From hessiess at hessiess.com  Sun May  3 03:34:14 2009
From: hessiess at hessiess.com (hessiess at hessiess.com)
Date: Sun, 3 May 2009 02:34:14 +0100 (BST)
Subject: [C++-sig] Wrapping functions that return a raw pointer
Message-ID: <a58abb7de34f9f3cfc374d97ceb534d0.squirrel@www.hessiess.dyndns.org>

How can I wrap a function which returns a raw pointer? for example the
following generates tonnes of errors on compile:

struct test
{
    int *get()
    {
	return new int(1);
    }
};

BOOST_PYTHON_MODULE(test)
{
    using namespace boost::python;

    class_<test>("test")
	.def("get", &test::get);
}

Thanks


From diepen at astron.nl  Sun May  3 16:36:28 2009
From: diepen at astron.nl (Ger van Diepen)
Date: Sun, 03 May 2009 16:36:28 +0200
Subject: [C++-sig] numpy.int32 -> double?
Message-ID: <49FDC80C020000A900007D11@server7.nfra.nl>

We also had problems with passing numpy scalars to C++ using Boost (1.35 and before). Sometimes it succeeded, sometimes it failed depending on the data type and if your are on a 32-bit or 64-bit system.
AFAIK Boost does not handle numpy scalar data types, so you have to do the conversion yourself. We've written such conversion functions (see pyrap.googlecode.com).

Cheers,
Ger van Diepen

>>> Bruce Sherwood <bruce.sherwood at gmail.com> 04/28/09 7:52 PM >>>
I'm not sure this is the right forum for my question; it's quite possible
that this is a numpy question.

VPython (vpython.org) is the name of Python plus the 3D Visual module, which
is mostly written in C++ and connected to Python using the Boost libraries.
Visual imports numpy.

When I build Visual for Python 2.6 and numpy 1.3.0, I get a failure on
something that works on Python 2.5 with numpy 1.2.1. There is a
vector(double,double,double) class in Visual which now fails if handed
numpy.int32 arguments (works fine with numpy.float64 arguments). I was using
Boost 1.35 with Python 2.5 and am using Boost 1.38 with Python 2.6.

Any ideas or suggestions? Thanks.

Bruce Sherwood


From renatox at gmail.com  Sun May  3 21:27:10 2009
From: renatox at gmail.com (Renato Araujo)
Date: Sun, 3 May 2009 16:27:10 -0300
Subject: [C++-sig] Wrapping functions that return a raw pointer
In-Reply-To: <a58abb7de34f9f3cfc374d97ceb534d0.squirrel@www.hessiess.dyndns.org>
References: <a58abb7de34f9f3cfc374d97ceb534d0.squirrel@www.hessiess.dyndns.org>
Message-ID: <95291a80905031227v417de992ja262d0bbd93c3b0f@mail.gmail.com>

take a look in this documentation about return police and
ResultConverterGenerator.


http://www.boost.org/doc/libs/1_39_0/libs/python/doc/v2/reference.html#models_of_call_policies

For this specific case you can use this one:
http://www.boost.org/doc/libs/1_39_0/libs/python/doc/v2/manage_new_object.html


BR

On Sat, May 2, 2009 at 10:34 PM,  <hessiess at hessiess.com> wrote:
> How can I wrap a function which returns a raw pointer? for example the
> following generates tonnes of errors on compile:
>
> struct test
> {
> ? ?int *get()
> ? ?{
> ? ? ? ?return new int(1);
> ? ?}
> };
>
> BOOST_PYTHON_MODULE(test)
> {
> ? ?using namespace boost::python;
>
> ? ?class_<test>("test")
> ? ? ? ?.def("get", &test::get);
> }
>
> Thanks
>
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
>



-- 
Renato Araujo Oliveira Filho

From pcc482719 at gmail.com  Mon May  4 17:43:48 2009
From: pcc482719 at gmail.com (Peter C. Chapin)
Date: Mon, 4 May 2009 15:43:48 +0000 (UTC)
Subject: [C++-sig] Building Boost.Python against Stackless Python
Message-ID: <Xns9C017750C5AAApcc482719gmailcom@80.91.229.13>

Hello!

I'm interested in building the Boost/Python integration support using Boost 
1.38, Visual C++ v9 (2008 version), and Stackless Python 2.6.1. I'm on a 
Windows XP machine.

Currently the build fails with an error about not being able to find 
stackless.h as included from C:\Python26\include\Python.h. In fact, 
stackless.h is in C:\Python26\include\Stackless. I assume I need to inform 
the Boost build system to search this additional include directory. 
However, I'm not clear on how to do that (I'm new to Boost).

I searched the archive of this group and found a posting from August of 
2008 on exactly this issue. However, there was no reply. I'm hoping I'll be 
a bit luckier. :-)

Thanks for any advice you can give.

Peter


From rogeeff at gmail.com  Tue May  5 08:59:45 2009
From: rogeeff at gmail.com (Gennadiy Rozental)
Date: Tue, 5 May 2009 06:59:45 +0000 (UTC)
Subject: [C++-sig] how to deep copy boost python object
Message-ID: <loom.20090505T065834-270@post.gmane.org>

This looks like trivial question from FAQ, but I can seem to fond the answer.

Any pointers?

Genandiy


From sipickles at googlemail.com  Tue May  5 11:29:09 2009
From: sipickles at googlemail.com (Simon Pickles)
Date: Tue, 05 May 2009 10:29:09 +0100
Subject: [C++-sig] Extended python system needs access to cPickle in c++
Message-ID: <4A0006E5.4070602@googlemail.com>

Hi,

I have an app with a python core, then c++ extension modules.

I'd like to be able to use cPickle to pack structures, especially 
boost::python::tuples, in c++.

Is there a way I can expose a python module in the c++ extensions?

I thought about passing a module as an arg to a c++ function, as a 
boost::python::object:

// cModule

void DoStuff(object pickleModule)
{
    tuple t = make_tuple("Spam",42);
    object pickleDumps = pickleModule.attr("dumps");
    object s = pickleDumps(t);

    // Send s to other process
}



# python
import cModule
import cPickle

cModule.DoStuff(cPickle)
#

Am I barking up the wrong tree?

Many thanks

Simon

From hessiess at hessiess.com  Tue May  5 11:54:08 2009
From: hessiess at hessiess.com (hessiess at hessiess.com)
Date: Tue, 5 May 2009 10:54:08 +0100 (BST)
Subject: [C++-sig] Wrapping functions that return a raw pointer
In-Reply-To: <95291a80905031227v417de992ja262d0bbd93c3b0f@mail.gmail.com>
References: <a58abb7de34f9f3cfc374d97ceb534d0.squirrel@www.hessiess.dyndns.org>
	<95291a80905031227v417de992ja262d0bbd93c3b0f@mail.gmail.com>
Message-ID: <1ba66d9635e00804969a0a1181d35011.squirrel@www.hessiess.dyndns.org>

Thanks
> take a look in this documentation about return police and
> ResultConverterGenerator.
>
>
> http://www.boost.org/doc/libs/1_39_0/libs/python/doc/v2/reference.html#models_of_call_policies
>
> For this specific case you can use this one:
> http://www.boost.org/doc/libs/1_39_0/libs/python/doc/v2/manage_new_object.html
>
>
> BR
>
> On Sat, May 2, 2009 at 10:34 PM,  <hessiess at hessiess.com> wrote:
>> How can I wrap a function which returns a raw pointer? for example the
>> following generates tonnes of errors on compile:
>>
>> struct test
>> {
>> ? ?int *get()
>> ? ?{
>> ? ? ? ?return new int(1);
>> ? ?}
>> };
>>
>> BOOST_PYTHON_MODULE(test)
>> {
>> ? ?using namespace boost::python;
>>
>> ? ?class_<test>("test")
>> ? ? ? ?.def("get", &test::get);
>> }
>>
>> Thanks
>>
>> _______________________________________________
>> Cplusplus-sig mailing list
>> Cplusplus-sig at python.org
>> http://mail.python.org/mailman/listinfo/cplusplus-sig
>>
>
>
>
> --
> Renato Araujo Oliveira Filho
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
>



From ndbecker2 at gmail.com  Tue May  5 13:08:10 2009
From: ndbecker2 at gmail.com (Neal Becker)
Date: Tue, 05 May 2009 07:08:10 -0400
Subject: [C++-sig] how to deep copy boost python object
References: <loom.20090505T065834-270@post.gmane.org>
Message-ID: <gtp6mq$u3s$1@ger.gmane.org>

Gennadiy Rozental wrote:

> This looks like trivial question from FAQ, but I can seem to fond the
> answer.
> 
> Any pointers?
> 
> Genandiy

Maybe overide __copy__?



From wladwig at wdtinc.com  Tue May  5 16:53:28 2009
From: wladwig at wdtinc.com (William Ladwig)
Date: Tue, 5 May 2009 09:53:28 -0500
Subject: [C++-sig] Extended python system needs access to cPickle in c++
In-Reply-To: <4A0006E5.4070602@googlemail.com>
References: <4A0006E5.4070602@googlemail.com>
Message-ID: <765CBD9053EA2B438625895F30D1856F02262A53B9@storm.wdtinc.com>

You can import python modules in C++ using the import function.  See here:

http://www.boost.org/doc/libs/1_39_0/libs/python/doc/v2/import.html

You can also make your C++ extension classes 'pickleable' using boost python's pickle suite:

http://www.boost.org/doc/libs/1_39_0/libs/python/doc/v2/pickle.html


Bill

-----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 Simon Pickles
Sent: Tuesday, May 05, 2009 4:29 AM
To: cplusplus-sig at python.org
Subject: [C++-sig] Extended python system needs access to cPickle in c++

Hi,

I have an app with a python core, then c++ extension modules.

I'd like to be able to use cPickle to pack structures, especially 
boost::python::tuples, in c++.

Is there a way I can expose a python module in the c++ extensions?

I thought about passing a module as an arg to a c++ function, as a 
boost::python::object:

// cModule

void DoStuff(object pickleModule)
{
    tuple t = make_tuple("Spam",42);
    object pickleDumps = pickleModule.attr("dumps");
    object s = pickleDumps(t);

    // Send s to other process
}



# python
import cModule
import cPickle

cModule.DoStuff(cPickle)
#

Am I barking up the wrong tree?

Many thanks

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

From rogeeff at gmail.com  Tue May  5 22:39:28 2009
From: rogeeff at gmail.com (Gennadiy Rozental)
Date: Tue, 5 May 2009 20:39:28 +0000 (UTC)
Subject: [C++-sig] how to deep copy boost python object
References: <loom.20090505T065834-270@post.gmane.org>
	<gtp6mq$u3s$1@ger.gmane.org>
Message-ID: <loom.20090505T203728-985@post.gmane.org>

Neal Becker <ndbecker2 <at> gmail.com> writes:

> 
> Gennadiy Rozental wrote:
> 
> > This looks like trivial question from FAQ, but I can seem to fond the
> > answer.
> > 
> > Any pointers?
> > 
> > Genandiy
> 
> Maybe overide __copy__?


How is it involved? Here is an example:

bp::object o = foo();

// Here i want to make deep copy of o
// bp::object o_copy = o is not correct
// what I need is something like:

bp::object o_copy = bp::deep_copy( o )

Gennadiy


From jeremykie at gmail.com  Wed May  6 19:23:56 2009
From: jeremykie at gmail.com (Jeremy Kie)
Date: Wed, 6 May 2009 11:23:56 -0600
Subject: [C++-sig] wrapping objects which inherit from std::map
Message-ID: <f49e5f5e0905061023t4abd8ddasd3e0a1e74dde6a47@mail.gmail.com>

Hi,

I have the following:

class foo : public std::map<std::string, std::string> {
}

BOOST_PYTHON_MODULE(foo_ext)
{
  using namespace boost::python;
  class_<foo, bases<std::map>>("foo");
}

The compiler complains that template arguments 1 and 2 are invalid.  I'm
certain my syntax is the problem.  Can't anyone shed some light and help me
correct my syntax?

Thanks,

Jeremy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20090506/8cf8d708/attachment.htm>

From seefeld at sympatico.ca  Wed May  6 19:29:59 2009
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Wed, 06 May 2009 13:29:59 -0400
Subject: [C++-sig] wrapping objects which inherit from std::map
In-Reply-To: <f49e5f5e0905061023t4abd8ddasd3e0a1e74dde6a47@mail.gmail.com>
References: <f49e5f5e0905061023t4abd8ddasd3e0a1e74dde6a47@mail.gmail.com>
Message-ID: <4A01C917.3090800@sympatico.ca>

Jeremy Kie wrote:
> Hi,
>
> I have the following:
>
> class foo : public std::map<std::string, std::string> {
> }

You lack a semicolon here. (I also don't think it's a good idea to 
derive from standard containers, but that is an entirely different subject).

>
> BOOST_PYTHON_MODULE(foo_ext)
> {
>   using namespace boost::python;
>   class_<foo, bases<std::map>>("foo");

There are three errors here, let's start with the simplest:

You use '>>' in this line, which is a single token. You need to insert a 
whitespace there to make it two tokens, e.g. '> >'.
Next, you use 'std::map' where you ought to use a type (such as 
'std::map<std::string, std::string>'.
Finally, you only can declare bases that are themselves already exported 
to python via class_, so you need to start by exporting the base type 
std::map<std::string, std::string> prior to 'foo'.

Hope this helps,
       Stefan


-- 

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


From jeremykie at gmail.com  Wed May  6 19:44:18 2009
From: jeremykie at gmail.com (Jeremy Kie)
Date: Wed, 6 May 2009 11:44:18 -0600
Subject: [C++-sig] wrapping objects which inherit from std::map
In-Reply-To: <4A01C917.3090800@sympatico.ca>
References: <f49e5f5e0905061023t4abd8ddasd3e0a1e74dde6a47@mail.gmail.com>
	<4A01C917.3090800@sympatico.ca>
Message-ID: <f49e5f5e0905061044l6e1f1a45h2dfbce0173bf774e@mail.gmail.com>

Thanks for the direction Stefan.

Apologies for the typos. What I was missing was the declaration of the
"std::map" prior to my foo declaration.  Yes, I agree with you regarding
derivation from standard containers.  My portion of the project is written
in python, I'm trying to work with libraries written by another member of
the team.

With the addition of the "std::map" declaration, my compiler errors
disappeared.

Thanks,

Jeremy


On Wed, May 6, 2009 at 11:29 AM, Stefan Seefeld <seefeld at sympatico.ca>wrote:

> Jeremy Kie wrote:
>
>> Hi,
>>
>> I have the following:
>>
>> class foo : public std::map<std::string, std::string> {
>> }
>>
>
> You lack a semicolon here. (I also don't think it's a good idea to derive
> from standard containers, but that is an entirely different subject).
>
>
>> BOOST_PYTHON_MODULE(foo_ext)
>> {
>>  using namespace boost::python;
>>  class_<foo, bases<std::map>>("foo");
>>
>
> There are three errors here, let's start with the simplest:
>
> You use '>>' in this line, which is a single token. You need to insert a
> whitespace there to make it two tokens, e.g. '> >'.
> Next, you use 'std::map' where you ought to use a type (such as
> 'std::map<std::string, std::string>'.
> Finally, you only can declare bases that are themselves already exported to
> python via class_, so you need to start by exporting the base type
> std::map<std::string, std::string> prior to 'foo'.
>
> Hope this helps,
>      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
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20090506/7e1ec1c6/attachment.htm>

From seefeld at sympatico.ca  Wed May  6 19:56:26 2009
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Wed, 06 May 2009 13:56:26 -0400
Subject: [C++-sig] wrapping objects which inherit from std::map
In-Reply-To: <f49e5f5e0905061044l6e1f1a45h2dfbce0173bf774e@mail.gmail.com>
References: <f49e5f5e0905061023t4abd8ddasd3e0a1e74dde6a47@mail.gmail.com>	<4A01C917.3090800@sympatico.ca>
	<f49e5f5e0905061044l6e1f1a45h2dfbce0173bf774e@mail.gmail.com>
Message-ID: <4A01CF4A.7080502@sympatico.ca>

Jeremy Kie wrote:
>
> With the addition of the "std::map" declaration, my compiler errors 
> disappeared.

That is very curious, as Python only discovers whether the base class 
has been exported or not at runtime. Thus it can't possibly result in a 
compiler error.

