#include <boost/python.hpp>

using namespace boost::python;

class MyClass {

public:

  int qq(int x1, int x2, int x3, int x4, int x5,
         int x6, int x7, int x8, int x9, int x10,
         int x11, int x12, int x13, int x14, int x15,
         int x16, int x17, int x18, int x19, int x20,
	 int x21, int x22, int x23, int x24, float x25 /*,float x26*/);

};

BOOST_PYTHON_MODULE(mytest) 
{

class_< MyClass >("MyClass", init<>() )
  .def("qq",&MyClass::qq);

}

Any ideas?  

cheers,

Eilif



-------- Original-Nachricht --------
> Datum: Mon, 26 Oct 2009 12:20:14 -0500
> Von: William Ladwig <wladwig at wdtinc.com>
> An: Development of Python/C++ integration <cplusplus-sig at python.org>
> Betreff: Re: [C++-sig] function with >15 args yields get_signature error

> There is a config macro BOOST_PYTHON_MAX_ARITY which may work for you. 
> I'm not sure how high you can increase this number though.  Refer to:
> 
> http://www.boost.org/doc/libs/1_40_0/libs/python/doc/v2/configuration.html
> 
> 
> 
> -----Original Message-----
> From: cplusplus-sig-bounces+wladwig=wdtinc.com at python.org
> [mailto:cplusplus-sig-bounces+wladwig=wdtinc.com at python.org] On Behalf Of Eilif Mueller
> Sent: Monday, October 26, 2009 12:02 PM
> To: cplusplus-sig at python.org
> Subject: [C++-sig] function with >15 args yields get_signature error
> 
> Hi,
> 
> Wrapping a function f with 16 arguments:
> 
> int f(int x1, int x2, int x3, int x4, int x5,
>       int x6, int x7, int x8, int x9, int x10,
>       int x11, int x12, int x13, int x14, int x15,
>       int x16) {
> 
>   return x1;
> 
> }
> 
> BOOST_PYTHON_MODULE(test)
> {
> 
>   def("f",f);
> 
> }
> 
> 
> yields
> 
> /usr/include/boost/python/make_function.hpp: In function
> 'boost::python::api::object boost::python::make_function(F) [with F = int (*)(int, int, int,
> int, int, int, int, int, int, int, int, int, int, int, int, int)]':
> /usr/include/boost/python/def.hpp:82:   instantiated from
> 'boost::python::api::object boost::python::detail::make_function1(T, ...) [with T = int
> (*)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int,
> int)]'
> /usr/include/boost/python/def.hpp:91:   instantiated from 'void
> boost::python::def(const char*, Fn) [with Fn = int (*)(int, int, int, int, int, int,
> int, int, int, int, int, int, int, int, int, int)]'
> test_module.cpp:20:   instantiated from here
> /usr/include/boost/python/make_function.hpp:104: error: no matching
> function for call to 'get_signature(int (*&)(int, int, int, int, int, int, int,
> int, int, int, int, int, int, int, int, int))'
> make: *** [stream.o] Error 1
> 
> 
> whereas all is fine if that last arg "int x16" is removed.  All of
> gcc-4.2, gcc-4.3 and gcc-4.4 seem to exhibit the same behaviour.  Same effect for
> libboost-python1.38-dev on Ubuntu karmic and libboost-python1.35-dev on
> Ubuntu jaunty.
> 
> I need that 16th arg and more ... about 32 args, I think.
> 
> Thanks for any help you can offer.
> 
> cheers,
> 
> Eilif
> 
> 
> 
> 
> 
>  
> -- 
> Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5
> -
> sicherer, schneller und einfacher! http://portal.gmx.net/de/go/atbrowser
> 
> -- 
> Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5
> -
> sicherer, schneller und einfacher! http://portal.gmx.net/de/go/atbrowser
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig

-- 
Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5 -
sicherer, schneller und einfacher! http://portal.gmx.net/de/go/atbrowser

From craig.bosma at gmail.com  Mon Oct 26 22:03:20 2009
From: craig.bosma at gmail.com (Craig S. Bosma)
Date: Mon, 26 Oct 2009 16:03:20 -0500
Subject: [C++-sig] RuntimeError on import from use of bases<>
Message-ID: <03DA9677-860D-4D33-B6FD-AEF598CEC391@gmail.com>

Hi,

I'm having a difficult time debugging an issue with boost.python  
bindings under OS X 10.6 -- the same bindings have been working fine  
under windows.

