From markus at relix.de  Tue Aug  1 02:23:09 2006
From: markus at relix.de (Markus Heller)
Date: Tue, 1 Aug 2006 02:23:09 +0200
Subject: [C++-sig] Working with python unicode strings without wstring
In-Reply-To: <7465b6170607311127y1d5f6962l4c2dd662146b8eb7@mail.gmail.com>
References: <BAY108-DAV765F689103A7BED1FE3D5B4B10@phx.gbl>
	<200607311933.51670.markus@relix.de>
	<7465b6170607311127y1d5f6962l4c2dd662146b8eb7@mail.gmail.com>
Message-ID: <200608010223.09974.markus@relix.de>

Dear Roman,

> May be you should use another set of API?
> http://docs.python.org/api/unicodeObjects.html
Thanks for your advice. I'm just reading the documentation :-)

> It is also possible that your locale is wrong.
So what is the influence of the locale setting (here: de_DE at euro) on the 
interpretation of a German umlaut character? Python gets it right, only the 
handover to the C++ std::string variable crashes. 

Regards
Markus



From mharidev at qualcomm.com  Tue Aug  1 03:01:36 2006
From: mharidev at qualcomm.com (Haridev, Meghana)
Date: Mon, 31 Jul 2006 18:01:36 -0700
Subject: [C++-sig] Pyplusplus: Removing generation of
	bp::implicitly_convertible code?
Message-ID: <5383A81D5EB8F14FBDB487D9B5FBEA880167CDFF@NAEX11.na.qualcomm.com>

I'm using pyplusplus to generate boost.python wrappers for the following
piece of code:

 

//test.h 

class Y

{  };

 

class X

{

    public:

    X(const Y *yptr = 0){  }

};

 

Pyplusplus generates the corresponding boost.python code for it:

 

// This file has been generated by pyplusplus.

#include "boost/python.hpp"

#include "test.h"

namespace bp = boost::python;

 

BOOST_PYTHON_MODULE(TEST){

    bp::class_< Y >( "Y" );

    bp::class_< X >( "X", bp::init< bp::optional< Y const * > >((
bp::arg("yptr")=bp::object() )) );

    bp::implicitly_convertible< Y const *, X >();

}

 

Two Questions:

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

Question 1: How can I make pyplusplus to NOT generate the implicitly
convertible boost code? I do not want to take advantage of the
conversion facility. Basically, I want the pyplusplus generated code to
look like this:

BOOST_PYTHON_MODULE(TEST){

    bp::class_< Y >( "Y" );

    bp::class_< X >( "X", bp::init< bp::optional< Y const * > >((
bp::arg("yptr")=bp::object() )) );

   //bp::implicitly_convertible< Y const *, X >(); - Do not want to
generate

}

-----

OR

------

Question 2: Is there any way I can expose only the default no-arg
constructor for class X i.e X( ) and not expose the constructor X(const
Y *yptr = 0)? That is, expose class X as: bp::class_< X >( "X",
bp::init< >( ) ). That would automatically take care of not generating
the bp::implicitly_convertible code also. Generated code should look
like this:

BOOST_PYTHON_MODULE(TEST){

    bp::class_< Y >( "Y" );

    bp::class_< X >( "X", bp::init< >( ) ); //Expose just default
constructor

   //bp::implicitly_convertible< Y const *, X >(); - Will not be
generated

}

I also tried excluding constructors for class X, but it excludes all the
constructors. I want to be able to select and just exclude the X(const Y
*yptr = 0) constructor. Is it possible to selectively exclude
constructors from being exposed to C++ using pyplusplus?

 

Thanks,

-Meghana.

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060731/fce95002/attachment.htm>

From roman.yakovenko at gmail.com  Tue Aug  1 08:17:39 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Tue, 1 Aug 2006 09:17:39 +0300
Subject: [C++-sig] Working with python unicode strings without wstring
In-Reply-To: <200608010223.09974.markus@relix.de>
References: <BAY108-DAV765F689103A7BED1FE3D5B4B10@phx.gbl>
	<200607311933.51670.markus@relix.de>
	<7465b6170607311127y1d5f6962l4c2dd662146b8eb7@mail.gmail.com>
	<200608010223.09974.markus@relix.de>
Message-ID: <7465b6170607312317y7be98c4fh87d650a74e0589ce@mail.gmail.com>

On 8/1/06, Markus Heller <markus at relix.de> wrote:
> Dear Roman,
>
> > May be you should use another set of API?
> > http://docs.python.org/api/unicodeObjects.html
> Thanks for your advice. I'm just reading the documentation :-)
>
> > It is also possible that your locale is wrong.
> So what is the influence of the locale setting (here: de_DE at euro) on the
> interpretation of a German umlaut character? Python gets it right, only the
> handover to the C++ std::string variable crashes.

I could be wrong, but it seems to me that default C++ locale is C
locale, and it
treats only ASCII characters. So, may be you need to set global locale
to the right
one?

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


From roman.yakovenko at gmail.com  Tue Aug  1 08:37:02 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Tue, 1 Aug 2006 09:37:02 +0300
Subject: [C++-sig] Pyplusplus: Removing generation of
	bp::implicitly_convertible code?
In-Reply-To: <5383A81D5EB8F14FBDB487D9B5FBEA880167CDFF@NAEX11.na.qualcomm.com>
References: <5383A81D5EB8F14FBDB487D9B5FBEA880167CDFF@NAEX11.na.qualcomm.com>
Message-ID: <7465b6170607312337j86a4bdfi9e767165c2a4186b@mail.gmail.com>

On 8/1/06, Haridev, Meghana <mharidev at qualcomm.com> wrote:
> I'm using pyplusplus to generate boost.python wrappers for the following
> piece of code:
>
> //test.h
> class Y
> {  };
> class X
> {
>     public:
>     X(const Y *yptr = 0){  }
> };
> Pyplusplus generates the corresponding boost.python code for it:
>
> // This file has been generated by pyplusplus.
> #include "boost/python.hpp"
> #include "test.h"
> namespace bp = boost::python;
> BOOST_PYTHON_MODULE(TEST){
>     bp::class_< Y >( "Y" );
>     bp::class_< X >( "X", bp::init< bp::optional< Y const * > >((
> bp::arg("yptr")=bp::object() )) );
>     bp::implicitly_convertible< Y const *, X >();
> }
> Two Questions:
>
> ----------------------
>
> Question 1: How can I make pyplusplus to NOT generate the implicitly
> convertible boost code? I do not want to take advantage of the conversion
> facility. Basically, I want the pyplusplus generated code to look like this:

mb = module_builder_t( ... )
mb.build_code_creator( ..., create_castinig_constructor=False, ... )

>
> Question 2: Is there any way I can expose only the default no-arg
> constructor for class X i.e X( ) and not expose the constructor X(const Y
> *yptr = 0)? That is, expose class X as: bp::class_< X >( "X", bp::init< >( )
> ). That would automatically take care of not generating the
> bp::implicitly_convertible code also. Generated code should look like this:
>
> BOOST_PYTHON_MODULE(TEST){
>
>     bp::class_< Y >( "Y" );
>
>     bp::class_< X >( "X", bp::init< >( ) ); //Expose just default
> constructor
>
>    //bp::implicitly_convertible< Y const *, X >(); - Will not be generated
>
> }
>
> I also tried excluding constructors for class X, but it excludes all the
> constructors. I want to be able to select and just exclude the X(const Y
> *yptr = 0) constructor. Is it possible to selectively exclude constructors
> from being exposed to C++ using pyplusplus?

Class X does not have empty constructor. I assume that it has.

Before you read my explanation, please read this tutorials:
http://language-binding.net/pygccxml/query_interface.html

from pygccxml import declarations

mb = module_builder_t( ... )

def criteria( constructor ):
    if len( constructor.arguments ) != 1:
        return False
    arg0_type = constructor.arguments[0].type
    if not declarations.is_pointer( arg0_type ):
        return
    tmp = declarations.remove_pointer( arg0_type ):
    if not declarations.is_const( tmp ):
         return False
    return declarations.is_same( declarations.remove_const( tmp ),
mb.class_( "Y" ) )

X = mb.class_( "X" )
constructor_from_Y = X.constructor( criteria )

Ther is another approach.
constructor_from_Y = X.constructor( args_type=[ "Y const *" ] )
-----------------------------------------------------^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

This query will return constructor, that takes 1 argument "Y const *".
The problem
with this approach is that this string should be equal to arg0.type.decl_string.

And third approach it to select all constructors that take 1 argument:

cs = X.constructors( [None] )


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


From j.reid at mail.cryst.bbk.ac.uk  Tue Aug  1 10:49:59 2006
From: j.reid at mail.cryst.bbk.ac.uk (John Reid)
Date: Tue, 01 Aug 2006 09:49:59 +0100
Subject: [C++-sig] Boost.Python shared_ptr container gotchas
Message-ID: <ean4jp$v77$1@sea.gmane.org>

Hi,

I use boost::shared_ptr to manage lifetimes of objects. I've run into a 
couple of problems using these inside containers with Boost.Python. I 
thought I could share my experiences and perhaps get some useful 
feedback or help others avoid these problems.

1) Exposing a vector< shared_ptr< T > > through the indexing suite was 
giving me headaches. I couldn't work out why I couldn't access the 
elements from python. I kept getting a runtime error about no python 
class for boost::shared_ptr< T > even after I was registering the 
pointer type. It turns out that my problem was the vector indexing suite 
was trying to wrap a reference to the boost::shared_ptr< T > in a python 
object. I needed to set the NoProxy template parameter on 
vector_indexing_suite to true and that solved all those issues. Once I 
understood what was happening the answer was clear. Until that moment 
there was a lot of head-scratching. The documentation is factually 
comprehensive but I wonder whether it couldn't have more use cases that 
explain why certain things are necessary.

2) This was a slightly more insidious problem. I also use the 
shared_ptrs as keys in associative containers. I ensure that equal 
objects are always identical (i.e. share the same address) so that I can 
just compare addresses rather than contents. The trouble was that 
operator< for shared_ptrs seemed to be comparing more than just the 
address of the pointee. If a shared_ptr made a round trip through python 
and back again it would not be equal (not greater nor less than) to the 
original shared_ptr under the operator< relationship. I fixed this by 
changing the ordering on my map to:

struct smart_ptr_less_than
{
	template< typename T >
	bool operator()( T p1, T p2 ) const
	{
		return p1.get() < p2.get();
	}
};

AFAIK this is defines a valid weak ordering. I'm not sure why operator< 
on shared_ptrs is not defined like this. Something to do with the 
deleter stored for shared_ptrs perhaps?

I hope this proves useful to someone or that I can be corrected and 
learn something in the process.

Regards,
John.



From roman.yakovenko at gmail.com  Tue Aug  1 11:16:08 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Tue, 1 Aug 2006 12:16:08 +0300
Subject: [C++-sig] Boost.Python shared_ptr container gotchas
In-Reply-To: <ean4jp$v77$1@sea.gmane.org>
References: <ean4jp$v77$1@sea.gmane.org>
Message-ID: <7465b6170608010216h794cc164re96a6ceaef6e1e60@mail.gmail.com>

On 8/1/06, John Reid <j.reid at mail.cryst.bbk.ac.uk> wrote:
> Hi,
>
> I use boost::shared_ptr to manage lifetimes of objects. I've run into a
> couple of problems using these inside containers with Boost.Python. I
> thought I could share my experiences and perhaps get some useful
> feedback or help others avoid these problems.
>
> 1) Exposing a vector< shared_ptr< T > > through the indexing suite was
> giving me headaches. I couldn't work out why I couldn't access the
> elements from python. I kept getting a runtime error about no python
> class for boost::shared_ptr< T > even after I was registering the
> pointer type. It turns out that my problem was the vector indexing suite
> was trying to wrap a reference to the boost::shared_ptr< T > in a python
> object. I needed to set the NoProxy template parameter on
> vector_indexing_suite to true and that solved all those issues. Once I
> understood what was happening the answer was clear. Until that moment
> there was a lot of head-scratching. The documentation is factually
> comprehensive but I wonder whether it couldn't have more use cases that
> explain why certain things are necessary.

For next code:

struct item{};

std::vector< boost::shared_ptr< item > > get_items(){
    return std::vector< boost::shared_ptr< item > >();
}

pyplusplus will generate next registration code for the vector:

  bp::vector_indexing_suite< ::std::vector<boost::shared_ptr<item>,
std::allocator<boost::shared_ptr<item> > >, true >()

Do you see "true" at the end? :-)

pyplusplus tries to help a user as much as it can. It can serve as a
user guide and
as an inspector: http://tinyurl.com/mbd4r ( development version ).

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


From j.reid at mail.cryst.bbk.ac.uk  Tue Aug  1 11:59:48 2006
From: j.reid at mail.cryst.bbk.ac.uk (John Reid)
Date: Tue, 01 Aug 2006 10:59:48 +0100
Subject: [C++-sig] Boost.Python shared_ptr container gotchas
In-Reply-To: <7465b6170608010216h794cc164re96a6ceaef6e1e60@mail.gmail.com>
References: <ean4jp$v77$1@sea.gmane.org>
	<7465b6170608010216h794cc164re96a6ceaef6e1e60@mail.gmail.com>
Message-ID: <ean8ml$cf9$1@sea.gmane.org>

Roman Yakovenko wrote:
> pyplusplus will generate next registration code for the vector:
> 
>   bp::vector_indexing_suite< ::std::vector<boost::shared_ptr<item>,
> std::allocator<boost::shared_ptr<item> > >, true >()
> 
> Do you see "true" at the end? :-)
> 
> pyplusplus tries to help a user as much as it can. It can serve as a
> user guide and
> as an inspector: http://tinyurl.com/mbd4r ( development version ).
> 

That sounds good to me. I did try to use pyplusplus but setting up 
gccxml on my windows box proved to be such a headache I gave up and went 
back to Boost.Python. I will probably try in the future when I have more 
time. It wasn't clear from the pyplusplus documentation which version of 
gccxml was required.

Also I forgot to mention in my first post that I think Boost.Python is a 
fantastic library overall. I don't just want to mention the problems 
I've had. :-)

John.



From roman.yakovenko at gmail.com  Tue Aug  1 12:22:11 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Tue, 1 Aug 2006 13:22:11 +0300
Subject: [C++-sig] Boost.Python shared_ptr container gotchas
In-Reply-To: <ean8ml$cf9$1@sea.gmane.org>
References: <ean4jp$v77$1@sea.gmane.org>
	<7465b6170608010216h794cc164re96a6ceaef6e1e60@mail.gmail.com>
	<ean8ml$cf9$1@sea.gmane.org>
Message-ID: <7465b6170608010322r6b8bf4bcu8bb0e87817368b28@mail.gmail.com>

On 8/1/06, John Reid <j.reid at mail.cryst.bbk.ac.uk> wrote:
> Roman Yakovenko wrote:
> > pyplusplus will generate next registration code for the vector:
> >
> >   bp::vector_indexing_suite< ::std::vector<boost::shared_ptr<item>,
> > std::allocator<boost::shared_ptr<item> > >, true >()
> >
> > Do you see "true" at the end? :-)
> >
> > pyplusplus tries to help a user as much as it can. It can serve as a
> > user guide and
> > as an inspector: http://tinyurl.com/mbd4r ( development version ).
> >
>
> That sounds good to me. I did try to use pyplusplus but setting up
> gccxml on my windows box proved to be such a headache I gave up and went
> back to Boost.Python. I will probably try in the future when I have more
> time. It wasn't clear from the pyplusplus documentation which version of
> gccxml was required.

I can help you with that.

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


From sbalasub at qualcomm.com  Wed Aug  2 05:27:24 2006
From: sbalasub at qualcomm.com (Balasubramanyam, Shivakumar)
Date: Tue, 1 Aug 2006 20:27:24 -0700
Subject: [C++-sig] Problem exposing templates
Message-ID: <A70971E99D54CE4BBED4AD2D97C60D8801B3D1@NAEX15.na.qualcomm.com>

Hi,

I have installed the latest cvs version of gccxml, pyplusplus, pygccxml

I am attaching the code to this mail that has the problem.

Files:

Stack.h:
	Has the class template 

BP_Stack.cpp 
	Contains the boost.python code generated by py++

Sconstruct-lib:
	Builds the stack library

Sconstruct:
	Builds the boost-python-stack library

 <<genCode.py>>  <<sconstruct>>  <<sconstruct-lib>>  <<Stack.h>>  
<<BP_Stack.cpp>> 


The compilation fails at line number 29 and 40.

The problem the functional definition generated with
"_GLOBAL__I_Stack.h******

{ //::_GLOBAL__D_Stack.hLZJ4nd
    
        typedef void ( *function_ptr_t )(  );
        
        bp::def( 
            "_GLOBAL__D_Stack.hLZJ4nd"
            , function_ptr_t( &::_GLOBAL__D_Stack.hLZJ4nd )
            , bp::default_call_policies() );
    
    }

    { //::_GLOBAL__I_Stack.hLZJ4nd
    
        typedef void ( *function_ptr_t )(  );
        
        bp::def( 
            "_GLOBAL__I_Stack.hLZJ4nd"
            , function_ptr_t( &::_GLOBAL__I_Stack.hLZJ4nd )
            , bp::default_call_policies() );
    
}

Is there any ways else to explicitly instantiate the template?

The class compiles and runs fine if I remove those two declarations.

Thanks,
Shiva


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060801/912ec513/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: genCode.py
Type: application/octet-stream
Size: 808 bytes
Desc: genCode.py
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060801/912ec513/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: sconstruct
Type: application/octet-stream
Size: 318 bytes
Desc: sconstruct
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060801/912ec513/attachment-0001.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: sconstruct-lib
Type: application/octet-stream
Size: 149 bytes
Desc: sconstruct-lib
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060801/912ec513/attachment-0002.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Stack.h
Type: application/octet-stream
Size: 656 bytes
Desc: Stack.h
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060801/912ec513/attachment-0003.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: BP_Stack.cpp
Type: application/octet-stream
Size: 1241 bytes
Desc: BP_Stack.cpp
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060801/912ec513/attachment-0004.obj>

From roman.yakovenko at gmail.com  Wed Aug  2 07:02:04 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Wed, 2 Aug 2006 08:02:04 +0300
Subject: [C++-sig] Problem exposing templates
In-Reply-To: <A70971E99D54CE4BBED4AD2D97C60D8801B3D1@NAEX15.na.qualcomm.com>
References: <A70971E99D54CE4BBED4AD2D97C60D8801B3D1@NAEX15.na.qualcomm.com>
Message-ID: <7465b6170608012202ne6bf42bhc6f80832ef368cdb@mail.gmail.com>

On 8/2/06, Balasubramanyam, Shivakumar <sbalasub at qualcomm.com> wrote:
> Hi,

Good morning.

> Is there any ways else to explicitly instantiate the template?

Yes, you can write C++ code that uses that functions and feed it to gccxml.

P.S. Also gccxml is involved in the process it has nothing to do with
the problem:
http://gccxml.org/HTML/FAQ.html

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


From jan at langernetz.de  Wed Aug  2 15:28:56 2006
From: jan at langernetz.de (Jan Langer)
Date: Wed, 02 Aug 2006 15:28:56 +0200
Subject: [C++-sig] [boost.python] conversion between string and unsigned
	char *
Message-ID: <eaq9av$ld0$1@sea.gmane.org>

Hello,
I have a library I want to expose to python by using boost.python and
pyplusplus. This library works completely with unsigned char * for
strings. My current approach is

// library

typedef unsigned char * StringT;
StringT F_ApiGetString (F_ObjHandleT docId, F_ObjHandleT objId, IntT
propNum);

// boost.python code

struct StringT_converter
{
  static PyObject *convert (StringT s)
  {
    return bp::incref (bp::str (reinterpret_cast <char *> (s)).ptr ());
  }
};

BOOST_PYTHON_MODULE(pyfdk){
    bp::to_python_converter <StringT, StringT_converter> ();

    bp::def("F_ApiGetString"
            , &::F_ApiGetString
            , ( bp::arg("arg0"), bp::arg("arg1"), bp::arg("arg2") )
            , bp::return_value_policy< bp::return_by_value,
bp::default_call_policies >() );
}

But somehow it does not work:
TypeError: No to_python (by-value) converter found for C++ type: char

Actually, I must free the memory of the returned StringT, but I am not
sure if I can savely change the converter like this:

struct StringT_converter
{
  static PyObject *convert (StringT s)
  {
    PyObject *p = bp::incref (bp::str (reinterpret_cast <char *>
(s)).ptr ());
    F_ApiDeallocate(s);
    return p;
  }
};

I hope someone can help me.
Thanks,
Jan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 252 bytes
Desc: OpenPGP digital signature
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060802/4dc64d6a/attachment.pgp>

From jan at langernetz.de  Wed Aug  2 15:42:11 2006
From: jan at langernetz.de (Jan Langer)
Date: Wed, 02 Aug 2006 15:42:11 +0200
Subject: [C++-sig] Pyplusplus - anonymous union
In-Reply-To: <7465b6170607201021j2ff8e92cy154d70e95c565455@mail.gmail.com>
References: <e9nq1d$smm$1@sea.gmane.org>
	<7465b6170607201021j2ff8e92cy154d70e95c565455@mail.gmail.com>
Message-ID: <eaqa3p$oid$1@sea.gmane.org>

Roman Yakovenko wrote:
>>I'm using pyplusplus to generate python-bindings for a C API.
> 
> mb = module_builder_t( ... )
> #The code in subversion does it for you
> mb.classes( name='' ).exclude()
> 
> Next code is copied from Python-OGRE bindings:
> 
> [...]
> 
> This should work.
> 
> P.S. You project is 3rd project that need this feature, may be this is a time
> to implement it :-)

Thank you. Actually this behavior does not work like I want, but gave me
more insight into the workings of pyplusplus. I now use a different
strategy. First, I generate a union wrapper to give the anonymous union
a name. Then, I provide getter and setter functions for each union
member and getter/setter for the union itself. The access to the union
members and the union is then accomplished by the .add_property calls. I
paste the relevant boost.python code below. However, I omit the
pyplusplus code, but in case someone is interested I can send it, too.

Jan

//-------------------------------------------------
// library code

typedef struct
        {
         IntT offset;
         IntT dataType;
         union {
                StringT sdata;
                IntT idata;
                } u;
        } F_TextItemT;

// boost.python code

union F_TextItemT_union
{
  StringT sdata;
  IntT idata;
};

StringT get_F_TextItemT_union_sdata (union_wrapper <F_TextItemT_union>
const &t)
{
  return t.u.sdata;
}
void set_F_TextItemT_union_sdata (union_wrapper <F_TextItemT_union> &t,
StringT const &v)
{
  t.u.sdata = v;
}

IntT get_F_TextItemT_union_idata (union_wrapper <F_TextItemT_union>
const &t)
{
  return t.u.idata;
}
void set_F_TextItemT_union_idata (union_wrapper <F_TextItemT_union> &t,
IntT const &v)
{
  t.u.idata = v;
}

union_wrapper <F_TextItemT_union> get_F_TextItemT_u (F_TextItemT const &t)
{
  union_wrapper <F_TextItemT_union> uw;
  assert(sizeof(uw.u) == sizeof(uw.u));
  std::memcpy(&uw.u,&t.u,sizeof(uw.u));
  return uw;
}
void set_F_TextItemT_u (F_TextItemT &t, union_wrapper
<F_TextItemT_union> const &uw)
{
  assert(sizeof(uw.u) == sizeof(t.u));
  std::memcpy(&t.u,&uw.u,sizeof(uw.u));
}

BOOST_PYTHON_MODULE(pyfdk){
    if( true ){
        typedef bp::class_< F_TextItemT > F_TextItemT_exposer_t;
        F_TextItemT_exposer_t F_TextItemT_exposer =
F_TextItemT_exposer_t( "F_TextItemT" );
        bp::scope F_TextItemT_scope( F_TextItemT_exposer );
        bp::class_ <union_wrapper <F_TextItemT_union > >
("F_TextItemT_union")
          .add_property ("sdata",
                         bp::make_function
(&get_F_TextItemT_union_sdata,bp::return_value_policy
<bp::return_by_value> ()),
                         &set_F_TextItemT_union_sdata)
          .add_property ("idata",
                         bp::make_function
(&get_F_TextItemT_union_idata,bp::return_value_policy
<bp::return_by_value> ()),
                         &set_F_TextItemT_union_idata)
        ;
        F_TextItemT_exposer.def_readwrite( "offset", &F_TextItemT::offset );
        F_TextItemT_exposer.def_readwrite( "dataType",
&F_TextItemT::dataType );
        F_TextItemT_exposer.add_property("u",
                     &get_F_TextItemT_u,
                     &set_F_TextItemT_u);
    }
}
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 252 bytes
Desc: OpenPGP digital signature
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060802/a5623158/attachment.pgp>

From mkirk at pdi.com  Wed Aug  2 18:32:35 2006
From: mkirk at pdi.com (Mark Kirk)
Date: Wed, 2 Aug 2006 16:32:35 +0000 (UTC)
Subject: [C++-sig] wrapping abstract bases whose virtual functions should
	*not* be overridden in Python
Message-ID: <loom.20060802T181334-10@post.gmane.org>

Hi...

I'm using using boost::python for the first time to wrap a C++ lib.  I'm
extremely impressed with it, and I offer a warm thank-you to Dave and others on
the boost::python team for creating it. 

I've been working with it for a couple of days, and I've read the tutorial and
other bits of the Boost docs;  I think most of it has sunk in.  However, I'm
having trouble with something:

I have an abstract base class (one+ pure virtuals).  I do *not* want any of
these virtual functions overridden from within Python;  I'm quite happy with the
implementations in the C++ derived classes.  This is an in-house lib, and the
C++ implementation is and should remain the canonical implementation.  In the
docs, there are a couple of sections which address abstract bases ("Class
Virtual Functions" subsection in the tutorial and the wrapper.hpp docs).  Both
state the procedure outlined is for wrapping bases for which you wish the
ability to override the virtuals from *within* Python.  Again, I explicitly do
*not* want this ability.  

So, how do I correctly wrap my abstract base?  I get errors complaining about
the the virtual functions being abstract.  Must I still use the wrapper<>
procedure in my situation?  If so, how do I prevent overriding of my virtuals in
Python?  Or, is there another way entirely?