Regards,
       Stefan


-- 

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


From jeremykie at gmail.com  Wed May  6 20:02:23 2009
From: jeremykie at gmail.com (Jeremy Kie)
Date: Wed, 6 May 2009 12:02:23 -0600
Subject: [C++-sig] wrapping objects which inherit from std::map
In-Reply-To: <4A01CF4A.7080502@sympatico.ca>
References: <f49e5f5e0905061023t4abd8ddasd3e0a1e74dde6a47@mail.gmail.com>
	<4A01C917.3090800@sympatico.ca>
	<f49e5f5e0905061044l6e1f1a45h2dfbce0173bf774e@mail.gmail.com>
	<4A01CF4A.7080502@sympatico.ca>
Message-ID: <f49e5f5e0905061102g4836026kc63b8547ebcf32b6@mail.gmail.com>

I think you are right.  I removed the declaration of the std::map and it
compiled without any errors.  The problem must have been related to the
missing template types.

regards,

Jeremy

On Wed, May 6, 2009 at 11:56 AM, Stefan Seefeld <seefeld at sympatico.ca>wrote:

> Jeremy Kie wrote:
>
>>
>> With the addition of the "std::map" declaration, my compiler errors
>> disappeared.
>>
>
> That is very curious, as Python only discovers whether the base class has
> been exported or not at runtime. Thus it can't possibly result in a compiler
> error.
>
> 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
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20090506/bf38dfcd/attachment.htm>

From jeremykie at gmail.com  Thu May  7 00:10:22 2009
From: jeremykie at gmail.com (Jeremy Kie)
Date: Wed, 6 May 2009 16:10:22 -0600
Subject: [C++-sig] boost python class inherits boost::asio::ip::tcp::socket
Message-ID: <f49e5f5e0905061510i241c4312sb606fbd22806d67b@mail.gmail.com>

Hello,

The library I am attempting to extend Python to contains a class which
inherits from boost::asio::ip::tcp::socket.

class tcp_client_socket : public boost::asio::ip::tcp::socket {
// class members and functions go here...
};

My question is whether or not it is a good idea to wrap this.  My naive
approach would be to something like:

BOOST_PYTHON_MODULE(tcp_client_socket_ext)
{
  using namespace boost::python;
  class_<tcp_client_socket, bases<boost::asio::ip::tcp::socket>
>("tcp_client_socket");
}

I did an initial compile and received many compiler errors.  If anyone has
done this before, what is the correct way to wrap such an object.

Any help is greatly appreciated.

Regards,

-Jeremy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20090506/42ececf5/attachment-0001.htm>

From eudoxos at arcig.cz  Thu May  7 19:21:21 2009
From: eudoxos at arcig.cz (=?windows-1252?Q?V=E1clav_=8Amilauer?=)
Date: Thu, 07 May 2009 19:21:21 +0200
Subject: [C++-sig] hijacking getattr/setattr
Message-ID: <4A031891.1070700@arcig.cz>

Hello,

I have  wrapper c++ classes that are exposed to python that wrap "real" 
c++ classes with data. Wrapper classes have some attributes that .def'ed 
and .add_property'd in the glue code. Wrapped classes have also 
attributes that are registered by name (for serialization purposes) and 
I can retrieve them.

To be able to get both wrapper attributes and attributes of the wrapped 
class, I use getitem/setitem for the latter ones. That makes the python 
code a bit ugly and difficult to read:

  b=O.bodies[44]
  b.geom['radius']=5  ## .geom returns wrapped class holding geometry 
attributes, is .add_property'ed
  b.phys['mass']=8     ## .phys returns wrapped class holding physical 
attributes, is .add_property'ed
  print b['id']           ## get an attribute that exists as such in the 
wrapped class

I would like to unify both approaches so that if I say b.id, the pyBody 
object (wrapping a shared_ptr<Body> object) will first look if the 
attribute 'id' is within attributes of the wrapper class; return it if 
it is, look in the wrapped object attributes if it is not; return the 
wrapped object attribute of that name if it exists, otherwise throw. 
(The same for b.geom.radius, b.phys.mass)

I simulated that with pure python code (attached), but it is not clear 
to me how to put this in boost::python.

Can I get some pointers? I can overrider __getattr__ and __setattr__ of 
the wrapper class, but I don't know how to simulate the default behavior 
of getattr/setattr, once the key wasn't found in the wrapped class, for 
example.

Thanks,

Vaclav

From eudoxos at arcig.cz  Thu May  7 23:38:59 2009
From: eudoxos at arcig.cz (=?UTF-8?Q?V=C3=A1clav_=C5=A0milauer?=)
Date: Thu, 07 May 2009 23:38:59 +0200
Subject: [C++-sig] hijacking getattr/setattr
In-Reply-To: <4A031891.1070700@arcig.cz>
References: <4A031891.1070700@arcig.cz>
Message-ID: <1241732339.4433.1.camel@flux>

Forgotten attachment (sorry): python hijack.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: a.py
Type: text/x-python
Size: 1276 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20090507/f3207db0/attachment.py>

From rwgk at yahoo.com  Sun May 10 05:13:37 2009
From: rwgk at yahoo.com (Ralf W. Grosse-Kunstleve)
Date: Sat, 9 May 2009 20:13:37 -0700 (PDT)
Subject: [C++-sig] how to deep copy boost python object
In-Reply-To: <loom.20090505T065834-270@post.gmane.org>
References: <loom.20090505T065834-270@post.gmane.org>
Message-ID: <377808.11921.qm@web111416.mail.gq1.yahoo.com>


A comprehensive approach is to make the object picklable.
As a side-effect, copy.deepcopy() will also work.

http://www.boost.org/doc/libs/1_39_0/libs/python/doc/v2/pickle.html

(In many cases - in my experience - making an object picklable is very simple.)



----- Original Message ----
From: Gennadiy Rozental <rogeeff at gmail.com>
To: cplusplus-sig at python.org
Sent: Monday, May 4, 2009 11:59:45 PM
Subject: [C++-sig] how to deep copy boost python object

This looks like trivial question from FAQ, but I can seem to fond the answer.

Any pointers?

Genandiy


From security at BankofAmerica.com  Fri May  8 16:44:59 2009
From: security at BankofAmerica.com (security at BankofAmerica.com)
Date: Fri, 08 May 2009 10:44:59 -0400
Subject: [C++-sig] Online Banking Alert
Message-ID: <E1M2RJr-0006mz-Q4@w1.pressplug.com>

An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20090508/5af3b73e/attachment.htm>

From rogeeff at gmail.com  Mon May 11 07:46:32 2009
From: rogeeff at gmail.com (Gennadiy Rozental)
Date: Mon, 11 May 2009 05:46:32 +0000 (UTC)
Subject: [C++-sig] how to deep copy boost python object
References: <loom.20090505T065834-270@post.gmane.org>
	<377808.11921.qm@web111416.mail.gq1.yahoo.com>
Message-ID: <loom.20090511T054237-113@post.gmane.org>

Ralf W. Grosse-Kunstleve <rwgk <at> yahoo.com> writes:

> 
> 
> A comprehensive approach is to make the object picklable.
> As a side-effect, copy.deepcopy() will also work.

1. Are you saying that copy.deepcopy does not work without object being
picklable? What will happend if I try to use it for non picklable object?

2. What if object IS picklable. How will invoke deepcopy operation on bp::object?

Gennadiy


From Matthew.Scouten at tradingtechnologies.com  Mon May 11 18:31:41 2009
From: Matthew.Scouten at tradingtechnologies.com (Matthew Scouten (TT))
Date: Mon, 11 May 2009 11:31:41 -0500
Subject: [C++-sig] how to deep copy boost python object
In-Reply-To: <loom.20090511T054237-113@post.gmane.org>
References: <loom.20090505T065834-270@post.gmane.org><377808.11921.qm@web111416.mail.gq1.yahoo.com>
	<loom.20090511T054237-113@post.gmane.org>
Message-ID: <32490DFF7774554A85D65D23A9F0F9380AAB7A38@chiex01>

If you make an object pickleable, you get deepcopy for free, because the
deepcopy function will use a pickle-unpickle cycle as a fall back if it
cannot find any other way to copy. This has the disadvantage that you
must come up with a sensible pickled representation, and you pay for a
round trip through a string. This is especially annoying since all c++
objects are already deepcopyable by default: just use a call copy-ctor.

There is another way to tell python how to deepcopy an object: give it a
__deepcopy__ function (and a __copy__ function). 
This is some code I got from Hans Meine a while back:

#define PYTHON_ERROR(TYPE, REASON) \
{ \
    PyErr_SetString(TYPE, REASON); \
    throw bp::error_already_set(); \
}

template<class T>
inline PyObject * managingPyObject(T *p)
{
    return typename bp::manage_new_object::apply<T *>::type()(p);
}

template<class Copyable>
bp::object
generic__copy__(bp::object copyable)
{
    Copyable *newCopyable(new Copyable(bp::extract<const Copyable
&>(copyable)));
    bp::object
result(bp::detail::new_reference(managingPyObject(newCopyable)));

    bp::extract<bp::dict>(result.attr("__dict__"))().update(
        copyable.attr("__dict__"));

    return result;
}

template<class Copyable>
bp::object
generic__deepcopy__(bp::object copyable, bp::dict memo)
{
    bp::object copyMod = bp::import("copy");
    bp::object deepcopy = copyMod.attr("deepcopy");

    Copyable *newCopyable(new Copyable(bp::extract<const Copyable
&>(copyable)));
    bp::object
result(bp::detail::new_reference(managingPyObject(newCopyable)));

    // HACK: copyableId shall be the same as the result of id(copyable)
in Python -
    // please tell me that there is a better way! (and which ;-p)
    int copyableId = (int)(copyable.ptr());
    memo[copyableId] = result;

    bp::extract<bp::dict>(result.attr("__dict__"))().update(
        deepcopy(bp::extract<bp::dict>(copyable.attr("__dict__"))(),
memo));

    return result;
}

To use it:

class_<foo>(foo)
   .def("__copy__", &generic__copy__< foo >)
   .def("__deepcopy__", &generic__deepcopy__< foo >)
   .def(init< const foo & >())


This has the advantage that you can use it on any existing class that
has a sensible cctor. 









-----Original Message-----
From:
cplusplus-sig-bounces+matthew.scouten=tradingtechnologies.com at python.org
[mailto:cplusplus-sig-bounces+matthew.scouten=tradingtechnologies.com at py
thon.org] On Behalf Of Gennadiy Rozental
Sent: Monday, May 11, 2009 12:47 AM
To: cplusplus-sig at python.org
Subject: Re: [C++-sig] how to deep copy boost python object

Ralf W. Grosse-Kunstleve <rwgk <at> yahoo.com> writes:

> 
> 
> A comprehensive approach is to make the object picklable.
> As a side-effect, copy.deepcopy() will also work.

1. Are you saying that copy.deepcopy does not work without object being
picklable? What will happend if I try to use it for non picklable
object?

2. What if object IS picklable. How will invoke deepcopy operation on
bp::object?

Gennadiy

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

From bruce.sherwood at gmail.com  Wed May 13 03:38:43 2009
From: bruce.sherwood at gmail.com (Bruce Sherwood)
Date: Tue, 12 May 2009 19:38:43 -0600
Subject: [C++-sig] Manual converter numpy.int32 -> double
Message-ID: <eab3aea70905121838n2ae835b6if1a38c3dbe487900@mail.gmail.com>

Some time ago I wrote that I was having trouble installing a manual
converter to handle numpy.int32 -> double, something that used to work
automatically but no longer does. I  got crucial help from David Scherer
(the original creator of VPython). Just in case someone might be interested,
here is a routine that works.

struct double_from_int
{
double_from_int()
{
py::converter::registry::push_back(
&convertible,
&construct,
py::type_id<double>());
}

static void* convertible( PyObject* obj)
{
PyObject* newobj = PyNumber_Float(obj);
if (!PyString_Check(obj) && newobj) {
Py_DECREF(newobj);
return obj;
} else {
if (newobj) {
Py_DECREF(newobj);
}
PyErr_Clear();
return 0;
}
}

static void construct(
PyObject* _obj,
py::converter::rvalue_from_python_stage1_data* data)
{
PyObject* newobj = PyNumber_Float(_obj);
//void* storage = (
//            (py::converter::rvalue_from_python_storage<double>*)
//            data)->storage.bytes;
//*(double*)(storage) = py::extract<double>(newobj);

        double* storage = (double*)(
            (py::converter::rvalue_from_python_storage<double>*)
            data)->storage.bytes;
        *storage = py::extract<double>(newobj);
Py_DECREF(newobj);
data->convertible = storage;
}
};

Bruce Sherwood
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20090512/0298008e/attachment.htm>

From bmiller at soe.sony.com  Wed May 13 23:43:27 2009
From: bmiller at soe.sony.com (Brandon)
Date: Wed, 13 May 2009 21:43:27 +0000 (UTC)
Subject: [C++-sig] [boost::python] Exposing a pointer to a vector to Python.
Message-ID: <loom.20090513T205632-832@post.gmane.org>

All,

I'm currently exposing much of our C++ API to Python and have had much success
with everything up to here.  I'm running into issues trying to expose a pointer
to a vector which holds pointers to instances of a class.

The question is:  is there more I have to do to export a pointer to a vector of
a class type we've defined?  Or is there something simpler that I'm not seeing?
 I've tried exposing a pointer to a vector, but that didn't turn out well.  I
was getting value_type undefined errors and could not find any postings on them.
 Perhaps that way was correct but I didn't not have the whole piece in place.

Thanks in advance!

CODE: (all of my code and sample types replace our actual data types)
-------------------------------------------------------------------------

C++
-------

class sample {
public:
   std::vector<Derived*> *vec;
};


class_<std::vector<Derived *> >("std_vector_Derived")
        .def(vector_indexing_suite<vector<Derived *>())
;

class_<sample>("sample")
   .def_readwrite("vec", &sample::vec)
;

Python
--------
inst = EXPOSED_MODULE.EXPOSED_FUNCT( )      
#EXPOSED_MODULE/FUNCT are replacements for the real name...
#inst is an instance of "sample" returned from a exposed c++ function after it
was populated it with data.

print inst.vec[0] 
#Assume vector was populated from C++ side and verified to be populated in C++.
#Problem occurs here because of type...

STDERR (PYTHON)
---------------
TypeError: No to_python(by-value) converter found for C++ type: class
std::vector<class Derived*, class std::allocator<class Derived*> > *









From hansroessler at yahoo.de  Thu May 14 15:15:30 2009
From: hansroessler at yahoo.de (Hans Roessler)
Date: Thu, 14 May 2009 06:15:30 -0700 (PDT)
Subject: [C++-sig] operator+(A,
	B) without B::B() default constructor in boost.python
Message-ID: <130401.91415.qm@web24608.mail.ird.yahoo.com>


Hello,
how can I wrap the following overloaded operator+ using boost.python?
The code as is fails with 
>...
>pytest.cpp: In function ?void init_module_pytest()?:
>pytest.cpp:19: error: expected primary-expression before ?(? token
>...
If I try ".def(self+B())" instead of ".def(self+other<B>())", the error is
>...
>pytest.cpp: In function ?void init_module_pytest()?:
>pytest.cpp:17: error: no matching function for call to ?B::B()?
>...

best wishes
Hans


pytest.cpp:

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

class A{
public:
};
class B{
public:
  B(int i){};
};

int operator+(const A& a,const B& b){return 42;};

BOOST_PYTHON_MODULE(pytest)
{
  class_<A>("A")
    .def(self+other<B>())   //.def(self+B())            //line 17
    ;
  class_<B,boost::noncopyable>("B",init<int>)    //line 19
    ;
}


      

From magnus at clara.is  Thu May 14 20:41:02 2009
From: magnus at clara.is (=?ISO-8859-1?Q?Magn=FAs_Bergur_Magn=FAsson?=)
Date: Thu, 14 May 2009 18:41:02 +0000
Subject: [C++-sig] Question regarding importing libraries to weave
Message-ID: <20504a380905141141j4ac4793ej306603590deb21d@mail.gmail.com>

Hello,

My name is Magnus and I am working as a software developer for a company in
Iceland. We use weave inline with python to boost our processing capacity.
However there is one factor that is limiting us. We cannot figure out how to
import libraries to Weave such as just <math.h> or <string.h>, what is the
right procedure for that? Also, what is the most effective tool for text
processing in python? We have been using weave inline and moving char by
char but is there a more effective way and could string.h speed things up?

Regards, Magnus Magnusson
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20090514/75b4c02f/attachment.htm>

From pbloom at crystald.com  Thu May 14 23:01:18 2009
From: pbloom at crystald.com (Philip Bloom)
Date: Thu, 14 May 2009 14:01:18 -0700
Subject: [C++-sig] Trouble with custom container class
Message-ID: <2DEFEB25B65B1D4C962D79E3CCBAFC7F06E883FF@mpkexc01.eidos.com>

Hello,

I'm having some issues with retrieving objects from a custom container
class.  The objects in the container are otherwise wrapped objects
(PythonDTPNodeHandle *) that the python wrapper knows how to convert,
but what I get back from the container with __getitem__ seems to not
know how to convert it.    I suspect that I'm doing something wrong on
what I'm feeding Py++ and thus the generated wrapping is bad, but I
don't see anything wrong with the final wrapping for getitem or
operator[].  I've no idea why it wouldn't be able to convert, for this
specific usage.  Any advise/help would be well welcome.  I've included
the whole wrapper at the bottom.