On import of my module, I get an exception caught by python:

 > RuntimeError: unidentifiable C++ exception

which I've traced down as far as the use of a bases<> entry in my  
class_ definition. If I remove the bases<> entry, it works, but of  
course removing all inheiritance from the module isn't an option.  
Furthermore, it seems that any (so far) use of bases<> in this module  
causes the problem, while in another module (wrapping another library)  
classes deriving from others via bases<> work normally.

Any thoughts on what could case this type of failure?

Craig

From Matthew.Scouten at tradingtechnologies.com  Mon Oct 26 22:03:18 2009
From: Matthew.Scouten at tradingtechnologies.com (Matthew Scouten (TT))
Date: Mon, 26 Oct 2009 16:03:18 -0500
Subject: [C++-sig] function with >15 args yields get_signature error
In-Reply-To: <20091026205913.294100@gmx.net>
References: <20091026170149.143750@gmx.net><765CBD9053EA2B438625895F30D1856F02DC395454@storm.wdtinc.com>
	<20091026205913.294100@gmx.net>
Message-ID: <32490DFF7774554A85D65D23A9F0F9380BAC6C8B@chiex01>

Yeah. Find the guy who wrote a function with over 25 arguments. Bring a baseball bat. Persuade him of his error.

-----Original Message-----
From: cplusplus-sig-bounces+matthew.scouten=tradingtechnologies.com at python.org [mailto:cplusplus-sig-bounces+matthew.scouten=tradingtechnologies.com at python.org] On Behalf Of Eilif Mueller
Sent: Monday, October 26, 2009 3:59 PM
To: Development of Python/C++ integration
Subject: Re: [C++-sig] function with >15 args yields get_signature error

Hi,

Thanks!  -DBOOST_PYTHON_MAX_ARITY=34 did indeed help ... for simple functions it did the trick, and it did increase my limit for member functions, but only to 25.  For over 25 args for a member functions of an exposed class I get the following compiler error:

/usr/include/boost/python/detail/invoke.hpp:81: error: must use ?.*? or ?->*? to call pointer-to-member function in ?f (...)?


#include <boost/python.hpp>

using namespace boost::python;

class MyClass {

public:

  int qq(int x1, int x2, int x3, int x4, int x5,
         int x6, int x7, int x8, int x9, int x10,
         int x11, int x12, int x13, int x14, int x15,
         int x16, int x17, int x18, int x19, int x20,
	 int x21, int x22, int x23, int x24, float x25 /*,float x26*/);

};

BOOST_PYTHON_MODULE(mytest) 
{

class_< MyClass >("MyClass", init<>() )
  .def("qq",&MyClass::qq);

}

Any ideas?  

cheers,

Eilif



-------- Original-Nachricht --------
> Datum: Mon, 26 Oct 2009 12:20:14 -0500
> Von: William Ladwig <wladwig at wdtinc.com>
> An: Development of Python/C++ integration <cplusplus-sig at python.org>
> Betreff: Re: [C++-sig] function with >15 args yields get_signature error