Thanks for your help, and let me know if you need clarification.

mark




From roman.yakovenko at gmail.com  Wed Aug  2 19:26:55 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Wed, 2 Aug 2006 20:26:55 +0300
Subject: [C++-sig] Pyplusplus - anonymous union
In-Reply-To: <eaqa3p$oid$1@sea.gmane.org>
References: <e9nq1d$smm$1@sea.gmane.org>
	<7465b6170607201021j2ff8e92cy154d70e95c565455@mail.gmail.com>
	<eaqa3p$oid$1@sea.gmane.org>
Message-ID: <7465b6170608021026i7fa376a7lcf2d99e984cb4ecd@mail.gmail.com>

On 8/2/06, Jan Langer <jan at langernetz.de> wrote:
> Roman Yakovenko wrote:
> >>I'm using pyplusplus to generate python-bindings for a C API.
> >
> > mb = module_builder_t( ... )
> > #The code in subversion does it for you
> > mb.classes( name='' ).exclude()
> >
> > Next code is copied from Python-OGRE bindings:
> >
> > [...]
> >
> > This should work.
> >
> > P.S. You project is 3rd project that need this feature, may be this is a time
> > to implement it :-)
>
> Thank you. Actually this behavior does not work like I want, but gave me
> more insight into the workings of pyplusplus. I now use a different
> strategy. First, I generate a union wrapper to give the anonymous union
> a name. Then, I provide getter and setter functions for each union
> member and getter/setter for the union itself. The access to the union
> members and the union is then accomplished by the .add_property calls. I
> paste the relevant boost.python code below. However, I omit the
> pyplusplus code, but in case someone is interested I can send it, too.
>
> Jan
>
> //-------------------------------------------------
> // library code
>
> typedef struct
>         {
>          IntT offset;
>          IntT dataType;
>          union {
>                 StringT sdata;
>                 IntT idata;
>                 } u;
>         } F_TextItemT;
>
> // boost.python code
>
> union F_TextItemT_union
> {
>   StringT sdata;
>   IntT idata;
> };
>
> StringT get_F_TextItemT_union_sdata (union_wrapper <F_TextItemT_union>
> const &t)
> {
>   return t.u.sdata;
> }
> void set_F_TextItemT_union_sdata (union_wrapper <F_TextItemT_union> &t,
> StringT const &v)
> {
>   t.u.sdata = v;
> }
>
> IntT get_F_TextItemT_union_idata (union_wrapper <F_TextItemT_union>
> const &t)
> {
>   return t.u.idata;
> }
> void set_F_TextItemT_union_idata (union_wrapper <F_TextItemT_union> &t,
> IntT const &v)
> {
>   t.u.idata = v;
> }
>
> union_wrapper <F_TextItemT_union> get_F_TextItemT_u (F_TextItemT const &t)
> {
>   union_wrapper <F_TextItemT_union> uw;
>   assert(sizeof(uw.u) == sizeof(uw.u));
>   std::memcpy(&uw.u,&t.u,sizeof(uw.u));
>   return uw;
> }
> void set_F_TextItemT_u (F_TextItemT &t, union_wrapper
> <F_TextItemT_union> const &uw)
> {
>   assert(sizeof(uw.u) == sizeof(t.u));
>   std::memcpy(&t.u,&uw.u,sizeof(uw.u));
> }
>
> BOOST_PYTHON_MODULE(pyfdk){
>     if( true ){
>         typedef bp::class_< F_TextItemT > F_TextItemT_exposer_t;
>         F_TextItemT_exposer_t F_TextItemT_exposer =
> F_TextItemT_exposer_t( "F_TextItemT" );
>         bp::scope F_TextItemT_scope( F_TextItemT_exposer );
>         bp::class_ <union_wrapper <F_TextItemT_union > >
> ("F_TextItemT_union")
>           .add_property ("sdata",
>                          bp::make_function
> (&get_F_TextItemT_union_sdata,bp::return_value_policy
> <bp::return_by_value> ()),
>                          &set_F_TextItemT_union_sdata)
>           .add_property ("idata",
>                          bp::make_function
> (&get_F_TextItemT_union_idata,bp::return_value_policy
> <bp::return_by_value> ()),
>                          &set_F_TextItemT_union_idata)
>         ;
>         F_TextItemT_exposer.def_readwrite( "offset", &F_TextItemT::offset );
>         F_TextItemT_exposer.def_readwrite( "dataType",
> &F_TextItemT::dataType );
>         F_TextItemT_exposer.add_property("u",
>                      &get_F_TextItemT_u,
>                      &set_F_TextItemT_u);
>     }
> }
>

Cool! Can you post pyplusplus code? I will create a FAQ entry.

Thank you

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


From roman.yakovenko at gmail.com  Wed Aug  2 19:35:20 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Wed, 2 Aug 2006 20:35:20 +0300
Subject: [C++-sig] wrapping abstract bases whose virtual functions
	should *not* be overridden in Python
In-Reply-To: <loom.20060802T181334-10@post.gmane.org>
References: <loom.20060802T181334-10@post.gmane.org>
Message-ID: <7465b6170608021035s3b582d53h88472908776ac993@mail.gmail.com>

On 8/2/06, Mark Kirk <mkirk at pdi.com> wrote:
> Hi...
>
> I have an abstract base class (one+ pure virtuals).  I do *not* want any of
> these virtual functions overridden from within Python;  I'm quite happy with the
> implementations in the C++ derived classes.  This is an in-house lib, and the
> C++ implementation is and should remain the canonical implementation.  In the
> docs, there are a couple of sections which address abstract bases ("Class
> Virtual Functions" subsection in the tutorial and the wrapper.hpp docs).  Both
> state the procedure outlined is for wrapping bases for which you wish the
> ability to override the virtuals from *within* Python.  Again, I explicitly do
> *not* want this ability.
> So, how do I correctly wrap my abstract base?  I get errors complaining about
> the the virtual functions being abstract.  Must I still use the wrapper<>
> procedure in my situation?  If

No. I think you can just say that class could not be created( no_init )

http://www.boost.org/libs/python/doc/tutorial/doc/html/python/exposing.html

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


From jan at langernetz.de  Wed Aug  2 20:06:08 2006
From: jan at langernetz.de (Jan Langer)
Date: Wed, 02 Aug 2006 20:06:08 +0200
Subject: [C++-sig] Pyplusplus - anonymous union
In-Reply-To: <7465b6170608021026i7fa376a7lcf2d99e984cb4ecd@mail.gmail.com>
References: <e9nq1d$smm$1@sea.gmane.org>	<7465b6170607201021j2ff8e92cy154d70e95c565455@mail.gmail.com>	<eaqa3p$oid$1@sea.gmane.org>
	<7465b6170608021026i7fa376a7lcf2d99e984cb4ecd@mail.gmail.com>
Message-ID: <eaqpin$nnr$1@sea.gmane.org>

Roman Yakovenko wrote:
> On 8/2/06, Jan Langer <jan at langernetz.de> wrote:
>>Roman Yakovenko wrote:
>>>>I'm using pyplusplus to generate python-bindings for a C API.
>>>
>>>mb = module_builder_t( ... )
>>>#The code in subversion does it for you
>>>mb.classes( name='' ).exclude()
>>>
>>>Next code is copied from Python-OGRE bindings:
>>>
>>>[...]
>>>
>>>This should work.
>>>
>>>P.S. You project is 3rd project that need this feature, may be this is a time
>>>to implement it :-)
>>
>>Thank you. Actually this behavior does not work like I want, but gave me
>>more insight into the workings of pyplusplus. I now use a different
>>strategy. First, I generate a union wrapper to give the anonymous union
>>a name. Then, I provide getter and setter functions for each union
>>member and getter/setter for the union itself. The access to the union
>>members and the union is then accomplished by the .add_property calls. I
>>paste the relevant boost.python code below. However, I omit the
>>pyplusplus code, but in case someone is interested I can send it, too.
> 
> [...]
>
> Cool! Can you post pyplusplus code? I will create a FAQ entry.

Hello Roman,
actually, I am not very happy with my code, because I probably do many
things not the right pyplusplus way and not even the right (safe) c++
way (see my other post today). Anyway, I attached the script and I
think I don't need to explain in further detail :-[. The std::memcpy
part of it works for this library, but it probably does not for other
more advanced libraries. In case it is relevant, I am trying to wrap the
FrameMaker FDK library.

The union_wrapper.hpp file basically just contains the following code:

template <typename Union>
struct union_wrapper
{
  typedef Union union_type;
  Union u;
};
-------------- next part --------------
A non-text attachment was scrubbed...
Name: build.py
Type: application/x-python
Size: 9583 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060802/2ee3641e/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 252 bytes
Desc: OpenPGP digital signature
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060802/2ee3641e/attachment.pgp>

From sbalasub at qualcomm.com  Thu Aug  3 02:43:56 2006
From: sbalasub at qualcomm.com (Balasubramanyam, Shivakumar)
Date: Wed, 2 Aug 2006 17:43:56 -0700
Subject: [C++-sig] [GCC-XML] (__GLOBAL function was Problem exposing
	templates)
In-Reply-To: <7465b6170608012202ne6bf42bhc6f80832ef368cdb@mail.gmail.com>
Message-ID: <A70971E99D54CE4BBED4AD2D97C60D8801B3DA@NAEX15.na.qualcomm.com>

>-----Original Message-----
>From: gccxml-bounces+sbalasub=qualcomm.com at gccxml.org [mailto:gccxml-
>bounces+sbalasub=qualcomm.com at gccxml.org] On Behalf Of Roman Yakovenko
>Sent: Tuesday, August 01, 2006 10:02 PM
>To: Development of Python/C++ integration
>Cc: gccxml at gccxml.org
>Subject: [GCC-XML] Re: [C++-sig] Problem exposing templates
>
>On 8/2/06, Balasubramanyam, Shivakumar <sbalasub at qualcomm.com> wrote:
>> Hi,
>
>Good morning.
>
>> Is there any ways else to explicitly instantiate the template?
>
>Yes, you can write C++ code that uses that functions and feed it to
gccxml.
>
>P.S. Also gccxml is involved in the process it has nothing to do with
>the problem:
>http://gccxml.org/HTML/FAQ.html

[shiva] Roman, I understood the second part of my question, which was
using typedef for explicit instantiation.

Now I declared using "typedef Stack<int>::Stack Stack_Int;"

I still see pyplusplus generate the _GLOBAL function (see below) which
results in compilation error? 

    { //::_GLOBAL__D_Stack.hLZJ4nd
    
        typedef void ( *function_ptr_t )(  );
        
        bp::def( 
            "_GLOBAL__D_Stack.hLZJ4nd"
            , function_ptr_t( &::_GLOBAL__D_Stack.hLZJ4nd )
            , bp::default_call_policies() );
    
    }

    { //::_GLOBAL__I_Stack.hLZJ4nd
    
        typedef void ( *function_ptr_t )(  );
        
        bp::def( 
            "_GLOBAL__I_Stack.hLZJ4nd"
            , function_ptr_t( &::_GLOBAL__I_Stack.hLZJ4nd )
            , bp::default_call_policies() );
    
    }

Thanks,
Shiva





From roman.yakovenko at gmail.com  Thu Aug  3 06:51:02 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 3 Aug 2006 07:51:02 +0300
Subject: [C++-sig] [GCC-XML] (__GLOBAL function was Problem exposing
	templates)
In-Reply-To: <A70971E99D54CE4BBED4AD2D97C60D8801B3DA@NAEX15.na.qualcomm.com>
References: <7465b6170608012202ne6bf42bhc6f80832ef368cdb@mail.gmail.com>
	<A70971E99D54CE4BBED4AD2D97C60D8801B3DA@NAEX15.na.qualcomm.com>
Message-ID: <7465b6170608022151o704a3cabna565cc798b2ff74a@mail.gmail.com>

On 8/3/06, Balasubramanyam, Shivakumar <sbalasub at qualcomm.com> wrote:
> [shiva] Roman, I understood the second part of my question, which was
> using typedef for explicit instantiation.
>
> Now I declared using "typedef Stack<int>::Stack Stack_Int;"

This does not couse to template instantiation.

Please write:

Stack_Int s;

Also I will fix py++ to skip those declarations

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


From bjorn at 4roald.org  Thu Aug  3 07:34:55 2006
From: bjorn at 4roald.org (=?ISO-8859-1?Q?Bj=F8rn_Roald?=)
Date: Thu, 03 Aug 2006 07:34:55 +0200
Subject: [C++-sig] [GCC-XML] (__GLOBAL function was Problem exposing
	templates)
In-Reply-To: <7465b6170608022151o704a3cabna565cc798b2ff74a@mail.gmail.com>
References: <7465b6170608012202ne6bf42bhc6f80832ef368cdb@mail.gmail.com>
	<A70971E99D54CE4BBED4AD2D97C60D8801B3DA@NAEX15.na.qualcomm.com>
	<7465b6170608022151o704a3cabna565cc798b2ff74a@mail.gmail.com>
Message-ID: <44D18AFF.7060903@4roald.org>

Roman Yakovenko wrote:
> On 8/3/06, Balasubramanyam, Shivakumar <sbalasub at qualcomm.com> wrote:
>> [shiva] Roman, I understood the second part of my question, which was
>> using typedef for explicit instantiation.
>>
>> Now I declared using "typedef Stack<int>::Stack Stack_Int;"
>
> This does not couse to template instantiation.
>
> Please write:
>
> Stack_Int s;
Why do you need the typdef?

Stack<int>::Stack Stack_Int;

Should do it!  Or am I missing some other purpose of the typedef?

 Bjorn


From roman.yakovenko at gmail.com  Thu Aug  3 13:20:12 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 3 Aug 2006 14:20:12 +0300
Subject: [C++-sig] [boost.python] conversion between string and unsigned
	char *
In-Reply-To: <eaq9av$ld0$1@sea.gmane.org>
References: <eaq9av$ld0$1@sea.gmane.org>
Message-ID: <7465b6170608030420p36156610n4c525ec931f100f0@mail.gmail.com>

On 8/2/06, Jan Langer <jan at langernetz.de> wrote:
> Hello,
> I have a library I want to expose to python by using boost.python and
> pyplusplus. This library works completely with unsigned char * for
> strings. My current approach is
>
> // library
>
> typedef unsigned char * StringT;
> StringT F_ApiGetString (F_ObjHandleT docId, F_ObjHandleT objId, IntT
> propNum);
>
> // boost.python code
>
> struct StringT_converter
> {
>   static PyObject *convert (StringT s)
>   {
>     return bp::incref (bp::str (reinterpret_cast <char *> (s)).ptr ());
>   }
> };
>
> BOOST_PYTHON_MODULE(pyfdk){
>     bp::to_python_converter <StringT, StringT_converter> ();

I think this line does not do what you expect. You are trying to
register custom
converter for C++ fundamental type. I could be wrong, but it seems to
me that boost.python
uses different approach ( type conversion ) when it comes to
fundamental C++ types.


If I right, than pyplusplus can help you.
1. For every function, that uses StringT generate small wrapper,
replacing StringT with
   std::string or custom string class(
http://boost.org/libs/python/doc/v2/faq.html#custom_string )

2. Export those wrappers. Consider to run pyplusplus on generated
files, instead of
   adding code to the module builder.

> I hope someone can help me.

Does it help?

> Thanks,
> Jan


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


From jan at langernetz.de  Thu Aug  3 17:21:33 2006
From: jan at langernetz.de (Jan Langer)
Date: Thu, 03 Aug 2006 17:21:33 +0200
Subject: [C++-sig] [boost.python] conversion between string and unsigned
 char *
In-Reply-To: <7465b6170608030420p36156610n4c525ec931f100f0@mail.gmail.com>
References: <eaq9av$ld0$1@sea.gmane.org>
	<7465b6170608030420p36156610n4c525ec931f100f0@mail.gmail.com>
Message-ID: <eat4a4$uqo$1@sea.gmane.org>

Roman Yakovenko wrote:
> On 8/2/06, Jan Langer <jan at langernetz.de> wrote:
>>I have a library I want to expose to python by using boost.python and
>>pyplusplus. This library works completely with unsigned char * for
>>strings. My current approach is
>>
>>// library
>>
>>typedef unsigned char * StringT;
>>StringT F_ApiGetString (F_ObjHandleT docId, F_ObjHandleT objId, IntT
>>propNum);
>>
>>// boost.python code
>>
>>struct StringT_converter
>>{
>>  static PyObject *convert (StringT s)
>>  {
>>    return bp::incref (bp::str (reinterpret_cast <char *> (s)).ptr ());
>>  }
>>};
>>
>>BOOST_PYTHON_MODULE(pyfdk){
>>    bp::to_python_converter <StringT, StringT_converter> ();
> 
> I think this line does not do what you expect. You are trying to
> register custom
> converter for C++ fundamental type. I could be wrong, but it seems to
> me that boost.python
> uses different approach ( type conversion ) when it comes to
> fundamental C++ types.

Actually, that was my intention, and I wanted to achieve it with the
to_python_converter. But I probably did not understand the concept
correctly.

> If I right, than pyplusplus can help you.
> 1. For every function, that uses StringT generate small wrapper,
> replacing StringT with
>    std::string or custom string class(
> http://boost.org/libs/python/doc/v2/faq.html#custom_string )

Okay, thank you, I will try to do that.

> 2. Export those wrappers. Consider to run pyplusplus on generated
> files, instead of
>    adding code to the module builder.

Can you please explain in more detail. Do you mean a two-stage
generation process with pyplusplus?

Regards,
Jan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 252 bytes
Desc: OpenPGP digital signature
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060803/2516ab2d/attachment.pgp>

From roman.yakovenko at gmail.com  Thu Aug  3 19:34:09 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 3 Aug 2006 20:34:09 +0300
Subject: [C++-sig] [boost.python] conversion between string and unsigned
	char *
In-Reply-To: <eat4a4$uqo$1@sea.gmane.org>
References: <eaq9av$ld0$1@sea.gmane.org>
	<7465b6170608030420p36156610n4c525ec931f100f0@mail.gmail.com>
	<eat4a4$uqo$1@sea.gmane.org>
Message-ID: <7465b6170608031034n69ae9ab7r8b02000587151aae@mail.gmail.com>

On 8/3/06, Jan Langer <jan at langernetz.de> wrote:
> >>BOOST_PYTHON_MODULE(pyfdk){
> >>    bp::to_python_converter <StringT, StringT_converter> ();
> >
> > I think this line does not do what you expect. You are trying to
> > register custom
> > converter for C++ fundamental type. I could be wrong, but it seems to
> > me that boost.python
> > uses different approach ( type conversion ) when it comes to
> > fundamental C++ types.
>
> Actually, that was my intention, and I wanted to achieve it with the
> to_python_converter. But I probably did not understand the concept
> correctly.

No, I think that you understand it right, boost.python treats fundamental type
in special way and you can not change this.

> > If I right, than pyplusplus can help you.
> > 1. For every function, that uses StringT generate small wrapper,
> > replacing StringT with
> >    std::string or custom string class(
> > http://boost.org/libs/python/doc/v2/faq.html#custom_string )
>
> Okay, thank you, I will try to do that.
>
> > 2. Export those wrappers. Consider to run pyplusplus on generated
> > files, instead of
> >    adding code to the module builder.
>
> Can you please explain in more detail. Do you mean a two-stage
> generation process with pyplusplus?

Yes. First run you generate wrappers, second run you expose those wrappers.

> Regards,
> Jan

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


From roberto.c.aguilar at gmail.com  Thu Aug  3 21:25:32 2006
From: roberto.c.aguilar at gmail.com (Roberto Aguilar)
Date: Thu, 3 Aug 2006 12:25:32 -0700
Subject: [C++-sig] [Boost.Python] Using bbv2 to build simple extension
Message-ID: <e190e3480608031225q429ae0d0q51a32bef04c2a572@mail.gmail.com>

Hello,

I'm trying to setup my build environment to properly build a python
extension outside of the boost tree using Boost.Build v2.

So far I have this tree setup:

top/
  |
  +-- Jamroot
  |
  +-- lib/
  |    |
  |    +-- pytest
  |    |
  |    `-- Jamfile
  |    `-- python.cpp

My Jamroot file looks like:
----
.boost-root = /Users/berto/apps/opt/boost_1_33_1 ;

use-project /boost : $(.boost-root) ;

using python : 2.4 : /sw ;

project myproj
	: requirements <link>static
	;
----

My Jamfile:
----
project :
	usage-requirements <include>.
	;

[...]

python-extension fileformats_py
		 :
		 python.cpp
		 :
		 <library>/boost/python//boost_python
		 ;
----

In the jamfile I also have some library declarations, but they're
irrelevant to this.

When I run bjam, I get this:

[berto at drjekyll][1108]$ bjam
Building Boost.Regex with the optional Unicode/ICU support disabled.
Please refer to the Boost.Regex documentation for more information
(and if you don't know what ICU is then you probably don't need it).
error: Unable to find file or target named
error:     '/python//python_for_extensions
error: referred from project aterror:     '.'

My guess is that I'm missing some sort of declaration in my Jamroot file.

Any help is much appreciated!

Thanks!
-Roberto.


From investtcartier at yahoo.com  Thu Aug  3 21:52:01 2006
From: investtcartier at yahoo.com (Kevin Jones)
Date: Thu, 3 Aug 2006 12:52:01 -0700 (PDT)
Subject: [C++-sig] Converter with return_value_policy<manage_new_object>
	semantics...
Message-ID: <20060803195202.92517.qmail@web55501.mail.re4.yahoo.com>

Hi all,

I have a function T foo(). The problem is that T is a
variant type and by default is being converted to a
python object using my supplied converter.

I would like to rewrite the converter to use
return_value_policy<manage_new_object> semantics on a
case by case basis for each variant type.

Any suggestions?

Best regards,

Kevin


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 


From roman.yakovenko at gmail.com  Thu Aug  3 22:04:33 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 3 Aug 2006 23:04:33 +0300
Subject: [C++-sig] Converter with return_value_policy<manage_new_object>
	semantics...
In-Reply-To: <20060803195202.92517.qmail@web55501.mail.re4.yahoo.com>
References: <20060803195202.92517.qmail@web55501.mail.re4.yahoo.com>
Message-ID: <7465b6170608031304g53da4120t3535692f2edf0127@mail.gmail.com>

On 8/3/06, Kevin Jones <investtcartier at yahoo.com> wrote:
> Hi all,
>
> I have a function T foo(). The problem is that T is a
> variant type and by default is being converted to a
> python object using my supplied converter.
>
> I would like to rewrite the converter to use
> return_value_policy<manage_new_object> semantics on a
> case by case basis for each variant type.
>
> Any suggestions?

You can create custom call policies.


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


From mharidev at qualcomm.com  Fri Aug  4 03:14:05 2006
From: mharidev at qualcomm.com (Haridev, Meghana)
Date: Thu, 3 Aug 2006 18:14:05 -0700
Subject: [C++-sig] Pyplusplus: Exporting arrays of non-copyable type?
Message-ID: <5383A81D5EB8F14FBDB487D9B5FBEA880167D603@NAEX11.na.qualcomm.com>

Hi Folks,

 

I'm using pyplusplus to generate boost.python wrappers for the following
piece of C++ code:

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

//test.cpp

 

class Base {

protected:

    Base() { }

private:

   // Copy constructor and assignment operator should not

   // be used.

   Base(const Base&);

   const Base& operator=(const Base&);

};

 

class Derv: public Base {

public:

   int x;

};

 

class Test: public Base {

public:

    Derv dataArray[20];

};

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

 

In class Test, the dataArray if of type Derv which is a non-copyable
type (because of private copy constructor and assignment operator of
Base class). In other words, this is not allowed:

 

>> obj = Test()

>> obj.dataArray[0] = Derv() #Not allowed

 

But the pyplusplus generated code (register_array_1() in
__array_1.pypp.hpp) exposes a set_item method for the data array though
it's type is non-copyable and hence the code does not compile with the
following error message:

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

test.cpp:8: error: 'const Base& Base::operator=(const Base&)' is private

test.cpp:11: error: within this context

__array_1.pypp.hpp: In member function 'void
pyplusplus::containers::static_sized::array_1_t<TItemType,
size>::set_item(long unsigned int, typename boost::call_traits<const
TItemType>::param_type) [with TItemType = Derv, long unsigned int size =
20ul]':

__array_1.pypp.hpp:73: warning: synthesized method 'Derv&
Derv::operator=(const Derv&)' first required here

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

 

Please see attached source/generated files.

 

Is there any way I can make pyplusplus to NOT generate the set_item
method for arrays of non-copyable type? (maybe by setting some flag?)

 

Thanks,

-Meghana.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060803/e76f561f/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: test.cpp
Type: application/octet-stream
Size: 300 bytes
Desc: test.cpp
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060803/e76f561f/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: generate_code.py
Type: application/octet-stream
Size: 1066 bytes
Desc: generate_code.py
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060803/e76f561f/attachment-0001.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: PYPP_YMI.cpp
Type: application/octet-stream
Size: 1591 bytes
Desc: PYPP_YMI.cpp
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060803/e76f561f/attachment-0002.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: __array_1.pypp.hpp
Type: application/octet-stream
Size: 3149 bytes
Desc: __array_1.pypp.hpp
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060803/e76f561f/attachment-0003.obj>

From roman.yakovenko at gmail.com  Fri Aug  4 14:48:19 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Fri, 4 Aug 2006 14:48:19 +0200
Subject: [C++-sig] Pyplusplus: Exporting arrays of non-copyable type?
In-Reply-To: <5383A81D5EB8F14FBDB487D9B5FBEA880167D603@NAEX11.na.qualcomm.com>
References: <5383A81D5EB8F14FBDB487D9B5FBEA880167D603@NAEX11.na.qualcomm.com>
Message-ID: <7465b6170608040548q65f0eb99u56f19dad10e6474@mail.gmail.com>

On 8/4/06, Haridev, Meghana <mharidev at qualcomm.com> wrote:
> I'm using pyplusplus to generate boost.python wrappers for the following
> piece of C++ code:
> ...
>
> In class Test, the dataArray if of type Derv which is a non-copyable type
> (because of private copy constructor and assignment operator of Base class).
> In other words, this is not allowed:

> But the pyplusplus generated code (register_array_1() in __array_1.pypp.hpp)
> exposes a set_item method for the data array though it's type is
> non-copyable and hence the code does not compile with the following error
> message:
>
>
> Is there any way I can make pyplusplus to NOT generate the set_item method
> for arrays of non-copyable type? (maybe by setting some flag?)

Unfortunately no. I did not thought about this. You have 2 choices:
1. To add this functionality to pyplusplus
2. To exclude the variable from generation and to add hand written code.