 

Example use:

          Node=Test.getRootNode() #Returns a PythonDTPNodeHandle *
object, which is converted into a PythonDTPNodeHandle refrence in
python.
          #print Node - Which is wrapped fine, converting naturally.

    array=test.getCDCArray() #Returns the container object.  This is a
wrapped cdc::array<PythonDTPNodeHandle *> * object.

    #help(array) - Verified Python sees it as cdcArray(super), the
container object

    print "Array has: %s elements"%array.size() # Array is empty

    node.GetItemChildren(array) #Function call passed the Array.  The
function expects a cdc::array<PythonDTPNodeHandle *> * object.

    print "Array has: %s elements"%array.size() # Array has 125
elements.

    print array[0] #Causes error

    anode=array.__getitem__(0) #Causes error.

 

Output:

Array has: 0 elements

Array has: 125 elements

Traceback (most recent call last):

  File "D:\stgame\cdc\tools\cdcPython\DTPXMLNode\example.py", line 43,
in <module>

    print array[0]

TypeError: No to_python (by-value) converter found for C++ type: class
PythonDTPNodeHandle *

The getitem wrapped:
        { //::cdc::ArrayImpl< PythonDTPNodeHandle* >::operator[]

        

            typedef cdc::ArrayImpl< PythonDTPNodeHandle* >
exported_class_t;

            typedef ::PythonDTPNodeHandle * & (
exported_class_t::*__getitem___function_type )( ::cdc::uint32 ) ;

            

            super_exposer.def( 

                "__getitem__"

                , __getitem___function_type( &::cdc::ArrayImpl<
PythonDTPNodeHandle* >::operator[] )

                , ( bp::arg("i") )

                , bp::return_internal_reference< >() );

        

        }

 



My Python Wrapper:
// This file has been generated by Py++.

 

//CDC DTP Autogenerated Python Interface

 

#include "boost/python.hpp"

 

#include "pythondtpscriptinterfacewrapper.h"

 

#include "cdcsys/array.h"

 

namespace bp = boost::python;

 

struct PythonDTPInterface_wrapper : PythonDTPInterface, bp::wrapper<
PythonDTPInterface > {

 

    PythonDTPInterface_wrapper(PythonDTPInterface const & arg )

    : PythonDTPInterface( arg )

      , bp::wrapper< PythonDTPInterface >(){

        // copy constructor

        

    }

 

    PythonDTPInterface_wrapper(char const * szGameRoot )

    : PythonDTPInterface( szGameRoot )

      , bp::wrapper< PythonDTPInterface >(){

        // constructor

    

    }

 

    PythonDTPInterface_wrapper(char const * szGameRoot, char const *
datFile, char const * dtpFile )

    : PythonDTPInterface( szGameRoot, datFile, dtpFile )

      , bp::wrapper< PythonDTPInterface >(){

        // constructor

    

    }

 

    virtual void NotifyDATFileChanged( char const * pDATItemPath ) {

        if( bp::override func_NotifyDATFileChanged = this->get_override(
"NotifyDATFileChanged" ) )

            func_NotifyDATFileChanged( pDATItemPath );

        else

            this->PythonDTPInterface::NotifyDATFileChanged( pDATItemPath
);

    }

    

    

    void default_NotifyDATFileChanged( char const * pDATItemPath ) {

        PythonDTPInterface::NotifyDATFileChanged( pDATItemPath );

    }

 

};

 

BOOST_PYTHON_MODULE(DTPScriptInterface){

    { //::PythonDTPInterface

        typedef bp::class_< PythonDTPInterface_wrapper >
PythonDTPInterface_exposer_t;

        PythonDTPInterface_exposer_t PythonDTPInterface_exposer =
PythonDTPInterface_exposer_t( "PythonDTPInterface", bp::init< char const
* >(( bp::arg("szGameRoot") )) );

        bp::scope PythonDTPInterface_scope( PythonDTPInterface_exposer
);

        bp::implicitly_convertible< char const *, PythonDTPInterface
>();

        PythonDTPInterface_exposer.def( bp::init< char const *, char
const *, char const * >(( bp::arg("szGameRoot"), bp::arg("datFile"),
bp::arg("dtpFile") )) );

        { //::PythonDTPInterface::DumbTestFunction

        

            typedef int (
::PythonDTPInterface::*DumbTestFunction_function_type )(  ) ;

            

            PythonDTPInterface_exposer.def( 

                "DumbTestFunction"

                , DumbTestFunction_function_type(
&::PythonDTPInterface::DumbTestFunction ) );

        

        }

        { //::PythonDTPInterface::GetRootNode

        

            typedef ::PythonDTPNodeHandle * (
::PythonDTPInterface::*GetRootNode_function_type )(  ) ;

            

            PythonDTPInterface_exposer.def( 

                "GetRootNode"

                , GetRootNode_function_type(
&::PythonDTPInterface::GetRootNode )

                , bp::return_value_policy< bp::manage_new_object >() );

        

        }

        { //::PythonDTPInterface::NotifyDATFileChanged

        

            typedef void (
::PythonDTPInterface::*NotifyDATFileChanged_function_type )( char const
* ) ;

            typedef void (
PythonDTPInterface_wrapper::*default_NotifyDATFileChanged_function_type
)( char const * ) ;

            

            PythonDTPInterface_exposer.def( 

                "NotifyDATFileChanged"

                ,
NotifyDATFileChanged_function_type(&::PythonDTPInterface::NotifyDATFileC
hanged)

                ,
default_NotifyDATFileChanged_function_type(&PythonDTPInterface_wrapper::
default_NotifyDATFileChanged)

                , ( bp::arg("pDATItemPath") ) );

        

        }

        { //::PythonDTPInterface::OpenDATFile

        

            typedef bool (
::PythonDTPInterface::*OpenDATFile_function_type )( char const *,char
const *,bool ) ;

            

            PythonDTPInterface_exposer.def( 

                "OpenDATFile"

                , OpenDATFile_function_type(
&::PythonDTPInterface::OpenDATFile )

                , ( bp::arg("datFilenameAndPath"),
bp::arg("dtpFilenameAndPath"), bp::arg("bSavePrevious")=(bool)(true) )
);

        

        }

        { //::PythonDTPInterface::Release

        

            typedef void ( ::PythonDTPInterface::*Release_function_type
)(  ) ;

            

            PythonDTPInterface_exposer.def( 

                "Release"

                , Release_function_type( &::PythonDTPInterface::Release
) );

        

        }

        { //::PythonDTPInterface::SaveDATFile

        

            typedef void (
::PythonDTPInterface::*SaveDATFile_function_type )(  ) ;

            

            PythonDTPInterface_exposer.def( 

                "SaveDATFile"

                , SaveDATFile_function_type(
&::PythonDTPInterface::SaveDATFile ) );

        

        }

        { //::PythonDTPInterface::getCDCArray

        

            typedef ::cdc::Array< PythonDTPNodeHandle* > * (
::PythonDTPInterface::*getCDCArray_function_type )(  ) ;

            

            PythonDTPInterface_exposer.def( 

                "getCDCArray"

                , getCDCArray_function_type(
&::PythonDTPInterface::getCDCArray )

                , bp::return_value_policy< bp::manage_new_object >() );

        

        }

        { //::PythonDTPInterface::instantiate

        

            typedef void (
::PythonDTPInterface::*instantiate_function_type )(  ) ;

            

            PythonDTPInterface_exposer.def( 

                "instantiate"

                , instantiate_function_type(
&::PythonDTPInterface::instantiate ) );

        

        }

    }

 

    bp::class_< PythonDTPNodeHandle, boost::noncopyable >(
"PythonDTPNodeHandle", bp::no_init )    

        .def( 

            "AddChild"

            , (::PythonDTPNodeHandle * ( ::PythonDTPNodeHandle::* )(
char const * ) )( &::PythonDTPNodeHandle::AddChild )

            , ( bp::arg("szName") )

            , bp::return_value_policy< bp::manage_new_object >() )    

        .def( 

            "DumbNodeTestFunction"

            , (int ( ::PythonDTPNodeHandle::* )(  ) )(
&::PythonDTPNodeHandle::DumbNodeTestFunction ) )    

        .def( 

            "GetChild"

            , (::PythonDTPNodeHandle * ( ::PythonDTPNodeHandle::* )(
char const * ) )( &::PythonDTPNodeHandle::GetChild )

            , ( bp::arg("szName") )

            , bp::return_value_policy< bp::manage_new_object >() )    

        .def( 

            "GetItemChildren"

            , (void ( ::PythonDTPNodeHandle::* )( ::cdc::Array<
PythonDTPNodeHandle* > * ) )( &::PythonDTPNodeHandle::GetItemChildren )

            , ( bp::arg("pAllItemChildren") ) )    

        .def( 

            "GetName"

            , (char const * ( ::PythonDTPNodeHandle::* )(  ) )(
&::PythonDTPNodeHandle::GetName ) )    

        .def( 

            "GetTagName"

            , (char const * ( ::PythonDTPNodeHandle::* )(  ) )(
&::PythonDTPNodeHandle::GetTagName ) )    

        .def( 

            "GetValue"

            , (char const * ( ::PythonDTPNodeHandle::* )(  ) )(
&::PythonDTPNodeHandle::GetValue ) )    

        .def( 

            "Release"

            , (void ( ::PythonDTPNodeHandle::* )(  ) )(
&::PythonDTPNodeHandle::Release ) )    

        .def( 

            "RemoveChild"

            , (bool ( ::PythonDTPNodeHandle::* )( char const * ) )(
&::PythonDTPNodeHandle::RemoveChild )

            , ( bp::arg("szName") ) )    

        .def( 

            "SetName"

            , (bool ( ::PythonDTPNodeHandle::* )( char const * ) )(
&::PythonDTPNodeHandle::SetName )

            , ( bp::arg("szNewName") ) )    

        .def( 

            "SetValue"

            , (bool ( ::PythonDTPNodeHandle::* )( char const * ) )(
&::PythonDTPNodeHandle::SetValue )

            , ( bp::arg("szNewValue") ) );

 

    { //::cdc::ArrayImpl< PythonDTPNodeHandle* >

        typedef bp::class_< cdc::ArrayImpl< PythonDTPNodeHandle* > >
super_exposer_t;

        super_exposer_t super_exposer = super_exposer_t( "super",
bp::init< cdc::MemTag >(( bp::arg("tag") )) );

        bp::scope super_scope( super_exposer );

        bp::implicitly_convertible< cdc::MemTag, cdc::ArrayImpl<
PythonDTPNodeHandle* > >();

        super_exposer.def( bp::init< cdc::ArrayImpl<
PythonDTPNodeHandle* > const & >(( bp::arg("x") )) );

        { //::cdc::ArrayImpl< PythonDTPNodeHandle* >::capacity

        

            typedef cdc::ArrayImpl< PythonDTPNodeHandle* >
exported_class_t;

            typedef ::cdc::uint32 (
exported_class_t::*capacity_function_type )(  ) const;

            

            super_exposer.def( 

                "capacity"

                , capacity_function_type( &::cdc::ArrayImpl<
PythonDTPNodeHandle* >::capacity ) );

        

        }

        { //::cdc::ArrayImpl< PythonDTPNodeHandle* >::clear

        

            typedef cdc::ArrayImpl< PythonDTPNodeHandle* >
exported_class_t;

            typedef void ( exported_class_t::*clear_function_type )(  )
;

            

            super_exposer.def( 

                "clear"

                , clear_function_type( &::cdc::ArrayImpl<
PythonDTPNodeHandle* >::clear ) );

        

        }

        { //::cdc::ArrayImpl< PythonDTPNodeHandle* >::empty

        

            typedef cdc::ArrayImpl< PythonDTPNodeHandle* >
exported_class_t;

            typedef bool ( exported_class_t::*empty_function_type )(  )
const;

            

            super_exposer.def( 

                "empty"

                , empty_function_type( &::cdc::ArrayImpl<
PythonDTPNodeHandle* >::empty ) );

        

        }

        { //::cdc::ArrayImpl< PythonDTPNodeHandle* >::operator=

        

            typedef cdc::ArrayImpl< PythonDTPNodeHandle* >
exported_class_t;

            typedef ::cdc::ArrayImpl< PythonDTPNodeHandle* > & (
exported_class_t::*assign_function_type )( ::cdc::ArrayImpl<
PythonDTPNodeHandle* > const & ) ;

            

            super_exposer.def( 

                "assign"

                , assign_function_type( &::cdc::ArrayImpl<
PythonDTPNodeHandle* >::operator= )

                , ( bp::arg("x") )

                , bp::return_self< >() );

        

        }

        { //::cdc::ArrayImpl< PythonDTPNodeHandle* >::operator[]

        

            typedef cdc::ArrayImpl< PythonDTPNodeHandle* >
exported_class_t;

            typedef ::PythonDTPNodeHandle * & (
exported_class_t::*__getitem___function_type )( ::cdc::uint32 ) ;

            

            super_exposer.def( 

                "__getitem__"

                , __getitem___function_type( &::cdc::ArrayImpl<
PythonDTPNodeHandle* >::operator[] )

                , ( bp::arg("i") )

                , bp::return_internal_reference< >() );

        

        }

        { //::cdc::ArrayImpl< PythonDTPNodeHandle* >::operator[]

        

            typedef cdc::ArrayImpl< PythonDTPNodeHandle* >
exported_class_t;

            typedef ::PythonDTPNodeHandle * const & (
exported_class_t::*__getitem___function_type )( ::cdc::uint32 ) const;

            

            super_exposer.def( 

                "__getitem__"

                , __getitem___function_type( &::cdc::ArrayImpl<
PythonDTPNodeHandle* >::operator[] )

                , ( bp::arg("i") )

                , bp::return_value_policy< bp::copy_const_reference >()
);

        

        }

        { //::cdc::ArrayImpl< PythonDTPNodeHandle* >::pop_back

        

            typedef cdc::ArrayImpl< PythonDTPNodeHandle* >
exported_class_t;

            typedef void ( exported_class_t::*pop_back_function_type )(
) ;

            

            super_exposer.def( 

                "pop_back"

                , pop_back_function_type( &::cdc::ArrayImpl<
PythonDTPNodeHandle* >::pop_back ) );

        

        }

        { //::cdc::ArrayImpl< PythonDTPNodeHandle* >::push_back

        

            typedef cdc::ArrayImpl< PythonDTPNodeHandle* >
exported_class_t;

            typedef void ( exported_class_t::*push_back_function_type )(
::PythonDTPNodeHandle * const & ) ;

            

            super_exposer.def( 

                "push_back"

                , push_back_function_type( &::cdc::ArrayImpl<
PythonDTPNodeHandle* >::push_back )

                , ( bp::arg("val") ) );

        

        }

        { //::cdc::ArrayImpl< PythonDTPNodeHandle* >::reserve

        

            typedef cdc::ArrayImpl< PythonDTPNodeHandle* >
exported_class_t;

            typedef void ( exported_class_t::*reserve_function_type )(
::cdc::uint32 ) ;

            

            super_exposer.def( 

                "reserve"

                , reserve_function_type( &::cdc::ArrayImpl<
PythonDTPNodeHandle* >::reserve )

                , ( bp::arg("n") ) );

        

        }

        { //::cdc::ArrayImpl< PythonDTPNodeHandle* >::shrink_to_size

        

            typedef cdc::ArrayImpl< PythonDTPNodeHandle* >
exported_class_t;

            typedef void (
exported_class_t::*shrink_to_size_function_type )(  ) ;

            

            super_exposer.def( 

                "shrink_to_size"

                , shrink_to_size_function_type( &::cdc::ArrayImpl<
PythonDTPNodeHandle* >::shrink_to_size ) );

        

        }

        { //::cdc::ArrayImpl< PythonDTPNodeHandle* >::size

        

            typedef cdc::ArrayImpl< PythonDTPNodeHandle* >
exported_class_t;

            typedef ::cdc::uint32 (
exported_class_t::*size_function_type )(  ) const;

            

            super_exposer.def( 

                "size"

                , size_function_type( &::cdc::ArrayImpl<
PythonDTPNodeHandle* >::size ) );

        

        }

        { //::cdc::ArrayImpl< PythonDTPNodeHandle* >::swap

        

            typedef cdc::ArrayImpl< PythonDTPNodeHandle* >
exported_class_t;

            typedef void ( exported_class_t::*swap_function_type )(
::cdc::ArrayImpl< PythonDTPNodeHandle* > & ) ;

            

            super_exposer.def( 

                "swap"

                , swap_function_type( &::cdc::ArrayImpl<
PythonDTPNodeHandle* >::swap )

                , ( bp::arg("x") ) );

        

        }

    }

 