> There is a config macro BOOST_PYTHON_MAX_ARITY which may work for you. 
> I'm not sure how high you can increase this number though.  Refer to:
> 
> http://www.boost.org/doc/libs/1_40_0/libs/python/doc/v2/configuration.html
> 
> 
> 
> -----Original Message-----
> From: cplusplus-sig-bounces+wladwig=wdtinc.com at python.org
> [mailto:cplusplus-sig-bounces+wladwig=wdtinc.com at python.org] On Behalf Of Eilif Mueller
> Sent: Monday, October 26, 2009 12:02 PM
> To: cplusplus-sig at python.org
> Subject: [C++-sig] function with >15 args yields get_signature error
> 
> Hi,
> 
> Wrapping a function f with 16 arguments:
> 
> int f(int x1, int x2, int x3, int x4, int x5,
>       int x6, int x7, int x8, int x9, int x10,
>       int x11, int x12, int x13, int x14, int x15,
>       int x16) {
> 
>   return x1;
> 
> }
> 
> BOOST_PYTHON_MODULE(test)
> {
> 
>   def("f",f);
> 
> }
> 
> 
> yields
> 
> /usr/include/boost/python/make_function.hpp: In function
> 'boost::python::api::object boost::python::make_function(F) [with F = int (*)(int, int, int,
> int, int, int, int, int, int, int, int, int, int, int, int, int)]':
> /usr/include/boost/python/def.hpp:82:   instantiated from
> 'boost::python::api::object boost::python::detail::make_function1(T, ...) [with T = int
> (*)(int, int, int, int, int, int, int, int, int, int, int, int, int, int, int,
> int)]'
> /usr/include/boost/python/def.hpp:91:   instantiated from 'void
> boost::python::def(const char*, Fn) [with Fn = int (*)(int, int, int, int, int, int,
> int, int, int, int, int, int, int, int, int, int)]'
> test_module.cpp:20:   instantiated from here
> /usr/include/boost/python/make_function.hpp:104: error: no matching
> function for call to 'get_signature(int (*&)(int, int, int, int, int, int, int,
> int, int, int, int, int, int, int, int, int))'
> make: *** [stream.o] Error 1
> 
> 
> whereas all is fine if that last arg "int x16" is removed.  All of
> gcc-4.2, gcc-4.3 and gcc-4.4 seem to exhibit the same behaviour.  Same effect for
> libboost-python1.38-dev on Ubuntu karmic and libboost-python1.35-dev on
> Ubuntu jaunty.
> 
> I need that 16th arg and more ... about 32 args, I think.
> 
> Thanks for any help you can offer.
> 
> cheers,
> 
> Eilif
> 
> 
> 
> 
> 
>  
> -- 
> Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5
> -
> sicherer, schneller und einfacher! http://portal.gmx.net/de/go/atbrowser
> 
> -- 
> Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5
> -
> sicherer, schneller und einfacher! http://portal.gmx.net/de/go/atbrowser
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig

-- 
Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5 -
sicherer, schneller und einfacher! http://portal.gmx.net/de/go/atbrowser
_______________________________________________
Cplusplus-sig mailing list
Cplusplus-sig at python.org
http://mail.python.org/mailman/listinfo/cplusplus-sig

From troy at resophonic.com  Mon Oct 26 22:30:57 2009
From: troy at resophonic.com (troy d. straszheim)
Date: Mon, 26 Oct 2009 17:30:57 -0400
Subject: [C++-sig] function with >15 args yields get_signature error
In-Reply-To: <32490DFF7774554A85D65D23A9F0F9380BAC6C8B@chiex01>
References: <20091026170149.143750@gmx.net><765CBD9053EA2B438625895F30D1856F02DC395454@storm.wdtinc.com>	<20091026205913.294100@gmx.net>
	<32490DFF7774554A85D65D23A9F0F9380BAC6C8B@chiex01>
Message-ID: <4AE61511.4070006@resophonic.com>

Matthew Scouten (TT) wrote:
> Yeah. Find the guy who wrote a function with over 25 arguments. Bring
> a baseball bat. Persuade him of his error.

+1.  "Clue-by-four" time.