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


From lee.will at gmail.com  Fri Aug  4 16:30:03 2006
From: lee.will at gmail.com (Will Lee)
Date: Fri, 4 Aug 2006 09:30:03 -0500
Subject: [C++-sig]  Overriding operator= in boost python
Message-ID: <7f03db650608040730ga70b414ieb0e265ea2db2bf@mail.gmail.com>

Hi,

Is there any way to wrap the operator= for a C++ assignment operator in
boost python?  The documentation on operators doesn't seem to allow that (
http://www.boost.org/libs/python/doc/v2/operators.html).  Essentially, I
would like to do the python equalivalent of:

class A;

class B : public A;

I then have:

B& operator=(const A& a)
{
// assign A to B
}

So that I can do (say bHolder is a class that contains a B instance):

a = A();
bHolder.b = a;

Thanks,

Will
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060804/85d780ca/attachment.htm>

From seefeld at sympatico.ca  Fri Aug  4 16:46:31 2006
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Fri, 04 Aug 2006 10:46:31 -0400
Subject: [C++-sig] Overriding operator= in boost python
In-Reply-To: <7f03db650608040730ga70b414ieb0e265ea2db2bf@mail.gmail.com>
References: <7f03db650608040730ga70b414ieb0e265ea2db2bf@mail.gmail.com>
Message-ID: <44D35DC7.7050001@sympatico.ca>

Will Lee wrote:
> Hi,
> 
> Is there any way to wrap the operator= for a C++ assignment operator in
> boost python?  The documentation on operators doesn't seem to allow that (
> http://www.boost.org/libs/python/doc/v2/operators.html).  Essentially, I
> would like to do the python equalivalent of:
> 
> class A;
> 
> class B : public A;
> 
> I then have:
> 
> B& operator=(const A& a)
> {
> // assign A to B
> }
> 
> So that I can do (say bHolder is a class that contains a B instance):
> 
> a = A();
> bHolder.b = a;

The statement

c.b = a

in python will call c's '__setattr__' method, not, as in C++, c.b's
assignment operator. This is because the above changes the binding
of 'c.b', not the state of the object 'c.b' refers to.

Regards,
		Stefan

-- 

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


From gsculpt at googlemail.com  Sat Aug  5 11:57:05 2006
From: gsculpt at googlemail.com (Geoffrey French)
Date: Sat, 5 Aug 2006 10:57:05 +0100
Subject: [C++-sig] Calling list.__len__() from C++ causes crash with >= 100
	elements
Message-ID: <562c63b10608050257v64ada668u82e3d359c8054ecf@mail.gmail.com>

Hi,

I am currently running Python 2.4.3 and Boost 1.33.1-5 on Fedora Core 5.
I have prepared some test code which seems to cause a seg fault when
C++ code attempts to get the length of a list which has 100 or more
elements.
Valgrind seems to report it as an attempt to dereference a NULL pointer.
An application that I have written (http://gsculpt.sf.net) displays the same
behavior when compiled for Fedora Core 5, and Windows.

Here is the code for a C++ module to demonstrate this:


======================== boostpythonbug.cpp ========================
#include <iostream>

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



void printListLength(boost::python::object elements)
{
    boost::python::extract<int> lenExtract( elements.attr( "__len__" )() );
    if ( lenExtract.check() )
    {
        int numElements = lenExtract;
        std::cout << numElements << " elements" << std::endl;
    }
}

void printValue(boost::python::object obj)
{
    boost::python::extract<int> vExtract( obj.attr( "value" )() );
    if ( vExtract.check() )
    {
        int v = vExtract;
        std::cout << "value = " << v << std::endl;
    }
}

BOOST_PYTHON_MODULE(ListLength)
{
    def( "printListLength", &printListLength );
    def( "printValue", &printValue );
}
======================== boostpythonbug.cpp ========================

To compile:
$ c++ boostpythonbug.cpp --shared -o ListLength.so
-lpython2.4-lboost_python -I/usr/include/python2.4



The python module below tests this module:

======================== boostpythonbug.py ========================
from ListLength import *

class Test (object):
    def __init__(self, x):
        self.x = x

    def value(self):
        return self.x

    def __len__(self):
        return self.x



printValue( Test( 5 ) )
printValue( Test( 99 ) )
printValue( Test( 100 ) )
printValue( Test( 101 ) )

printListLength( Test( 5 ) )
printListLength( Test( 99 ) )
printListLength( Test( 100 ) )
printListLength( Test( 101 ) )

printListLength( [] )
printListLength( [ 0 ] )
printListLength( range( 0, 5 ) )
printListLength( range( 0, 99 ) )
printListLength( range( 0, 100 ) )
======================== boostpythonbug.py ========================


Here is the output:

$ python boostpythonbug.py
value = 5
value = 99
value = 100
value = 101
5 elements
99 elements
100 elements
101 elements
0 elements
1 elements
5 elements
99 elements
Segmentation fault (core dumped)



Getting integer values from a generic function call works fine,
as demonstrated by printValue(), and printListLength() invoked
on a 'Test' object.
It only seems to affect lists.

Defining printListLength as:
void printListLength(boost::python::object elements)
OR
void printListLength(boost::python::list elements)
does not prevent the seg-faulting.
The second definition means you can only pass lists though,
i.e. the cases where a 'Test' object is passed will not work.


Is this a boost::python bug, or have I screwed up some how?
Thank you for your time.



Regards

Geoffrey French
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060805/1ddab512/attachment.htm>

From jorge.vargas at gmail.com  Sat Aug  5 18:57:34 2006
From: jorge.vargas at gmail.com (Jorge Vargas)
Date: Sat, 5 Aug 2006 12:57:34 -0400
Subject: [C++-sig] boost.org broken link to this list
Message-ID: <32822fe60608050957u58768f8fodef47526a58744db@mail.gmail.com>

Hello I'm new here and I was looking for boost support when I find out their
link is broken.


http://www.boost.org/more/mailing_lists.htm#projects

broken link - http://www.python.org/community/sigs/c++-sig/
new link - http://www.python.org/community/sigs/current/c++-sig/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060805/5ee018ec/attachment.htm>

From jorge.vargas at gmail.com  Sat Aug  5 19:32:41 2006
From: jorge.vargas at gmail.com (Jorge Vargas)
Date: Sat, 5 Aug 2006 13:32:41 -0400
Subject: [C++-sig] can't build boost.python under msys/windows
Message-ID: <32822fe60608051032o4993e584r864d5cb1d75a0807@mail.gmail.com>

hi

Could someone stop what I'm doing wrong?

$ echo $PATH
.:/usr/local/bin:/mingw/bin:/bin:/c/WINDOWS/system32:/c/WINDOWS:/c/WINDOWS/System32/Wbem:/c/Program
Files/Subversion/bin:/c/Python24:/c/Python24/Scripts

$ bjam "-sTOOLS=mgw" stage
$ bjam "-sTOOLS=mingw" stage

both commands fail doing a lot of lookups for vc-7_1 and the .Net framework
(I have those install)
CALL "C:\Program Files\Microsoft Visual Studio .NET
2003\VC7\bin\VCVARS32.BAT" >nul
...skipped
<@boost!libs!python!build\libboost_python.lib\vc-7_1\debug\threading-multi>libboost_python.CMD
for lack of
<@boost!libs!python!build\libboost_python.lib\vc-7_1\debug\threading-multi>
numeric.obj...

------

so I take them out of the path

$ echo $PATH
.:/usr/local/bin:/mingw/bin:/bin:/c/Python24:/c/Python24/Scripts

$ bjam "-sTOOLS=mgw" stage


and it fails to do anything
$ bjam "-sTOOLS=mgw" stage
...found 944 targets...
...updating 121 targets...
vc-C++
..\..\..\bin\boost\libs\python\build\boost_python.dll\vc-7_1\debug\threading-multi\numeric.obj
spawn: No such file or directory

so basically bjam is ignoring the TOOLS flag

I have to say I downloaded the binary for windows (
boost-jam-3.1.13-1-ntx86.zip)
and all boostJam related links seems to be broken
http://www.boost.org/doc/html/jam/building.html
http://www.boost.org/tools/jam/index.html

what I'm actually trying to do is porting this
https://develop.participatoryculture.org/projects/democracy/wiki/WindowsBuildDocs
to compile with free tools so at the moment what I need is a replacement for
boost_python-vc71-mt-1_33.dll
that works with mingw does anyone distributes this as binary?

thanks for the help.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060805/37adbd2d/attachment.htm>

From markus at relix.de  Sun Aug  6 01:58:12 2006
From: markus at relix.de (Markus Heller)
Date: Sun, 6 Aug 2006 01:58:12 +0200
Subject: [C++-sig] Working with python unicode strings without wstring
In-Reply-To: <7465b6170607312317y7be98c4fh87d650a74e0589ce@mail.gmail.com>
References: <BAY108-DAV765F689103A7BED1FE3D5B4B10@phx.gbl>
	<200608010223.09974.markus@relix.de>
	<7465b6170607312317y7be98c4fh87d650a74e0589ce@mail.gmail.com>
Message-ID: <200608060158.12985.markus@relix.de>

Dear Roman,

> > > May be you should use another set of API?
> > > http://docs.python.org/api/unicodeObjects.html

well, I have done some experiments with the API you're suggesting. Though, I 
have a number of problems here:

I see that Py_UNICODE refers to one single character / symbol which may be any 
symbol whatsoever in its numeric representation. 

I see that when I retrieve my unicoded Python string with 

PyContainer = PyList_GetItem(Tuple, 1);

and when I convert the PyObject "PyContainer" into a PyUnicodeObject, I do it 
with a simple cast, which works fine. Then, I'd like to get hold of the 
individual symbols, also in their numeric (not encoded) representation. But 
how can I do this? Initially, I wanted to get everything into the form of a 
stl string, but a stl vector will also do fine. 

Thanks very much for your help!

Markus



From rwgk at yahoo.com  Sun Aug  6 08:08:52 2006
From: rwgk at yahoo.com (Ralf W. Grosse-Kunstleve)
Date: Sat, 5 Aug 2006 23:08:52 -0700 (PDT)
Subject: [C++-sig] boost.org broken link to this list
In-Reply-To: <32822fe60608050957u58768f8fodef47526a58744db@mail.gmail.com>
Message-ID: <20060806060852.51367.qmail@web31505.mail.mud.yahoo.com>

--- Jorge Vargas <jorge.vargas at gmail.com> wrote:

> Hello I'm new here and I was looking for boost support when I find out their
> link is broken.
> 
> 
> http://www.boost.org/more/mailing_lists.htm#projects
> 
> broken link - http://www.python.org/community/sigs/c++-sig/
> new link - http://www.python.org/community/sigs/current/c++-sig/

Thanks for pointing this out! However, the broken link was fixed a while ago in
the boost CVS. Unfortunately we have to wait for the next Boost release before
the fix is visible in the main page.


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 


From rwgk at yahoo.com  Sun Aug  6 08:48:47 2006
From: rwgk at yahoo.com (Ralf W. Grosse-Kunstleve)
Date: Sat, 5 Aug 2006 23:48:47 -0700 (PDT)
Subject: [C++-sig] Calling list.__len__() from C++ causes crash with >=
	100 elements
In-Reply-To: <562c63b10608050257v64ada668u82e3d359c8054ecf@mail.gmail.com>
Message-ID: <20060806064847.33668.qmail@web31512.mail.mud.yahoo.com>

--- Geoffrey French <gsculpt at googlemail.com> wrote:
> void printListLength(boost::python::object elements)
> {
>     boost::python::extract<int> lenExtract( elements.attr( "__len__" )() );
>     if ( lenExtract.check() )
>     {
>         int numElements = lenExtract;
>         std::cout << numElements << " elements" << std::endl;
>     }
> }

I can reproduce your problem under FC3 and FC5 (both x86_64). I am not sure if
this is a bug or a feature. In any case, this workaround does the trick for me:

void printListLength(boost::python::object elements)
{
    boost::python::object len_result = elements.attr( "__len__" )();
    boost::python::extract<int> lenExtract( len_result );
    if ( lenExtract.check() )
    {
        int numElements = lenExtract;
        std::cout << numElements << " elements" << std::endl;
    }
}

BTW: With the latest boost CVS you can simply write

  std::cout << len(elements) << " elements" << std::endl;


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 


From roman.yakovenko at gmail.com  Sun Aug  6 11:45:56 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Sun, 6 Aug 2006 12:45:56 +0300
Subject: [C++-sig] Working with python unicode strings without wstring
In-Reply-To: <200608060158.12985.markus@relix.de>
References: <BAY108-DAV765F689103A7BED1FE3D5B4B10@phx.gbl>
	<200608010223.09974.markus@relix.de>
	<7465b6170607312317y7be98c4fh87d650a74e0589ce@mail.gmail.com>
	<200608060158.12985.markus@relix.de>
Message-ID: <7465b6170608060245x2c5b2f52w9e443ad5dabe3226@mail.gmail.com>

On 8/6/06, Markus Heller <markus at relix.de> wrote:
> Dear Roman,
>
> > > > May be you should use another set of API?
> > > > http://docs.python.org/api/unicodeObjects.html
>
> well, I have done some experiments with the API you're suggesting. Though, I
> have a number of problems here:
>
> I see that Py_UNICODE refers to one single character / symbol which may be any
> symbol whatsoever in its numeric representation.
>
> I see that when I retrieve my unicoded Python string with
>
> PyContainer = PyList_GetItem(Tuple, 1);
>
> and when I convert the PyObject "PyContainer" into a PyUnicodeObject, I do it
> with a simple cast, which works fine. Then, I'd like to get hold of the
> individual symbols, also in their numeric (not encoded) representation. But
> how can I do this? Initially, I wanted to get everything into the form of a
> stl string, but a stl vector will also do fine.
>
> Thanks very much for your help!

I did not help you. I think you will be more productive if you will
ask this question
on Python user group.

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


From jorge.vargas at gmail.com  Sun Aug  6 15:42:52 2006
From: jorge.vargas at gmail.com (Jorge Vargas)
Date: Sun, 6 Aug 2006 09:42:52 -0400
Subject: [C++-sig] boost.org broken link to this list
In-Reply-To: <20060806060852.51367.qmail@web31505.mail.mud.yahoo.com>
References: <32822fe60608050957u58768f8fodef47526a58744db@mail.gmail.com>
	<20060806060852.51367.qmail@web31505.mail.mud.yahoo.com>
Message-ID: <32822fe60608060642g39ff4648qf86462c17b858059@mail.gmail.com>

On 8/6/06, Ralf W. Grosse-Kunstleve <rwgk at yahoo.com> wrote:
>
> --- Jorge Vargas <jorge.vargas at gmail.com> wrote:
>
> > Hello I'm new here and I was looking for boost support when I find out
> their
> > link is broken.
> >
> >
> > http://www.boost.org/more/mailing_lists.htm#projects
> >
> > broken link - http://www.python.org/community/sigs/c++-sig/
> > new link - http://www.python.org/community/sigs/current/c++-sig/
>
> Thanks for pointing this out! However, the broken link was fixed a while
> ago in
> the boost CVS. Unfortunately we have to wait for the next Boost release
> before
> the fix is visible in the main page.


ohhh i see autogenerated docs :)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060806/67d8add5/attachment.htm>

From dave at boost-consulting.com  Sun Aug  6 19:54:08 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Sun, 06 Aug 2006 13:54:08 -0400
Subject: [C++-sig] Calling list.__len__() from C++ causes crash with >=
	100 elements
References: <562c63b10608050257v64ada668u82e3d359c8054ecf@mail.gmail.com>
Message-ID: <uzmehyclb.fsf@boost-consulting.com>

"Geoffrey French" <gsculpt at googlemail.com> writes:

> Here is the code for a C++ module to demonstrate this:
>

<snip>

> Is this a boost::python bug, or have I screwed up some how?
> Thank you for your time.

What is the *smallest* code that demonstrates it?

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From dave at boost-consulting.com  Sun Aug  6 19:55:14 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Sun, 06 Aug 2006 13:55:14 -0400
Subject: [C++-sig] boost.org broken link to this list
References: <32822fe60608050957u58768f8fodef47526a58744db@mail.gmail.com>
	<20060806060852.51367.qmail@web31505.mail.mud.yahoo.com>
Message-ID: <uu04pycjh.fsf@boost-consulting.com>

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

> --- Jorge Vargas <jorge.vargas at gmail.com> wrote:
>
>> Hello I'm new here and I was looking for boost support when I find out their
>> link is broken.
>> 
>> 
>> http://www.boost.org/more/mailing_lists.htm#projects
>> 
>> broken link - http://www.python.org/community/sigs/c++-sig/
>> new link - http://www.python.org/community/sigs/current/c++-sig/
>
> Thanks for pointing this out! However, the broken link was fixed a while ago in
> the boost CVS. Unfortunately we have to wait for the next Boost release before
> the fix is visible in the main page.

Ralf, you _can_ update the site directly:
http://boost.org/more/updating_the_website.html


-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From cablepuff at gmail.com  Sun Aug  6 21:03:24 2006
From: cablepuff at gmail.com (chun ping wang)
Date: Sun, 6 Aug 2006 12:03:24 -0700
Subject: [C++-sig] boost python compiling question.
Message-ID: <7253f6b30608061203v135e867ex3db9e7e2dad0d565@mail.gmail.com>

Hmm k i have a simple question, how would I build boost.python given this
following information.

I am using dev-cpp and its install in "c:\dev-cpp" folder.

IN it contains the mingw32 sub directory

My project folder is "c:\cs classes\cs386\c++".

In it contains FiniteAutomta.cpp which maps to python module...

my boost folder is "C:\boost_1_33_1" in it contains the bjam.exe file.

PYTHON_ROOT =  "C:\python24"

Thanks for the help. I look at the tutorial but is confused on how I can
build my project.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060806/790029cc/attachment.htm>

From michael_gruner at arcor.de  Mon Aug  7 13:58:55 2006
From: michael_gruner at arcor.de (michael_gruner at arcor.de)
Date: Mon, 7 Aug 2006 13:58:55 +0200 (CEST)
Subject: [C++-sig] [boost.python] void ptr problem
Message-ID: <20872957.1154951935894.JavaMail.ngmail@webmail14>

Hello, 


I'm trying to use boost.python to wrap a C function that takes a pointer on a void pointer as an argument. But I'm going wrong to use the example from the FAQ. Could anyone help me on this? 


my code looks like this: 


struct void_; 
BOOST_PYTHON_OPAQUE_SPECIALIZED_TYPE_ID(void_); 


int myfunc(char * mychar, void ** mypointer); 


int myfunc_wrp(char * mychar, void_ ** mypointer){ 
        return myfunc(mychar,(void ** ) mypointer); 
} 


... 


BOOST_PYTHON_MODULE(mymodule){ 


        def("myfunc", &myfunc_wrp); 
} 


later in python I do: 


import mymodule 


mystring = " " 
mypointerobject = 0 


mymodule.myfunc(mystring,mypointerobject) 


and get: 
Boost.Python.ArgumentError: Python argument types in mymodule.myfunc(str, int) 
did not match C++ signature: 
        myfunc(char *, struct void_ * *) 



Are there any advices to solve this problem? 


thank you for your answers. 


Kind regards, 


Michael 





From richjohnson905 at msn.com  Mon Aug  7 15:07:17 2006
From: richjohnson905 at msn.com (RICH JOHNSON)
Date: Mon, 07 Aug 2006 08:07:17 -0500
Subject: [C++-sig] ImportError: timemodule.so
Message-ID: <BAY107-F200BD0104F20E4EF95793498570@phx.gbl>

I am new to boost python and am having difficulty importing python modules 
into a little C++ program.  Any insight into the problem would be 
appreciated.

OS: Linux Redhat
gcc: 3.2.3
python 2.2
boost 1.33.1

>From the command line I can do:
[rjohnson at pitwar7 rjohnson]$ python
Python 2.2.3 (#1, Feb 15 2005, 02:41:06)
[GCC 3.2.3 20030502 (Red Hat Linux 3.2.3-49)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>>from time import time,ctime
>>>print 'Today is', ctime(time())
Today is Fri Aug  4 10:47:20 2006
>>>

But the C++ program below outputs:
25
20
Traceback (most recent call last):
File "<string>", line 1, in ?
ImportError: /usr/lib/python2.2/lib-dynload/timemodule.so: undefined symbol: 
PyDict_SetItemString


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

int main(int, char **)
{
using namespace boost::python;

Py_Initialize();

try
{
  PyRun_SimpleString("result = 5 ** 2");
     object module((handle<>(borrowed(PyImport_AddModule("__main__")))));
  object dictionary = module.attr("__dict__");
  object result = dictionary["result"];
  int result_value = extract<int>(result);
     std::cout << result_value << std::endl;
     dictionary["result"] = 20;

  PyRun_SimpleString("print result");

  // Here is the same command that ran successfully at the command line 
above but fails from  C++
  PyRun_SimpleString("from time import time,ctime\n"
                   "print 'Today is',ctime(time())\n");
}
catch (error_already_set)
{
  PyErr_Print();
}

Py_Finalize();
return 0;
}

My LD_LIBRARY_PATH is /usr/local/lib:/usr/lib/python2.2/config
PYTHONPATH is /usr/lib/python2.2/config




From richjohnson905 at msn.com  Mon Aug  7 16:17:03 2006
From: richjohnson905 at msn.com (RICH JOHNSON)
Date: Mon, 07 Aug 2006 09:17:03 -0500
Subject: [C++-sig] ImportError: timemodule.so
In-Reply-To: <BAY107-F200BD0104F20E4EF95793498570@phx.gbl>
Message-ID: <BAY107-F638217720774995D8D5B398570@phx.gbl>

More info on my import error:

In my Jamfile, my import statement when I build boost_python-gcc was:
import python ;

I am currently trying to rebuild my boost_python.so with:
import python time ;

... thinking this may be the problem.

----Original Message Follows----
From: "RICH JOHNSON" <richjohnson905 at msn.com>
Reply-To: Development of Python/C++ integration <c++-sig at python.org>
To: c++-sig at python.org
Subject: [C++-sig] ImportError: timemodule.so
Date: Mon, 07 Aug 2006 08:07:17 -0500

I am new to boost python and am having difficulty importing python modules
into a little C++ program.  Any insight into the problem would be
appreciated.

OS: Linux Redhat
gcc: 3.2.3
python 2.2
boost 1.33.1

 >From the command line I can do:
[rjohnson at pitwar7 rjohnson]$ python
Python 2.2.3 (#1, Feb 15 2005, 02:41:06)
[GCC 3.2.3 20030502 (Red Hat Linux 3.2.3-49)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
 >>>from time import time,ctime
 >>>print 'Today is', ctime(time())
Today is Fri Aug  4 10:47:20 2006
 >>>

But the C++ program below outputs:
25
20
Traceback (most recent call last):
File "<string>", line 1, in ?
ImportError: /usr/lib/python2.2/lib-dynload/timemodule.so: undefined symbol:
PyDict_SetItemString


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

int main(int, char **)
{
using namespace boost::python;

Py_Initialize();

try
{
   PyRun_SimpleString("result = 5 ** 2");
      object module((handle<>(borrowed(PyImport_AddModule("__main__")))));
   object dictionary = module.attr("__dict__");
   object result = dictionary["result"];
   int result_value = extract<int>(result);
      std::cout << result_value << std::endl;
      dictionary["result"] = 20;

   PyRun_SimpleString("print result");

   // Here is the same command that ran successfully at the command line
above but fails from  C++
   PyRun_SimpleString("from time import time,ctime\n"
                    "print 'Today is',ctime(time())\n");
}
catch (error_already_set)
{
   PyErr_Print();
}

Py_Finalize();
return 0;
}

My LD_LIBRARY_PATH is /usr/local/lib:/usr/lib/python2.2/config
PYTHONPATH is /usr/lib/python2.2/config


_______________________________________________
C++-sig mailing list
C++-sig at python.org
http://mail.python.org/mailman/listinfo/c++-sig




From dgregor at cs.indiana.edu  Mon Aug  7 15:59:03 2006
From: dgregor at cs.indiana.edu (Doug Gregor)
Date: Mon, 7 Aug 2006 09:59:03 -0400
Subject: [C++-sig] Boost.Python: Caching Python object instances
Message-ID: <13F95FA6-11E4-4937-B80A-A51DF3080A36@cs.indiana.edu>

Hello,

I am using Boost.Python to expose the Boost Graph Library in Python.  
The vertices in the graph types are small structures that are always  
passed by value. A vertex looks something like this:

	struct Vertex {
	  int index;
	  Graph* graph;
	};

This class is exposed to python in the normal way:

	class_<vertex_descriptor>();

Now, each time that we convert a Vertex from C++ to Python,  
Boost.Python will create a new PyObject for it. This behavior has  
surprised us a few times, and contributes to some performance  
problems. So, I would like to cache the PyObject instances in the  
Graph object itself:

	class Graph {
	public:
	  std::vector<boost::python::object> py_vertices;
	};

When I add a new vertex, I want to create a Python object for it and  
store that object in the py_vertices vector. Then, when Boost.Python  
needs to convert a vertex_descriptor value "v" to a Python object,  
I'd like to return "py_vertices[v.index]" instead of creating a new  
object.

With lots of hackery, I've attained some limited success in this area.

My approach is to register a to_python_vertex for the vertex type  
with this function object:

   struct cached_vertex_object {
     static PyObject* convert(const Vertex& v) {
         PyObject* result = v.graph->py_vertices[v.index].ptr();
         Py_INCREF(result);
         return result;
     }
   };

   to_python_converter<Vertex, cached_vertex_object>();

Of course, now I have the problem of creating the  
boost::python::object values to put in the cache! Digging around in  
the source a bit, I came up with this:

   typedef class_cref_wrapper<Vertex,
                              make_instance<Vertex,  
value_holder<Vertex> > >
     Converter;

   Vertex result(this->py_vertices.size(), this);
   boost::python::object vertex_object(handle<>(Converter::convert 
(result)));
   this->py_vertices.push_back(result);

This actually does work, but I get a sharp rebuke from Boost.Python  
whenever my extension module gets loaded:

/u/dgregor/lib/python/boost/graph/__init__.py:10: RuntimeWarning: to- 
Python converter for boost::graph::python::basic_descriptor<void*,  
boost::undirectedS> already registered; second conversion method  
ignored.
   from _graph import *

I have two questions:

	1) Is there a better way to cache Python object instances with  
Boost.Python?

	2) I've filtered out those Boost.Python warnings: will they come  
back to bite me eventually?