    { //::cdc::Array< PythonDTPNodeHandle* >

        typedef bp::class_< cdc::Array< PythonDTPNodeHandle* >,
bp::bases< cdc::ArrayImpl< PythonDTPNodeHandle* > > >
cdcArray_exposer_t;

        cdcArray_exposer_t cdcArray_exposer = cdcArray_exposer_t(
"cdcArray", bp::init< cdc::MemTag >(( bp::arg("tag") )) );

        bp::scope cdcArray_scope( cdcArray_exposer );

        bp::implicitly_convertible< cdc::MemTag, cdc::Array<
PythonDTPNodeHandle* > >();

    }

}

 


______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email 
______________________________________________________________________
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20090514/fbea0cc2/attachment-0001.htm>

From walter at mathematik.hu-berlin.de  Fri May 15 11:02:23 2009
From: walter at mathematik.hu-berlin.de (Sebastian Walter)
Date: Fri, 15 May 2009 11:02:23 +0200
Subject: [C++-sig] operator+(A,
 B) without B::B() default constructor in boost.python
In-Reply-To: <130401.91415.qm@web24608.mail.ird.yahoo.com>
References: <130401.91415.qm@web24608.mail.ird.yahoo.com>
Message-ID: <4A0D2F9F.4000208@mathematik.hu-berlin.de>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hans Roessler schrieb:

Hello Hans,

in my experience .def(self+other) often didn't quite do what I expected.
You could try something like that:

int int__add__A_B(const A &lhs, const B &rhs){
     return operator+(A,B);
}

.def("__add__", &int__add__A_B)


i.e. first define a non-overloaded function int_add__A_B and then bind it to __add__ (which is the Python operator+ function).

regards,
Sebastian


> Hello,
> how can I wrap the following overloaded operator+ using boost.python?
> The code as is fails with 
>> ...
>> pytest.cpp: In function ?void init_module_pytest()?:
>> pytest.cpp:19: error: expected primary-expression before ?(? token
>> ...
> If I try ".def(self+B())" instead of ".def(self+other<B>())", the error is
>> ...
>> pytest.cpp: In function ?void init_module_pytest()?:
>> pytest.cpp:17: error: no matching function for call to ?B::B()?
>> ...
> 
> best wishes
> Hans
> 
> 
> pytest.cpp:
> 
> #include <boost/python.hpp>
> using namespace boost::python;
> 
> class A{
> public:
> };
> class B{
> public:
>   B(int i){};
> };
> 
> int operator+(const A& a,const B& b){return 42;};
> 
> BOOST_PYTHON_MODULE(pytest)
> {
>   class_<A>("A")
>     .def(self+other<B>())   //.def(self+B())            //line 17
>     ;
>   class_<B,boost::noncopyable>("B",init<int>)    //line 19
>     ;
> }
> 
> 
>       
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.4-svn0 (GNU/Linux)
Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org

iD8DBQFKDS+f9PBA5IG0h0ARAnfLAJ9rGkoAUY9c2FUgJTIcWx+AFN1akACdHg7s
MAVRLstiqwAtaf9RmDsLXko=
=b8b/
-----END PGP SIGNATURE-----

From sipickles at googlemail.com  Fri May 15 17:40:10 2009
From: sipickles at googlemail.com (Simon Pickles)
Date: Fri, 15 May 2009 16:40:10 +0100
Subject: [C++-sig] Trouble reinstalling boost
Message-ID: <4A0D8CDA.7080606@googlemail.com>

Hi,

I had a nice build system set up on my VM ubuntu 8.10 server until I 
broke the VM. :) After rebuild the VM I am obviously reinstalling all 
the packages.

Boost Python is causing me trouble. I built it from svn using bjam, 
boost 1_40_0. I can compile my python extension without error but when I 
run it the python code that uses it, I get:

ImportError: libboost_python-gcc43-d-1_40.so.1.40.0: cannot open shared 
object file: No such file or directory

I have this file on my system, and have copied it to /usr/local/lib

simon at fenixurth:~/WorkingCopies/server/extended$ locate 
libboost_python-gcc43-d-1_40.so.1.40.0
/home/simon/source/boost/bin.v2/libs/python/build/gcc-4.3.2/debug/libboost_python-gcc43-d-1_40.so.1.40.0
/home/simon/source/boost/stage/lib/libboost_python-gcc43-d-1_40.so.1.40.0
/home/simon/WorkingCopies/server/extended/libboost_python-gcc43-d-1_40.so.1.40.0
/usr/local/lib/libboost_python-gcc43-d-1_40.so.1.40.0


Why is this not found?

Thanks

Simon


From vdeleo at dsf.unica.it  Fri May 15 22:05:38 2009
From: vdeleo at dsf.unica.it (Vincenzo De Leo)
Date: Fri, 15 May 2009 22:05:38 +0200
Subject: [C++-sig] Trouble reinstalling boost
In-Reply-To: <4A0D8CDA.7080606@googlemail.com>
References: <4A0D8CDA.7080606@googlemail.com>
Message-ID: <b19b11170905151305kb9fc205q6fe8c5c66fb27a42@mail.gmail.com>

Hi,maybe you have to do "ldconfig" before using the shared library.

Vincenzo

On Fri, May 15, 2009 at 5:40 PM, Simon Pickles <sipickles at googlemail.com>wrote:

> Hi,
>
> I had a nice build system set up on my VM ubuntu 8.10 server until I broke
> the VM. :) After rebuild the VM I am obviously reinstalling all the
> packages.
>
> Boost Python is causing me trouble. I built it from svn using bjam, boost
> 1_40_0. I can compile my python extension without error but when I run it
> the python code that uses it, I get:
>
> ImportError: libboost_python-gcc43-d-1_40.so.1.40.0: cannot open shared
> object file: No such file or directory
>
> I have this file on my system, and have copied it to /usr/local/lib
>
> simon at fenixurth:~/WorkingCopies/server/extended$ locate
> libboost_python-gcc43-d-1_40.so.1.40.0
>
> /home/simon/source/boost/bin.v2/libs/python/build/gcc-4.3.2/debug/libboost_python-gcc43-d-1_40.so.1.40.0
> /home/simon/source/boost/stage/lib/libboost_python-gcc43-d-1_40.so.1.40.0
>
> /home/simon/WorkingCopies/server/extended/libboost_python-gcc43-d-1_40.so.1.40.0
> /usr/local/lib/libboost_python-gcc43-d-1_40.so.1.40.0
>
>
> Why is this not found?
>
> Thanks
>
> Simon
>
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20090515/3d5256a2/attachment.htm>

From lists_ravi at lavabit.com  Sat May 16 03:08:08 2009
From: lists_ravi at lavabit.com (Ravi)
Date: Fri, 15 May 2009 21:08:08 -0400
Subject: [C++-sig] Question regarding importing libraries to weave
In-Reply-To: <20504a380905141141j4ac4793ej306603590deb21d@mail.gmail.com>
References: <20504a380905141141j4ac4793ej306603590deb21d@mail.gmail.com>
Message-ID: <200905152108.11885.lists_ravi@lavabit.com>

On Thursday 14 May 2009 2:41:02 pm Magn?s Bergur Magn?sson wrote:
> My name is Magnus and I am working as a software developer for a company in
> Iceland. We use weave inline with python to boost our processing capacity.
> However there is one factor that is limiting us. We cannot figure out how
> to import libraries to Weave such as just <math.h> or <string.h>, what is
> the right procedure for that? Also, what is the most effective tool for
> text processing in python? We have been using weave inline and moving char
> by char but is there a more effective way and could string.h speed things
> up?

This is not a boost.python question; I doubt we have very many users of weave 
here. You may have better luck asking on the numpy lists (numpy-
discussion at scipy.org).

Regards,
Ravi



From hansroessler at yahoo.de  Sat May 16 11:48:36 2009
From: hansroessler at yahoo.de (Hans Roessler)
Date: Sat, 16 May 2009 02:48:36 -0700 (PDT)
Subject: [C++-sig] operator+(A,
	B) without B::B() default constructor in boost.python
In-Reply-To: <4A0D2F9F.4000208@mathematik.hu-berlin.de>
References: <130401.91415.qm@web24608.mail.ird.yahoo.com>
	<4A0D2F9F.4000208@mathematik.hu-berlin.de>
Message-ID: <467728.69018.qm@web24610.mail.ird.yahoo.com>


Thank you for suggesting the workaround. But I would like to understand _why_ the operator overloading doesn't work the Boost.Python way in my example (or what I have done wrong).
It can be quite tedious to write a wrapper like __add__ for each operator and pair of classes, and I thought the ".def(self <op> other)" syntax was implemented for exactly this purpose?

Hans



----- Urspr?ngliche Mail ----
> Von: Sebastian Walter <walter at mathematik.hu-berlin.de>
> An: Development of Python/C++ integration <cplusplus-sig at python.org>
> Gesendet: Freitag, den 15. Mai 2009, 11:02:23 Uhr
> Betreff: Re: [C++-sig] operator+(A, B) without B::B() default constructor in boost.python
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hans Roessler schrieb:
> 
> Hello Hans,
> 
> in my experience .def(self+other) often didn't quite do what I expected.
> You could try something like that:
> 
> int int__add__A_B(const A &lhs, const B &rhs){
>      return operator+(A,B);
> }
> 
> .def("__add__", &int__add__A_B)
> 
> 
> i.e. first define a non-overloaded function int_add__A_B and then bind it to 
> __add__ (which is the Python operator+ function).
> 
> regards,
> Sebastian
> 
> 
> > Hello,
> > how can I wrap the following overloaded operator+ using boost.python?
> > The code as is fails with 
> >> ...
> >> pytest.cpp: In function ?void init_module_pytest()?:
> >> pytest.cpp:19: error: expected primary-expression before ?(? token
> >> ...
> > If I try ".def(self+B())" instead of ".def(self+other())", the error is
> >> ...
> >> pytest.cpp: In function ?void init_module_pytest()?:
> >> pytest.cpp:17: error: no matching function for call to ?B::B()?
> >> ...
> > 
> > best wishes
> > Hans
> > 
> > 
> > pytest.cpp:
> > 
> > #include 
> > using namespace boost::python;
> > 
> > class A{
> > public:
> > };
> > class B{
> > public:
> >   B(int i){};
> > };
> > 
> > int operator+(const A& a,const B& b){return 42;};
> > 
> > BOOST_PYTHON_MODULE(pytest)
> > {
> >   class_("A")
> >     .def(self+other())   //.def(self+B())            //line 17
> >     ;
> >   class_("B",init)    //line 19
> >     ;
> > }
> > 
> > 
> >      
> > _______________________________________________
> > Cplusplus-sig mailing list
> > Cplusplus-sig at python.org
> > http://mail.python.org/mailman/listinfo/cplusplus-sig
> 
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v2.0.4-svn0 (GNU/Linux)
> Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org
> 
> iD8DBQFKDS+f9PBA5IG0h0ARAnfLAJ9rGkoAUY9c2FUgJTIcWx+AFN1akACdHg7s
> MAVRLstiqwAtaf9RmDsLXko=
> =b8b/
> -----END PGP SIGNATURE-----
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> http://mail.python.org/mailman/listinfo/cplusplus-sig



      

From roman.yakovenko at gmail.com  Sun May 17 07:32:59 2009
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Sun, 17 May 2009 08:32:59 +0300
Subject: [C++-sig] Trouble with custom container class
In-Reply-To: <2DEFEB25B65B1D4C962D79E3CCBAFC7F06E883FF@mpkexc01.eidos.com>
References: <2DEFEB25B65B1D4C962D79E3CCBAFC7F06E883FF@mpkexc01.eidos.com>
Message-ID: <7465b6170905162232g73f616eds74a2d25bc8b87e72@mail.gmail.com>

On Fri, May 15, 2009 at 12:01 AM, Philip Bloom <pbloom at crystald.com> wrote:
> Hello,

Hello.

> I?m having some issues with retrieving objects from a custom container
> class.? The objects in the container are otherwise wrapped objects
> (PythonDTPNodeHandle *) that the python wrapper knows how to convert, but
> what I get back from the container with __getitem__ seems to not know how to
> convert it.? ??I suspect that I?m doing something wrong on what I?m feeding
> Py++ and thus the generated wrapping is bad, but I don?t see anything wrong
> with the final wrapping for getitem or operator[].? I?ve no idea why it
> wouldn?t be able to convert, for this specific usage.? Any advise/help would
> be well welcome.? I?ve included the whole wrapper at the bottom.

Too much information. Can you create stand-alone, small and complete
test case, that shows your problem?

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

From walter at mathematik.hu-berlin.de  Mon May 18 10:17:24 2009
From: walter at mathematik.hu-berlin.de (Sebastian Walter)
Date: Mon, 18 May 2009 10:17:24 +0200
Subject: [C++-sig] operator+(A,
 B) without B::B() default constructor in boost.python
In-Reply-To: <467728.69018.qm@web24610.mail.ird.yahoo.com>
References: <130401.91415.qm@web24608.mail.ird.yahoo.com>	<4A0D2F9F.4000208@mathematik.hu-berlin.de>
	<467728.69018.qm@web24610.mail.ird.yahoo.com>
Message-ID: <4A111994.9010508@mathematik.hu-berlin.de>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

That, you will have to ask the devs.
The question is however:
what is quicker: wrapping 20 functions manually with copy/paste or trying to understand boost::python resp. wait for an answer from the devs.
It is my impression, that you have to know a lot about boost::python to understand why something does not work as it should.

If someone knows the answer, I'd be also be interested, of course ;)




Hans Roessler schrieb:
> Thank you for suggesting the workaround. But I would like to understand _why_ the operator overloading doesn't work the Boost.Python way in my example (or what I have done wrong).
> It can be quite tedious to write a wrapper like __add__ for each operator and pair of classes, and I thought the ".def(self <op> other)" syntax was implemented for exactly this purpose?
> 
> Hans
> 
> 
> 
> ----- Urspr?ngliche Mail ----
>> Von: Sebastian Walter <walter at mathematik.hu-berlin.de>
>> An: Development of Python/C++ integration <cplusplus-sig at python.org>
>> Gesendet: Freitag, den 15. Mai 2009, 11:02:23 Uhr
>> Betreff: Re: [C++-sig] operator+(A, B) without B::B() default constructor in boost.python
>>
> Hans Roessler schrieb:
> 
> Hello Hans,
> 
> in my experience .def(self+other) often didn't quite do what I expected.
> You could try something like that:
> 
> int int__add__A_B(const A &lhs, const B &rhs){
>      return operator+(A,B);
> }
> 
> .def("__add__", &int__add__A_B)
> 
> 
> i.e. first define a non-overloaded function int_add__A_B and then bind it to 
> __add__ (which is the Python operator+ function).
> 
> regards,
> Sebastian
> 
> 
>>>> Hello,
>>>> how can I wrap the following overloaded operator+ using boost.python?
>>>> The code as is fails with 
>>>>> ...
>>>>> pytest.cpp: In function void init_module_pytest():
>>>>> pytest.cpp:19: error: expected primary-expression before ( token
>>>>> ...
>>>> If I try ".def(self+B())" instead of ".def(self+other())", the error is
>>>>> ...
>>>>> pytest.cpp: In function void init_module_pytest():
>>>>> pytest.cpp:17: error: no matching function for call to B::B()
>>>>> ...
>>>> best wishes
>>>> Hans
>>>>
>>>>
>>>> pytest.cpp:
>>>>
>>>> #include 
>>>> using namespace boost::python;
>>>>
>>>> class A{
>>>> public:
>>>> };
>>>> class B{
>>>> public:
>>>>   B(int i){};
>>>> };
>>>>
>>>> int operator+(const A& a,const B& b){return 42;};
>>>>
>>>> BOOST_PYTHON_MODULE(pytest)
>>>> {
>>>>   class_("A")
>>>>     .def(self+other())   //.def(self+B())            //line 17
>>>>     ;
>>>>   class_("B",init)    //line 19
>>>>     ;
>>>> }
>>>>
>>>>
>>>>      
>>>> _______________________________________________
>>>> 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

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

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.4-svn0 (GNU/Linux)
Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org

iD8DBQFKERmU9PBA5IG0h0ARAgFMAKCDx+DGod3m8c+nE4auuR0+dZjtIQCeIXHf
S6X++I+Ilw74MqtKzvYWiCo=
=dcH2
-----END PGP SIGNATURE-----

From roman.yakovenko at gmail.com  Tue May 19 12:33:58 2009
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Tue, 19 May 2009 13:33:58 +0300
Subject: [C++-sig] pygccxml and multidimensional arrays
In-Reply-To: <7465b6170904231123x7c7fbb73j9b95aee05786c757@mail.gmail.com>
References: <Acm9P9dc6WsxloGaTu+5DirkER6/7Q==@lmco.com>
	<5A2986C73A1DCC4EB5CC1C22D82700090FD12E26@emss02m11.us.lmco.com>
	<7465b6170904152204q5e0381cdmb6643e3bc5de918b@mail.gmail.com>
	<5A2986C73A1DCC4EB5CC1C22D82700090FDAB895@emss02m11.us.lmco.com>
	<5A2986C73A1DCC4EB5CC1C22D82700090FF26840@emss02m11.us.lmco.com>
	<7465b6170904231011xf63c219ge7f398e70ff1bfbd@mail.gmail.com>
	<5A2986C73A1DCC4EB5CC1C22D82700090FF61E05@emss02m11.us.lmco.com>
	<7465b6170904231121o17cee0acm47d847328c3d1dce@mail.gmail.com>
	<7465b6170904231123x7c7fbb73j9b95aee05786c757@mail.gmail.com>
Message-ID: <7465b6170905190333p6c1262b4hd63899b902a81179@mail.gmail.com>

On Thu, Apr 23, 2009 at 9:23 PM, Roman Yakovenko
<roman.yakovenko at gmail.com> wrote:
> On Thu, Apr 23, 2009 at 9:21 PM, Roman Yakovenko
> <roman.yakovenko at gmail.com> wrote:
>> On Thu, Apr 23, 2009 at 8:17 PM, Davidson, Josh <josh.davidson at lmco.com> wrote:
>>> I'm using information provided by pygccxml to generate Construct objects
>>> (http://construct.wikispaces.com/) that are capable of reading and
>>> writing to data structures stored in shared memory. ?Right now, I just
>>> manually reverse the indices any time I detect an array type before I
>>> generate a Construct.
>>
>> Okey.
>
> Sorry, I forgot to add, can you open bug on SourceForge project page?