> -----Original Message----- From:
> cplusplus-sig-bounces+matthew.scouten=tradingtechnologies.com at python.org
> [mailto:cplusplus-sig-bounces+matthew.scouten=tradingtechnologies.com at python.org]
> On Behalf Of Eilif Mueller Sent: Monday, October 26, 2009 3:59 PM To:
> Development of Python/C++ integration Subject: Re: [C++-sig] function
> with >15 args yields get_signature error
> 
> Hi,
> 
> Thanks!  -DBOOST_PYTHON_MAX_ARITY=34 did indeed help ... for simple
> functions it did the trick, and it did increase my limit for member
> functions, but only to 25.  For over 25 args for a member functions
> of an exposed class I get the following compiler error:
> 
> /usr/include/boost/python/detail/invoke.hpp:81: error: must use ?.*?
> or ?->*? to call pointer-to-member function in ?f (...)?
> 

I believe this is because BOOST_PYTHON_MAX_ARITY doesn't have the 
ability to change the hardcoded limit of 25 parameters in 
boost/type_traits/detail/is_mem_fun_ptr_impl.hpp, which boost.python 
uses.  This is why I recommended that workaround.

-t

From micdestefano at gmail.com  Tue Oct 27 09:10:52 2009
From: micdestefano at gmail.com (Michele De Stefano)
Date: Tue, 27 Oct 2009 09:10:52 +0100
Subject: [C++-sig] Python converters for raw pointers: request for help
In-Reply-To: <slrnhebt1v.gpl.rfritz333@160.86.161.25.in-addr.arpa>
References: <c410adc90910230138k77675a29n6e2fe435d2c2b21b@mail.gmail.com>
	<slrnhe9rgf.cq9.rfritz333@D-128-95-103-160.dhcp4.washington.edu>
	<c410adc90910260129q4ebfc9b6g684b9970851300e0@mail.gmail.com>
	<c410adc90910260138o53c05697j8ce45cb4fe62c506@mail.gmail.com>
	<slrnhebt1v.gpl.rfritz333@160.86.161.25.in-addr.arpa>
Message-ID: <c410adc90910270110k756260fdmd05ff23909728b47@mail.gmail.com>

That's not what I'd like to do. I'd like to make a "from-Python"
converter. I don't want to make a wrapper that takes a
boost::python::object and then call extract ...

The problem in my converter (see the previous e-mails) is that Python
does not even enter into the "convertible" method ...


2009/10/26 Randolph Fritz <rfritz333 at gmail.com>:
> I think what you want is here:
> ?<http://www.boost.org/doc/libs/1_39_0/libs/python/doc/tutorial/doc/html/python/object.html#python.extracting_c___objects>
>
> and here:
> ?<http://www.boost.org/doc/libs/1_40_0/libs/python/doc/v2/extract.html>
>
> Take a look at that--see if it helps.
>
> --
> Randolph Fritz
> ?design machine group, architecture department, university of washington
> rfritz at u.washington.edu -or- rfritz333 at gmail.com
>
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
>



-- 
Michele De Stefano
http://www.linkedin.com/in/micdestefano
http://xoomer.virgilio.it/michele_de_stefano

From rfritz333 at gmail.com  Tue Oct 27 16:39:12 2009
From: rfritz333 at gmail.com (Randolph Fritz)
Date: Tue, 27 Oct 2009 15:39:12 +0000 (UTC)
Subject: [C++-sig] Python converters for raw pointers: request for help
References: <c410adc90910230138k77675a29n6e2fe435d2c2b21b@mail.gmail.com>
	<slrnhe9rgf.cq9.rfritz333@D-128-95-103-160.dhcp4.washington.edu>
	<c410adc90910260129q4ebfc9b6g684b9970851300e0@mail.gmail.com>
	<c410adc90910260138o53c05697j8ce45cb4fe62c506@mail.gmail.com>
	<slrnhebt1v.gpl.rfritz333@160.86.161.25.in-addr.arpa>
	<c410adc90910270110k756260fdmd05ff23909728b47@mail.gmail.com>
Message-ID: <slrnhee510.lej.rfritz333@D-69-91-145-139.dhcp4.washington.edu>

On 2009-10-27, Michele De Stefano <micdestefano at gmail.com> wrote:
> That's not what I'd like to do. I'd like to make a "from-Python"
> converter. I don't want to make a wrapper that takes a
> boost::python::object and then call extract ...
>
> The problem in my converter (see the previous e-mails) is that Python
> does not even enter into the "convertible" method ...
>

Hunh.  Well, I'm out of ideas, then.  Sorry.