	Cheers,
	Doug


From amohr at pixar.com  Mon Aug  7 18:46:18 2006
From: amohr at pixar.com (Alex Mohr)
Date: Mon, 07 Aug 2006 09:46:18 -0700
Subject: [C++-sig] Calling list.__len__() from C++ causes crash with >=
 100 elements
In-Reply-To: <20060806064847.33668.qmail@web31512.mail.mud.yahoo.com>
References: <20060806064847.33668.qmail@web31512.mail.mud.yahoo.com>
Message-ID: <44D76E5A.5010905@pixar.com>

>> void printListLength(boost::python::object elements)
>> {
>>     boost::python::extract<int> lenExtract( elements.attr( "__len__" )() );
>>     if ( lenExtract.check() )
>>     {
>>         int numElements = lenExtract;
>>         std::cout << numElements << " elements" << std::endl;
>>     }
>> }
> 
> I can reproduce your problem under FC3 and FC5 (both x86_64). I am not sure if
> this is a bug or a feature. In any case, this workaround does the trick for me:

It is apparently a feature.  We've had a similar bug before as well.

You're creating a c++ temporary holding a new python object when you do 
"elements.attr("__len__")()" and passing it off to the extractor, and 
then you attempt to invoke the extractor later after the c++ temporary 
has died.

The problem is that the extractor does not keep a reference to the 
python object it was constructed with (and is documented accordingly). 
So, by the time you invoke the extractor, the python object it's got is 
long since dead, and you get a crash.

The fix posted in the parent message is the correct one.  You need to 
make sure the object you pass to the extractor lives at least as long as 
the last invocation of the extractor.

Here's the documentation:

http://www.boost.org/libs/python/doc/v2/extract.html#extract-spec

Note this sentence:
"Stores a pointer to the Python object managed by its constructor 
argument. In particular, the reference count of the object is not 
incremented. The onus is on the user to be sure it is not destroyed 
before the extractor's conversion function is called."

When we came across this, we wondered why it would be bad for the 
extractor to take a reference to its argument.  David -- can you shed 
some light on this?

Alex


From tcickovs at nd.edu  Tue Aug  8 02:45:07 2006
From: tcickovs at nd.edu (tcickovs at nd.edu)
Date: Mon,  7 Aug 2006 20:45:07 -0400
Subject: [C++-sig] C++ and Python variables referencing the same location
Message-ID: <1154997907.44d7de939c3de@webmail.nd.edu>

Hello,
I have been working on something for the past few days and have not been able to
find much online help on it.  Conceptually, what I want to do is very simple: I
would like to declare a C++ variable, then use the Python/C API (or some other
available utility) to define a Python variable, and then somehow get the two to
reference the same memory location so that when I modify one, the other changes.
So, for example, in my C++ file if I have this:

object main_module((
                           handle<>(borrowed(PyImport_AddModule("__main__")))));
int y = 2;  // y = 2
main_module.attr("x") = y;  // x = 2, y = 2
main_module.attr("x") = 3;  // x = 3, y = 2 (does not change)
y = 4;    // x = 3 (does not change), y = 4

Modifications to x and y happen independently because one is on the C++ runtime
stack and the other is on the Python interpreter's heap.  I'd like it to be if
I modify x, then y changes; and vice-versa.

Now, I saw in the Python/C API there was a way to modify information in the
Python interpreter heap from C++:

int* y = PyMem_New(int, 1);
*y = 4;  // Changes data in the heap

However, I need to somehow get x to reference the location in the heap that I
just allocated.  Is there a way?

Or is there a completely alternative solution that I can use?  I feel like
someone must have done this before, but cannot find much documentation.

Appreciate any help at all.
-Trevor








From sbalasub at qualcomm.com  Tue Aug  8 06:10:31 2006
From: sbalasub at qualcomm.com (Balasubramanyam, Shivakumar)
Date: Mon, 7 Aug 2006 21:10:31 -0700
Subject: [C++-sig] boost.python version that contains indexing suite.
Message-ID: <A70971E99D54CE4BBED4AD2D97C60D8801B407@NAEX15.na.qualcomm.com>

Hi,

Can someone provide a tar file for the boost.python code that contains
the indexing_suite_v2. The last time I tried, I had the compilation
failures and could go no further.

I followed the steps listed in
http://home.clara.net/raoulgough/boost/index.html, but with the latest
version of boost.

cvs -z3 -d:pserver:anonymous at boost.cvs.sourceforge.net:/cvsroot/boost co
-P boost
cvs update -d -r indexing_v2 boost/python/suite/indexing
cvs update -d -r indexing_v2 libs/python

Any ideas?

Thanks,
Shiva
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060807/0d5b84d0/attachment.htm>

From sbalasub at qualcomm.com  Tue Aug  8 06:40:25 2006
From: sbalasub at qualcomm.com (Balasubramanyam, Shivakumar)
Date: Mon, 7 Aug 2006 21:40:25 -0700
Subject: [C++-sig] STL list data member.
Message-ID: <A70971E99D54CE4BBED4AD2D97C60D8801B408@NAEX15.na.qualcomm.com>

Hi,

I have some classes which has std::list as public data members that need
to be exposed.

What is the good way to do it?

The short term solution I can think of is to add a wrapper getter/setter
method to convert to/from boost::python::list.

I could probably use the decl_string from py++ to check if the
member_data is "::std::list" and automate the wrapper generation.

Does anyone else have other solutions?

Thanks,
Shiva
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060807/5d3fc3d3/attachment.htm>

From sbalasub at qualcomm.com  Tue Aug  8 06:57:16 2006
From: sbalasub at qualcomm.com (Balasubramanyam, Shivakumar)
Date: Mon, 7 Aug 2006 21:57:16 -0700
Subject: [C++-sig] [GCC-XML] (__GLOBAL function was Problem exposing
	templates)
In-Reply-To: <44D18AFF.7060903@4roald.org>
Message-ID: <A70971E99D54CE4BBED4AD2D97C60D8801B409@NAEX15.na.qualcomm.com>

Roman, Bjorn,

The template instantiations happens in both declarations below.

1. typedef Stack<int>::Stack Stack_int;

	This would result in a template instantiation with user defined class name "Stack_int"

2. template class Stack<int> as instructed in http://www.gccxml.org/HTML/FAQ.html

	This would result in a template instantiation with class name picked by gcc "Stack_less_int_grate_"


Though I used step 1, I think it makes sense to use step 2 and probably rename the class name in py++ or pyapp to Stack_Int or StackOfInt. 
This would be generic way for naming the template instantiations,


BTW, I used the following code to get rid of _GLOBAL functions,
***
for decl in mb.free_functions():
	if(decl.name.startswith('_GLOBAL')):
		mb.free_functions(decl.name).exclude()
***

HTH,
Shiva
>-----Original Message-----
>From: Bj?rn Roald [mailto:bjorn at 4roald.org]
>Sent: Wednesday, August 02, 2006 10:35 PM
>To: Roman Yakovenko
>Cc: Balasubramanyam, Shivakumar; gccxml at gccxml.org; Development of
>Python/C++ integration
>Subject: Re: [GCC-XML] (__GLOBAL function was Problem exposing templates)
>
>Roman Yakovenko wrote:
>> On 8/3/06, Balasubramanyam, Shivakumar <sbalasub at qualcomm.com> wrote:
>>> [shiva] Roman, I understood the second part of my question, which was
>>> using typedef for explicit instantiation.
>>>
>>> Now I declared using "typedef Stack<int>::Stack Stack_Int;"
>>
>> This does not couse to template instantiation.
>>
>> Please write:
>>
>> Stack_Int s;
>Why do you need the typdef?
>
>Stack<int>::Stack Stack_Int;
>
>Should do it!  Or am I missing some other purpose of the typedef?
>
> Bjorn


From roman.yakovenko at gmail.com  Tue Aug  8 07:17:12 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Tue, 8 Aug 2006 08:17:12 +0300
Subject: [C++-sig] [GCC-XML] (__GLOBAL function was Problem exposing
	templates)
In-Reply-To: <A70971E99D54CE4BBED4AD2D97C60D8801B409@NAEX15.na.qualcomm.com>
References: <44D18AFF.7060903@4roald.org>
	<A70971E99D54CE4BBED4AD2D97C60D8801B409@NAEX15.na.qualcomm.com>
Message-ID: <7465b6170608072217g2def47c2ld47f8965423a4cbe@mail.gmail.com>

On 8/8/06, Balasubramanyam, Shivakumar <sbalasub at qualcomm.com> wrote:
> Roman, Bjorn,
>
> The template instantiations happens in both declarations below.
>
> 1. typedef Stack<int>::Stack Stack_int;
>
>         This would result in a template instantiation with user defined class name "Stack_int"
>
> 2. template class Stack<int> as instructed in http://www.gccxml.org/HTML/FAQ.html
>
>         This would result in a template instantiation with class name picked by gcc "Stack_less_int_grate_"
>

Please, read this document, you will understand why there is a
difference in class aliases:

http://svn.sourceforge.net/viewvc/pygccxml/pyplusplus_dev/docs/documentation/hints.rest?view=markup

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


From roman.yakovenko at gmail.com  Tue Aug  8 07:24:43 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Tue, 8 Aug 2006 08:24:43 +0300
Subject: [C++-sig] boost.python version that contains indexing suite.
In-Reply-To: <A70971E99D54CE4BBED4AD2D97C60D8801B407@NAEX15.na.qualcomm.com>
References: <A70971E99D54CE4BBED4AD2D97C60D8801B407@NAEX15.na.qualcomm.com>
Message-ID: <7465b6170608072224p6f9eb935r2744bf071db53c1@mail.gmail.com>

On 8/8/06, Balasubramanyam, Shivakumar <sbalasub at qualcomm.com> wrote:
> Hi,
>
> Can someone provide a tar file for the boost.python code that contains the
> indexing_suite_v2. The last time I tried, I had the compilation failures and
> could go no further.
>

pyplusplus documentation for indexing suite v2:

http://svn.sourceforge.net/viewvc/pygccxml/pyplusplus_dev/docs/documentation/containers.rest?view=markup

source code and instructions for indexing suite v2:

http://svn.sourceforge.net/viewvc/pygccxml/pyplusplus_dev/indexing_suite_v2/

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


From quncao at gmail.com  Tue Aug  8 11:22:29 2006
From: quncao at gmail.com (Qun Cao)
Date: Tue, 8 Aug 2006 02:22:29 -0700
Subject: [C++-sig] Can Boost.Python help me?
Message-ID: <1f527df80608080222r61821e97u8b0dddb66a560d53@mail.gmail.com>

Hi Everyone,

I am a beginner on cross language development. My problem at hand is
to build a python interface for a C++ application built on top of a 3D
game engine.  The purpose of this python interface is providing a
convenient scripting toolkit for the application.  One example is that
a user can write a python script like:
	   player = Player()
	   game.loadPlayer(player)
	   player.moveTo(location)
To configure the game environment.

I am trying to figure out how to make this happen.  I only have the
vague idea that this concerns embedding/extending python with C++, but
after reading Python Doc, SWIG, BOOST.Python, I am still not sure how
to architecture it.

Since the main program is still going to be the C++ application, I
guess we need to embedding the python scripts in the C++ code.  So at
initialization stage, the python script needs to be loaded into the
C++ code.  And this code can be simple, like
	player = Player()
	game.loadPlayer(player)
	
	
But for this to work, the python code needs to know the Player class,
is it right? Does that mean I need to build a python wrapper class for
Player and "import Player" in the python code?  But because this
application is built on top of a game engine, Player class inherits
many classes from there, I cannot possibly wrapping them all, right?
Also, some global objects are probably needed in this code of adding
players, how can the python code access them?

I know I probably don't have a grasp of basics here,  please kindly
enlighten me!

Btw, if you can point me to any source code of non-trivial projects
utilizing Boost.Python, that would be very helpful.  I found the
examples on the tutorials are far too simple.

Thank you very much,
Qun Cao


From roman.yakovenko at gmail.com  Tue Aug  8 11:35:09 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Tue, 8 Aug 2006 12:35:09 +0300
Subject: [C++-sig] Can Boost.Python help me?
In-Reply-To: <1f527df80608080222r61821e97u8b0dddb66a560d53@mail.gmail.com>
References: <1f527df80608080222r61821e97u8b0dddb66a560d53@mail.gmail.com>
Message-ID: <7465b6170608080235k47be9af5n20306dccda1e91da@mail.gmail.com>

On 8/8/06, Qun Cao <quncao at gmail.com> wrote:
> Hi Everyone,
>
> I am a beginner on cross language development. My problem at hand is
> to build a python interface for a C++ application built on top of a 3D
> game engine.  The purpose of this python interface is providing a
> convenient scripting toolkit for the application.

Fortunately you are not the first one, and I hope not the last:

http://language-binding.net/pyplusplus/quotes.html#who-is-using-pyplusplus
1. Python OGRE
   * http://lakin.weckers.net/index_ogre_python.html
   * http://www.ogre3d.org/phpBB2addons/viewtopic.php?t=1478&sid=df5ad0fff81d1f746f527ebcdd4aa1db

2. http://cgkit.sourceforge.net/ - contains Python bindings for Maya C++ SDK

3. PyOpenSG - https://realityforge.vrsource.org/view/PyOpenSG/WebHome
    The goal of PyOpenSG is to provide python bindings for the OpenSG
scene graph.

> Since the main program is still going to be the C++ application, I
> guess we need to embedding the python scripts in the C++ code.

I think cgkit project is dealing with this problem too.
>
> But for this to work, the python code needs to know the Player class,
> is it right?

Right

> Does that mean I need to build a python wrapper class for
> Player and "import Player" in the python code?  But because this
> application is built on top of a game engine, Player class inherits
> many classes from there, I cannot possibly wrapping them all, right?

It depends on how much functionality you want to export.

> Also, some global objects are probably needed in this code of adding
> players, how can the python code access them?

Boost.Python provides the functionality.

> I know I probably don't have a grasp of basics here,  please kindly
> enlighten me!
>
> Btw, if you can point me to any source code of non-trivial projects
> utilizing Boost.Python, that would be very helpful.  I found the
> examples on the tutorials are far too simple.

Those are tutorials, they should be simple, right :-) ?

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


From gsculpt at googlemail.com  Tue Aug  8 22:42:02 2006
From: gsculpt at googlemail.com (Geoffrey French)
Date: Tue, 8 Aug 2006 21:42:02 +0100
Subject: [C++-sig] Calling list.__len__() from C++ causes crash with >= 100
	elements
Message-ID: <562c63b10608081342p690d2472y82f38ed504526b9d@mail.gmail.com>

I tried to code the smallest module to demonstrate the problem. I
presented it as a module so
that people could get it working and running easily, to aid any
debugging that might have been
done.
Sorry if I got the balance between too much code and too little, somewhat off.


Regards

Geoffrey French



"Geoffrey French" <gsculpt at
googlemail.com<http://mail.python.org/mailman/listinfo/c++-sig>>
> writes:
>
> >* Here is the code for a C++ module to demonstrate this:*
> >
> **
> <snip>
>
> >* Is this a boost::python bug, or have I screwed up some how?*
> >* Thank you for your time.*
>
> What is the *smallest* code that demonstrates it?
>
> --
> Dave Abrahams
> Boost Consulting
> www.boost-consulting.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060808/d64c2b8f/attachment.htm>

From lee.will at gmail.com  Tue Aug  8 23:48:38 2006
From: lee.will at gmail.com (Will Lee)
Date: Tue, 8 Aug 2006 16:48:38 -0500
Subject: [C++-sig]  Wrapping shared_ptr of virtual classes
Message-ID: <7f03db650608081448w44d9d332hae6cca39efd9cfe3@mail.gmail.com>

I have seen this post (
http://aspn.activestate.com/ASPN/Mail/Message/cpp-sig/2262721) a while a go
but I couldn't find the response that solves this problem.  Essentially, I
couldn't find a way to wrap an abstract class that is being returned as a
shared_ptr.  For the following scenario:

class A has virtual method f
class B inherits from A that implements f
class F that has a static method: shared_ptr<A> getInstance()

Essentially I want to call F.getInstance().f() in python, which will call
B::f

All the documentation points to creating a wrapper class, so I have done so:

struct AWrapper : A, wrapper<A>
(would wrap function "f" that uses this->get_override() call to call the
overridden function)

My question is, like what is in the post, how do I declare class A?

I don't think I can declare the AWrapper, since F.getInstance() returns
shared_ptr<A> and not shared_ptr<AWrapper>:

class_<AWrapper, shared_ptr<AWrapper>, boost::noncopyable>...

I don't think I can declare with A either, since how can I tie the wrapper
function calls?

class_<A, shared_ptr<A>, boost::noncopyable>

Then this just gets me confused in what it would do (I got a convert error
saying that there is no to_python converter found):

class_<A, AWrapper, shared_ptr<A>, boost::noncopyable>

...

Your help is greatly appreciated.

Will
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060808/d9685b72/attachment.htm>

From jorge.vargas at gmail.com  Wed Aug  9 00:55:48 2006
From: jorge.vargas at gmail.com (Jorge Vargas)
Date: Tue, 8 Aug 2006 18:55:48 -0400
Subject: [C++-sig] can't build boost.python under msys/windows
In-Reply-To: <32822fe60608051032o4993e584r864d5cb1d75a0807@mail.gmail.com>
References: <32822fe60608051032o4993e584r864d5cb1d75a0807@mail.gmail.com>
Message-ID: <32822fe60608081555l6032f232n91c55689f3d3c362@mail.gmail.com>

anyone?

On 8/5/06, Jorge Vargas <jorge.vargas at gmail.com> wrote:
>
> hi
>
> Could someone stop what I'm doing wrong?
>
> $ echo $PATH
> .:/usr/local/bin:/mingw/bin:/bin:/c/WINDOWS/system32:/c/WINDOWS:/c/WINDOWS/System32/Wbem:/c/Program
> Files/Subversion/bin:/c/Python24:/c/Python24/Scripts
>
> $ bjam "-sTOOLS=mgw" stage
> $ bjam "-sTOOLS=mingw" stage
>
> both commands fail doing a lot of lookups for vc-7_1 and the .Net
> framework (I have those install)
> CALL "C:\Program Files\Microsoft Visual Studio .NET
> 2003\VC7\bin\VCVARS32.BAT" >nul
> ...skipped <@boost!libs!python!build\libboost_python.lib\vc-7_1\debug\threading-multi>libboost_python.CMD
> for lack of <@boost
> !libs!python!build\libboost_python.lib\vc-7_1\debug\threading-multi>
> numeric.obj.. .
>
> ------
>
> so I take them out of the path
>
> $ echo $PATH
> .:/usr/local/bin:/mingw/bin:/bin:/c/Python24:/c/Python24/Scripts
>
> $ bjam "-sTOOLS=mgw" stage
>
>
> and it fails to do anything
> $ bjam "-sTOOLS=mgw" stage
> ...found 944 targets...
> ...updating 121 targets...
> vc-C++ ..\..\..\bin\boost
> \libs\python\build\boost_python.dll\vc-7_1\debug\threading-multi\numeric.obj
> spawn: No such file or directory
>
> so basically bjam is ignoring the TOOLS flag
>
> I have to say I downloaded the binary for windows (boost-
> jam-3.1.13-1-ntx86.zip)
> and all boostJam related links seems to be broken
> http://www.boost.org/doc/html/jam/building.html
> http://www.boost.org/tools/jam/index.html
>
> what I'm actually trying to do is porting this
>
> https://develop.participatoryculture.org/projects/democracy/wiki/WindowsBuildDocs
> to compile with free tools so at the moment what I need is a replacement
> for boost_python-vc71-mt-1_33.dll
> that works with mingw does anyone distributes this as binary?
>
> thanks for the help.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060808/38df14f6/attachment.htm>

From roman.yakovenko at gmail.com  Wed Aug  9 08:11:23 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Wed, 9 Aug 2006 09:11:23 +0300
Subject: [C++-sig] Wrapping shared_ptr of virtual classes
In-Reply-To: <7f03db650608081448w44d9d332hae6cca39efd9cfe3@mail.gmail.com>
References: <7f03db650608081448w44d9d332hae6cca39efd9cfe3@mail.gmail.com>
Message-ID: <7465b6170608082311r6482e291k30c50e50b9a080a7@mail.gmail.com>

On 8/9/06, Will Lee <lee.will at gmail.com> wrote:
> I have seen this post
> (http://aspn.activestate.com/ASPN/Mail/Message/cpp-sig/2262721)
> a while a go but I couldn't find the response that solves this problem.
> Essentially, I couldn't find a way to wrap an abstract class that is being
> returned as a shared_ptr.

I hope next links will help you:

Python code:
http://svn.sourceforge.net/viewvc/pygccxml/pyplusplus_dev/unittests/smart_pointers_tester.py?view=markup

C++ code:
http://svn.sourceforge.net/viewvc/pygccxml/pyplusplus_dev/unittests/data/smart_pointers_to_be_exported.hpp?view=markup
http://svn.sourceforge.net/viewvc/pygccxml/pyplusplus_dev/unittests/data/smart_pointers_to_be_exported.cpp?view=markup

Generated code attached.

Also, it could be nice to get the answer, why user can not specify
HeldType for abstract class.


-- 
Roman Yakovenko
C++ Python language binding
http://www.language-binding.net/
-------------- next part --------------
// This file has been generated by pyplusplus.

// Copyright 2004 Roman Yakovenko.
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)

#include "boost/python.hpp"

#include "smart_pointers_to_be_exported.hpp"

namespace bp = boost::python;

struct base_wrapper : smart_pointers::base, bp::wrapper< smart_pointers::base > {

    base_wrapper( )
    : smart_pointers::base( )
      , bp::wrapper< smart_pointers::base >()
    {   // Normal constructor
    
    }

    virtual int get_base_value(  ) {
        if( bp::override func_get_base_value = this->get_override( "get_base_value" ) )
            return func_get_base_value(  );
        else
            return smart_pointers::base::get_base_value(  );
    }
    
    
    int default_get_base_value(  ) {
        return smart_pointers::base::get_base_value( );
    }

    virtual int get_some_value(  ){
        bp::override func_get_some_value = this->get_override( "get_some_value" );
        return func_get_some_value(  );
    }

};

struct data_wrapper : smart_pointers::data, bp::wrapper< smart_pointers::data > {

    data_wrapper(smart_pointers::data const & arg )
    : smart_pointers::data( arg )
      , bp::wrapper< smart_pointers::data >(){
        // copy constructor
        
    }

    data_wrapper( )
    : smart_pointers::data( )
      , bp::wrapper< smart_pointers::data >()
    {   // Normal constructor
    
    }

    virtual int get_some_value(  ) {
        if( bp::override func_get_some_value = this->get_override( "get_some_value" ) )
            return func_get_some_value(  );
        else
            return smart_pointers::data::get_some_value(  );
    }
    
    
    int default_get_some_value(  ) {
        return smart_pointers::data::get_some_value( );
    }

    virtual int get_value(  ) {
        if( bp::override func_get_value = this->get_override( "get_value" ) )
            return func_get_value(  );
        else
            return smart_pointers::data::get_value(  );
    }
    
    
    int default_get_value(  ) {
        return smart_pointers::data::get_value( );
    }

};

BOOST_PYTHON_MODULE(smart_pointers){
    { //::smart_pointers::base
        typedef bp::class_< base_wrapper, boost::noncopyable > base_exposer_t;
        base_exposer_t base_exposer = base_exposer_t( "base", "documentation" );
        bp::scope base_scope( base_exposer );
        base_exposer.def( bp::init< >("documentation")[bp::default_call_policies()] );
        { //::smart_pointers::base::get_base_value
        
            typedef int ( ::smart_pointers::base::*function_ptr_t )(  ) ;
            typedef int ( base_wrapper::*default_function_ptr_t )(  ) ;
            
            base_exposer.def( 
                "get_base_value"
                , function_ptr_t(&::smart_pointers::base::get_base_value)
                , default_function_ptr_t(&base_wrapper::default_get_base_value)
                , bp::default_call_policies() );
        
        }
        { //::smart_pointers::base::get_some_value
        
            typedef int ( ::smart_pointers::base::*function_ptr_t )(  ) ;
            
            base_exposer.def( 
                "get_some_value"
                , bp::pure_virtual( function_ptr_t(&::smart_pointers::base::get_some_value) )
                , bp::default_call_policies()
                , "documentation" );
        
        }
        base_exposer.def_readwrite( "base_value", &smart_pointers::base::base_value, "documentation" );
        bp::register_ptr_to_python< std::auto_ptr< smart_pointers::base > >();
        bp::register_ptr_to_python< boost::shared_ptr< smart_pointers::base > >();
    }

    { //::smart_pointers::data
        typedef bp::class_< data_wrapper, bp::bases< smart_pointers::base >, boost::shared_ptr< smart_pointers::data > > data_exposer_t;
        data_exposer_t data_exposer = data_exposer_t( "data", "documentation" );
        bp::scope data_scope( data_exposer );
        data_exposer.def( bp::init< >("documentation")[bp::default_call_policies()] );
        { //::smart_pointers::data::get_some_value
        
            typedef int ( ::smart_pointers::data::*function_ptr_t )(  ) ;
            typedef int ( data_wrapper::*default_function_ptr_t )(  ) ;
            
            data_exposer.def( 
                "get_some_value"
                , function_ptr_t(&::smart_pointers::data::get_some_value)
                , default_function_ptr_t(&data_wrapper::default_get_some_value)
                , bp::default_call_policies() );
        
        }
        { //::smart_pointers::data::get_value
        
            typedef int ( ::smart_pointers::data::*function_ptr_t )(  ) ;
            typedef int ( data_wrapper::*default_function_ptr_t )(  ) ;
            
            data_exposer.def( 
                "get_value"
                , function_ptr_t(&::smart_pointers::data::get_value)
                , default_function_ptr_t(&data_wrapper::default_get_value)
                , bp::default_call_policies() );
        
        }
        data_exposer.def_readwrite( "value", &smart_pointers::data::value, "documentation" );
        bp::register_ptr_to_python< std::auto_ptr< smart_pointers::data > >();
        bp::implicitly_convertible< std::auto_ptr< smart_pointers::data >, std::auto_ptr< smart_pointers::base > >();
        bp::implicitly_convertible< boost::shared_ptr< smart_pointers::data >, boost::shared_ptr< smart_pointers::base > >();
    }