Josh, I think the bug is fixed now:
http://pygccxml.svn.sourceforge.net/viewvc/pygccxml?view=rev&revision=1735

Can you verify this?

Thanks

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

From james at goldwater.org.uk  Tue May 19 14:40:06 2009
From: james at goldwater.org.uk (james at goldwater.org.uk)
Date: Tue, 19 May 2009 13:40:06 +0100
Subject: [C++-sig] py++ and indexing_v2
Message-ID: <20090519124005.GN25013@eccehomo.vm.bytemark.co.uk>

When I use py++ (1.0.0) with indexing_suite_version = 2 to wrap std::set<string> etc, it generates code which doesn't compile out of the box.  The code from the boost sandbox puts various definitions in the boost::python::indexing_v2 namespace, and the header files in boost/python/suite/indexing_v2.

However, py++ generates code that assumes that the sandbox code is in the boost trunk namespace (bp::indexing) and filestructure (boost/python/suite/indexing).  It's easy to either fix-up the generated code, or edit the sandbox code and merge it into trunk, but I'd rather keep to published code rather than commit the project to our own version of boost or add another step to the code generation.

Is there a way of telling py++ to use the different namespace and directory?  Or am I misunderstanding some things?

Thanks for any help

James.

-- 
Ring the bells that still can ring...
Forget your perfect offering.
There is a crack in everything
That's how the light gets in.

From wladwig at wdtinc.com  Tue May 19 17:34:49 2009
From: wladwig at wdtinc.com (William Ladwig)
Date: Tue, 19 May 2009 10:34:49 -0500
Subject: [C++-sig] operator+(A,
 B) without B::B() default constructor in boost.python
In-Reply-To: <130401.91415.qm@web24608.mail.ird.yahoo.com>
References: <130401.91415.qm@web24608.mail.ird.yahoo.com>
Message-ID: <765CBD9053EA2B438625895F30D1856F0226436168@storm.wdtinc.com>

It looks to me like you're missing a '()' next to your 'init<>' in the class wrapper for B.

Try changing this:

  class_<B,boost::noncopyable>("B",init<int>)    //line 19
    ;

to this:

  class_<B,boost::noncopyable>("B",init<int>())    //line 19
    ;

The only operator that has given me problems is the << operator.  I've had pretty good success with the others.

Regards,
Bill

-----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 Hans Roessler
Sent: Thursday, May 14, 2009 8:16 AM
To: cplusplus-sig at python.org
Subject: [C++-sig] operator+(A, B) without B::B() default constructor in boost.python


Hello,
how can I wrap the following overloaded operator+ using boost.python?
The code as is fails with 
>...
>pytest.cpp: In function 'void init_module_pytest()':
>pytest.cpp:19: error: expected primary-expression before '(' token
>...
If I try ".def(self+B())" instead of ".def(self+other<B>())", the error is
>...
>pytest.cpp: In function 'void init_module_pytest()':
>pytest.cpp:17: error: no matching function for call to 'B::B()'
>...

best wishes
Hans


pytest.cpp:

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

class A{
public:
};
class B{
public:
  B(int i){};
};

int operator+(const A& a,const B& b){return 42;};

BOOST_PYTHON_MODULE(pytest)
{
  class_<A>("A")
    .def(self+other<B>())   //.def(self+B())            //line 17
    ;
  class_<B,boost::noncopyable>("B",init<int>)    //line 19
    ;
}


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

From roman.yakovenko at gmail.com  Tue May 19 20:24:27 2009
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Tue, 19 May 2009 21:24:27 +0300
Subject: [C++-sig] py++ and indexing_v2
In-Reply-To: <20090519124005.GN25013@eccehomo.vm.bytemark.co.uk>
References: <20090519124005.GN25013@eccehomo.vm.bytemark.co.uk>
Message-ID: <7465b6170905191124l54b4d151h72a4a5e6b869e315@mail.gmail.com>

On Tue, May 19, 2009 at 3:40 PM,  <james at goldwater.org.uk> wrote:
> When I use py++ (1.0.0) with indexing_suite_version = 2 to wrap std::set<string> etc, it generates code which doesn't compile out of the box. ?The code from the boost sandbox puts various definitions in the boost::python::indexing_v2 namespace, and the header files in boost/python/suite/indexing_v2.
>
> However, py++ generates code that assumes that the sandbox code is in the boost trunk namespace (bp::indexing) and filestructure (boost/python/suite/indexing). ?It's easy to either fix-up the generated code, or edit the sandbox code and merge it into trunk, but I'd rather keep to published code rather than commit the project to our own version of boost or add another step to the code generation.
>
> Is there a way of telling py++ to use the different namespace and directory?

Not exactly. I suggest you to use current SVN version. It is pretty
stable. I hope to release it pretty soon.
The main reason is describe here:
http://language-binding.net/pyplusplus/documentation/containers.html#indexing-suite-version-2-installation

The short version: Indexing Suite V2 was reworked and now you don't
need to patch Boost libraries and Py++ does all the work.

> Or am I misunderstanding some things?

No, the Py++ you use, assumes that you installed indexing suite V2 to
the location you specified.

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

From terencebbr at gmail.com  Wed May 20 05:15:47 2009
From: terencebbr at gmail.com (terence tan)
Date: Wed, 20 May 2009 11:15:47 +0800
Subject: [C++-sig] XXX undefined error throw by python,
	Exception caused by get_override? Safe to ignor or cause for
	concern?
Message-ID: <85c908b70905192015k5321a120q44e283b72f801c2c@mail.gmail.com>

I am implementing a interface between my C++ class and a python class that
requires the ability for the python class to have a Initialize/Update method
called  from C++ but still call the default methods of the original C++
class I am integrating. If no such methods are defined in the python class
then the default C++ methods are called. So I approach it by writing a
wrapper.

Right now python is throw an exception XXX undefined error saying that the
Initialize(or Update method) is undefined. I believe it is the get_override
method. Am I approaching this the wrong way? The Initialize and Update
function are virtual.

struct BaseGameWrap : BaseGame, wrapper<BaseGame>
{
  void Initialize()
  {
    if (override Initialize = this->get_override("Initialize"))
    {
      BaseGame::Initialize();
#ifdef WIN32
      return call<void>(Initialize.ptr());
#else
      return Initialize();
#endif
}

     // This supresses a XXX undefined error thrown by
     // by get_override when Initialize is not found. Unless a better way
     // is found to handle this.
     PyErr_Clear();
     return BaseGame::Initialize();
}

};

void exportBaseGame()
{
class_<BaseGameWrap,boost::noncopyable,bases<Game>>("BaseGame")
.def("Run",&BaseGame::Run)
.def("GetComponents",&BaseGame::GetComponents,return_value_policy<reference_existing_object>())
.def("Exit",&BaseGame::Exit);
;
}


Terence
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20090520/bac13736/attachment.htm>

From hansroessler at yahoo.de  Thu May 21 15:39:58 2009
From: hansroessler at yahoo.de (Hans Roessler)
Date: Thu, 21 May 2009 06:39:58 -0700 (PDT)
Subject: [C++-sig] variable argument number and
	BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS
Message-ID: <759109.42277.qm@web24605.mail.ird.yahoo.com>


Hello,
I try to wrap a class with a function info([int arg]) that takes either zero or one argument (exact colde below). For some reason I can't get BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS to work.

The examples at http://www.boost.org/doc/libs/1_39_0/libs/python/doc/tutorial/doc/html/python/functions.html#python.overloading and
http://www.boost.org/doc/libs/1_39_0/libs/python/doc/v2/overloads.html#BOOST_PYTHON_FUNCTION_OVERLOADS-spec only show the situation where either
- the *member* function has up to three arguments with default values assigned or
- three *free* functions have one to three  arguments.

I don't find an example where a class has several *member* functions with the same name, that take a variable number of (non-default!) arguments.

My attempt below gives the error
"pytest.cpp:17: error: no matching function for call to ?boost::python::class_<A, boost::python::detail::not_specified, boost::python::detail::not_specified, boost::python::detail::not_specified>::def(const char [5], <unresolved overloaded function type>, A_info_overloads)?"

Do I have to replace the ".def("info",&A::info, ..." with something like  ".def("foo", (void(*)(int,int))0, ..." as shown in the Auto-Overloading section in the tutorial? If yes, how must this function signature look like? It doesn't work like this.

Hope someone can help
Hans

=== pytest.py ===
#include <boost/python.hpp>
#include <stdio.h>

using namespace boost::python;

class A{
public:
  void info() {printf("This is an A\n");}
  void info(int arg) {printf("Unnecessary arg.\n");}
};

BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS(A_info_overloads, info, 0,1)

BOOST_PYTHON_MODULE(test)
{
  class_<A>("A")
      .def("info",&A::info,A_info_overloads()) //line 17
    ;
}



      

From wladwig at wdtinc.com  Thu May 21 18:30:46 2009
From: wladwig at wdtinc.com (William Ladwig)
Date: Thu, 21 May 2009 11:30:46 -0500
Subject: [C++-sig] variable argument number
	and	BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS
In-Reply-To: <759109.42277.qm@web24605.mail.ird.yahoo.com>
References: <759109.42277.qm@web24605.mail.ird.yahoo.com>
Message-ID: <765CBD9053EA2B438625895F30D1856F022643655E@storm.wdtinc.com>

I haven't tried this myself, but I think all you need to do to wrap the member function is this:

.def("info", (void(A::*)(int))0, A_info_overloads());

Hope this helps,
Bill

-----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 Hans Roessler
Sent: Thursday, May 21, 2009 8:40 AM
To: cplusplus-sig at python.org
Subject: [C++-sig] variable argument number and BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS


Hello,
I try to wrap a class with a function info([int arg]) that takes either zero or one argument (exact colde below). For some reason I can't get BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS to work.

The examples at http://www.boost.org/doc/libs/1_39_0/libs/python/doc/tutorial/doc/html/python/functions.html#python.overloading and
http://www.boost.org/doc/libs/1_39_0/libs/python/doc/v2/overloads.html#BOOST_PYTHON_FUNCTION_OVERLOADS-spec only show the situation where either
- the *member* function has up to three arguments with default values assigned or
- three *free* functions have one to three  arguments.

I don't find an example where a class has several *member* functions with the same name, that take a variable number of (non-default!) arguments.

My attempt below gives the error
"pytest.cpp:17: error: no matching function for call to 'boost::python::class_<A, boost::python::detail::not_specified, boost::python::detail::not_specified, boost::python::detail::not_specified>::def(const char [5], <unresolved overloaded function type>, A_info_overloads)'"

Do I have to replace the ".def("info",&A::info, ..." with something like  ".def("foo", (void(*)(int,int))0, ..." as shown in the Auto-Overloading section in the tutorial? If yes, how must this function signature look like? It doesn't work like this.

Hope someone can help
Hans

=== pytest.py ===
#include <boost/python.hpp>
#include <stdio.h>

using namespace boost::python;

class A{
public:
  void info() {printf("This is an A\n");}
  void info(int arg) {printf("Unnecessary arg.\n");}
};

BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS(A_info_overloads, info, 0,1)

BOOST_PYTHON_MODULE(test)
{
  class_<A>("A")
      .def("info",&A::info,A_info_overloads()) //line 17
    ;
}



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

From Matthew.Scouten at tradingtechnologies.com  Thu May 21 18:45:58 2009
From: Matthew.Scouten at tradingtechnologies.com (Matthew Scouten (TT))
Date: Thu, 21 May 2009 11:45:58 -0500
Subject: [C++-sig] variable argument
	numberand	BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS
In-Reply-To: <765CBD9053EA2B438625895F30D1856F022643655E@storm.wdtinc.com>
References: <759109.42277.qm@web24605.mail.ird.yahoo.com>
	<765CBD9053EA2B438625895F30D1856F022643655E@storm.wdtinc.com>
Message-ID: <32490DFF7774554A85D65D23A9F0F9380AC0994B@chiex01>

Well I think that's probably the 'proper' way to do things, but I
usually use the stupid-but-obvious ways.

Overloaded functions:
Class A
{
	void info_1(int i){//Do Stuff}
	void info_2(){//Do Stuff}
}

class_<A>("A")
      .def("info",&A::info_1)
	.def("info",&A::info_2)

Boost::python will sort out which to call. If bp can't correctly
distinguish the types, there may be surprises. (bool vs. int) 

Functions with defaults:
Class A
{
	void info(int i = 0){//Do Stuff}
}

class_<A>("A")
      .def("info",
            &A::info, 
            (arg('i') = 0) 
          )






-----Original Message-----
From:
cplusplus-sig-bounces+matthew.scouten=tradingtechnologies.com at python.org
[mailto:cplusplus-sig-bounces+matthew.scouten=tradingtechnologies.com at py
thon.org] On Behalf Of William Ladwig
Sent: Thursday, May 21, 2009 11:31 AM
To: Development of Python/C++ integration
Subject: Re: [C++-sig] variable argument numberand
BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS

I haven't tried this myself, but I think all you need to do to wrap the
member function is this:

.def("info", (void(A::*)(int))0, A_info_overloads());

Hope this helps,
Bill

-----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 Hans Roessler
Sent: Thursday, May 21, 2009 8:40 AM
To: cplusplus-sig at python.org
Subject: [C++-sig] variable argument number and
BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS


Hello,
I try to wrap a class with a function info([int arg]) that takes either
zero or one argument (exact colde below). For some reason I can't get
BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS to work.

The examples at
http://www.boost.org/doc/libs/1_39_0/libs/python/doc/tutorial/doc/html/p
ython/functions.html#python.overloading and
http://www.boost.org/doc/libs/1_39_0/libs/python/doc/v2/overloads.html#B
OOST_PYTHON_FUNCTION_OVERLOADS-spec only show the situation where either
- the *member* function has up to three arguments with default values
assigned or
- three *free* functions have one to three  arguments.

I don't find an example where a class has several *member* functions
with the same name, that take a variable number of (non-default!)
arguments.

My attempt below gives the error
"pytest.cpp:17: error: no matching function for call to
'boost::python::class_<A, boost::python::detail::not_specified,
boost::python::detail::not_specified,
boost::python::detail::not_specified>::def(const char [5], <unresolved
overloaded function type>, A_info_overloads)'"

Do I have to replace the ".def("info",&A::info, ..." with something like
".def("foo", (void(*)(int,int))0, ..." as shown in the Auto-Overloading
section in the tutorial? If yes, how must this function signature look
like? It doesn't work like this.

Hope someone can help
Hans

=== pytest.py ===
#include <boost/python.hpp>
#include <stdio.h>

using namespace boost::python;

class A{
public:
  void info() {printf("This is an A\n");}
  void info(int arg) {printf("Unnecessary arg.\n");}
};

BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS(A_info_overloads, info, 0,1)

BOOST_PYTHON_MODULE(test)
{
  class_<A>("A")
      .def("info",&A::info,A_info_overloads()) //line 17
    ;
}



      
_______________________________________________
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

From hansroessler at yahoo.de  Thu May 21 23:00:25 2009
From: hansroessler at yahoo.de (Hans Roessler)
Date: Thu, 21 May 2009 14:00:25 -0700 (PDT)
Subject: [C++-sig] variable argument number and
	BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS
In-Reply-To: <765CBD9053EA2B438625895F30D1856F022643655E@storm.wdtinc.com>
References: <759109.42277.qm@web24605.mail.ird.yahoo.com>
	<765CBD9053EA2B438625895F30D1856F022643655E@storm.wdtinc.com>
Message-ID: <289047.3550.qm@web24606.mail.ird.yahoo.com>


Thanks, ".def("info", (void(A::*)(int))0, A_info_overloads());" does it.