-- 
Randolph Fritz
  design machine group, architecture department, university of washington
rfritz at u.washington.edu -or- rfritz333 at gmail.com


From micdestefano at gmail.com  Wed Oct 28 00:30:32 2009
From: micdestefano at gmail.com (Michele De Stefano)
Date: Wed, 28 Oct 2009 00:30:32 +0100
Subject: [C++-sig] New mds-utils release (1.2.0)
Message-ID: <c410adc90910271630o165871b2g47c18f197462e98c@mail.gmail.com>

mds-utils 1.2.0 have been released.

Release Notes:

	- Improved python/fileobj.hpp: now there are 3 helper file objects
	  and they are copy-constructible. Derivation from boost::python::object
	  has been mantained.
	- Added utilities for indexing support in Python extensions.
	- Added boost::python converters for some boost::numeric::ublas classes.

-------------------
What is mds-utils ?
-------------------

mds-utils is a library intended to become a collection of several C++ utilities.
It makes heavy usage of the Boost C++ libraries.

Release 1.2.0 contains:
	- a tool for detecting machine endianity.
	- some useful classes that allow to treat the old C FILE pointer as a
C++ stream.
	- C++ classes that help on treating Python file objects as C++ streams.
	- simple utilities for indexing support in Python extensions.
	- new C++ to-Python and from-Python converters for some Boost uBlas objects.

Project home: http://code.google.com/p/mds-utils/

-- 
Michele De Stefano
http://www.linkedin.com/in/micdestefano
http://xoomer.virgilio.it/michele_de_stefano

From walter at mathematik.hu-berlin.de  Wed Oct 28 20:31:33 2009
From: walter at mathematik.hu-berlin.de (Sebastian Walter)
Date: Wed, 28 Oct 2009 20:31:33 +0100
Subject: [C++-sig] How to call reshape((M,N),order='F') ?
In-Reply-To: <c410adc90910260928j388fe761tf331912d282f484d@mail.gmail.com>
References: <c410adc90910260928j388fe761tf331912d282f484d@mail.gmail.com>
Message-ID: <8867253f08604f5750d48ec4c93030f6.squirrel@webmail.mathematik.hu-berlin.de>


> Hello to everyone.
>
> I have this problem.
>
> Assume I'm building a to-Python converter for a 3x2 matrix type,
> FORTRAN storage (these are not the real dimensions of my problem, but
> I give these numbers only to simplify the discussion).
>
> Into the "convert" C++ method, first, I build a list with all the
> matrix elements (lout);
> then, I instantiate a boost::python::numeric::array in this way (note:
> I had previously used
> set_module_and_type("numpy","ndarray"), for using the NumPy's array):
>
> boost::python::numeric::array   v(lout)
>
> Finally, I'd like to call "reshape" on this array, in this way
>
> v.attr("reshape")(make_tuple(3,2),"F");

don't know the answer to your question, but you could do the following:

set the strides of the underlying PyArrayObject manually to match column
major format, i.e.
PyArrayObject  *v_array_obj = reinterpret_cast<PyArrayObject*> ( v.ptr() );
v_array_obj->strides[0] = sizeof(double);
v_array_obj->strides[1] = 3 * sizeof(double);

hope that helps,

Sebastian


>
> But it does not work. It builds, but, on runtime, I obtain the error:
>
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
> TypeError: an integer is required
>
> I tried to build and re-shape a matrix entirely from Python and I
> discovered that, when using reshape,
> you are forced to give the second parameter as
>
> order = 'C' or 'F'
>
> Now the question is: is it possible to specify a named parameter from
> the C++ code ?
> If the answer is "yes", how ?
>
> Thank you in advance,
> Michele
>
> --
> Michele De Stefano
> http://www.linkedin.com/in/micdestefano
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
>