    { //::smart_pointers::const_ref_auto
    
        typedef int ( *function_ptr_t )( ::smart_pointers::data_a_ptr const & );
        
        bp::def( 
            "const_ref_auto"
            , function_ptr_t( &::smart_pointers::const_ref_auto )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::const_ref_auto_base_value
    
        typedef int ( *function_ptr_t )( ::smart_pointers::base_a_ptr const & );
        
        bp::def( 
            "const_ref_auto_base_value"
            , function_ptr_t( &::smart_pointers::const_ref_auto_base_value )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::const_ref_auto_some_value
    
        typedef int ( *function_ptr_t )( ::smart_pointers::base_a_ptr const & );
        
        bp::def( 
            "const_ref_auto_some_value"
            , function_ptr_t( &::smart_pointers::const_ref_auto_some_value )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::const_ref_shared
    
        typedef int ( *function_ptr_t )( ::smart_pointers::data_s_ptr const & );
        
        bp::def( 
            "const_ref_shared"
            , function_ptr_t( &::smart_pointers::const_ref_shared )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::const_ref_shared_base_value
    
        typedef int ( *function_ptr_t )( ::smart_pointers::base_s_ptr const & );
        
        bp::def( 
            "const_ref_shared_base_value"
            , function_ptr_t( &::smart_pointers::const_ref_shared_base_value )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::const_ref_shared_some_value
    
        typedef int ( *function_ptr_t )( ::smart_pointers::base_s_ptr const & );
        
        bp::def( 
            "const_ref_shared_some_value"
            , function_ptr_t( &::smart_pointers::const_ref_shared_some_value )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::create_auto
    
        typedef ::smart_pointers::data_a_ptr ( *function_ptr_t )(  );
        
        bp::def( 
            "create_auto"
            , function_ptr_t( &::smart_pointers::create_auto )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::create_shared
    
        typedef ::smart_pointers::data_s_ptr ( *function_ptr_t )(  );
        
        bp::def( 
            "create_shared"
            , function_ptr_t( &::smart_pointers::create_shared )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::ref_auto
    
        typedef int ( *function_ptr_t )( ::smart_pointers::data_a_ptr & );
        
        bp::def( 
            "ref_auto"
            , function_ptr_t( &::smart_pointers::ref_auto )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::ref_auto_base_value
    
        typedef int ( *function_ptr_t )( ::smart_pointers::base_a_ptr & );
        
        bp::def( 
            "ref_auto_base_value"
            , function_ptr_t( &::smart_pointers::ref_auto_base_value )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::ref_auto_some_value
    
        typedef int ( *function_ptr_t )( ::smart_pointers::base_a_ptr & );
        
        bp::def( 
            "ref_auto_some_value"
            , function_ptr_t( &::smart_pointers::ref_auto_some_value )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::ref_shared
    
        typedef int ( *function_ptr_t )( ::smart_pointers::data_s_ptr & );
        
        bp::def( 
            "ref_shared"
            , function_ptr_t( &::smart_pointers::ref_shared )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::ref_shared_base_value
    
        typedef int ( *function_ptr_t )( ::smart_pointers::base_s_ptr & );
        
        bp::def( 
            "ref_shared_base_value"
            , function_ptr_t( &::smart_pointers::ref_shared_base_value )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::ref_shared_some_value
    
        typedef int ( *function_ptr_t )( ::smart_pointers::base_s_ptr & );
        
        bp::def( 
            "ref_shared_some_value"
            , function_ptr_t( &::smart_pointers::ref_shared_some_value )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::val_auto
    
        typedef int ( *function_ptr_t )( ::smart_pointers::data_a_ptr );
        
        bp::def( 
            "val_auto"
            , function_ptr_t( &::smart_pointers::val_auto )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::val_auto_base_value
    
        typedef int ( *function_ptr_t )( ::smart_pointers::base_a_ptr );
        
        bp::def( 
            "val_auto_base_value"
            , function_ptr_t( &::smart_pointers::val_auto_base_value )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::val_auto_some_value
    
        typedef int ( *function_ptr_t )( ::smart_pointers::base_a_ptr );
        
        bp::def( 
            "val_auto_some_value"
            , function_ptr_t( &::smart_pointers::val_auto_some_value )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::val_shared
    
        typedef int ( *function_ptr_t )( ::smart_pointers::data_s_ptr );
        
        bp::def( 
            "val_shared"
            , function_ptr_t( &::smart_pointers::val_shared )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::val_shared_base_value
    
        typedef int ( *function_ptr_t )( ::smart_pointers::base_s_ptr );
        
        bp::def( 
            "val_shared_base_value"
            , function_ptr_t( &::smart_pointers::val_shared_base_value )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }

    { //::smart_pointers::val_shared_some_value
    
        typedef int ( *function_ptr_t )( ::smart_pointers::base_s_ptr );
        
        bp::def( 
            "val_shared_some_value"
            , function_ptr_t( &::smart_pointers::val_shared_some_value )
            , ( bp::arg("a") )
            , bp::default_call_policies()
            , "documentation" );
    
    }
}

From roman.yakovenko at gmail.com  Wed Aug  9 08:16:56 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Wed, 9 Aug 2006 09:16:56 +0300
Subject: [C++-sig] can't build boost.python under msys/windows
In-Reply-To: <32822fe60608081555l6032f232n91c55689f3d3c362@mail.gmail.com>
References: <32822fe60608051032o4993e584r864d5cb1d75a0807@mail.gmail.com>
	<32822fe60608081555l6032f232n91c55689f3d3c362@mail.gmail.com>
Message-ID: <7465b6170608082316s520d3dof3bab1eead77ce02@mail.gmail.com>

I suggest you to ask the question to bjam mailing list. I am sure they
will be able to
help you.

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


From eduardo.picado at gmail.com  Wed Aug  9 14:15:28 2006
From: eduardo.picado at gmail.com (Eduardo Picado)
Date: Wed, 9 Aug 2006 14:15:28 +0200
Subject: [C++-sig] boost python liking problems
In-Reply-To: <uslklk1rv.fsf@boost-consulting.com>
References: <6c75ff8f0607280627t1e86f90bicefbfd1344567fe3@mail.gmail.com>
	<uslklk1rv.fsf@boost-consulting.com>
Message-ID: <6c75ff8f0608090515r296fd051w4a0ea08bb300020a@mail.gmail.com>

On 7/28/06, David Abrahams <dave at boost-consulting.com> wrote:
> "Eduardo Picado" <eduardo.picado at gmail.com> writes:
>
> > I'm trying to use Boost Python to wrap a 3rd party library. I have the
> > headers, DLLs and a .lib file.
> > When I try to compile my .cpp files I receive linkage errors:
>
> Surely this is when you try to link your shared objects...
>
> > -------------
> > Creating library
> > bin\telismaPyste\TASRNlSmlResult.pyd\msvc\debug\threading-multi\TASRNlSmlResult.lib
> > and object bin\telismaPyste\TASRNlSmlResult.pyd\msvc\debug\threading-multi\TASRNlSmlResult.exp
> > TASRNlSmlResult.obj : error LNK2001: unresolved external symbol
> > "public: char const * __thiscall TASRNlSmlResult::str(void)const "
> > (?str at TASRNlSmlResult@@QBEPBDXZ)
>
> Are you sure you didn't just forget to define the member function
>
>   TASRNlSmlResult::str() const
>
> as denoted in the error message?  The message gives the impression
> that it's supposed to be part of your code.  If not,
>
> extension TASRNlSmlResult
> : # sources
>   whatever.cpp
>
>   # requirements and dependencies for Boost.Python extensions
>   <template>@boost/libs/python/build/extension
>
>   : # additional requirements
>     <library-file>path/to/library/containing/that/function.lib
>   ;

Dave,

Thanks for your reply; the appropriate <library-file> directive solved
my problem here.

BTW, it seems to me that it is much easier to compile my modules with
Boost Python using VC8 instead of VC6 (that gives many "fatal error
C1001: INTERNAL COMPILER ERROR", per example). Is this a fact? Should
I forget using VC 6?


Regards,

Eduardo Picado

>
> HTH,
>
> --
> Dave Abrahams
> Boost Consulting
> www.boost-consulting.com
>
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
>


-- 
Eduardo PICADO
MSN eduardo.picado at gmail.com
portable +33 6 32 47 10 10 (NEW)
fix +33 1 45 03 52 54
|??<?????>??|

From w.richert at gmx.net  Wed Aug  9 16:32:21 2006
From: w.richert at gmx.net (Willi Richert)
Date: Wed, 9 Aug 2006 16:32:21 +0200
Subject: [C++-sig] boost.python newbie hassles
Message-ID: <200608091632.22195.w.richert@gmx.net>

Hi,

As I am exploring whether I can let my students use it in our robotics group I 
like it more and more. However, there are some little problems for newbies:

1. the web doc is out of date. 
http://www.cs.brown.edu/people/jwicks/boost/libs/python/doc/tutorial/doc/html/python/hello.html 
e.g. is very different from the doc in the tar.bz2 on sourceforge. A simple 
copy of the tar version to the web would be enough I think and save lots of 
trials.

2. boost.python projects have to be located in the boost directory tree, other 
than printed at the tutorial docs - as far as I have tried. If I move the 
tutorial greet example to another directory and update path-global in 
Jamrules I always get: 

 wr@[ikarus]:~/workspace/Boost.Python/src> bjam
Failed to find the project root for directory '.'.
Did not find a project-root.jam file there or in any of its parent 
directories.
Please consult the documentation at 'http://www.boost.org'.

I'm using ubuntur 6.06. There I have installed 
boost-build
libboost-python1.33.1
libboost-python-dev
bjam

In order to get the tutorial compiled I have to use the sf.net downloaded 
version. The above mentioned packages do not suffice. Shouldn't they?

My version of bjam is
Boost.Jam  Version 3.1.11. OS=LINUX

Regards,
wr


From jamesp at trdlnk.com  Thu Aug 10 17:14:12 2006
From: jamesp at trdlnk.com (James Pelletier)
Date: Thu, 10 Aug 2006 10:14:12 -0500
Subject: [C++-sig] problem with iterator and non-destructible container
Message-ID: <44DB4D44.6030008@trdlnk.com>

I'm having trouble wrapping a container object that has a private 
destructor. As far as I can tell, the object has all of the necessary 
data to be used with the boost::python::iterator object, but when I try 
to compile I get an error message saying that the container's destructor 
is private. Why would the iterator want to destroy my container? Am I 
missing something?

I'm using gcc-3.4 on solaris 10 on intel and boost-1.32.  I couldn't 
find any differences in header files between boost 1.32 and 1.33 in this 
area.

I've whittled it down to a small example below that gives me the same 
error message.
The code compiles cleanly if I take out the .def("__iter__"...) line.

-------

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

using namespace boost::python;

class Y
{
	Y() {}
	~Y() {}

	Y( Y const & );
	Y & operator=( Y const & );

	std::vector<int> v_;
public:
	typedef std::vector<int>::const_iterator const_iterator;
	const_iterator begin() const { return v_.begin(); }
	const_iterator end() const { return v_.end(); }
};

BOOST_PYTHON_MODULE(boostpy)
{
	class_<Y, boost::noncopyable>("Y", no_init)
		.def( "__iter__", iterator<Y const>() )
		;
}

------

test.C: In static member function `static void 
boost::python::detail::value_destroyer< false>::execute(const volatile 
T*) [with T = Y]':
/opt/app/g++lib6/boost-1.32/include/boost-1_32/boost/python/detail/destroy.hpp:90: 
   instantiated from `void 
boost::python::detail::destroy_referent_impl(void*, T&(*)()) [with T = 
const Y]'
/opt/app/g++lib6/boost-1.32/include/boost-1_32/boost/python/detail/destroy.hpp:101: 
   instantiated from `void 
boost::python::detail::destroy_referent(void*, T (*)()) [with T = const Y&]'
/opt/app/g++lib6/boost-1.32/include/boost-1_32/boost/python/converter/rvalue_from_python_data.hpp:135: 
   instantiated from 
`boost::python::converter::rvalue_from_python_data<T>::~rvalue_from_python_data() 
[with T = const Y&]'
/opt/app/g++lib6/boost-1.32/include/boost-1_32/boost/preprocessor/iteration/detail/local.hpp:34: 
   instantiated from `PyObject* 
boost::python::detail::caller_arity<1u>::impl<F, Policies, 
Sig>::operator()(PyObject*, PyObject*) [with F = 
boost::python::objects::detail::py_iter_<const Y, 
__gnu_cxx::__normal_iterator<const int*, std::vector<int, 
std::allocator<int> > >, 
boost::_bi::protected_bind_t<boost::_bi::bind_t<__gnu_cxx::__normal_iterator<const 
int*, std::vector<int, std::allocator<int> > >, 
__gnu_cxx::__normal_iterator<const int*, std::vector<int, 
std::allocator<int> > > (*)(const Y&), boost::_bi::list1<boost::arg<1> > 
 > >, 
boost::_bi::protected_bind_t<boost::_bi::bind_t<__gnu_cxx::__normal_iterator<const 
int*, std::vector<int, std::allocator<int> > >, 
__gnu_cxx::__normal_iterator<const int*, std::vector<int, 
std::allocator<int> > > (*)(const Y&), boost::_bi::list1<boost::arg<1> > 
 > >, boost::python::return_value_policy<boost::python::return_by_value, 
boost::python::default_call_policies> >, Policies = 
boost::python::default_call_policies, Sig = 
boost::mpl::vector2<boost::python::objects::iterator_range<boost::python::return_value_policy<boost::python::return_by_value, 
boost::python::default_call_policies>, 
__gnu_cxx::__normal_iterator<const int*, std::vector<int, 
std::allocator<int> > > >, boost::python::back_reference<const Y&> >]'
/opt/app/g++lib6/boost-1.32/include/boost-1_32/boost/python/object/py_function.hpp:38: 
   instantiated from `PyObject* 
boost::python::objects::caller_py_function_impl<Caller>::operator()(PyObject*, 
PyObject*) [with Caller = 
boost::python::detail::caller<boost::python::objects::detail::py_iter_<const 
Y, __gnu_cxx::__normal_iterator<const int*, std::vector<int, 
std::allocator<int> > >, 
boost::_bi::protected_bind_t<boost::_bi::bind_t<__gnu_cxx::__normal_iterator<const 
int*, std::vector<int, std::allocator<int> > >, 
__gnu_cxx::__normal_iterator<const int*, std::vector<int, 
std::allocator<int> > > (*)(const Y&), boost::_bi::list1<boost::arg<1> > 
 > >, 
boost::_bi::protected_bind_t<boost::_bi::bind_t<__gnu_cxx::__normal_iterator<const 
int*, std::vector<int, std::allocator<int> > >, 
__gnu_cxx::__normal_iterator<const int*, std::vector<int, 
std::allocator<int> > > (*)(const Y&), boost::_bi::list1<boost::arg<1> > 
 > >, boost::python::return_value_policy<boost::python::return_by_value, 
boost::python::default_call_policies> >, 
boost::python::default_call_policies, 
boost::mpl::vector2<boost::python::objects::iterator_range<boost::python::return_value_policy<boost::python::return_by_value, 
boost::python::default_call_policies>, 
__gnu_cxx::__normal_iterator<const int*, std::vector<int, 
std::allocator<int> > > >, boost::python::back_reference<const Y&> > >]'
test.C:37:   instantiated from here
test.C:10: error: `Y::~Y()' is private
/opt/app/g++lib6/boost-1.32/include/boost-1_32/boost/python/detail/destroy.hpp:33: 
error: within this context


From duranlef at iro.umontreal.ca  Thu Aug 10 20:48:03 2006
From: duranlef at iro.umontreal.ca (=?ISO-8859-1?Q?Fran=E7ois_Duranleau?=)
Date: Thu, 10 Aug 2006 14:48:03 -0400 (EDT)
Subject: [C++-sig] boost.python newbie hassles
In-Reply-To: <200608091632.22195.w.richert@gmx.net>
References: <200608091632.22195.w.richert@gmx.net>
Message-ID: <Pine.LNX.4.63.0608101443250.24635@brocoli.iro.umontreal.ca>

On Wed, 9 Aug 2006, Willi Richert wrote:

> Hi,
>
> As I am exploring whether I can let my students use it in our robotics group I
> like it more and more. However, there are some little problems for newbies:
>
> 1. the web doc is out of date.
> http://www.cs.brown.edu/people/jwicks/boost/libs/python/doc/tutorial/doc/html/python/hello.html
> e.g. is very different from the doc in the tar.bz2 on sourceforge. A simple
> copy of the tar version to the web would be enough I think and save lots of
> trials.

Maybe you should try the official web page from boost.org:
http://www.boost.org/libs/python/doc/tutorial/doc/html/python/hello.html

> 2. boost.python projects have to be located in the boost directory tree, other
> than printed at the tutorial docs - as far as I have tried. If I move the
> tutorial greet example to another directory and update path-global in
> Jamrules I always get:
>
> wr@[ikarus]:~/workspace/Boost.Python/src> bjam
> Failed to find the project root for directory '.'.
> Did not find a project-root.jam file there or in any of its parent
> directories.
> Please consult the documentation at 'http://www.boost.org'.

[I will let other bjam experts reply on this.]

> I'm using ubuntur 6.06. There I have installed
> boost-build
> libboost-python1.33.1
> libboost-python-dev
> bjam
>
> In order to get the tutorial compiled I have to use the sf.net downloaded
> version. The above mentioned packages do not suffice. Shouldn't they?
>
> My version of bjam is
> Boost.Jam  Version 3.1.11. OS=LINUX

If the packages for boost with Ubuntu and Debian are the same, you also 
need the package boost-build.

-- 
Fran?ois Duranleau
LIGUM, Universit? de Montr?al

From dave at boost-consulting.com  Fri Aug 11 00:52:24 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 10 Aug 2006 18:52:24 -0400
Subject: [C++-sig] Calling list.__len__() from C++ causes crash with >=
	100 elements
References: <20060806064847.33668.qmail@web31512.mail.mud.yahoo.com>
	<44D76E5A.5010905@pixar.com>
Message-ID: <u3bc46w5z.fsf@boost-consulting.com>

Alex Mohr <amohr at pixar.com> writes:

> When we came across this, we wondered why it would be bad for the 
> extractor to take a reference to its argument.  David -- can you shed 
> some light on this?

I'm afraid I can't.  If the rationale isn't documented, it's lost
forever :(.

Maybe that should be changed.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From dave at boost-consulting.com  Fri Aug 11 00:54:23 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 10 Aug 2006 18:54:23 -0400
Subject: [C++-sig] C++ and Python variables referencing the same location
References: <1154997907.44d7de939c3de@webmail.nd.edu>
Message-ID: <uwt9g5hi8.fsf@boost-consulting.com>

tcickovs at nd.edu writes:

> Hello,
> I have been working on something for the past few days and have not been able to
> find much online help on it.  Conceptually, what I want to do is very simple: I
> would like to declare a C++ variable, then use the Python/C API (or some other
> available utility) to define a Python variable, and then somehow get the two to
> reference the same memory location so that when I modify one, the other changes.
> So, for example, in my C++ file if I have this:
>
> object main_module((
>                            handle<>(borrowed(PyImport_AddModule("__main__")))));
> int y = 2;  // y = 2
> main_module.attr("x") = y;  // x = 2, y = 2
> main_module.attr("x") = 3;  // x = 3, y = 2 (does not change)
> y = 4;    // x = 3 (does not change), y = 4

You can't do it with assignment syntax, since in Python that always
rebinds.

> Modifications to x and y happen independently because one is on the C++ runtime
> stack and the other is on the Python interpreter's heap.  I'd like it to be if
> I modify x, then y changes; and vice-versa.
>
> Now, I saw in the Python/C API there was a way to modify information in the
> Python interpreter heap from C++:
>
> int* y = PyMem_New(int, 1);
> *y = 4;  // Changes data in the heap
>
> However, I need to somehow get x to reference the location in the heap that I
> just allocated.  Is there a way?

No; for another reason, Python ints are immutable.  Changing the
stored value would break _everything_.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From dave at boost-consulting.com  Fri Aug 11 00:58:02 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 10 Aug 2006 18:58:02 -0400
Subject: [C++-sig] can't build boost.python under msys/windows
References: <32822fe60608051032o4993e584r864d5cb1d75a0807@mail.gmail.com>
Message-ID: <ur6zo5hc5.fsf@boost-consulting.com>

"Jorge Vargas" <jorge.vargas at gmail.com> writes:

> Could someone stop what I'm doing wrong?
>
> $ echo $PATH
> .:/usr/local/bin:/mingw/bin:/bin:/c/WINDOWS/system32:/c/WINDOWS:/c/WINDOWS/System32/Wbem:/c/Program
> Files/Subversion/bin:/c/Python24:/c/Python24/Scripts
>
> $ bjam "-sTOOLS=mgw" stage

What gave you an idea that would work?  Does our documentation mention
"mgw" somewhere?

> $ bjam "-sTOOLS=mingw" stage
>
> both commands fail doing a lot of lookups for vc-7_1 and the .Net framework
> (I have those install)

You are doing this from the msys shell, right?  That's not supported.
Use the good-old Windows command shell, "CMD," and you should be all
set.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From dave at boost-consulting.com  Fri Aug 11 01:07:42 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 10 Aug 2006 19:07:42 -0400
Subject: [C++-sig] boost.python newbie hassles
References: <200608091632.22195.w.richert@gmx.net>
	<Pine.LNX.4.63.0608101443250.24635@brocoli.iro.umontreal.ca>
Message-ID: <ulkpw5gw1.fsf@boost-consulting.com>

Fran?ois Duranleau <duranlef at iro.umontreal.ca> writes:

> On Wed, 9 Aug 2006, Willi Richert wrote:
>
>> Hi,
>>
>> As I am exploring whether I can let my students use it in our robotics group I
>> like it more and more. However, there are some little problems for newbies:
>>
>> 1. the web doc is out of date.
>> http://www.cs.brown.edu/people/jwicks/boost/libs/python/doc/tutorial/doc/html/python/hello.html
>> e.g. is very different from the doc in the tar.bz2 on sourceforge. A simple
>> copy of the tar version to the web would be enough I think and save lots of
>> trials.
>
> Maybe you should try the official web page from boost.org:
> http://www.boost.org/libs/python/doc/tutorial/doc/html/python/hello.html

What he said

>> 2. boost.python projects have to be located in the boost directory tree, other
>> than printed at the tutorial docs - as far as I have tried. If I move the
>> tutorial greet example to another directory and update path-global in
>> Jamrules I always get:
>>
>> wr@[ikarus]:~/workspace/Boost.Python/src> bjam
>> Failed to find the project root for directory '.'.
>> Did not find a project-root.jam file there or in any of its parent
>> directories.
>> Please consult the documentation at 'http://www.boost.org'.

see the libs/python/example/ subdirectory of your boost distribution
for an example that should be transplantable.  I'm not sure why it's
looking for project-root.jam.  It's almost as though you edited your
boost-build.jam to reference the tools/build/v2 subdirectory instead
of tools/build/v1.  That won't work for you :)

>> I'm using ubuntur 6.06. There I have installed
>> boost-build
>> libboost-python1.33.1
>> libboost-python-dev
>> bjam

Can't vouch for any of those packages.  I suggest contacting the
package maintainers, or downloading a boost distro from sourceforge
and starting from there.

>> In order to get the tutorial compiled I have to use the sf.net downloaded
>> version. The above mentioned packages do not suffice. Shouldn't they?

Well they should, maybe, depending on the intentions of the packager.
Really that's out of our hands, since we didn't make them.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From roberto.c.aguilar at gmail.com  Fri Aug 11 01:22:16 2006
From: roberto.c.aguilar at gmail.com (Roberto Aguilar)
Date: Thu, 10 Aug 2006 16:22:16 -0700
Subject: [C++-sig] boost.python newbie hassles
In-Reply-To: <ulkpw5gw1.fsf@boost-consulting.com>
References: <200608091632.22195.w.richert@gmx.net>
	<Pine.LNX.4.63.0608101443250.24635@brocoli.iro.umontreal.ca>
	<ulkpw5gw1.fsf@boost-consulting.com>
Message-ID: <e190e3480608101622l4e364b19y55c57802dd1e61f4@mail.gmail.com>

On 8/10/06, David Abrahams <dave at boost-consulting.com> wrote:
[...]
> >> 2. boost.python projects have to be located in the boost directory tree, other
> >> than printed at the tutorial docs - as far as I have tried. If I move the
> >> tutorial greet example to another directory and update path-global in
> >> Jamrules I always get:
> >>
> >> wr@[ikarus]:~/workspace/Boost.Python/src> bjam
> >> Failed to find the project root for directory '.'.
> >> Did not find a project-root.jam file there or in any of its parent
> >> directories.
> >> Please consult the documentation at 'http://www.boost.org'.
>
> see the libs/python/example/ subdirectory of your boost distribution
> for an example that should be transplantable.  I'm not sure why it's
> looking for project-root.jam.  It's almost as though you edited your
> boost-build.jam to reference the tools/build/v2 subdirectory instead
> of tools/build/v1.  That won't work for you :)

Are you saying that boost.python does not work with bbv2?  I see
Jamfile.v2 files in the example directories.  But, if I try compiling
outside the example directory, my project will not build.  I am trying
to use bbv2, so if that's not possible then that's my problem.

Also, if building with bbv2 is possible, is there some documentation
on the process?
-Roberto.


From dave at boost-consulting.com  Fri Aug 11 01:23:13 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 10 Aug 2006 19:23:13 -0400
Subject: [C++-sig] problem with iterator and non-destructible container
References: <44DB4D44.6030008@trdlnk.com>
Message-ID: <ufyg45g66.fsf@boost-consulting.com>

James Pelletier <jamesp at trdlnk.com> writes:

> I'm having trouble wrapping a container object that has a private 
> destructor. As far as I can tell, the object has all of the necessary 
> data to be used with the boost::python::iterator object, but when I try 
> to compile I get an error message saying that the container's destructor 
> is private. Why would the iterator want to destroy my container? 

Heh, good question.

> Am I missing something?

Yeah, but I can't blame you.

> I'm using gcc-3.4 on solaris 10 on intel and boost-1.32.  I couldn't 
> find any differences in header files between boost 1.32 and 1.33 in this 
> area.
>
> I've whittled it down to a small example below that gives me the same 
> error message.
> The code compiles cleanly if I take out the .def("__iter__"...) line.
>
> -------
>
> #include <vector>
> #include <boost/python.hpp>
>
> using namespace boost::python;
>
> class Y
> {
> 	Y() {}
> 	~Y() {}
>
> 	Y( Y const & );
> 	Y & operator=( Y const & );
>
> 	std::vector<int> v_;
> public:
> 	typedef std::vector<int>::const_iterator const_iterator;
> 	const_iterator begin() const { return v_.begin(); }
> 	const_iterator end() const { return v_.end(); }
> };
>
> BOOST_PYTHON_MODULE(boostpy)
> {
> 	class_<Y, boost::noncopyable>("Y", no_init)
> 		.def( "__iter__", iterator<Y const>() )
                                     ^^^^^
This is the problem.
> 		;
> }

When you do that, Boost.Python internally builds a function taking 

 Y const& 

and wraps that.  T const& accepts "rvalue conversions" which
essentially mean there doesn't need to be a pre-existing T; it can be
converted from elsewhere... and thus it needs to be destroyed.  This
is the same problem mentioned in
http://news.gmane.org/find-root.php?message_id=%3cuejxga6gu.fsf%40boost%2dconsulting.com%3e,
and probably it's about time we implement the solution I propose
there.  In the meantime, you could try

        .def( "__iter__", range(&Y::begin, &Y::end))

and if that doesn't work, the horrible

    Y::const_iterator y_begin(Y& y) 
    { return const_cast<Y const&>(y).begin(); }

    Y::const_iterator y_end(Y& y) 
    { return const_cast<Y const&>(y).end(); }

        ...