Could somebody explain what the meaning of this "(void(A::*)(int))0" construct is? Is it a function pointer? Why the 0?
(Just now I am fine with the proposed solution, but if I even understand it, you will read fewer stupid question from me in this list ;-)


Hans



----- Urspr?ngliche Mail ----
> Von: William Ladwig <wladwig at wdtinc.com>
> An: Development of Python/C++ integration <cplusplus-sig at python.org>
> Gesendet: Donnerstag, den 21. Mai 2009, 18:30:46 Uhr
> Betreff: Re: [C++-sig] variable argument number and BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS
> 
> I haven't tried this myself, but I think all you need to do to wrap the member 
> function is this:
> 
> .def("info", (void(A::*)(int))0, A_info_overloads());
> 
> Hope this helps,
> Bill



      

From hansroessler at yahoo.de  Thu May 21 23:08:03 2009
From: hansroessler at yahoo.de (Hans Roessler)
Date: Thu, 21 May 2009 14:08:03 -0700 (PDT)
Subject: [C++-sig] iterate a boost::python::list
In-Reply-To: <765CBD9053EA2B438625895F30D1856F022643655E@storm.wdtinc.com>
References: <759109.42277.qm@web24605.mail.ird.yahoo.com>
	<765CBD9053EA2B438625895F30D1856F022643655E@storm.wdtinc.com>
Message-ID: <940734.2985.qm@web24607.mail.ird.yahoo.com>


How can I iterate in C++ over a boost::python::list?

According to http://www.boost.org/doc/libs/1_39_0/libs/python/doc/v2/list.html it has no begin() and end() functions.
Neither a size() function that would allow to call pop(size()-1). Neither an isEmpty() function that would allow to call pop(0) until the list is empty.

I'm sure I'm missing something trivial. Can somebody help me?

Thanks
Hans



      

From wladwig at wdtinc.com  Thu May 21 23:32:08 2009
From: wladwig at wdtinc.com (William Ladwig)
Date: Thu, 21 May 2009 16:32:08 -0500
Subject: [C++-sig] variable argument number
	and	BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS
In-Reply-To: <289047.3550.qm@web24606.mail.ird.yahoo.com>
References: <759109.42277.qm@web24605.mail.ird.yahoo.com>
	<765CBD9053EA2B438625895F30D1856F022643655E@storm.wdtinc.com>
	<289047.3550.qm@web24606.mail.ird.yahoo.com>
Message-ID: <765CBD9053EA2B438625895F30D1856F022643666D@storm.wdtinc.com>

It's making a NULL pointer to a member function of A with an integer argument and a void return type.  

I'm not sure what's going on under the hood with why the NULL is required, but I suspect that it has something to do with the fact that 

void info();
void info(int i);

require two function pointers, but 

void info(int i=0) 

can only use one function pointer.  I suspect that they're trying to handle both cases with the same interface.  I'm just guessing though, so I could be wrong about this.

Regards,
Bill


-----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 Hans Roessler
Sent: Thursday, May 21, 2009 4:00 PM
To: Development of Python/C++ integration
Subject: Re: [C++-sig] variable argument number and BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS


Thanks, ".def("info", (void(A::*)(int))0, A_info_overloads());" does it.

Could somebody explain what the meaning of this "(void(A::*)(int))0" construct is? Is it a function pointer? Why the 0?
(Just now I am fine with the proposed solution, but if I even understand it, you will read fewer stupid question from me in this list ;-)


Hans



----- Urspr?ngliche Mail ----
> Von: William Ladwig <wladwig at wdtinc.com>
> An: Development of Python/C++ integration <cplusplus-sig at python.org>
> Gesendet: Donnerstag, den 21. Mai 2009, 18:30:46 Uhr
> Betreff: Re: [C++-sig] variable argument number and BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS
> 
> I haven't tried this myself, but I think all you need to do to wrap the member 
> function is this:
> 
> .def("info", (void(A::*)(int))0, A_info_overloads());
> 
> Hope this helps,
> Bill



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

From rwgk at yahoo.com  Fri May 22 00:45:09 2009
From: rwgk at yahoo.com (Ralf W. Grosse-Kunstleve)
Date: Thu, 21 May 2009 15:45:09 -0700 (PDT)
Subject: [C++-sig] iterate a boost::python::list
In-Reply-To: <940734.2985.qm@web24607.mail.ird.yahoo.com>
References: <759109.42277.qm@web24605.mail.ird.yahoo.com>
	<765CBD9053EA2B438625895F30D1856F022643655E@storm.wdtinc.com>
	<940734.2985.qm@web24607.mail.ird.yahoo.com>
Message-ID: <998525.34120.qm@web111403.mail.gq1.yahoo.com>


It should work like this (untested):

  boost::python::ssize_t n = boost::python::len(your_list);
  for(boost::python::ssize_t i=0;i<n;i++) {
    boost::python::object elem = your_list[i];
  }



----- Original Message ----
From: Hans Roessler <hansroessler at yahoo.de>
To: Development of Python/C++ integration <cplusplus-sig at python.org>
Sent: Thursday, May 21, 2009 2:08:03 PM
Subject: [C++-sig] iterate a boost::python::list


How can I iterate in C++ over a boost::python::list?

According to http://www.boost.org/doc/libs/1_39_0/libs/python/doc/v2/list.html it has no begin() and end() functions.
Neither a size() function that would allow to call pop(size()-1). Neither an isEmpty() function that would allow to call pop(0) until the list is empty.

I'm sure I'm missing something trivial. Can somebody help me?

Thanks
Hans


From andrea.bedini at gmail.com  Fri May 22 11:50:56 2009
From: andrea.bedini at gmail.com (Andrea Bedini)
Date: Fri, 22 May 2009 11:50:56 +0200
Subject: [C++-sig] iterate a boost::python::list
In-Reply-To: <940734.2985.qm@web24607.mail.ird.yahoo.com>
References: <759109.42277.qm@web24605.mail.ird.yahoo.com>
	<765CBD9053EA2B438625895F30D1856F022643655E@storm.wdtinc.com> 
	<940734.2985.qm@web24607.mail.ird.yahoo.com>
Message-ID: <f3917ab0905220250n3c719815n4e9bd3302c016677@mail.gmail.com>

On Thu, May 21, 2009 at 11:08 PM, Hans Roessler <hansroessler at yahoo.de>wrote:

>
> How can I iterate in C++ over a boost::python::list?


you can use stl_input_iterator:

stl_input_iterator<T> begin(o), end;

gives you a pair of stl compatible iterators assuming you can extract type T
from objects obtained by iterating over o.

see
http://www.boost.org/doc/libs/1_39_0/libs/python/doc/v2/stl_iterator.html

-- 
Andrea Bedini <andrea.bedini at gmail.com>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20090522/ce452123/attachment.htm>

From zookog at gmail.com  Mon May 25 01:24:06 2009
From: zookog at gmail.com (Zooko O'Whielacronx)
Date: Sun, 24 May 2009 17:24:06 -0600
Subject: [C++-sig] RTTI across shared library boundaries
Message-ID: <cd6401a0905241624p480de314ma8205b82584f2fa7@mail.gmail.com>

Dear cplusplus-sig Folks:

I'm the maintainer of pycryptopp [1], a library whose main but not
sole user is Tahoe-LAFS [2].  I've recently stumbled across the
problem of RTTI crossing shared library boundaries, which seems to be
a well-known problem e.g. [3] but without, as far as I can tell, a
well-known solution.

Pycryptopp is mostly just Python wrappers for the Crypto++ library
[4].  The current status is the pycryptopp builds and passes all of
its unit tests [*, **] by building Python modules such as aes.so and
rsa.so from a combination of Crypto++ object files and pycryptopp
object files.  However, we're in the process of getting pycryptopp and
Tahoe-LAFS included in Debian and Fedora, and those two Linux
distributions have a policy that code which re-uses a separate library
has to dynamically link to the distribution-provided library instead
of bundling a copy of that library.  This is so that the distribution
maintainers can easily control the combination of libraries included
in their distribution -- for example if they want to upgrade Crypto++
or apply a patch to Crypto++ (such as a security patch), they need do
so only for the one copy of the shared library, and not for each
package which uses it.

So, I changed the pycryptopp setup.py so that if you pass the option
"--disable-embedded-cryptopp" to the "build" command it will stop
using its own internal copy of Crypto++ and instead simply link to
-lcryptopp.  Now the trouble starts.  An exception thrown by
libcryptopp.so cannot be caught by its specific type
(CryptoPP::InvalidKeyLength) in aes.so.  Investigating this leads me
to the well-known problem of RTTI comparison across shared library
boundaries, and the potential work-around of using the RTLD_GLOBAL in
dlopen().  Trying that work-around makes this problem go away, but
then if I load more than one .so which dynamically links to
libcryptopp.so, the second and later ones that get loaded are messed
up in a way that quickly leads to a crash (see the valgrind-generated
stack trace in [5] to see what I mean).

There is another problem with the same root cause, which is that
Crypto++ uses RTTI for a named-argument feature, see [6] for details.

I'm now considering a few ways forward:

1.  Persuade Debian and Fedora to accept pycryptopp and Tahoe-LAFS
using Crypto++ code compiled directly into the pycryptopp .so files
instead of dynamically linked.

2.  Refactor pycryptopp so that there is only one .so file, named for
example _pycryptopp.so, which is dynamically linked to libcryptopp.so,
and the separate modules for aes, sha256, rsa, ecdsa, etc. would each
import a subset of the Python names defined by _pycryptopp.so, and
then use RTLD_GLOBAL to load _pycryptopp.so.  This would, I think,
solve all currently known issues, but it does mean for example that if
anybody ever imports both pycryptopp and another Python module that
links to libcryptopp.so into the same Python process that one of them
will be screwed up and the process will quickly crash.

3.  Resign myself to working-around the lack of portable RTTI crossing
shared library boundaries in the pycryptopp source code.  Brian Warner
has already submitted patches for pycryptopp (see [6]) to work-around
the two known problems by (a) not catching CryptoPP::InvalidKeyLength
exception by its specific type and instead catching any type of
exception, and (b) not providing the hex-encoding feature which
happens to exercise Crypto++'s named-arguments feature.  I could
accept those two patches and resign myself to a fate of being unable
to safely use some ill-understood subset of the Crypto++ API.

4.  Figure out how to build an aes.so that has the relevant RTTI
symbols marked as "these must be satisfied by some other dynamic
library".  I'm not sure if this is possible, but it seems to how
things are done on Windows.  I read this page from the gcc wiki [7]
and experimented quite a bit with it.  When I started, using "nm" on
libcryptopp.so would show this:

$ nm -C /usr/local/lib/libcryptopp.so | grep "typeinfo for
CryptoPP::InvalidKeyLength"
00000000008747b0 V typeinfo for CryptoPP::InvalidKeyLength

And on my aes.so, it would show this:

$ nm -C ./pycryptopp/cipher/_aes.so | grep "typeinfo for
CryptoPP::InvalidKeyLength"
0000000000214cf0 V typeinfo for CryptoPP::InvalidKeyLength

After extensive exploration of the new gcc visibility features, I
finally managed to build an aes.so like this:

$ nm -C ./pycryptopp/cipher/aes.so | grep "typeinfo for
CryptoPP::InvalidKeyLength"
0000000000214cf0 d typeinfo for CryptoPP::InvalidKeyLength

Oops!  In other words, I managed to make the typeinfo symbol private
to aes.so instead of dynamic, thus guaranteeing that the exception
won't be caught even if I *do* specify RTLD_GLOBAL.  It sort of seems
like gcc offers the rough equivalent of Microsoft's "dllexport"
attribute, but not the rough equivalent of Microsoft's "dllimport"
attribute -- something that would, for example, force the symbol to
appear as "U" -- undefined -- in the .so's symbol table so that the
symbol's value would *have* to be provided by another DSO (in this
case libcryptopp.so) at load-time.  On the other hand, maybe if I
changed the libcryptopp.so so that the symbol was marked as non-weak,
such as "T", instead of its current type of weak -- "V" -- then maybe
the loader would have rewritten the value of the weak-symbol in aes.so
and the exception would have been caught.  I don't see how to do that,
either.


Okay, here's the question: do you know of any alternative besides
these four, and if not, which of these four do you recommend?


Thank you very much.

Regards,

Zooko Wilcox-O'Hearn

[*] Actually it fails one of the unit tests consistently on Mac OS
10.5/Intel, but not on Mac OS 10.4 or on any of the other platforms.
The failure *does* have something to do with RTTI since it is a
failure to downcast, but other than that I don't have any reason to
believe that it is related to the rest of this message, and I haven't
investigated it yet.  See the pycryptopp buildbot on Mac OS 10.5:
http://allmydata.org/buildbot-pycryptopp/builders/mac-i386-osx-10.5-faust

[**] Oh, and there's a mysterious problem on ARMv5 CPU in which a
memory buffer seems to be shifted by one byte, also probably
unrelated: http://allmydata.org/buildbot-pycryptopp/builders/zandr-linkstation

[1] http://allmydata.org/trac/pycryptopp
[2] http://allmydata.org/trac/tahoe
[3] http://mail.python.org/pipermail/python-dev/2002-May/024075.html
[4] http://cryptopp.com
[5] http://groups.google.com/group/cryptopp-users/browse_thread/thread/eb815f228db50380
[6] http://allmydata.org/trac/pycryptopp/ticket/9
[7] http://gcc.gnu.org/wiki/Visibility
---
store your data: $10/month -- http://allmydata.com/?tracking=zsig
I am available for work -- http://zooko.com/r?sum?.html

From hansroessler at yahoo.de  Mon May 25 16:08:47 2009
From: hansroessler at yahoo.de (Hans Roessler)
Date: Mon, 25 May 2009 07:08:47 -0700 (PDT)
Subject: [C++-sig] wrapping operator=
Message-ID: <831751.9082.qm@web24613.mail.ird.yahoo.com>


Hello!
I want to wrap with boost.python a C++ class "DataCube" which has overloaded the operator=.

The  constructor DataCube::DataCube(double x) allocates huge amounts of memory and fills it with x's.
The  DataCube::operator=(double x) just overwrites the already allocated memory with x's.

Now in C++ these commands first allocate memory, which is filled with 0.'s, and then overwrite the memory with 42.'s:
>DataCube data(0.)
>data=42.

In Python these commands first build the DataCube as desired, but then set data=42. (now data is a float), where the reference to the DataCube is lost:
>data=DataCube(0.)
>data=42.

I have circumvented this by replacing the second line with
> data.assign(42.)
with a function assign which is defined appropriately, but I would prefer to use the same assignment as in C++.

Finally the question:
Can I define the DataCube class in Python, so that the data variable above will behave as in C++, when I write "data=42." ?
In other words, is there any possibility in Python that "x=y" does NOT make x a reference to y?

Thank you
Hans



      

From wladwig at wdtinc.com  Mon May 25 18:54:53 2009
From: wladwig at wdtinc.com (William Ladwig)
Date: Mon, 25 May 2009 11:54:53 -0500
Subject: [C++-sig] wrapping operator=
In-Reply-To: <831751.9082.qm@web24613.mail.ird.yahoo.com>
References: <831751.9082.qm@web24613.mail.ird.yahoo.com>
Message-ID: <765CBD9053EA2B438625895F30D1856F0226274C69@storm.wdtinc.com>

Generally how that situation is handled on the Python side is to use the container emulation API, so that assigning all values would look like this to a user of your class:

data[:] = 42.

To do this, you just need to write a small c++ wrapper to expose the __setitem__(self, key, value) function, which accepts (or extracts) a slice object for the key and then calls your class operator= function.  Or, you could just stick with the assign function that you created, which may be better if your class isn't supposed to look like a container.  This may actually be less error prone.

Here is the Python container emulation documentation:

http://docs.python.org/reference/datamodel.html#emulating-container-types

Other people may have better suggestions.  

Regards,
Bill

________________________________________
From: cplusplus-sig-bounces+wladwig=wdtinc.com at python.org [cplusplus-sig-bounces+wladwig=wdtinc.com at python.org] On Behalf Of Hans Roessler [hansroessler at yahoo.de]
Sent: Monday, May 25, 2009 9:08 AM
To: cplusplus-sig at python.org
Subject: [C++-sig] wrapping operator=

Hello!
I want to wrap with boost.python a C++ class "DataCube" which has overloaded the operator=.

The  constructor DataCube::DataCube(double x) allocates huge amounts of memory and fills it with x's.
The  DataCube::operator=(double x) just overwrites the already allocated memory with x's.

Now in C++ these commands first allocate memory, which is filled with 0.'s, and then overwrite the memory with 42.'s:
>DataCube data(0.)
>data=42.

In Python these commands first build the DataCube as desired, but then set data=42. (now data is a float), where the reference to the DataCube is lost:
>data=DataCube(0.)
>data=42.

I have circumvented this by replacing the second line with
> data.assign(42.)
with a function assign which is defined appropriately, but I would prefer to use the same assignment as in C++.