-- 
Sebastian Walter

Institut fuer Mathematik, Humboldt-Universitaet    Tel: +49 (30) 2093-5869
Rudower Chaussee 25, Adlershof, Berlin             Fax: +49 (30) 2093-5859
Post: Unter den Linden 6, D-10099 Berlin         Email: 
walter at mathematik.hu-berlin.de
WWW: http://www.mathematik.hu-berlin.de/~walter


From micdestefano at gmail.com  Thu Oct 29 08:58:11 2009
From: micdestefano at gmail.com (Michele De Stefano)
Date: Thu, 29 Oct 2009 08:58:11 +0100
Subject: [C++-sig] How to call reshape((M,N),order='F') ?
In-Reply-To: <8867253f08604f5750d48ec4c93030f6.squirrel@webmail.mathematik.hu-berlin.de>
References: <c410adc90910260928j388fe761tf331912d282f484d@mail.gmail.com>
	<8867253f08604f5750d48ec4c93030f6.squirrel@webmail.mathematik.hu-berlin.de>
Message-ID: <c410adc90910290058j5bef2b3r7a790a3f81e4a077@mail.gmail.com>

Thank you Sebastian.

I discovered also that the problem could be solved using the global
numpy.reshape.

That function, does not complain if the last parameter is simply "F"
and not "storage=F".

Anyway the issue is still open.

If anyone knows it, is it possible to explicitly specify a named
parameter through a call to "attr" ?

2009/10/28 Sebastian Walter <walter at mathematik.hu-berlin.de>:
>
>> Hello to everyone.
>>
>> I have this problem.
>>
>> Assume I'm building a to-Python converter for a 3x2 matrix type,
>> FORTRAN storage (these are not the real dimensions of my problem, but
>> I give these numbers only to simplify the discussion).
>>
>> Into the "convert" C++ method, first, I build a list with all the
>> matrix elements (lout);
>> then, I instantiate a boost::python::numeric::array in this way (note:
>> I had previously used
>> set_module_and_type("numpy","ndarray"), for using the NumPy's array):
>>
>> boost::python::numeric::array ? v(lout)
>>
>> Finally, I'd like to call "reshape" on this array, in this way
>>
>> v.attr("reshape")(make_tuple(3,2),"F");
>
> don't know the answer to your question, but you could do the following:
>
> set the strides of the underlying PyArrayObject manually to match column
> major format, i.e.
> PyArrayObject ?*v_array_obj = reinterpret_cast<PyArrayObject*> ( v.ptr() );
> v_array_obj->strides[0] = sizeof(double);
> v_array_obj->strides[1] = 3 * sizeof(double);
>
> hope that helps,
>
> Sebastian
>
>
>>
>> But it does not work. It builds, but, on runtime, I obtain the error:
>>
>> Traceback (most recent call last):
>> ? File "<stdin>", line 1, in ?
>> TypeError: an integer is required
>>
>> I tried to build and re-shape a matrix entirely from Python and I
>> discovered that, when using reshape,
>> you are forced to give the second parameter as
>>
>> order = 'C' or 'F'
>>
>> Now the question is: is it possible to specify a named parameter from
>> the C++ code ?
>> If the answer is "yes", how ?
>>
>> Thank you in advance,
>> Michele
>>
>> --
>> Michele De Stefano
>> http://www.linkedin.com/in/micdestefano
>> _______________________________________________
>> Cplusplus-sig mailing list
>> Cplusplus-sig at python.org
>> http://mail.python.org/mailman/listinfo/cplusplus-sig
>>
>
>
> --
> Sebastian Walter
>
> Institut fuer Mathematik, Humboldt-Universitaet ? ?Tel: +49 (30) 2093-5869
> Rudower Chaussee 25, Adlershof, Berlin ? ? ? ? ? ? Fax: +49 (30) 2093-5859
> Post: Unter den Linden 6, D-10099 Berlin ? ? ? ? Email:
> walter at mathematik.hu-berlin.de
> WWW: http://www.mathematik.hu-berlin.de/~walter
>
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
>