        .def("__iter__", range(y_begin, y_end))

will definitely work for you.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From jorge.vargas at gmail.com  Fri Aug 11 03:59:00 2006
From: jorge.vargas at gmail.com (Jorge Vargas)
Date: Thu, 10 Aug 2006 21:59:00 -0400
Subject: [C++-sig] can't build boost.python under msys/windows
In-Reply-To: <ur6zo5hc5.fsf@boost-consulting.com>
References: <32822fe60608051032o4993e584r864d5cb1d75a0807@mail.gmail.com>
	<ur6zo5hc5.fsf@boost-consulting.com>
Message-ID: <32822fe60608101859q65aac75w7f930b134792a3d4@mail.gmail.com>

On 8/10/06, David Abrahams <dave at boost-consulting.com> wrote:
>
> "Jorge Vargas" <jorge.vargas at gmail.com> writes:
>
> > Could someone stop what I'm doing wrong?
> >
> > $ echo $PATH
> >
> .:/usr/local/bin:/mingw/bin:/bin:/c/WINDOWS/system32:/c/WINDOWS:/c/WINDOWS/System32/Wbem:/c/Program
> > Files/Subversion/bin:/c/Python24:/c/Python24/Scripts
> >
> > $ bjam "-sTOOLS=mgw" stage
>
> What gave you an idea that would work?  Does our documentation mention
> "mgw" somewhere?


I got that from http://www.boost.org/more/getting_started.html#Tools but now
that you mention it I went over a read it and I notice I got past into the
naming convention section without noticing it.

> $ bjam "-sTOOLS=mingw" stage
> >
> > both commands fail doing a lot of lookups for vc-7_1 and the .Net
> framework
> > (I have those install)
>
> You are doing this from the msys shell, right?  That's not supported.


oh i see since I didn't saw any reference to the shell prompt (c:\ or $) I
assume "any shell will do". maybe a warning note should be added because
there is no reference to this either on boost.build or boost-python docs.
and the error is very cryptic, if you haven't point it out I will probably
never had fix it.

Use the good-old Windows command shell, "CMD," and you should be all
> set.


I'll test that tonight, thanks.

--
> Dave Abrahams
> Boost Consulting
> www.boost-consulting.com
>
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060810/1c742974/attachment.htm>

From 5p4ced0ut at gmail.com  Fri Aug 11 06:38:26 2006
From: 5p4ced0ut at gmail.com (Blake Skinner)
Date: Thu, 10 Aug 2006 20:38:26 -0800
Subject: [C++-sig] weird runtime error I'm getting with PyRun_SimpleFile()
Message-ID: <4dd037b70608102138v4ae9ebeay7be371a61ee4ba8b@mail.gmail.com>

Here's my C++ code, I'm building with VS 2005, under multi-threaded:

#include <Python/python.h>
#include <iostream>
#include <cstdio>
using namespace std;

int main(int argc, char **argv)
{
    Py_Initialize();

    PyObject *main_module = PyImport_AddModule("__main__");
    PyObject *main_dict = PyModule_GetDict(main_module);

    FILE *f = fopen("test.py", "r");
    PyRun_SimpleFile(f, "test.py");

    if (PyDict_GetItemString(main_dict, "num") == NULL)
        cout << "No" << endl;
    else
        cout << "Yes" << endl;

    Py_Finalize();
    return 0;

}

When I run it, I get an error:

Unhandled exception at 0x7c918fea in Embed_practice.exe: 0xC0000005: Access
violation writing location 0x00000010

from a file called _file.c, the debugger calls attention to this method
specificially:

/***
* _lock_file - Lock a FILE
*
*Purpose:
*       Assert the lock for a stdio-level file
*
*Entry:
*       pf = __piob[] entry (pointer to a FILE or _FILEX)
*
*Exit:
*
*Exceptions:
*
*******************************************************************************/
void __cdecl _lock_file (
        FILE *pf
        )
{
        /*
         * The way the FILE (pointed to by pf) is locked depends on whether
         * it is part of _iob[] or not
         */
        if ( (pf >= _iob) && (pf <= (&_iob[_IOB_ENTRIES-1])) )
            /*
             * FILE lies in _iob[] so the lock lies in _locktable[].
             */
            _lock( _STREAM_LOCKS + (int)(pf - _iob) );
        else
            /*
             * Not part of _iob[]. Therefore, *pf is a _FILEX and the
             * lock field of the struct is an initialized critical
             * section.
             */
            EnterCriticalSection( &(((_FILEX *)pf)->lock) );
}

Any help solving this would be appreciated.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060810/24f66e1a/attachment.htm>

From seefeld at sympatico.ca  Fri Aug 11 09:18:56 2006
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Fri, 11 Aug 2006 03:18:56 -0400
Subject: [C++-sig] weird runtime error I'm getting with
	PyRun_SimpleFile()
In-Reply-To: <4dd037b70608102138v4ae9ebeay7be371a61ee4ba8b@mail.gmail.com>
References: <4dd037b70608102138v4ae9ebeay7be371a61ee4ba8b@mail.gmail.com>
Message-ID: <44DC2F60.9000804@sympatico.ca>

Blake Skinner wrote:
> Here's my C++ code, I'm building with VS 2005, under multi-threaded:
> 
> #include <Python/python.h>
> #include <iostream>
> #include <cstdio>
> using namespace std;

Please note that this list is about the python<->C++ integration
provided by boost.python, not any language binding (even though the
list name may suggest something else).


> 
> int main(int argc, char **argv)
> {
>    Py_Initialize();
> 
>    PyObject *main_module = PyImport_AddModule("__main__");
>    PyObject *main_dict = PyModule_GetDict(main_module);
> 
>    FILE *f = fopen("test.py", "r");
>    PyRun_SimpleFile(f, "test.py");
> 
>    if (PyDict_GetItemString(main_dict, "num") == NULL)
>        cout << "No" << endl;
>    else
>        cout << "Yes" << endl;
> 
>    Py_Finalize();
>    return 0;
> 
> }
> 
> When I run it, I get an error:
> 
> Unhandled exception at 0x7c918fea in Embed_practice.exe: 0xC0000005: Access
> violation writing location 0x00000010
> 
> from a file called _file.c, the debugger calls attention to this method
> specificially:

This sounds as if the definition of the FILE structure as seen by your
compiler isn't the same that was seen during compilation of the python support
library, resulting in an ABI incompatibility.
Make sure you use the same compiler for both, or avoid the explicit use
of FILE in your code (using PyFile_FromString and PyFile_FromString instead).

HTH,
		Stefan

-- 

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


From dave at boost-consulting.com  Fri Aug 11 17:06:11 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Fri, 11 Aug 2006 11:06:11 -0400
Subject: [C++-sig] boost.python newbie hassles
References: <200608091632.22195.w.richert@gmx.net>
	<Pine.LNX.4.63.0608101443250.24635@brocoli.iro.umontreal.ca>
	<ulkpw5gw1.fsf@boost-consulting.com>
	<e190e3480608101622l4e364b19y55c57802dd1e61f4@mail.gmail.com>
Message-ID: <u7j1f2ty4.fsf@boost-consulting.com>

"Roberto Aguilar" <roberto.c.aguilar at gmail.com> writes:

> On 8/10/06, David Abrahams <dave at boost-consulting.com> wrote:
> [...]
>> >> 2. boost.python projects have to be located in the boost directory tree, other
>> >> than printed at the tutorial docs - as far as I have tried. If I move the
>> >> tutorial greet example to another directory and update path-global in
>> >> Jamrules I always get:
>> >>
>> >> wr@[ikarus]:~/workspace/Boost.Python/src> bjam
>> >> Failed to find the project root for directory '.'.
>> >> Did not find a project-root.jam file there or in any of its parent
>> >> directories.
>> >> Please consult the documentation at 'http://www.boost.org'.
>>
>> see the libs/python/example/ subdirectory of your boost distribution
>> for an example that should be transplantable.  I'm not sure why it's
>> looking for project-root.jam.  It's almost as though you edited your
>> boost-build.jam to reference the tools/build/v2 subdirectory instead
>> of tools/build/v1.  That won't work for you :)
>
> Are you saying that boost.python does not work with bbv2?  

No I'm not saying that.  The user is issuing BBv1 invocation commands,
though.

> I see Jamfile.v2 files in the example directories.  But, if I try
> compiling outside the example directory, my project will not build.

What happens?

> I am trying to use bbv2, so if that's not possible then that's my
> problem.
>
> Also, if building with bbv2 is possible, is there some documentation
> on the process?

Not yet.


-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From thilinasepala at gmail.com  Mon Aug 14 11:10:55 2006
From: thilinasepala at gmail.com (thilina sepala dahanayake)
Date: Mon, 14 Aug 2006 15:10:55 +0600
Subject: [C++-sig] (no subject)
Message-ID: <e554344e0608140210s3638a833h19cbc4513f8a6c35@mail.gmail.com>

fuck you man don't send me any mail again
you are very very fool man
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060814/4b053243/attachment.htm>

From quncao at gmail.com  Tue Aug 15 09:46:53 2006
From: quncao at gmail.com (Qun Cao)
Date: Tue, 15 Aug 2006 00:46:53 -0700
Subject: [C++-sig] boost-python link errors
Message-ID: <1f527df80608150046g3c796663g24dedca7bc6b96cc@mail.gmail.com>

Hi Everyone,

I ran into some linking problem when I try to use MSVC7 IDE to compile
my python extension, I then tried to compile the getting_started1.cpp
(located on boost_1_33\libs\python\example), the same link error
happened.  My .dsp file was copied from
libs/python/build/VisualStudio.as instructed on the boost building and
testing manual on
http://www.boost.org/libs/python/doc/building.html#testing.
The only change I made was to add python include and libs path.  The
errors are like this:


Linking...
   Creating library .\../bin-stage/boost_python_debug.lib and object
.\../bin-stage/boost_python_debug.exp
getting_started2.obj : error LNK2019: unresolved external symbol
"public: virtual __thiscall
boost::python::error_already_set::~error_already_set(void)"
(??1error_already_set at python@boost@@UAE at XZ) referenced in function
"public: virtual void * __thiscall
boost::python::error_already_set::`scalar deleting
destructor'(unsigned int)"
(??_Gerror_already_set at python@boost@@UAEPAXI at Z)
getting_started1.obj : error LNK2019: unresolved external symbol
"public: virtual __thiscall
boost::python::error_already_set::~error_already_set(void)"
(??1error_already_set at python@boost@@UAE at XZ) referenced in function
"class std::ctype<char> const & __cdecl std::use_facet<class
std::ctype<char> >(class std::locale const &)"
(??$use_facet at V?$ctype at D@std@@@std@@YAABV?$ctype at D@0 at ABVlocale@0@@Z)
getting_started2.obj : error LNK2001: unresolved external symbol
"public: virtual unsigned int __thiscall
boost::python::objects::py_function_impl_base::max_arity(void)const "
(?max_arity at py_function_impl_base@objects at python@boost@@UBEIXZ)
getting_started1.obj : error LNK2019: unresolved external symbol
"public: virtual unsigned int __thiscall
boost::python::objects::py_function_impl_base::max_arity(void)const "
(?max_arity at py_function_impl_base@objects at python@boost@@UBEIXZ)
referenced in function $L65603


I went through the manual and could not find any hints on this, please help!

Thanks,
Qun


From roman.yakovenko at gmail.com  Tue Aug 15 09:55:42 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Tue, 15 Aug 2006 10:55:42 +0300
Subject: [C++-sig] boost-python link errors
In-Reply-To: <1f527df80608150046g3c796663g24dedca7bc6b96cc@mail.gmail.com>
References: <1f527df80608150046g3c796663g24dedca7bc6b96cc@mail.gmail.com>
Message-ID: <7465b6170608150055p6c19a343ha4a6ba6501a8af52@mail.gmail.com>

On 8/15/06, Qun Cao <quncao at gmail.com> wrote:
> Hi Everyone,
>
> I ran into some linking problem when I try to use MSVC7 IDE to compile
> my python extension, I then tried to compile the getting_started1.cpp
> (located on boost_1_33\libs\python\example), the same link error
> happened.  My .dsp file was copied from
> libs/python/build/VisualStudio.as instructed on the boost building and
> testing manual on
> http://www.boost.org/libs/python/doc/building.html#testing.
> The only change I made was to add python include and libs path.  The
> errors are like this:
>
>
> Linking...
> getting_started2.obj : error LNK2019: unresolved external symbol
> "public: virtual __thiscall
> boost::python::error_already_set::~error_already_set(void)"
> ...
> boost::python::objects::py_function_impl_base::max_arity(void)const "

> I went through the manual and could not find any hints on this, please help!

The dsp project is not maintained well. It is possible, that it is
missing some source files
or user wrong runtime library. Or may be you forgot to add
boost_python.lib to the
link line( or dependencies ).

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


From b.christophe at yahoo.fr  Tue Aug 15 11:26:35 2006
From: b.christophe at yahoo.fr (cbailly)
Date: Tue, 15 Aug 2006 11:26:35 +0200
Subject: [C++-sig] How argument passing works with boost python ?
Message-ID: <op.tebi6lolopqgqb@christop>

Hi,

    I have read tutorials/FAQS about boost python but still, I do not  
understand how arguments are passed and converted (automatically ?) to the  
C++ types... since the profile of the methods are not declared.

For instance, in the Hello world example (from the tutorial)

BOOST_PYTHON_MODULE(hello)
{
     class_<World>("World")
         .def("greet", &World::greet)
         .def("set", &World::set)
     ;
}

"set" method takes a std::string argument (not declared). This is really a  
mystery...  Is the std::string type compatible with the str type ? Where  
is the conversion made ?. What if the method set takes std::string &  
parameter instead ?

Can somebody help me understand the mechanism for argument passing ?  
Thanks a lot,

Chris


-- 
Utilisant le client e-mail r?volutionnaire d'Opera :  
http://www.opera.com/mail/



From roman.yakovenko at gmail.com  Tue Aug 15 12:26:05 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Tue, 15 Aug 2006 13:26:05 +0300
Subject: [C++-sig] How argument passing works with boost python ?
In-Reply-To: <op.tebi6lolopqgqb@christop>
References: <op.tebi6lolopqgqb@christop>
Message-ID: <7465b6170608150326g407b946am948d3b642e9a1747@mail.gmail.com>

On 8/15/06, cbailly <b.christophe at yahoo.fr> wrote:
> Hi,
>
>     I have read tutorials/FAQS about boost python but still, I do not
> understand how arguments are passed and converted (automatically ?) to the
> C++ types... since the profile of the methods are not declared.
>
> For instance, in the Hello world example (from the tutorial)
>
> BOOST_PYTHON_MODULE(hello)
> {
>      class_<World>("World")
>          .def("greet", &World::greet)
>          .def("set", &World::set)
>      ;
> }
>
> "set" method takes a std::string argument (not declared). This is really a
> mystery...  Is the std::string type compatible with the str type ? Where
> is the conversion made ?. What if the method set takes std::string &
> parameter instead ?

This link http://boost.org/libs/python/doc/v2/faq.html#custom_string
will help you to
understand what is going on in this specific case.

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


From b.christophe at yahoo.fr  Tue Aug 15 13:49:27 2006
From: b.christophe at yahoo.fr (cbailly)
Date: Tue, 15 Aug 2006 13:49:27 +0200
Subject: [C++-sig] How argument passing works with boost python ?
References: <op.tebi6lolopqgqb@christop>
	<7465b6170608150326g407b946am948d3b642e9a1747@mail.gmail.com>
Message-ID: <op.tebpspcgopqgqb@christop>

Thanks for your answer

Can you tell me if my understanding is ok:

Does it mean that each time a str python object is passed as an argument,  
then an automatic conversion is made to std::string type ? Regardless of  
the prototype of the method, just because converters have been registered.

str object -> alway converted to std::string ?
If a dict object is passed, no conversion is made ?

Sorry, it is still a bit confusing...

Chris



Le Tue, 15 Aug 2006 12:26:05 +0200, Roman Yakovenko  
<roman.yakovenko at gmail.com> a ?crit:

> On 8/15/06, cbailly <b.christophe at yahoo.fr> wrote:
>> Hi,
>>
>>     I have read tutorials/FAQS about boost python but still, I do not
>> understand how arguments are passed and converted (automatically ?) to  
>> the
>> C++ types... since the profile of the methods are not declared.
>>
>> For instance, in the Hello world example (from the tutorial)
>>
>> BOOST_PYTHON_MODULE(hello)
>> {
>>      class_<World>("World")
>>          .def("greet", &World::greet)
>>          .def("set", &World::set)
>>      ;
>> }
>>
>> "set" method takes a std::string argument (not declared). This is  
>> really a
>> mystery...  Is the std::string type compatible with the str type ? Where
>> is the conversion made ?. What if the method set takes std::string &
>> parameter instead ?
>
> This link http://boost.org/libs/python/doc/v2/faq.html#custom_string
> will help you to
> understand what is going on in this specific case.
>



-- 
Utilisant le client e-mail r?volutionnaire d'Opera :  
http://www.opera.com/mail/



From dave at boost-consulting.com  Tue Aug 15 14:42:22 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Tue, 15 Aug 2006 08:42:22 -0400
Subject: [C++-sig] How argument passing works with boost python ?
References: <op.tebi6lolopqgqb@christop>
	<7465b6170608150326g407b946am948d3b642e9a1747@mail.gmail.com>
	<op.tebpspcgopqgqb@christop>
Message-ID: <uodumqifl.fsf@boost-consulting.com>

cbailly <b.christophe at yahoo.fr> writes:

> Thanks for your answer
>
> Can you tell me if my understanding is ok:
>
> Does it mean that each time a str python object is passed as an
> argument, then an automatic conversion is made to std::string type ?
> Regardless of the prototype of the method, just because converters
> have been registered.

No: the prototype of the method is part of the type of the (member)
function pointer you pass to def().  The library uses this type
information to choose the eligible converters for each parameter.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From roman.yakovenko at gmail.com  Tue Aug 15 15:00:44 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Tue, 15 Aug 2006 16:00:44 +0300
Subject: [C++-sig] How argument passing works with boost python ?
In-Reply-To: <uodumqifl.fsf@boost-consulting.com>
References: <op.tebi6lolopqgqb@christop>
	<7465b6170608150326g407b946am948d3b642e9a1747@mail.gmail.com>
	<op.tebpspcgopqgqb@christop> <uodumqifl.fsf@boost-consulting.com>
Message-ID: <7465b6170608150600o51b0b5bdxcbef157768f382f2@mail.gmail.com>

On 8/15/06, David Abrahams <dave at boost-consulting.com> wrote:
> cbailly <b.christophe at yahoo.fr> writes:
>
> > Thanks for your answer
> >
> > Can you tell me if my understanding is ok:
> >
> > Does it mean that each time a str python object is passed as an
> > argument, then an automatic conversion is made to std::string type ?
> > Regardless of the prototype of the method, just because converters
> > have been registered.
>
> No: the prototype of the method is part of the type of the (member)
> function pointer you pass to def().  The library uses this type
> information to choose the eligible converters for each parameter.

Can you point to the source code?

Thanks

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


From b.christophe at yahoo.fr  Tue Aug 15 16:07:05 2006
From: b.christophe at yahoo.fr (cbailly)
Date: Tue, 15 Aug 2006 16:07:05 +0200
Subject: [C++-sig] How argument passing works with boost python ?
References: <op.tebi6lolopqgqb@christop>
	<7465b6170608150326g407b946am948d3b642e9a1747@mail.gmail.com>
	<op.tebpspcgopqgqb@christop> <uodumqifl.fsf@boost-consulting.com>
	<7465b6170608150600o51b0b5bdxcbef157768f382f2@mail.gmail.com>
Message-ID: <op.tebv53xaopqgqb@christop>

Ok, I think I've got it... It seemed a little bit magic for me to get the  
signature at compile time to call the right converters. It must be the  
magic of templates !

Thanks a lot for all your answers !.


Le Tue, 15 Aug 2006 15:00:44 +0200, Roman Yakovenko  
<roman.yakovenko at gmail.com> a ?crit:

> On 8/15/06, David Abrahams <dave at boost-consulting.com> wrote:
>> cbailly <b.christophe at yahoo.fr> writes:
>>
>> > Thanks for your answer
>> >
>> > Can you tell me if my understanding is ok:
>> >
>> > Does it mean that each time a str python object is passed as an
>> > argument, then an automatic conversion is made to std::string type ?
>> > Regardless of the prototype of the method, just because converters
>> > have been registered.
>>
>> No: the prototype of the method is part of the type of the (member)
>> function pointer you pass to def().  The library uses this type
>> information to choose the eligible converters for each parameter.
>
> Can you point to the source code?
>
> Thanks
>



-- 
Utilisant le client e-mail r?volutionnaire d'Opera :  
http://www.opera.com/mail/



From dave at boost-consulting.com  Tue Aug 15 18:33:51 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Tue, 15 Aug 2006 12:33:51 -0400
Subject: [C++-sig] How argument passing works with boost python ?
References: <op.tebi6lolopqgqb@christop>
	<7465b6170608150326g407b946am948d3b642e9a1747@mail.gmail.com>
	<op.tebpspcgopqgqb@christop> <uodumqifl.fsf@boost-consulting.com>
	<7465b6170608150600o51b0b5bdxcbef157768f382f2@mail.gmail.com>
Message-ID: <uveoum00g.fsf@boost-consulting.com>

"Roman Yakovenko" <roman.yakovenko at gmail.com> writes:

> On 8/15/06, David Abrahams <dave at boost-consulting.com> wrote:
>> cbailly <b.christophe at yahoo.fr> writes:
>>
>> > Thanks for your answer
>> >
>> > Can you tell me if my understanding is ok:
>> >
>> > Does it mean that each time a str python object is passed as an
>> > argument, then an automatic conversion is made to std::string type ?
>> > Regardless of the prototype of the method, just because converters
>> > have been registered.
>>
>> No: the prototype of the method is part of the type of the (member)
>> function pointer you pass to def().  The library uses this type
>> information to choose the eligible converters for each parameter.
>
> Can you point to the source code?

Start with make_function.hpp and work your way down from there.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From richjohnson905 at msn.com  Tue Aug 15 22:30:07 2006
From: richjohnson905 at msn.com (RICH JOHNSON)
Date: Tue, 15 Aug 2006 15:30:07 -0500
Subject: [C++-sig] ImportError: timemodule.so
In-Reply-To: <BAY107-F200BD0104F20E4EF95793498570@phx.gbl>
Message-ID: <BAY107-F205330B8006D42F1E85C43984F0@phx.gbl>

Hi Rich,

When you get the following error at runtime on a unix system:

Traceback (most recent call last):
File "<string>", line 1, in ?
ImportError: /usr/lib/python2.2/lib-dynload/timemodule.so: undefined symbol:
PyDict_SetItemString

it is possible that you may need to use the -export-dynamic -Xlinker flag.

See http://docs.python.org/ext/link-reqs.html for more information.

----Original Message Follows----
From: "RICH JOHNSON" <richjohnson905 at msn.com>
Reply-To: Development of Python/C++ integration <c++-sig at python.org>
To: c++-sig at python.org
Subject: [C++-sig] ImportError: timemodule.so
Date: Mon, 07 Aug 2006 08:07:17 -0500

I am new to boost python and am having difficulty importing python modules
into a little C++ program.  Any insight into the problem would be
appreciated.

OS: Linux Redhat
gcc: 3.2.3
python 2.2
boost 1.33.1

 >From the command line I can do:
[rjohnson at pitwar7 rjohnson]$ python
Python 2.2.3 (#1, Feb 15 2005, 02:41:06)
[GCC 3.2.3 20030502 (Red Hat Linux 3.2.3-49)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
 >>>from time import time,ctime
 >>>print 'Today is', ctime(time())
Today is Fri Aug  4 10:47:20 2006
 >>>

But the C++ program below outputs:
25
20
Traceback (most recent call last):
File "<string>", line 1, in ?
ImportError: /usr/lib/python2.2/lib-dynload/timemodule.so: undefined symbol:
PyDict_SetItemString


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

int main(int, char **)
{
using namespace boost::python;

Py_Initialize();

try
{
   PyRun_SimpleString("result = 5 ** 2");
      object module((handle<>(borrowed(PyImport_AddModule("__main__")))));
   object dictionary = module.attr("__dict__");
   object result = dictionary["result"];
   int result_value = extract<int>(result);
      std::cout << result_value << std::endl;
      dictionary["result"] = 20;

   PyRun_SimpleString("print result");

   // Here is the same command that ran successfully at the command line
above but fails from  C++
   PyRun_SimpleString("from time import time,ctime\n"
                    "print 'Today is',ctime(time())\n");
}
catch (error_already_set)
{
   PyErr_Print();
}

Py_Finalize();
return 0;
}

My LD_LIBRARY_PATH is /usr/local/lib:/usr/lib/python2.2/config
PYTHONPATH is /usr/lib/python2.2/config


_______________________________________________
C++-sig mailing list
C++-sig at python.org
http://mail.python.org/mailman/listinfo/c++-sig




From eduardo.picado at gmail.com  Wed Aug 16 18:25:52 2006
From: eduardo.picado at gmail.com (Eduardo Picado)
Date: Wed, 16 Aug 2006 18:25:52 +0200
Subject: [C++-sig] =?utf-8?q?Modules_using_Boost_Python_are_giving_?=
	=?utf-8?q?=E2=80=9Cerror_LNK2019=3A_unresolved_external_symbol?=
	=?utf-8?b?4oCd?=
Message-ID: <6c75ff8f0608160925n20150e4dq4acc39c434788c8c@mail.gmail.com>

Hi,

Some modules of the application I'm trying to build with Boost Python
are giving the same problem, for example:

# Jamfile:

project-root ;

import python ;

extension TASRRecognitionControl
:   TASRRecognitionControl.cpp
    # requirements and dependencies for Boost.Python extensions
    <template>@boost/libs/python/build/extension
    : # additional requirements
    <library-file>../telisma/lib/libtasr_so.lib
    ;


# TASRRecognitionControl.hpp (only showing relevant parts)

struct TASRRecognitionControlInternal;

class TASRRecognitionControl {
private:
    TASRRecognitionControlInternal * _ctrl;

    friend class TASREngineInternal;

public:
    TASRRecognitionControl(TASRRecognitionControlInternal * ctrl);

public:

    TASRRecognitionControl();

    TASRRecognitionControl(TASRRecognitionControl const & recoctrl);
    ~TASRRecognitionControl();


# TASRRecognitionControl.cpp (only showing relevant parts)

BOOST_PYTHON_MODULE(TASRRecognitionControl)
{
    class_< TASRRecognitionControl >("TASRRecognitionControl", init<  >())
        .def(init< TASRRecognitionControlInternal* >()) // HERE is the problem
        .def(init< const TASRRecognitionControl& >())


    ;

}


Compiling:

C:\boost_1_33_1\libs\python\example\telismaPyste>bjam -sTOOLS=vc-8_0
...found 1624 targets...
...updating 4 targets...
vc-C++ bin\telismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.obj
TASRRecognitionControl.cpp
vc-Link bin\telismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.pyd
bin\telismaPyste\
TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.lib
   Creating library
bin\telismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.lib
and o
bject bin\telismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.exp
TASRRecognitionControl.obj : error LNK2019: unresolved external symbol
"public: __thiscall TASRRecognitionControl::TASRRecogn
itionControl(struct TASRRecognitionControlInternal *)"
(??0TASRRecognitionControl@@QAE at PAUTASRRecognitionControlInternal@@@Z)
 referenced in function "public: __thiscall
boost::python::objects::value_holder<class
TASRRecognitionControl>::value_holder<
class TASRRecognitionControl><struct TASRRecognitionControlInternal
*>(struct _object *,struct TASRRecognitionControlInternal
 *)" (??$?0PAUTASRRecognitionControlInternal@@@?$value_holder at VTASRRecognitionControl@@@objects at python@boost@@QAE at PAU_object@
@PAUTASRRecognitionControlInternal@@@Z)
bin\telismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.pyd
: fatal error LNK1120: 1
unresolved externals
Microsoft (R) Manifest Tool version 5.2.3790.2014
Copyright (c) Microsoft Corporation 2005.
All rights reserved.

mt.exe : general error c10100b1: Failed to load file
"bin\telismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-mult
i\TASRRecognitionControl.pyd". Le fichier spicifii est introuvable.