Finally the question:
Can I define the DataCube class in Python, so that the data variable above will behave as in C++, when I write "data=42." ?
In other words, is there any possibility in Python that "x=y" does NOT make x a reference to y?

Thank you
Hans




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

From hansroessler at yahoo.de  Mon May 25 19:20:55 2009
From: hansroessler at yahoo.de (Hans Roessler)
Date: Mon, 25 May 2009 10:20:55 -0700 (PDT)
Subject: [C++-sig] wrapping operator=
In-Reply-To: <765CBD9053EA2B438625895F30D1856F0226274C69@storm.wdtinc.com>
References: <831751.9082.qm@web24613.mail.ird.yahoo.com>
	<765CBD9053EA2B438625895F30D1856F0226274C69@storm.wdtinc.com>
Message-ID: <76345.81452.qm@web24610.mail.ird.yahoo.com>


Thank you, William.
The __setitem__ emulation is definitely a better idea than the assign function, as the class is actually a container type.

But still I am looking for a way to allow the simple "data=42." syntax with the same result, because it would best copy the C++ syntax.

Hans



----- Urspr?ngliche Mail ----
> Von: William Ladwig <wladwig at wdtinc.com>
> An: Development of Python/C++ integration <cplusplus-sig at python.org>
> Gesendet: Montag, den 25. Mai 2009, 18:54:53 Uhr
> Betreff: Re: [C++-sig] wrapping operator=
> 
> Generally how that situation is handled on the Python side is to use the 
> container emulation API, so that assigning all values would look like this to a 
> user of your class:
> 
> data[:] = 42.
> 
> To do this, you just need to write a small c++ wrapper to expose the 
> __setitem__(self, key, value) function, which accepts (or extracts) a slice 
> object for the key and then calls your class operator= function.  Or, you could 
> just stick with the assign function that you created, which may be better if 
> your class isn't supposed to look like a container.  This may actually be less 
> error prone.
> 
> Here is the Python container emulation documentation:
> 
> http://docs.python.org/reference/datamodel.html#emulating-container-types
> 
> Other people may have better suggestions.  
> 
> Regards,
> Bill
> 
> ________________________________________
> From: cplusplus-sig-bounces+wladwig=wdtinc.com at python.org 
> [cplusplus-sig-bounces+wladwig=wdtinc.com at python.org] On Behalf Of Hans Roessler 
> [hansroessler at yahoo.de]
> Sent: Monday, May 25, 2009 9:08 AM
> To: cplusplus-sig at python.org
> Subject: [C++-sig] wrapping operator=
> 
> Hello!
> I want to wrap with boost.python a C++ class "DataCube" which has overloaded the 
> operator=.
> 
> The  constructor DataCube::DataCube(double x) allocates huge amounts of memory 
> and fills it with x's.
> The  DataCube::operator=(double x) just overwrites the already allocated memory 
> with x's.
> 
> Now in C++ these commands first allocate memory, which is filled with 0.'s, and 
> then overwrite the memory with 42.'s:
> >DataCube data(0.)
> >data=42.
> 
> In Python these commands first build the DataCube as desired, but then set 
> data=42. (now data is a float), where the reference to the DataCube is lost:
> >data=DataCube(0.)
> >data=42.
> 
> I have circumvented this by replacing the second line with
> > data.assign(42.)
> with a function assign which is defined appropriately, but I would prefer to use 
> the same assignment as in C++.
> 
> Finally the question:
> Can I define the DataCube class in Python, so that the data variable above will 
> behave as in C++, when I write "data=42." ?
> In other words, is there any possibility in Python that "x=y" does NOT make x a 
> reference to y?
> 
> Thank you
> Hans
> 
> 
> 
> 
> _______________________________________________
> 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



      

From seefeld at sympatico.ca  Mon May 25 19:58:17 2009
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Mon, 25 May 2009 13:58:17 -0400
Subject: [C++-sig] wrapping operator=
In-Reply-To: <76345.81452.qm@web24610.mail.ird.yahoo.com>
References: <831751.9082.qm@web24613.mail.ird.yahoo.com>	<765CBD9053EA2B438625895F30D1856F0226274C69@storm.wdtinc.com>
	<76345.81452.qm@web24610.mail.ird.yahoo.com>
Message-ID: <4A1ADC39.6050204@sympatico.ca>

Hans Roessler wrote:
> But still I am looking for a way to allow the simple "data=42." syntax with the same result, because it would best copy the C++ syntax.
>   
Would it, really ? I think it would be extremely confusing, if it would 
work. But it doesn't. "a = b" doesn't assign to an object, but to a 
variable. Variables in Python have reference semantics, so you'd simply 
reassign the name to a different object.

Regards,
       Stefan

-- 

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


From zookog at gmail.com  Wed May 27 17:42:48 2009
From: zookog at gmail.com (Zooko O'Whielacronx)
Date: Wed, 27 May 2009 09:42:48 -0600
Subject: [C++-sig] that problem with comparing types across DLLs
Message-ID: <cd6401a0905270842y7dae9b36t5aafdc5ab19ca474@mail.gmail.com>

(Note that several individuals, plus tahoe-dev at allmydata.org and
cplusplus-sig at python.org are Cc:'ed.  This is appropriate because this
is fundamentally a "cross project" issue and we need to talk to each
other until we understand what the other developers are planning.
Unfortunately, you're likely to have problems following-up to the
mailing lists if you aren't subscribed to them.)

On May 27, 2009, at 1:01 AM, Matthias Klose wrote:
>
> Well, maybe nobody did ask, because it's policy for distributions to build
> against shared libs? If there needs something to be fixed in the library
> package, you only have to take care of the library packages, not all packages
> which are linked against the static library (of course you would have to find
> out which these packages are). Plus you usually cannot link a python extension
> against a static lib because the static lib is built without -fPIC.

Thanks, Matthias, and thanks to Ruben for the earlier answers about Fedora.

It looks like option #1 from this list is right out:

http://mail.python.org/pipermail/cplusplus-sig/2009-May/014531.html

(Option #1 was to persuade distributions to support Crypto++ as a
static library.)

The option that would be technically the most general would be to
figure out how to annotate and compile Crypto++ as a shared lib that
has visible only the symbols that it needs to have visible, and then
use the RTLD_GLOBAL flag to dlopen() to make sure multiple .so's which
each depend on libcryptopp.so will get those symbols unified at
load-time so that cross-shared-library RTTI will work.  That is what
Niall Douglas recommends and has done successfully in the past.

However, I've already spent a couple of days trying to learn how to do
this and I haven't succeeded yet.  I'm not even sure that it is
possible.  Wei Dai, the author of Crypto++, has also tried in the past
to build such a shared lib of Crypto++ for Unix in the past, and he
too gave up.  That also means that he does not currently support using
Crypto++ as a shared library, which makes me uncomfortable, but I
guess that's the Linux distributions' problem more than mine.

I think that leaves option #2 from that list: refactor pycryptopp so
that it contains only a single shared library, and use RTLD_GLOBAL to
unify the symbols that are visible in the _pycryptopp.so with the
symbols in the the libcryptopp.so.  I think this is doable and that it
will solve all known problems for pycryptopp and Tahoe.  I think this
does mean that if in the future there is a different Python module
than pycryptopp which is dynamically linked to libcryptopp.so, that
anyone who imports both pycryptopp and that other Python module into
their Python interpreter will get a seg fault.  However, I suppose
this too is someone else's problem than mine.

So unless anyone else has a better suggestion, I'm going to proceed
with strategy #2.  I'll probably also continue to support the current
linking strategy of linking pycryptopp statically against
libcryptopp.a, as a build-time option for pycryptopp.  Therefore
people who have chosen the static-link-to-Crypto++ build-time option
(which excludes users of the Fedora, Debian, or Ubuntu packages of
pycryptopp) will be able to also import another Python module which
links to Crypto++.

Thanks, folks!

Regards,

Zooko

From zookog at gmail.com  Wed May 27 21:52:54 2009
From: zookog at gmail.com (Zooko O'Whielacronx)
Date: Wed, 27 May 2009 13:52:54 -0600
Subject: [C++-sig] that problem with comparing types across DLLs
In-Reply-To: <4A1D9269.16781.6A4D1E9@s_sourceforge.nedprod.com>
References: <cd6401a0905270842y7dae9b36t5aafdc5ab19ca474@mail.gmail.com>
	<4A1D9269.16781.6A4D1E9@s_sourceforge.nedprod.com>
Message-ID: <cd6401a0905271252u52f7cf6dvaef3cf286cbfaefe@mail.gmail.com>

On Wed, May 27, 2009 at 12:20 PM, Niall Douglas
<s_sourceforge at nedprod.com> wrote:
>
> If it turns out that GCC is incapable of emitting the correct symbol
> visibility declarations for a particular library, then an issue
> should be filed with GCC's bugtracker. They have been most responsive
> in the past with corner cases so I don't see it as an issue.

Thanks for the encouragement on that front.  I'll keep in mind that
gcc developers are responsive about these issues.  That helps.

> I am afraid I simply don't get this. Crypto++ seems to have a build
> option for Windows as a DLL. If it does then it ports very easily
> indeed to GCC symbol visibility. The only possible problem could be
> if he used a lot of dllexported non-class based static variables
> which you shouldn't in object orientatation.

As Wei Dai explained in this thread:

http://groups.google.com/group/cryptopp-users/browse_thread/thread/eb815f228db50380

He implemented the "CRYPTOPP_DLL" build of Crypto++ only because it
was a requirement for FIPS validation, and it doesn't include a lot of
the functionality of the library -- only that subset of the
functionality which was being validated.  Also, he attempted to port
the same shared-lib build to Unix and failed and gave up.

I don't know whether Crypto++ dllexports a lot of non-class-based
static variables.  It is a very big library with a big API and I don't
have a good understanding of its scope.

> If I'm reading between the lines here correctly, are you saying that
> your libcrypto++ uses vX.XX of a libaes library but you want that to
> interoperate which any other shared library which uses vY.YY of the
> same libaes library? Obviously should APIs move out of step you get
> all sorts of pernicious bugs.

Wei Dai is the author of Crypto++, and I'm the author of pycryptopp.
Pycryptopp is merely a few Python wrappers for certain Crypto++
features, plus a few small utilities that I need.  The current build
system of pycryptopp generates a module named aes.so, one named
ecdsa.so, one named rsa.so and one named sha256.so.  Each of these
requires different functionality from Crypto++.  The current build
system links aes.so with the Crypto++ object files which are built
from the Crypto++ .cpp files.  This works, but it is not allowed to
deploy as an official Debian or Fedora package when built like this.

If I instead link aes.so with "-lcryptopp" and use the Debian- or
Fedora- provided libcryptopp.so at load-time (using the default Python
behavior of RTLD_LOCAL), it fails, because RTTI symbols are not merged
at load-time, so an exception of type InvalidKeyLength, thrown from
libcryptopp.so will not be caught by the handler in aes.so which
attempts to catch exceptions of that type, because the RTTI symbol for
InvalidKeyLength in libcryptopp.so has a different value than the RTTI
symbol for InvalidKeyLength in aes.so.

If I then set the RTLD_GLOBAL flag, then loading aes.so (which
dynamically depends on libcryptopp.so) works, and the exception
handler can catch exceptions of the specific type.  But, if I then
load both aes.so and rsa.so into the same Python interpreter, it seg
faults.  Here is the valgrind stack trace of that event:

http://groups.google.com/group/cryptopp-users/attach/1a5553410c6976e5/crash.txt?part=2

So I am now engaged in refactoring pycryptopp so that instead of
separate aes.so and rsa.so (and so on), there is instead a single
_pycryptopp.so which provides all of the functionality, and then there
is an aes.py and an rsa.py which import the appropriate names from
_pycryptopp and leave out the other names.

The problem with this strategy, I think, is that if someone were to
load _pycryptopp.so using RTLD_GLOBAL, and then someone else were to
load libcryptopp.so into the same Python interpreter, it would seg
fault.  But maybe nobody will ever want to do that.

Thanks for you help!

Regards,

Zooko

From wjoergensen at yahoo.com  Thu May 28 22:32:30 2009
From: wjoergensen at yahoo.com (Werner Joergensen)
Date: Thu, 28 May 2009 13:32:30 -0700 (PDT)
Subject: [C++-sig] libboost_python-gcc43-mt.so or without "mt"?
Message-ID: <364933.85799.qm@web111810.mail.gq1.yahoo.com>


Dear list participants,

I have installed boost including the python lib using the boostrap.sh installation script, and now I have  the following libraries under /lib (linux system):
libboost_python-gcc43-mt-1_39.a
libboost_python-gcc43-mt-1_39.so
libboost_python-gcc43-mt-1_39.so.1.39.0
libboost_python-gcc43-mt.a
libboost_python-gcc43-mt.so

However when I build a python extension module with bjam, and when I import the newly built module in python, the problem is:
    $ ipython
    In [1]: import myModule
    ImportError: libboost_python-gcc43-1_39.so.1.39.0: cannot open shared object file: No such file or directory

Of course, I can copy and rename the mt-libraries to the expected library name. But is there a possibility to build the python extension so that the exististing libboost_python-gcc43-mt.so will be linked instead of some other name? Can I configure which libboost_python will be used somewhere in boost-build.jam or user-config.jam or Jamroot?

Thanks
Werner



      

From boliguayo1942 at hotmail.com  Fri May 29 05:55:06 2009
From: boliguayo1942 at hotmail.com (=?iso-8859-1?Q?Juan_Manuel_Martinez_Caama=F1o?=)
Date: Fri, 29 May 2009 00:55:06 -0300
Subject: [C++-sig] problems building a pyd under vc++ with distutils
Message-ID: <BLU149-DS42EDFDC15E2C38B5739EFAC510@phx.gbl>

I'm under vc++ 9.0 and I'm having problem when trying to build a simple *.pyd... (using distutils)
When I build the dll from the IDE it produces a DLL (not a *.pyd) and no compiler or linker errors.
But when I use distutils it just output a lot of errors...
I can't find the problem...

/*-----------------pydll.cpp:------------------*/
#include <boost/python.hpp> 

#include <string>

using namespace boost::python; 

__declspec(dllexport) std::string greet()

{ 

return "Hello Python"; 

}; 

BOOST_PYTHON_MODULE(hello) 

{ def("greet", greet); } 

--------------------------------------------------------------
#setup.py
from distutils.core import setup, Extension

module1 = Extension('hello',
                    sources = ["pydll.cpp"])

setup (name = 'pydll',
       version = '1.0',
       description = 'This is a demo package',
       ext_modules = [module1])
-------------------------------------------------------------

(under visual studio 2008, python 2.6, boost 1.38.0)

Thank you
Juan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20090529/3605991e/attachment.htm>

From wjoergensen at yahoo.com  Fri May 29 12:20:56 2009
From: wjoergensen at yahoo.com (Werner Joergensen)
Date: Fri, 29 May 2009 03:20:56 -0700 (PDT)
Subject: [C++-sig] [Py++] incomplete type and forward declaration
Message-ID: <573306.52699.qm@web111809.mail.gq1.yahoo.com>


How can I process such a file classA.hpp using a code generator like http://www.language-binding.net/pyplusplus/documentation/tutorials/module_builder/generate_code.py.html?

----------8<---- classA.hpp ------
class B;

class A{
public:
  B& dontCareAboutB(){ /*return C::getB();*/ };
  void doSomething(){};
};
----------8<----------

This class compiles fine with just "g++ -c classA.hpp", but applying the pyplusplus code generator followed by bjam throws several errors:
...
/opt/boost_1_39_0/boost/python/type_id.hpp:89: error: invalid use of incomplete type ?struct B?
./classA.hpp:1: error: forward declaration of ?struct B?
...

If I replace the first line with "class B{}", making B a comple, empty class, this example succeeds. But then I run into more troubles if I add the actual definition of the real class B from another header file to the python module.

Does a solution with pyplusplus exist, where I don't need to modify the classA.hpp header file?