-- 
Michele De Stefano
http://www.linkedin.com/in/micdestefano
http://xoomer.virgilio.it/michele_de_stefano

From amundson at fnal.gov  Fri Oct 30 22:47:45 2009
From: amundson at fnal.gov (James Amundson)
Date: Fri, 30 Oct 2009 16:47:45 -0500
Subject: [C++-sig] How do I wrap a static member function?
Message-ID: <4AEB5F01.2070503@fnal.gov>

Hi,

I have the following code:

-------------------------------------------------------------------------------
#include <boost/python.hpp>
#include <iostream>

class Foo
{
public:
     Foo() {} ;
     static void doit(const char * message) {
         std::cout << "doing it statically: " << message << "\n";
     };
     void nonstatic_doit(const char *message) {
         std::cout << "doing it nonstatically: " << message << "\n";
     };
};

using namespace boost::python;

BOOST_PYTHON_MODULE(foo)
{
     class_<Foo>("Foo",init<>())
     .def("doit", &Foo::doit)
     .def("nonstatic_doit", &Foo::nonstatic_doit)
     ;
}
-------------------------------------------------------------------------------

When I try out my foo module, I see this:

In [1]: from foo import Foo

In [2]: f = Foo()

In [3]: f.nonstatic_doit("hello")
doing it nonstatically: hello

In [4]: f.doit("hello")
---------------------------------------------------------------------------
ArgumentError                             Traceback (most recent call last)

/home/amundson/work/synergia2-devel_1_0/build/synergia2/components/foundation/<ipython 
console> in <module>()

ArgumentError: Python argument types in
     Foo.doit(Foo, str)
did not match C++ signature:
     doit(char const*)


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

How can I wrap doit?

Thanks,
Jim Amundson


From lists_ravi at lavabit.com  Sat Oct 31 04:43:31 2009
From: lists_ravi at lavabit.com (Ravi)
Date: Fri, 30 Oct 2009 23:43:31 -0400
Subject: [C++-sig] How do I wrap a static member function?
In-Reply-To: <4AEB5F01.2070503@fnal.gov>
References: <4AEB5F01.2070503@fnal.gov>
Message-ID: <200910302343.31622.lists_ravi@lavabit.com>

Use class_.staticmethod(...) as shown below:

On Friday 30 October 2009 17:47:45 James Amundson wrote:
> BOOST_PYTHON_MODULE(foo)
> {
>      class_<Foo>("Foo",init<>())
>      .def("doit", &Foo::doit)

       .staticmethod( "doit" )

>      .def("nonstatic_doit", &Foo::nonstatic_doit)
>      ;
> }
 
Regards,
Ravi


From lists_ravi at lavabit.com  Sat Oct 31 04:53:27 2009
From: lists_ravi at lavabit.com (Ravi)
Date: Fri, 30 Oct 2009 23:53:27 -0400
Subject: [C++-sig] Python converters for raw pointers: request for help
In-Reply-To: <c410adc90910230138k77675a29n6e2fe435d2c2b21b@mail.gmail.com>
References: <c410adc90910230138k77675a29n6e2fe435d2c2b21b@mail.gmail.com>
Message-ID: <200910302353.28002.lists_ravi@lavabit.com>

On Friday 23 October 2009 04:38:30 Michele De Stefano wrote:
> typedef struct _p_Mat*           Mat;

Creating a converter to/from a raw pointer is rather tricky (due to the way 
argument type deduction handles pointers & references seamlessly, IIUC). You 
should either expose _p_Mat or wrap the raw pointer in a struct and then 
expose converters to it:
  struct Mat { struct _p_Mat *ptr; };
If you come up with a better way, I'd be interested.