    CALL "C:\Program Files\Microsoft Visual Studio
8\Common7\Tools\..\..\VC\bin\vcvars32.BAT" >nul
    "C:\Program Files\Microsoft Visual Studio
8\Common7\Tools\..\..\VC\bin\link" /nologo /INCREMENTAL:NO /DEBUG /DLL
/subsyst
em:console  /out:"bin\telismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.pyd"
/IMPLI
B:"bin\telismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.lib"
/LIBPATH:"c:\Python24
\libs" /LIBPATH:"bin\boost\libs\python\build\boost_python.dll\vc-8_0\debug\threading-multi"
 "python24.lib" "boost_python.lib
" @"bin\telismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.CMD"
    mt -manifest
bin\telismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.pyd.manifest
 -outputresource:bin\telismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.pyd;#2

...failed vc-Link
bin\telismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.pyd
bin\tel
ismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.lib...
...failed updating 2 targets...
...updated 2 targets...



So, I don't know how to export the constructor
 .def(init< TASRRecognitionControlInternal* >())
that uses the class TASRRecognitionControlInternal that is declared
but not defined in my header.

Thanks very much,

Eduardo Picado


-- 
Eduardo PICADO
MSN eduardo.picado at gmail.com
portable +33 6 32 47 10 10 (NEW)
fix +33 1 45 03 52 54
|??<?????>??|

From roman.yakovenko at gmail.com  Wed Aug 16 19:40:16 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Wed, 16 Aug 2006 20:40:16 +0300
Subject: [C++-sig]
	=?windows-1252?q?Modules_using_Boost_Python_are_giving_?=
	=?windows-1252?q?=93error_LNK2019=3A_unresolved_external_symbol=94?=
In-Reply-To: <6c75ff8f0608160925n20150e4dq4acc39c434788c8c@mail.gmail.com>
References: <6c75ff8f0608160925n20150e4dq4acc39c434788c8c@mail.gmail.com>
Message-ID: <7465b6170608161040w5eb76694yb1b5bb4c074368c3@mail.gmail.com>

On 8/16/06, Eduardo Picado <eduardo.picado at gmail.com> wrote:
> Hi,
>
> So, I don't know how to export the constructor
>  .def(init< TASRRecognitionControlInternal* >())
> that uses the class TASRRecognitionControlInternal that is declared
> but not defined in my header.

I think that if you will add include to the header file
TASRRecognitionControlInternal defined in, you will be able to compile.

Also you will have to expose this class too. Otherwise, you will not be able
to call this function, from Python.

P.S.  It seems that  TASRRecognitionControlInternal is "private" class.
May you be you do not have to export it?

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


From eduardo.picado at gmail.com  Thu Aug 17 14:46:53 2006
From: eduardo.picado at gmail.com (Eduardo Picado)
Date: Thu, 17 Aug 2006 14:46:53 +0200
Subject: [C++-sig] =?utf-8?q?Modules_using_Boost_Python_are_giving_?=
	=?utf-8?q?=E2=80=9Cerror_LNK2019=3A_unresolved_external_symbol?=
	=?utf-8?b?4oCd?=
In-Reply-To: <7465b6170608161040w5eb76694yb1b5bb4c074368c3@mail.gmail.com>
References: <6c75ff8f0608160925n20150e4dq4acc39c434788c8c@mail.gmail.com>
	<7465b6170608161040w5eb76694yb1b5bb4c074368c3@mail.gmail.com>
Message-ID: <6c75ff8f0608170546u2b189b9fr95440e97d4890253@mail.gmail.com>

Roman,

Thank you very much for your attention.

On 8/16/06, Roman Yakovenko <roman.yakovenko at gmail.com> wrote:
> On 8/16/06, Eduardo Picado <eduardo.picado at gmail.com> wrote:
> > Hi,
> >
> > So, I don't know how to export the constructor
> >  .def(init< TASRRecognitionControlInternal* >())
> > that uses the class TASRRecognitionControlInternal that is declared
> > but not defined in my header.
>
> I think that if you will add include to the header file
> TASRRecognitionControlInternal defined in, you will be able to compile.
>

I'm already including the original header (TASRRecognitionControl.hpp)
- which declares TASRRecognitionControlInternal - in the wrapper
(TASRRecognitionControl.cpp), I have  just forgot to copy the line
here.

> Also you will have to expose this class too. Otherwise, you will not be able
> to call this function, from Python.
>
> P.S.  It seems that  TASRRecognitionControlInternal is "private" class.
> May you be you do not have to export it?

Exposing the class using Pyste the wrapper becomes:

# TASRRecognitionControl.cpp

// Boost Includes ==============================================================
#include <boost/python.hpp>
#include <boost/cstdint.hpp>

// Includes ====================================================================
#include <TASRRecognitionControl.hpp>

// Using =======================================================================
using namespace boost::python;

// Module ======================================================================
BOOST_PYTHON_MODULE(TASRRecognitionControl)
{
    class_< TASRRecognitionControlInternal, boost::noncopyable
>("TASRRecognitionControlInternal", no_init)
    ;

    class_< TASRRecognitionControl >("TASRRecognitionControl", init<  >())
        .def(init< TASRRecognitionControlInternal* >())
        .def(init< const TASRRecognitionControl& >())
        .def("setParameter", (TASRRecognitionControl&
(TASRRecognitionControl::*)(TASRSystem::Option, int)
throw(TASRValueOutOfRange,
TASRUnsupportedOption))&TASRRecognitionControl::setParameter,
return_internal_reference< 1 >())
        .def("setParameter", (TASRRecognitionControl&
(TASRRecognitionControl::*)(TASRSystem::OptionString, const
TASRString&)  throw(TASRUnsupportedOption))&TASRRecognitionControl::setParameter,
return_internal_reference< 1 >())
        .def("getParameter", (int
(TASRRecognitionControl::*)(TASRSystem::Option) const
throw(TASRUnsupportedOption))&TASRRecognitionControl::getParameter)
        .def("getParameter", (const char*
(TASRRecognitionControl::*)(TASRSystem::OptionString) const
throw(TASRUnsupportedOption))&TASRRecognitionControl::getParameter)
    ;

}


Compiling now I receive:


C:\boost_1_33_1\boost/type_traits/is_polymorphic.hpp(36) : error
C2504: 'TASRRecognitionControlInternal' : base class undefined
        C:\boost_1_33_1\boost/type_traits/is_polymorphic.hpp(48) : see
reference to class template instantiation
'boost::detail::is_polymorphic_imp1<T>::d2'
        with
        [
            T=TASRRecognitionControlInternal
        ]
        C:\boost_1_33_1\boost/type_traits/is_polymorphic.hpp(83) : see
reference to class template instantiation
'boost::detail::is_polymorphic_imp1<T>' bein
        with
        [
            T=TASRRecognitionControlInternal
        ]
        C:\boost_1_33_1\boost/type_traits/is_polymorphic.hpp(88) : see
reference to class template instantiation
'boost::detail::is_polymorphic_imp<T>' being
        with
        [
            T=TASRRecognitionControlInternal
        ]
        C:\boost_1_33_1\boost/mpl/if.hpp(63) : see reference to class
template instantiation 'boost::is_polymorphic<T>' being compiled
        with
        [
            T=TASRRecognitionControlInternal
        ]
        C:\boost_1_33_1\boost/python/object/inheritance.hpp(65) : see
reference to class template instantiation 'boost::mpl::if_<T1,T2,T3>'
being compiled
        with
        [
            T1=boost::is_polymorphic<TASRRecognitionControlInternal>,
            T2=boost::python::objects::polymorphic_id_generator<TASRRecognitionControlInternal>,
            T3=boost::python::objects::non_polymorphic_id_generator<TASRRecognitionControlInternal>
        ]
        C:\boost_1_33_1\boost/python/object/inheritance.hpp(72) : see
reference to class template instantiation
'boost::python::objects::dynamic_id_generator
        with
        [
            T=TASRRecognitionControlInternal
        ]
        C:\boost_1_33_1\boost/python/object/class_metadata.hpp(97) :
see reference to function template instantiation 'void
boost::python::objects::register_
        with
        [
            T=TASRRecognitionControlInternal
        ]
        C:\boost_1_33_1\boost/python/object/class_metadata.hpp(225) :
see reference to function template instantiation 'void
boost::python::objects::register
:python::bases<>>(T *,Bases)' being compiled
        with
        [
            T2=TASRRecognitionControlInternal,
            T=TASRRecognitionControlInternal,
            Bases=boost::python::bases<>
        ]
        C:\boost_1_33_1\boost/python/object/class_metadata.hpp(219) :
see reference to function template instantiation 'void
boost::python::objects::class_me
integral_constant<bool,false>>(T2 *,Callback)' being compiled
        with
        [
            T=TASRRecognitionControlInternal,
            X1=boost::noncopyable,
            X2=boost::python::detail::not_specified,
            X3=boost::python::detail::not_specified,
            T2=TASRRecognitionControlInternal,
            Callback=boost::integral_constant<bool,false>
        ]
        C:\boost_1_33_1\boost/python/object/class_metadata.hpp(217) :
while compiling class template member function 'void
boost::python::objects::class_meta
        with
        [
            T=TASRRecognitionControlInternal,
            X1=boost::noncopyable,
            X2=boost::python::detail::not_specified,
            X3=boost::python::detail::not_specified
        ]
        C:\boost_1_33_1\boost/python/class.hpp(174) : see reference to
class template instantiation
'boost::python::objects::class_metadata<T,X1,X2,X3>' bein
        with
        [
            T=TASRRecognitionControlInternal,
            X1=boost::noncopyable,
            X2=boost::python::detail::not_specified,
            X3=boost::python::detail::not_specified
        ]
        C:\boost_1_33_1\boost/python/class.hpp(635) : see reference to
class template instantiation 'boost::python::class_<W,X1>::id_vector'
being compiled
        with
        [
            W=TASRRecognitionControlInternal,
            X1=boost::noncopyable
        ]
        C:\boost_1_33_1\boost/python/class.hpp(635) : while compiling
class template member function
'boost::python::class_<W,X1>::class_(const char *,boost:
        with
        [
            W=TASRRecognitionControlInternal,
            X1=boost::noncopyable
        ]
        TASRRecognitionControl.cpp(15) : see reference to class
template instantiation 'boost::python::class_<W,X1>' being compiled
        with
        [
            W=TASRRecognitionControlInternal,
            X1=boost::noncopyable
        ]
C:\boost_1_33_1\boost/type_traits/is_polymorphic.hpp(28) : error
C2504: 'TASRRecognitionControlInternal' : base class undefined
        C:\boost_1_33_1\boost/type_traits/is_polymorphic.hpp(48) : see
reference to class template instantiation
'boost::detail::is_polymorphic_imp1<T>::d1'
        with
        [
            T=TASRRecognitionControlInternal
        ]

    CALL "C:\Program Files\Microsoft Visual Studio
8\Common7\Tools\..\..\VC\bin\vcvars32.BAT" >nul
    "C:\Program Files\Microsoft Visual Studio
8\Common7\Tools\..\..\VC\bin\cl" /Zm800 -nologo /EHsc -c
-DBOOST_PYTHON_DYNAMIC_LIB /Z7 /Od /Ob0 /EHsc /GR /MD
ste"  -I"C:\boost_1_33_1" -I"c:\Python24\include"
-Fo"bin\telismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.obj"
-

...failed vc-C++
bin\telismaPyste\TASRRecognitionControl.pyd\vc-8_0\debug\threading-multi\TASRRecognitionControl.obj...


------------------
that seems even worse...


-- 
Eduardo PICADO
MSN eduardo.picado at gmail.com
portable +33 6 32 47 10 10 (NEW)
fix +33 1 45 03 52 54
|??<?????>??|

From 5p4ced0ut at gmail.com  Thu Aug 17 06:26:38 2006
From: 5p4ced0ut at gmail.com (Blake Skinner)
Date: Thu, 17 Aug 2006 04:26:38 +0000 (UTC)
Subject: [C++-sig] =?utf-8?q?Under_what_conditions_would_PyRun=5FSimpleFil?=
	=?utf-8?q?e_return_false=3F?=
Message-ID: <loom.20060817T061627-965@post.gmane.org>

PyRun_SimpleFile() doesn't seem to like me.  First, passing a FILE* created by
fopen() in Visual Studio 2005 caused a runtime error, but I found a work around:

FILE* fp = PyFile_AsFile( PyFile_FromString(filename, "r") );
PyRun_SimpleString(fp, filename);

I'm mentioning this so you understand why I'm doing it in my code.

This worked fine, until I tried to construct a class around boost python, here's
the jist of it:

// constructor, Initializes python, creates the main module and gets the
// dictionary
pyEngine::pyEngine()
{
	Py_Initialize();

	// main module and main dictionary
	object main((
		handle<>(borrowed(PyImport_AddModule("__main__")))
		));
	_main = main;
	_mdict = _main.attr("__dict__");
}

// destructor
// finalizes python
pyEngine::~pyEngine()
{
	Py_Finalize();
}

//////////////////////////////////////////////////////////////////
// methods

// load a python module with the given filename
bool pyEngine::LoadModule(char *filename)
{
// originally more condensed, the 'if's were added for debugging
	PyObject *py_file = PyFile_FromString(filename, "r");
	if (!py_file)
	{
		fprintf(stderr, "Unable to locate module (pyEngine)\n");
		return false;
	}

	FILE *fp = PyFile_AsFile( py_file );
	if (!fp)
	{
		fprintf(stderr, "PyFile_AsFile returned NULL\n");
		return false;
	}

	if (!(PyRun_SimpleFile(fp, filename) == 1))
	{
		fprintf(stderr, "PyRun_SimpleFile returned 0\n");

		return false;
	}

	return true;
}

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

>From stderr I found that PyRun_SimpleFile was returning 0.  I checked the
documentation and it didn't mention what conditions that would happen for.  I'll
keep looking for it on my own and post if I solve it, but any help would be
appreciated.

Thanks.



From seefeld at sympatico.ca  Thu Aug 17 17:39:09 2006
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Thu, 17 Aug 2006 11:39:09 -0400
Subject: [C++-sig] Under what conditions would PyRun_SimpleFile return
 false?
In-Reply-To: <loom.20060817T061627-965@post.gmane.org>
References: <loom.20060817T061627-965@post.gmane.org>
Message-ID: <44E48D9D.3050103@sympatico.ca>

Blake Skinner wrote:
> PyRun_SimpleFile() doesn't seem to like me.  First, passing a FILE* created by
> fopen() in Visual Studio 2005 caused a runtime error, but I found a work around:
> 
> FILE* fp = PyFile_AsFile( PyFile_FromString(filename, "r") );
> PyRun_SimpleString(fp, filename);

I'm not sure what runtime error the above is a workaround for. There is some
ABI incompatibility between different versions of MSVC related to the FILE
structure, and one way to avoid tripping over that is not to hold FILE pointers
directly, but passing around python objects instead (PyFile_FromString to create
it and PyRun_File(PyFile_AsFile(...), ...) to use it).

> I'm mentioning this so you understand why I'm doing it in my code.
> 
> This worked fine, until I tried to construct a class around boost python, here's
> the jist of it:
> 
> // constructor, Initializes python, creates the main module and gets the
> // dictionary
> pyEngine::pyEngine()
> {
> 	Py_Initialize();
> 
> 	// main module and main dictionary
> 	object main((
> 		handle<>(borrowed(PyImport_AddModule("__main__")))
> 		));
> 	_main = main;
> 	_mdict = _main.attr("__dict__");
> }
> 
> // destructor
> // finalizes python
> pyEngine::~pyEngine()
> {
> 	Py_Finalize();

Using Py_Finalize() doesn't work with current boost.python versions.
(Search the archives to find more context around that.)


> }
> 
> //////////////////////////////////////////////////////////////////
> // methods
> 
> // load a python module with the given filename
> bool pyEngine::LoadModule(char *filename)
> {
> // originally more condensed, the 'if's were added for debugging
> 	PyObject *py_file = PyFile_FromString(filename, "r");
> 	if (!py_file)
> 	{
> 		fprintf(stderr, "Unable to locate module (pyEngine)\n");
> 		return false;
> 	}
> 
> 	FILE *fp = PyFile_AsFile( py_file );
> 	if (!fp)
> 	{
> 		fprintf(stderr, "PyFile_AsFile returned NULL\n");
> 		return false;
> 	}
> 
> 	if (!(PyRun_SimpleFile(fp, filename) == 1))
> 	{
> 		fprintf(stderr, "PyRun_SimpleFile returned 0\n");
> 
> 		return false;
> 	}
> 
> 	return true;
> }
> 
> ///////////////////////////////////////////////////////////////////////////////
> 
>>From stderr I found that PyRun_SimpleFile was returning 0.  I checked the
> documentation and it didn't mention what conditions that would happen for.  I'll
> keep looking for it on my own and post if I solve it, but any help would be
> appreciated.

In general, if a python C API function returns 0 it has an exception set that you
should query via PyErr_Occured(), PyErr_Fetch(), and the like.

The current CVS version of boost.python (including the branch that will end up
as version 1.34 if ever that gets released some day) contains a boost::python::exec
and boost::python::exec_file function that should provide all you need. Here is
the code for the latter:

// Execute python source code from file filename.
// global and local are the global and local scopes respectively,
// used during execution.
object exec_file(str filename, object global, object local)
{
  // should be 'char const *' but older python versions don't use 'const' yet.
  char *f = python::extract<char *>(filename);
  // Let python open the file to avoid potential binary incompatibilities.
  PyObject *pyfile = PyFile_FromString(f, "r");
  if (!pyfile) throw std::invalid_argument(std::string(f) + " : no such file");
  python::handle<> file(pyfile);
  PyObject* result = PyRun_File(PyFile_AsFile(file.get()),
                f,
                Py_file_input,
                global.ptr(), local.ptr());
  if (!result) throw_error_already_set();
  return object(detail::new_reference(result));
}

You may want to borrow from it. As you can see, if the call to PyRun_File fails
(i.e. returns 0), an exception is thrown that captures (or allows a handler to capture)
the associated python exception.

HTH,
		Stefan

-- 

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


From roman.yakovenko at gmail.com  Thu Aug 17 19:16:34 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 17 Aug 2006 20:16:34 +0300
Subject: [C++-sig]
	=?windows-1252?q?Modules_using_Boost_Python_are_giving_?=
	=?windows-1252?q?=93error_LNK2019=3A_unresolved_external_symbol=94?=
In-Reply-To: <6c75ff8f0608170546u2b189b9fr95440e97d4890253@mail.gmail.com>
References: <6c75ff8f0608160925n20150e4dq4acc39c434788c8c@mail.gmail.com>
	<7465b6170608161040w5eb76694yb1b5bb4c074368c3@mail.gmail.com>
	<6c75ff8f0608170546u2b189b9fr95440e97d4890253@mail.gmail.com>
Message-ID: <7465b6170608171016k406f7ee2tbc5ea87f8863882c@mail.gmail.com>

On 8/17/06, Eduardo Picado <eduardo.picado at gmail.com> wrote:
> I'm already including the original header (TASRRecognitionControl.hpp)
> - which declares TASRRecognitionControlInternal - in the wrapper
> (TASRRecognitionControl.cpp), I have  just forgot to copy the line
> here.
>

It is hard to see the problem. The generated code seems to be fine. It will help
if you will post small C++ code that reproduce the problem.

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


From 5p4ced0ut at gmail.com  Sat Aug 19 16:45:54 2006
From: 5p4ced0ut at gmail.com (Blake Skinner)
Date: Sat, 19 Aug 2006 14:45:54 +0000 (UTC)
Subject: [C++-sig]
	=?utf-8?q?Under_what_conditions_would_PyRun=5FSimpleFil?=
	=?utf-8?q?e_return_false=3F?=
References: <loom.20060817T061627-965@post.gmane.org>
	<44E48D9D.3050103@sympatico.ca>
Message-ID: <loom.20060819T164417-901@post.gmane.org>

Thanks for the help, but...

I'm sorry, but where do I get the global and local scopes from?  I didn't see
that mentioned in the boost python tutorial or the one on python.org.






From foolscreen at gmail.com  Mon Aug 21 11:26:38 2006
From: foolscreen at gmail.com (bEngO)
Date: Mon, 21 Aug 2006 18:26:38 +0900
Subject: [C++-sig] Embedding Python into a Shared Object
Message-ID: <39ae57c30608210226x5216bcecn1f40624288ecde8e@mail.gmail.com>

Dear all,

I am trying to embed Python into an application through a plug-in (
dynamically loaded shared object ).
And I do not know how I should link with Python when building my plug-in.
The plugin ( embedded Python interpreter ) works until I try to import a
module.
Import failed because some symbols ( like _Py_NoneStruct ) are undefined.

The Python doc ( http://docs.python.org/ext/link-reqs.html ) says I should
use the flags given by:

>>> import distutils.sysconfig
>>> distutils.sysconfig.get_config_var('LINKFORSHARED')

Which are ( i am using GNU GCC 4.1.0 on Linux ):

'-Xlinker -export-dynamic'

( I also tried -rdynamic and -Wl,-E which should be equivalent )

But this is doing nothing for me.
Is it because this linker flag is only relevant when building an executable,
and do nothing when building a shared object ?
Or is it a bug of GCC 4.1.0 ?
Or is it because I link against libpythonX.X.so instead of libpythonX.X.a ?
But I think I shouldn't link against libpythonX.X.a since it isn't Position
Independent Code ( GCC -fpic flag ) which is required when building a shared
object...

Please help :)
I searched the web and found people being stuck on the same problem but no
workaround...

Thanks in advance.
bEngo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060821/9e3a8292/attachment.htm>

From seefeld at sympatico.ca  Tue Aug 22 13:41:50 2006
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Tue, 22 Aug 2006 07:41:50 -0400
Subject: [C++-sig] Under what conditions would PyRun_SimpleFile return
 false?
In-Reply-To: <loom.20060819T164417-901@post.gmane.org>
References: <loom.20060817T061627-965@post.gmane.org>	<44E48D9D.3050103@sympatico.ca>
	<loom.20060819T164417-901@post.gmane.org>
Message-ID: <44EAED7E.7000406@sympatico.ca>

Blake Skinner wrote:
> Thanks for the help, but...

Sorry for the late reply.

> I'm sorry, but where do I get the global and local scopes from?  I didn't see
> that mentioned in the boost python tutorial or the one on python.org.

These scopes can be used to pass context information (global and local variables)
around, such as from one call to exec to the next. You can pass empty
dictionaries, even the same for global and local.

  dict global;
  object result = exec_file(script_name, global, global);

should just work.

HTH,
		Stefan

-- 

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


From eduardo.picado at gmail.com  Tue Aug 22 15:45:01 2006
From: eduardo.picado at gmail.com (Eduardo Picado)
Date: Tue, 22 Aug 2006 15:45:01 +0200
Subject: [C++-sig] Boost Python private members problem
Message-ID: <6c75ff8f0608220645l43694e6ar37a519df80e96f53@mail.gmail.com>

Hi,

I'm facing a problem with the last module I need to compile with Boost Python.
I receive this compilation error:

Compiling...(with vc-8_0)
TASRLogger.cpp
C:\Python24\include\TASRLogger.hpp(353) : error C2248:
'TASRLogger::TASRLogger' : cannot access private member declared in
class 'TASRLogger'
        C:\Python24\include\TASRLogger.hpp(37) : see declaration of
'TASRLogger::TASRLogger'
        C:\Python24\include\TASRLogger.hpp(33) : see declaration of 'TASRLogger'
        This diagnostic occurred in the compiler generated function
'TASRDefaultLogger::TASRDefaultLogger(const TASRDefaultLogger &)'

It's hard find a solution for the problem as I don't have the sources
for the dll. Otherwise I could at least change the member attributes
visibility.
Is there a workaround for it?
I tried to extract the relevant parts from the files.