Thanks
Werner



      

From hessiess at hessiess.com  Fri May 29 18:56:37 2009
From: hessiess at hessiess.com (hessiess at hessiess.com)
Date: Fri, 29 May 2009 17:56:37 +0100 (BST)
Subject: [C++-sig] Derived classes problem
Message-ID: <d9cae3843443da231295c573894028e3.squirrel@www.hessiess.dyndns.org>

I am trying to create a python wrapper for the quad-ren graphics
engine, which relies heavily on derived classes. The problem is
that I cannot create derived classes in python and pass them
to the C++ objects without getting the following error:

Traceback (most recent call last):
  File "test.py", line 79, in <module>
    quad3.set_event_receiver(event)
Boost.Python.ArgumentError: Python argument types in
    qr_scene_node.set_event_receiver(qr_quad, ev_receiver)
did not match C++ signature:
    set_event_receiver(qr_scene_node {lvalue}, qr_event_receiver*)

====================================================================
C++
====================================================================
#include <boost/python.hpp>
#include <quad-ren/quad-ren.h>

using namespace boost::python;

struct qr_scene_node_w : qr_scene_node, wrapper<qr_scene_node>
{
    void on_register()
    {
        this->get_override("on_register")();
    }

    void on_render()
    {
        this->get_override("on_render")();
    }

    void on_drop()
    {
        this->get_override("on_drop")();
    }
};

struct qr_event_receiver_w : qr_event_receiver, wrapper<qr_event_receiver>
{
    bool on_event(qr_event *event, qr_scene_node *node, float time_delta)
    {
        this->get_override("on_event")();
    }
};

BOOST_PYTHON_MODULE(qrpy)
{
    class_<vector2d_i>("vector2d_i")
        .def(init<float, float>())
        .def_readwrite("X", &vector2d_i::X)
        .def_readwrite("Y", &vector2d_i::Y);

    class_<vector2d_f>("vector2d_f")
        .def(init<float, float>())
        .def_readwrite("X", &vector2d_f::X)
        .def_readwrite("Y", &vector2d_f::Y);

    class_<qr_resource_manager>("qr_resource_manager");

    class_<qr_renderer>("qr_renderer")
        .def(init<vector2d_i, float, bool, vector2d_i>())
        .def("get_resource_manager", &qr_renderer::get_resource_manager,
            return_value_policy<reference_existing_object>())
        .def("run",                  &qr_renderer::run)
        .def("drop",                 &qr_renderer::drop)
        .def("set_total_layers",     &qr_renderer::set_total_layers)
        .def("set_window_title",     &qr_renderer::set_window_title)
        .def("get_time_delta",       &qr_renderer::get_time_delta)
        .def("get_ticks",            &qr_renderer::get_ticks)
        .def("get_fps",              &qr_renderer::get_fps)
        .def("set_camera_location",  &qr_renderer::set_camera_location)
        .def("get_camera_location",  &qr_renderer::get_camera_location)
        .def("get_pix_size",         &qr_renderer::get_pix_size)
        .def("set_bg_sprite",        &qr_renderer::set_bg_sprite)
        .def("get_bg_sprite",        &qr_renderer::get_bg_sprite,
            return_value_policy<reference_existing_object>());

    class_<qr_scene_node_w, boost::noncopyable>("qr_scene_node", no_init)
        .def("register_node",      &qr_scene_node_w::register_node)
        .def("drop",               &qr_scene_node_w::drop)
        .def("set_location",       &qr_scene_node_w::set_location)
        .def("get_location",       &qr_scene_node_w::get_location)
        .def("set_size",           &qr_scene_node_w::set_size)
        .def("get_size",           &qr_scene_node_w::get_size)
        .def("set_rotation",       &qr_scene_node_w::set_rotation)
        .def("get_rotation",       &qr_scene_node_w::get_rotation)
        .def("set_layer",          &qr_scene_node_w::set_layer)
        .def("get_layer",          &qr_scene_node_w::get_layer)
        .def("set_sprite",         &qr_scene_node_w::set_sprite)
        .def("get_sprite",         &qr_scene_node_w::get_sprite,
            return_value_policy<reference_existing_object>())
        .def("set_event_receiver", &qr_scene_node_w::set_event_receiver)
        .def("get_event_receiver", &qr_scene_node_w::get_event_receiver,
            return_value_policy<reference_existing_object>())
        .def("set_frame",          &qr_scene_node_w::set_frame)
        .def("get_frame",          &qr_scene_node_w::get_frame)

        .def("add_polygon",        &qr_scene_node_w::add_polygon)
        .def("clear_mesh",         &qr_scene_node_w::clear_mesh)
        .def("on_register",        &qr_scene_node_w::on_register)
        .def("on_render",          &qr_scene_node_w::on_render)
        .def("on_drop",            &qr_scene_node_w::on_drop);

    class_<qr_event_receiver_w, boost::noncopyable>("qr_event_receiver",
no_init)
        .def("on_event",           &qr_event_receiver_w::on_event);

    class_<qr_quad, bases<qr_scene_node> >("qr_quad")
        .def(init<qr_resource_manager* , bool>());
}

====================================================================
Python
====================================================================
import qrpy
false = bool(0)
true  = bool(1)

#create renderer + get resource manager
renderer = qrpy.qr_renderer(qrpy.vector2d_i(4, 3),
    10, false, qrpy.vector2d_i(0, 0))
res_man = renderer.get_resource_manager()

#event receiver
class ev_receiver(qrpy.qr_event_receiver):
    def __init__(self):
        return

    def on_event(self, event, node, time_delta):
        print event

event = ev_receiver()

quad = qrpy.qr_quad(res_man, 0)
quad.set_event_receiver(event) #error


From ndbecker2 at gmail.com  Fri May 29 20:27:55 2009
From: ndbecker2 at gmail.com (Neal Becker)
Date: Fri, 29 May 2009 14:27:55 -0400
Subject: [C++-sig] libboost_python-gcc43-mt.so or without "mt"?
References: <364933.85799.qm@web111810.mail.gq1.yahoo.com>
Message-ID: <gvp9fc$h50$1@ger.gmane.org>

Werner Joergensen wrote:

> 
> Dear list participants,
> 
> I have installed boost including the python lib using the boostrap.sh
> installation script, and now I have  the following libraries under /lib
> (linux system): libboost_python-gcc43-mt-1_39.a
> libboost_python-gcc43-mt-1_39.so libboost_python-gcc43-mt-1_39.so.1.39.0
> libboost_python-gcc43-mt.a
> libboost_python-gcc43-mt.so
> 
> However when I build a python extension module with bjam, and when I
> import the newly built module in python, the problem is:
>     $ ipython
>     In [1]: import myModule
>     ImportError: libboost_python-gcc43-1_39.so.1.39.0: cannot open shared
>     object file: No such file or directory
> 
> Of course, I can copy and rename the mt-libraries to the expected library
> name. But is there a possibility to build the python extension so that the
> exististing libboost_python-gcc43-mt.so will be linked instead of some
> other name? Can I configure which libboost_python will be used somewhere
> in boost-build.jam or user-config.jam or Jamroot?
> 
> Thanks
> Werner

Here's what I do

1. Get updated 1.39 from svn (the --layout=tagged option was just added, at 
my request).
Then,

I use:

bjam -sICU_PATH=/usr -sEXPAT_INCLUDE=/usr -sEXPAT_LIBPATH=/usr/lib64 --
layout=tagged threading=single,multi stage 

You will then have (sorry for ugly formatting):

 -rw-rw-r-- 1 nbecker nbecker  161294 2009-05-18 15:17 libboost_date_time.a
  -rw-rw-r-- 1 nbecker nbecker  161294 2009-05-18 15:17 libboost_date_time-
mt.a
  lrwxrwxrwx 1 nbecker nbecker      31 2009-05-18 15:17 libboost_date_time-
mt.so -> libboost_date_time-mt.so.1.40.0
  -rwxrwxr-x 1 nbecker nbecker   76680 2009-05-18 15:17 libboost_date_time-
mt.so.1.40.0
  lrwxrwxrwx 1 nbecker nbecker      28 2009-05-18 15:17 
libboost_date_time.so -> libboost_date_time.so.1.40.0
  -rwxrwxr-x 1 nbecker nbecker   76648 2009-05-18 15:17 
libboost_date_time.so.1.40.0
 ...




From wjoergensen at yahoo.com  Sat May 30 22:53:07 2009
From: wjoergensen at yahoo.com (Werner Joergensen)
Date: Sat, 30 May 2009 13:53:07 -0700 (PDT)
Subject: [C++-sig] [Py++] wrapping boost python types fails
Message-ID: <999875.14484.qm@web111808.mail.gq1.yahoo.com>


I narrowed down a failure with pyplusplus to this short example:

---------8<---- test.hpp -----
#include "boost/python.hpp"
--------->8---------

---------8<---- generator.py -----
import os
from pyplusplus import module_builder
mb = module_builder.module_builder_t(
    files=["test.hpp"],
        include_paths=["/usr/include/python2.6"]
    )

mb.build_code_creator( module_name='testmodule' )
mb.code_creator.user_defined_directories.append( os.path.abspath('.') )
mb.split_module("split")
mb.write_module( os.path.join( os.path.abspath('.'), 'generated.cpp' ) )
--------->8---------

Running "python generator.py" gives these errors:
---------8<---- output -----
...
pygccxml.parser.source_reader.gccxml_runtime_error_t: Error occured while running GCC-XML: In file included from /usr/local/include/boost-1_39/boost/python.hpp:60,
                 from test.hpp:1:
/usr/local/include/boost-1_39/boost/type.hpp:14: internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
--------->8---------

I need to use the python objects like list, str, tuple in my python extension modules. So can you help me to get the example running? I tried it on two different systems, one with gcc 4.2.4 and one with gcc 4.3.3. Other relevant packages are: gccxml 0.9 (cvs from 2008-05-30), boost 1.39.0, pygccxml 1.0, pyplusplus 1.0. Can you tell me which combination of these packages can compile the short example?

Thanks
Werner



      

From roman.yakovenko at gmail.com  Sun May 31 06:27:43 2009
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Sun, 31 May 2009 07:27:43 +0300
Subject: [C++-sig] [Py++] incomplete type and forward declaration
In-Reply-To: <573306.52699.qm@web111809.mail.gq1.yahoo.com>
References: <573306.52699.qm@web111809.mail.gq1.yahoo.com>
Message-ID: <7465b6170905302127u332ce2cenbd9ae4469d173bd4@mail.gmail.com>

On Fri, May 29, 2009 at 1:20 PM, Werner Joergensen
<wjoergensen at yahoo.com> wrote:
>
> How can I process such a file classA.hpp using a code generator like http://www.language-binding.net/pyplusplus/documentation/tutorials/module_builder/generate_code.py.html?
>
> ----------8<---- classA.hpp ------
> class B;
>
> class A{
> public:
> ?B& dontCareAboutB(){ /*return C::getB();*/ };
> ?void doSomething(){};
> };
> ----------8<----------
>
> This class compiles fine with just "g++ -c classA.hpp", but applying the pyplusplus code generator followed by bjam throws several errors:
> ...
> /opt/boost_1_39_0/boost/python/type_id.hpp:89: error: invalid use of incomplete type ?struct B?
> ./classA.hpp:1: error: forward declaration of ?struct B?
> ...
>
> If I replace the first line with "class B{}", making B a comple, empty class, this example succeeds. But then I run into more troubles if I add the actual definition of the real class B from another header file to the python module.
>
> Does a solution with pyplusplus exist, where I don't need to modify the classA.hpp header file?

I think, the solution exists. First of all you should understand, that
Boost.Python really needs to know( see definition ) of class B. So you
will have to include it "before" the generated code.

What you can do is to add the "B" header file to class A generate
code:  http://language-binding.net/pyplusplus/documentation/inserting_code.html#header-files

HTH





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

From roman.yakovenko at gmail.com  Sun May 31 06:30:17 2009
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Sun, 31 May 2009 07:30:17 +0300
Subject: [C++-sig] [Py++] wrapping boost python types fails
In-Reply-To: <999875.14484.qm@web111808.mail.gq1.yahoo.com>
References: <999875.14484.qm@web111808.mail.gq1.yahoo.com>
Message-ID: <7465b6170905302130h7d11a3a8pf70120bc435e8513@mail.gmail.com>

On Sat, May 30, 2009 at 11:53 PM, Werner Joergensen
<wjoergensen at yahoo.com> wrote:
>
> I narrowed down a failure with pyplusplus to this short example:
>
> ---------8<---- test.hpp -----
> #include "boost/python.hpp"
> --------->8---------
>
> ---------8<---- generator.py -----
> import os
> from pyplusplus import module_builder
> mb = module_builder.module_builder_t(
> ? ?files=["test.hpp"],
> ? ? ? ?include_paths=["/usr/include/python2.6"]
> ? ?)
>
> mb.build_code_creator( module_name='testmodule' )
> mb.code_creator.user_defined_directories.append( os.path.abspath('.') )
> mb.split_module("split")
> mb.write_module( os.path.join( os.path.abspath('.'), 'generated.cpp' ) )
> --------->8---------
>
> Running "python generator.py" gives these errors:
> ---------8<---- output -----
> ...
> pygccxml.parser.source_reader.gccxml_runtime_error_t: Error occured while running GCC-XML: In file included from /usr/local/include/boost-1_39/boost/python.hpp:60,
> ? ? ? ? ? ? ? ? from test.hpp:1:
> /usr/local/include/boost-1_39/boost/type.hpp:14: internal compiler error: Segmentation fault
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See <URL:http://gcc.gnu.org/bugs.html> for instructions.
> --------->8---------
>
> I need to use the python objects like list, str, tuple in my python extension modules. So can you help me to get the example running? I tried it on two different systems, one with gcc 4.2.4 and one with gcc 4.3.3. Other relevant packages are: gccxml 0.9 (cvs from 2008-05-30), boost 1.39.0, pygccxml 1.0, pyplusplus 1.0. Can you tell me which combination of these packages can compile the short example?

Unfortunately, gccxml is not able to parse code, which includes
"boost/python.hpp" header file. You will have to find someway to
exclude it from compilation.

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

From wjoergensen at yahoo.com  Sun May 31 20:30:07 2009
From: wjoergensen at yahoo.com (Werner Joergensen)
Date: Sun, 31 May 2009 11:30:07 -0700 (PDT)
Subject: [C++-sig] [Py++] wrapping boost python types fails
In-Reply-To: <7465b6170905302130h7d11a3a8pf70120bc435e8513@mail.gmail.com>
References: <999875.14484.qm@web111808.mail.gq1.yahoo.com>
	<7465b6170905302130h7d11a3a8pf70120bc435e8513@mail.gmail.com>
Message-ID: <298398.3395.qm@web111810.mail.gq1.yahoo.com>


>> I need to use the python objects like list, str, tuple in my python extension modules. So can you help me to get the example running? I tried it on two different systems, one with gcc 4.2.4 

>>and one with gcc 4.3.3. Other relevant packages are: gccxml 0.9 (cvs from 2008-05-30), boost 1.39.0, pygccxml 1.0, pyplusplus 1.0. Can you tell me which combination of these packages 
>>can compile the short example?

>Unfortunately, gccxml is not able to parse code, which includes
>"boost/python.hpp" header file. You will have to find someway to
>exclude it from compilation.

Good to know. But do you have an explanation, why gccxml is unable? Is there any chance I could work on a solution, or have experienced developers already given up?
Do you have some references of other attempts to solve the problem?

For a workaround: It is no problem to exclude the offending code from compilation. But then how can I wrap classes that expect for example a python list as input?

Thanks
Werner



      

From roman.yakovenko at gmail.com  Sun May 31 20:41:26 2009
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Sun, 31 May 2009 21:41:26 +0300
Subject: [C++-sig] [Py++] wrapping boost python types fails
In-Reply-To: <298398.3395.qm@web111810.mail.gq1.yahoo.com>
References: <999875.14484.qm@web111808.mail.gq1.yahoo.com>
	<7465b6170905302130h7d11a3a8pf70120bc435e8513@mail.gmail.com>
	<298398.3395.qm@web111810.mail.gq1.yahoo.com>
Message-ID: <7465b6170905311141y332ac035ic2dcc6e79b5c3e89@mail.gmail.com>

On Sun, May 31, 2009 at 9:30 PM, Werner Joergensen
<wjoergensen at yahoo.com> wrote:
>
>>> I need to use the python objects like list, str, tuple in my python extension modules. So can you help me to get the example running? I tried it on two different systems, one with gcc 4.2.4
>
>>>and one with gcc 4.3.3. Other relevant packages are: gccxml 0.9 (cvs from 2008-05-30), boost 1.39.0, pygccxml 1.0, pyplusplus 1.0. Can you tell me which combination of these packages
>>>can compile the short example?
>
>>Unfortunately, gccxml is not able to parse code, which includes
>>"boost/python.hpp" header file. You will have to find someway to
>>exclude it from compilation.
>
> Good to know. But do you have an explanation, why gccxml is unable?

Yes - Boost.Python is the art of template and preprocessor meta
programming. I guess, GCCXML fails to dump those template
instantiations.

> Is there any chance I could work on a solution, or have experienced developers already given up?

If you have experience with gcc code base, than I guess you will be
able to solve the problem. In any case, gccxml mailing list is a
better place to discuss this bug (
http://www.gccxml.org/mailman/listinfo/gccxml )

> Do you have some references of other attempts to solve the problem?

No

> For a workaround: It is no problem to exclude the offending code from compilation. But then how can I wrap classes that expect for example a python list as input?

Py++ doesn't support this use case, for the reason above. If you have
a lot of code, that uses Boost.Python classes, than we can work
together to patch Py++, otherwise you can use the following
functionality: http://www.language-binding.net/pyplusplus/documentation/inserting_code.html

HTH

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