Regards,
Ravi


From lists_ravi at lavabit.com  Sat Oct 31 05:02:37 2009
From: lists_ravi at lavabit.com (Ravi)
Date: Sat, 31 Oct 2009 00:02:37 -0400
Subject: [C++-sig] [python] python + phoenix
In-Reply-To: <4AD6820E.9040302@resophonic.com>
References: <200910101158.46996.lists_ravi@lavabit.com>
	<200910122104.46505.lists_ravi@lavabit.com>
	<4AD6820E.9040302@resophonic.com>
Message-ID: <200910310002.37065.lists_ravi@lavabit.com>

On Wednesday 14 October 2009 21:59:42 troy d. straszheim wrote:
> boost::function still works, and doesn't require as<>:
> 
>    boost::function<int(int, int)> myplus = std::plus<int>();
>    def("myplus", myplus);
> 
> and old-school function objects:
> 
>    def("otherplus", std::plus<int>())
> 
> I was surprised to find that it all plays nice with boost::phoenix:
> 
>    def("plus", as<int(int, int)>(arg1 + arg2));
> 
>    def("throw_if_gt_5",
>        as<void(int)>(if_(arg1 > 5)
>                     [ throw_(std::runtime_error("too big!!")) ]
>                     .else_
>                     [ std::cout << val("okay")                ]));
> 
> And boost::phoenix plays nice with shared_ptr, which plays nice with
> boost::python:
> 
>    struct X { int x; };   typedef boost::shared_ptr<X> XPtr;
> 
>    class_<X, XPtr>("X")
>      .def_readwrite("x", &X::x)
>      .def("add7", as<void(X*)>  (arg1->*&X::x      += 7))
>      .def("add8", as<void(XPtr)>(arg1->*&X::x      += 8))
>      .def("add9", as<void(X&)>  (bind(&X::x, arg1) += 9))
>    ;
> 
> the other little thing is overload resolution, for which I propose a
> little helper function "unload" (underload? deload? resolve?  I dunno):
> 
>    void foo(int);
>    void foo(double);
> 
>    def("foo", unload<void(int)>(foo));
> 
> Which is safer than
> 
>    def("foo", (void(*)(int)) foo);
> 
> and just as safe as, and less verbose than
> 
>    void(*fooint)(int) = foo;
>    def("foo", fooint);
> 
> So that's where I'm at...  what do you make of that interface?

This is very cool and pretty intuitive. My main concern is that one needs to 
build up the function type in generic code (as opposed to mpl vectors), but we 
do have function_types to help us.

You are going to have to propose this as several small patched, before Ralf & 
Dave agree to it.

Regards,
Ravi


From troy at resophonic.com  Sat Oct 31 06:24:16 2009
From: troy at resophonic.com (troy d. straszheim)
Date: Sat, 31 Oct 2009 01:24:16 -0400
Subject: [C++-sig] [python] python + phoenix
In-Reply-To: <200910310002.37065.lists_ravi@lavabit.com>
References: <200910101158.46996.lists_ravi@lavabit.com>	<200910122104.46505.lists_ravi@lavabit.com>	<4AD6820E.9040302@resophonic.com>
	<200910310002.37065.lists_ravi@lavabit.com>
Message-ID: <4AEBCA00.5010902@resophonic.com>

Ravi wrote:
> On Wednesday 14 October 2009 21:59:42 troy d. straszheim wrote:
>> 
>>    def("plus", as<int(int, int)>(arg1 + arg2));
>>
> 
> This is very cool and pretty intuitive. My main concern is that one needs to 
> build up the function type in generic code (as opposed to mpl vectors), but we 
> do have function_types to help us.

I take it that you have a use-case where it is difficult to specify

   as<R(A1,A2)>(thing)

and easy to specify

   as<mpl::vector<R, A1, A2> >(thing)

Could you elaborate?  I'm quite sure there is a way to provide (roughly) 
both interfaces:

   def("foo", as<int(double, string)>(functor()));

   def("foo", as<mpl::vector<int, double, std::string> >(functor()));

but I'm not sure I understand the problem.

> You are going to have to propose this as several small patched, 
 > before Ralf & Dave agree to it.

Shit.  I knew this would happen, I've written too much code again. 
Dave?   Ralf?

-t