The TASRLogger.hpp I'm trying to wrap:


class TASRLoggerInternal;
class TASRDefaultLoggerInternal;
struct GrammarsInternal;

/**
 * Base virtual class for session loggers.
 * The TASRLogger specifies all methods a session logger should implement.
 */

class TASRLogger {

    TASRLoggerInternal * _logger;

    TASRLogger(TASRLogger const &);

public:

    TASRLogger() throw (std::bad_alloc);

    virtual ~TASRLogger();
};

/**
 * Default implementation for simple session loggers.
 */
class TASRDefaultLogger : public TASRLogger {
    TASRDefaultLoggerInternal * data;

    TASRDefaultLogger(TASRDefaultLogger const &);

public:
    TASRDefaultLogger(size_t diskMinFreeSize = 10000, bool
hierarchical = true, TASRString const & root = "")
         throw (std::bad_alloc, TASRDirectoryAccessError);

    ~TASRDefaultLogger();
};


The TASRLogger.cpp:


// Boost Includes ==============================================================
#include <boost/python.hpp>
#include <boost/cstdint.hpp>

// Includes ====================================================================
#include <TASRLogger.hpp>

// Using =======================================================================
using namespace boost::python;

// Declarations ================================================================
namespace  {

struct TASRLogger_Wrapper: TASRLogger
{
    TASRLogger_Wrapper(PyObject* py_self_):
        TASRLogger(), py_self(py_self_) {}

    PyObject* py_self;
};

struct TASRDefaultLogger_Wrapper: TASRDefaultLogger
{
    TASRDefaultLogger_Wrapper(PyObject* py_self_, const TASRDefaultLogger& p0):
        TASRDefaultLogger(p0), py_self(py_self_) {}

    TASRDefaultLogger_Wrapper(PyObject* py_self_):
        TASRDefaultLogger(), py_self(py_self_) {}

    TASRDefaultLogger_Wrapper(PyObject* py_self_, size_t p0):
        TASRDefaultLogger(p0), py_self(py_self_) {}

    TASRDefaultLogger_Wrapper(PyObject* py_self_, size_t p0, bool p1):
        TASRDefaultLogger(p0, p1), py_self(py_self_) {}

    TASRDefaultLogger_Wrapper(PyObject* py_self_, size_t p0, bool p1,
const TASRString& p2):
        TASRDefaultLogger(p0, p1, p2), py_self(py_self_) {}

    PyObject* py_self;
};


}// namespace


// Module ======================================================================
BOOST_PYTHON_MODULE(TASRLogger)
{
    scope* TASRLogger_scope = new scope(
    class_< TASRLogger, boost::noncopyable, TASRLogger_Wrapper
>("TASRLogger", init<  >())
    );

    delete TASRLogger_scope;

    class_< TASRDefaultLogger, bases< TASRLogger > ,
TASRDefaultLogger_Wrapper >("TASRDefaultLogger", init< const
TASRDefaultLogger& >())
        .def(init< optional< size_t, bool, const TASRString& > >())
    ;

}


Thanks,

Eduardo Picado


-- 
Eduardo PICADO
MSN eduardo.picado at gmail.com
portable +33 6 32 47 10 10 (NEW)
fix +33 1 45 03 52 54
|??<?????>??|

From seefeld at sympatico.ca  Tue Aug 22 15:53:04 2006
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Tue, 22 Aug 2006 09:53:04 -0400
Subject: [C++-sig] Boost Python private members problem
In-Reply-To: <6c75ff8f0608220645l43694e6ar37a519df80e96f53@mail.gmail.com>
References: <6c75ff8f0608220645l43694e6ar37a519df80e96f53@mail.gmail.com>
Message-ID: <44EB0C40.8040405@sympatico.ca>

Eduardo Picado wrote:
> Hi,
> 
> I'm facing a problem with the last module I need to compile with Boost Python.
> I receive this compilation error:
> 
> Compiling...(with vc-8_0)
> TASRLogger.cpp
> C:\Python24\include\TASRLogger.hpp(353) : error C2248:
> 'TASRLogger::TASRLogger' : cannot access private member declared in
> class 'TASRLogger'
>         C:\Python24\include\TASRLogger.hpp(37) : see declaration of
> 'TASRLogger::TASRLogger'
>         C:\Python24\include\TASRLogger.hpp(33) : see declaration of 'TASRLogger'
>         This diagnostic occurred in the compiler generated function
> 'TASRDefaultLogger::TASRDefaultLogger(const TASRDefaultLogger &)'

I find confusing that the compiler claims the copy-constructor to be compiler-generated,
as you have declared it explicitly.

> It's hard find a solution for the problem as I don't have the sources
> for the dll. Otherwise I could at least change the member attributes
> visibility.
> Is there a workaround for it?

Well, the TASRDefaultLogger is non-copyable, and yet you try to copy it. You can't.
As you can't pass values around, you have to find a way to operate with pointers or
references.

HTH,
		Stefan

-- 

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


From dave at boost-consulting.com  Wed Aug 23 00:18:14 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Tue, 22 Aug 2006 18:18:14 -0400
Subject: [C++-sig] Boost Python private members problem
References: <6c75ff8f0608220645l43694e6ar37a519df80e96f53@mail.gmail.com>
Message-ID: <uveokbejd.fsf@boost-consulting.com>

"Eduardo Picado" <eduardo.picado at gmail.com> writes:

> I'm facing a problem with the last module I need to compile with Boost Python.
> I receive this compilation error:
>
> Compiling...(with vc-8_0)
> TASRLogger.cpp
> C:\Python24\include\TASRLogger.hpp(353) : error C2248:
> 'TASRLogger::TASRLogger' : cannot access private member declared in
> class 'TASRLogger'
>         C:\Python24\include\TASRLogger.hpp(37) : see declaration of
> 'TASRLogger::TASRLogger'
>         C:\Python24\include\TASRLogger.hpp(33) : see declaration of 'TASRLogger'
>         This diagnostic occurred in the compiler generated function
> 'TASRDefaultLogger::TASRDefaultLogger(const TASRDefaultLogger &)'
>
> It's hard find a solution for the problem as I don't have the sources
> for the dll. Otherwise I could at least change the member attributes
> visibility.

Actually you mean access I think.  In C++ you don't get direct control
over visibility.

> Is there a workaround for it?

This is strictly illegal, of course, but I think you have only one
choice: just change the access from private to public in the header
file.  The chance of breaking something is practically nil.

Well... actually you have another choice.  You could add a friend
declaration to the header and use the friend to do the access.  But
there's not much advantage in it.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From michael_gruner at arcor.de  Wed Aug 23 14:05:36 2006
From: michael_gruner at arcor.de (michael_gruner at arcor.de)
Date: Wed, 23 Aug 2006 14:05:36 +0200 (CEST)
Subject: [C++-sig] debug a python extension
Message-ID: <16708287.1156334736221.JavaMail.ngmail@webmail15>

Hello,

I have a binary dll file (compiled for debugging) which I built a wrapper for with boost.python. Within this binary there exists a derived class which destructor seems to have a problem. If I instantiate the class and del it afterwards Python crashes.

Do you have any advices how I may debug this within msvc or any other windows debugger?

Thanks for your help.

Michael


From meine at kogs.informatik.uni-hamburg.de  Wed Aug 23 15:31:01 2006
From: meine at kogs.informatik.uni-hamburg.de (Hans Meine)
Date: Wed, 23 Aug 2006 15:31:01 +0200
Subject: [C++-sig] Writing configure checks for libboost?
Message-ID: <200608231531.04767.meine@kogs.informatik.uni-hamburg.de>

Hi!

Does anyone have a useful configure check for the boost::python dynamic 
library (under Unix)?  Since Boost encodes build information, not only 
threading/debug status, but also toolset and library version into the 
filename, it seems to be rather hard to find reliably.

It appears sensible to link the specific filename to a generic one like 
"libboost_python.so*", as we did locally and several Linux distributions do 
too, but they differ in the chosen filenames as a quick survey shows:

Debian testing (http://ftp.debian.org/debian/pool/main/b/boost/) installs 
into /usr/lib:
   *libboost_python-gcc-1_33_1.so.1.33.1
   *libboost_python-gcc-mt-1_33_1.so.1.33.1
and obviously links via the "shlibs" file:
  libboost_python-gcc-1_33_1 1.33.1 libboost-python1.33.1
  libboost_python-gcc-mt-1_33_1 1.33.1 libboost-python1.33.1

Also, Ubuntu Breezy 
(http://packages.ubuntulinux.org/cgi-bin/search_contents.pl?searchmode=filelist&word=libboost-python1.33.0&version=breezy&arch=i386) 
installs the following:
  *libboost_python-gcc-mt-1_33.so.1.33.0  
(There are no links listed, but that could be a shortcoming of the listing.)

SuSE 10.1 boost-1.33.1-17.pm.2.i586.rpm installs into /usr/lib:
  @libboost_python-gcc-1_33.so.1.33.0
  @libboost_python-gcc-1_33_1.so.1.33.1
  *libboost_python.so.1.33.1
(Note that the regular file has the short name, and the links contain the 
toolset name.)

Fedora Core 3 
(http://rpmfind.net/linux/RPM/fedora/updates/testing/3/x86_64/boost-1.33.0-3.fc3.i386.html) 
installs into /usr/lib only:
  *libboost_python.so.1.33.0

(Yes, I may have overlooked some tricks in the installation procedures, since 
I only looked at the package contents.)

-- 
Ciao, /  /
     /--/
    /  / ANS


From eduardo.picado at gmail.com  Thu Aug 24 18:09:00 2006
From: eduardo.picado at gmail.com (Eduardo Picado)
Date: Thu, 24 Aug 2006 18:09:00 +0200
Subject: [C++-sig] Boost Python private members problem
In-Reply-To: <uveokbejd.fsf@boost-consulting.com>
References: <6c75ff8f0608220645l43694e6ar37a519df80e96f53@mail.gmail.com>
	<uveokbejd.fsf@boost-consulting.com>
Message-ID: <6c75ff8f0608240909g1bedf922h2e59d0a64542601@mail.gmail.com>

On 8/23/06, David Abrahams <dave at boost-consulting.com> wrote:
> "Eduardo Picado" <eduardo.picado at gmail.com> writes:
>
> > I'm facing a problem with the last module I need to compile with Boost Python.
> > I receive this compilation error:
> >
> > Compiling...(with vc-8_0)
> > TASRLogger.cpp
> > C:\Python24\include\TASRLogger.hpp(353) : error C2248:
> > 'TASRLogger::TASRLogger' : cannot access private member declared in
> > class 'TASRLogger'
> >         C:\Python24\include\TASRLogger.hpp(37) : see declaration of
> > 'TASRLogger::TASRLogger'
> >         C:\Python24\include\TASRLogger.hpp(33) : see declaration of 'TASRLogger'
> >         This diagnostic occurred in the compiler generated function
> > 'TASRDefaultLogger::TASRDefaultLogger(const TASRDefaultLogger &)'
> >
> > It's hard find a solution for the problem as I don't have the sources
> > for the dll. Otherwise I could at least change the member attributes
> > visibility.
>
> Actually you mean access I think.  In C++ you don't get direct control
> over visibility.
>

Yes.

> > Is there a workaround for it?
>
> This is strictly illegal, of course, but I think you have only one
> choice: just change the access from private to public in the header
> file.  The chance of breaking something is practically nil.
>

I had already tried it, and it gave me other errors.
Now I've seem it's a good solution, and the other errors really look
like a DLL version problem.

Thanks very much.


> Well... actually you have another choice.  You could add a friend
> declaration to the header and use the friend to do the access.  But
> there's not much advantage in it.
>
> --
> Dave Abrahams
> Boost Consulting
> www.boost-consulting.com
>
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
>


-- 
Eduardo PICADO
MSN eduardo.picado at gmail.com
portable +33 6 32 47 10 10 (NEW)
fix +33 1 45 03 52 54
|??<?????>??|

From roman.yakovenko at gmail.com  Thu Aug 24 20:56:45 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 24 Aug 2006 21:56:45 +0300
Subject: [C++-sig] Py++ 0.8.1
Message-ID: <7465b6170608241156o5119227ds30b3b215c937e781@mail.gmail.com>

I'm glad to announce the new version of Py++.

Download page: http://language-binding.net/pyplusplus/download.html

What is it?
Py++ is an object-oriented framework for creating a code generator for
Boost.Python library.

Project home page: http://language-binding.net/pyplusplus/pyplusplus.html

Changes:
1. Support for Boost.Python indexing suite version 2 was implemented. Indexing
    suite 2, provides support for almost all STD containers.
2. Py++ generates "stable" code. If header files were not changed,
Py++ will not
    change any file.
3. Support for huge classes was added. Py++ is able to split
registration code for
    the class to multiple cpp files.
4. User code could be added almost anywhere, without use of low level API.
5. Py++ will provide user with feedback: warnings about exported declarations.
6. Py++ now supports generation of documentation strings.
7. A lot of bugs were fixed.
8. Documentation was written.

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


From iluetkeb at gmail.com  Thu Aug 24 21:46:02 2006
From: iluetkeb at gmail.com (=?ISO-8859-1?Q?Ingo_L=FCtkebohle?=)
Date: Thu, 24 Aug 2006 21:46:02 +0200
Subject: [C++-sig] Embedding Python into a Shared Object
In-Reply-To: <39ae57c30608210226x5216bcecn1f40624288ecde8e@mail.gmail.com>
References: <39ae57c30608210226x5216bcecn1f40624288ecde8e@mail.gmail.com>
Message-ID: <fdd727f80608241246o6cabd919k1d1719855b146db9@mail.gmail.com>

Hi bEngO,

the flags you mention only work when used on the binary thats
ultimately loading the shared libraries (i.e., your plugin).

One kludge is to explicitly load libpython.so yourself before using
any modules.  When doing so, you can pass the correct options
(RTLD_GLOBAL, at least on Linux) to dlopen, so that the symbols are
globally available.

If anyone found a better solution, I'd be interested, too!

Cheers, Ingo

-- 
Dipl.-Inform. Ingo L?tkebohle
Faculty of Technology
Bielefeld University
http://www.techfak.uni-bielefeld.de/~iluetkeb/

Fingerprint 3187 4DEC 47E6 1B1E 6F4F  57D4 CD90 C164 34AD CE5B


From reddy at pixar.com  Fri Aug 25 18:31:53 2006
From: reddy at pixar.com (Martin Reddy)
Date: Fri, 25 Aug 2006 09:31:53 -0700 (PDT)
Subject: [C++-sig] Problems with Boost Python scope and
	-fno-elide-constructors
Message-ID: <Pine.LNX.4.63.0608250909100.17775@dalwhinnie.dynamic.pixar.com>


Hi there,

We recently came across an interesting problem where boost scopes do
not appear to be closed correctly when the code is compiled with the 
-fno-elide-constructors gcc option. We have worked around this with
a small modification to the boost/python/scope.hpp.

For example, take the following simple example:

void
wrapScope1()
{
     scope errorScope =
     class_<Scope1, Scope1*, boost::noncopyable>
         ( "Scope1", "", no_init )
         .def("sayHello", &SayHello);
}

void
wrapScope2()
{
     class_<Scope2, Scope2*, boost::noncopyable>
         ( "Scope2", "", no_init )
         .def("sayHello", &SayHello);
}

BOOST_PYTHON_MODULE( MyModule )
{
     wrapScope1();
     wrapScope2();
}

When compiled normally you get Scope1 and Scope2 at the same level, as you 
would expect. However, if you compile this code with -fno-elide-constructors 
(which we do to workaround gcov bugs) then Scope2 is nested inside of Scope1.

The problem is that -fno-elide-constructors causes 2 scope objects to get 
created: an initial temporary scope object is created, and then the copy 
constructor is called to copy this into the errorScope variable. The 
temporary scope is then immediately destroyed (popping the scope to the 
previous value), and then at the end of wrapScope1() errorScope is destroyed, 
popping the scope to the value of the temporary scope. To work around this, I 
made a small fix to scope.hpp to replace the m_previous_scope pointer with a 
vector that forms a stack of scope pointers. There may be more elegant fixes 
to this problem that this group may want to consider, but I figured I'd pass 
along the one potential change that works well for us.

I have put a complete repro case, along with the code changes that I made to 
address this, and a diff -c patch file, into a tarball at:

   http://www.martinreddy.net/tmp/boost-scope.tgz

(This problem occurred for us under gcc 4.0.x, boost 1.33.1, under Fedora 
Core.)

Cheers,

Martin.


---------------------------------------------------------------------------
Martin Reddy                                        Pixar Animation Studios
                                                     1200 Park Avenue
reddy at pixar.com                                     Emeryville, CA 94608
http://MartinReddy.net/                             Tel. (510) 752 4093


From ndbecker2 at gmail.com  Fri Aug 25 21:27:59 2006
From: ndbecker2 at gmail.com (Neal Becker)
Date: Fri, 25 Aug 2006 15:27:59 -0400
Subject: [C++-sig] boost python doc problem? (return_self)
Message-ID: <ecnj0q$8e7$1@sea.gmane.org>

The example in ref man for return_self, shows:
   class_<widget>("Widget")
      .def("sensitive", &Widget::get_sensitive)
      .def("sensitive", &Widget::set_sensitive, return_self<>())
      ;

But Widget::set_sensitive is void, it doesn't return anything.  What am I
missing here?



From quncao at alumni.usc.edu  Sat Aug 26 00:26:22 2006
From: quncao at alumni.usc.edu (Qun Cao)
Date: Fri, 25 Aug 2006 15:26:22 -0700
Subject: [C++-sig] how to convert C++ objects into python objects?
Message-ID: <1f527df80608251526i215c62edt98f3b641a44d58cc@mail.gmail.com>

Hi Everyone,
I have a question regarding type conversion when embedding python in C++.

First, I have created a python module of an C++ class using boost,

#include <classA.cpp>
BOOST_PYTHON_MODULE(PythonTest)
{
	class_<classA>("classA", init<MyString, MyString, MyPoint3>());
}

MyString, MyPoint3 are class types defined in the C++ libraries.
It compiles and generate PythonTest.dll.

Now, I want to call some Python code (object =classA(MyString,
MyString, MyPoint3)  from my main.cpp to create a classA object, how
do I convert
MyString, MyPoint3 variables into Python types?  Also since the python
module does not know anything about C++ objects MyString, MyPoint3; is
it
necessary to extending those classes into Python?

I saw some ways of convert native types such as int, string in
C/Python API, but nothing about generic type conversions.

Thanks for your help!
Qun


From quncao at alumni.usc.edu  Sat Aug 26 02:42:18 2006
From: quncao at alumni.usc.edu (Qun Cao)
Date: Fri, 25 Aug 2006 17:42:18 -0700
Subject: [C++-sig] A question about PyImport_AppendInittab
Message-ID: <1f527df80608251742x77912dc0gc4b47ac8ff602a64@mail.gmail.com>

Hi Everyone,

I am adopting the code in boost/libs/python/test/embedding.cpp for my
project.  Here is my question, in
embedding.cpp we have:

 if (PyImport_AppendInittab("embedded_hello", initembedded_hello) == -1)
        throw std::runtime_error("Failed to add embedded_hello to the
interpreter's "
                                 "builtin modules");

I noticed #define BOOST_PYTHON_MODULE BOOST_PYTHON_MODULE_INIT in
module.hpp file, so I guess "initembedded_hello" here is a pointer to
the module "embedd_hello".  However, in my project,
BOOST_PYTHON_MODULE{} definition code sits
in a different class class_wrapper.cpp than main.cpp, does that mean i
need to include class_wrapper.cpp in main.cpp?  That gives me wierd
problem in compilation though.

Also, generally, why main.cpp needs to know how a C++ wrapper module
is created?  Isn't it enough to tell it that there exists a python
module in class.dll and the embedded python interpreter will just
understand everything from there?

Thanks for any insights,
Qun


From seefeld at sympatico.ca  Sat Aug 26 03:04:45 2006
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Fri, 25 Aug 2006 21:04:45 -0400
Subject: [C++-sig] A question about PyImport_AppendInittab
In-Reply-To: <1f527df80608251742x77912dc0gc4b47ac8ff602a64@mail.gmail.com>
References: <1f527df80608251742x77912dc0gc4b47ac8ff602a64@mail.gmail.com>
Message-ID: <44EF9E2D.9020604@sympatico.ca>

Qun Cao wrote:
> Hi Everyone,
> 
> I am adopting the code in boost/libs/python/test/embedding.cpp for my
> project.  Here is my question, in
> embedding.cpp we have:
> 
>  if (PyImport_AppendInittab("embedded_hello", initembedded_hello) == -1)
>         throw std::runtime_error("Failed to add embedded_hello to the
> interpreter's "
>                                  "builtin modules");
> 
> I noticed #define BOOST_PYTHON_MODULE BOOST_PYTHON_MODULE_INIT in
> module.hpp file, so I guess "initembedded_hello" here is a pointer to
> the module "embedd_hello".

Correct. More specifically, 'initembedded_hello' is the symbol the python
interpreter will look for when it tries to import the 'embedded_hello'
module.

> However, in my project,
> BOOST_PYTHON_MODULE{} definition code sits
> in a different class class_wrapper.cpp than main.cpp, does that mean i
> need to include class_wrapper.cpp in main.cpp?  That gives me wierd
> problem in compilation though.

main.cpp typically contains the entry point for your application (i.e.
the int main(int, char **) function). It should be linked with the
python interpreter (i.e. libpython.lib or equivalent).
The 'embedded_hello' module will physically reside in a different
executable. However, for the module lookup mechanism to work you
have to follow some naming conventions. A module 'embedded_hello'
for example will be identified by the python interpreter as a shared
library named 'embedded_hello.pyd' or 'embedded_hello.dll', and it
should contain an 'initembedded_hello' function. (The latter
is best generated with the BOOST_PYTHON_MODULE macro.)

> Also, generally, why main.cpp needs to know how a C++ wrapper module
> is created?

It doesn't.

> Isn't it enough to tell it that there exists a python
> module in class.dll and the embedded python interpreter will just
> understand everything from there?

Yep. The embedding.cpp code you are looking at is a unit test that
lumps multiple things into a single file to make testing simple.
Real python extension modules won't be linked with the main
application, as that would be pretty pointless.

HTH,
		Stefan

-- 

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


From iluetkeb at gmail.com  Sat Aug 26 17:49:57 2006
From: iluetkeb at gmail.com (=?ISO-8859-1?Q?Ingo_L=FCtkebohle?=)
Date: Sat, 26 Aug 2006 17:49:57 +0200
Subject: [C++-sig] Writing configure checks for libboost?
In-Reply-To: <200608231531.04767.meine@kogs.informatik.uni-hamburg.de>
References: <200608231531.04767.meine@kogs.informatik.uni-hamburg.de>
Message-ID: <fdd727f80608260849h2ef0af64ma4cef8e21f6ee109@mail.gmail.com>

Hi Hans,

for the toolchain part, we hardcode a reasonably sane default
("gcc-mt-d") and then use a 'with'-option for configure to allow the
user to override our choice. As some things (e.g., whether to use mt
or not) cannot be automatically determined anyway, that seems like a
good compromise. Other than that, we look at the header directory to
figure out the latest boost-version installed.  I can pass on the
autoconf parts for these two things to you, if you're interested.

This procedure has worked fine for us on various Linux distributions.
For other unices, figuring the toolchain to use from the name of the
library should not be too hard but we haven't done that, yet.

best regards,

-- 
Dipl.-Inform. Ingo L?tkebohle
Faculty of Technology
Bielefeld University
http://www.techfak.uni-bielefeld.de/~iluetkeb/

Fingerprint 3187 4DEC 47E6 1B1E 6F4F  57D4 CD90 C164 34AD CE5B


From meine at kogs.informatik.uni-hamburg.de  Mon Aug 28 10:45:42 2006
From: meine at kogs.informatik.uni-hamburg.de (Hans Meine)
Date: Mon, 28 Aug 2006 10:45:42 +0200
Subject: [C++-sig] Writing configure checks for libboost?
In-Reply-To: <fdd727f80608260849h2ef0af64ma4cef8e21f6ee109@mail.gmail.com>
References: <200608231531.04767.meine@kogs.informatik.uni-hamburg.de>
	<fdd727f80608260849h2ef0af64ma4cef8e21f6ee109@mail.gmail.com>
Message-ID: <200608281045.43209.meine@kogs.informatik.uni-hamburg.de>

On Saturday, 26. August 2006 17:49, Ingo L?tkebohle wrote:
> for the toolchain part, we hardcode a reasonably sane default
> ("gcc-mt-d") and then use a 'with'-option for configure to allow the
> user to override our choice.

OK, that seems to be a pragmatical approach.  (Which most configure checks 
take anyways.)  I wonder about the -d though.  (I supposed that meant 
"debug", is that wrong?)

> As some things (e.g., whether to use mt 
> or not) cannot be automatically determined anyway, that seems like a
> good compromise. Other than that, we look at the header directory to
> figure out the latest boost-version installed.  I can pass on the
> autoconf parts for these two things to you, if you're interested.

Yes, I would be very much interested.  Maybe we can keep that on-list?

I would still like to hear other's opinions and approaches.

-- 
Ciao, /  /
     /--/
    /  / ANS


From grafikrobot at gmail.com  Mon Aug 28 18:19:32 2006
From: grafikrobot at gmail.com (Rene Rivera)
Date: Mon, 28 Aug 2006 11:19:32 -0500
Subject: [C++-sig] Writing configure checks for libboost?
In-Reply-To: <200608281045.43209.meine@kogs.informatik.uni-hamburg.de>
References: <200608231531.04767.meine@kogs.informatik.uni-hamburg.de>	<fdd727f80608260849h2ef0af64ma4cef8e21f6ee109@mail.gmail.com>
	<200608281045.43209.meine@kogs.informatik.uni-hamburg.de>
Message-ID: <44F31794.8040701@gmail.com>

Hans Meine wrote:
> I wonder about the -d though.  (I supposed that meant 
> "debug", is that wrong?)

No that's not wrong. All the variant tags are documented 
<http://boost.org/more/getting_started.html#Results>.



-- 
-- Grafik - Don't Assume Anything
-- Redshift Software, Inc. - http://redshift-software.com
-- rrivera/acm.org - grafik/redshift-software.com
-- 102708583/icq - grafikrobot/aim - grafikrobot/yahoo


From iluetkeb at gmail.com  Mon Aug 28 18:41:18 2006
From: iluetkeb at gmail.com (=?ISO-8859-1?Q?Ingo_L=FCtkebohle?=)
Date: Mon, 28 Aug 2006 18:41:18 +0200
Subject: [C++-sig] Writing configure checks for libboost?
In-Reply-To: <200608281045.43209.meine@kogs.informatik.uni-hamburg.de>
References: <200608231531.04767.meine@kogs.informatik.uni-hamburg.de>
	<fdd727f80608260849h2ef0af64ma4cef8e21f6ee109@mail.gmail.com>
	<200608281045.43209.meine@kogs.informatik.uni-hamburg.de>
Message-ID: <fdd727f80608280941s7e0046fbxcd56d2dc2c19f954@mail.gmail.com>

2006/8/28, Hans Meine <meine at kogs1.informatik.uni-hamburg.de>:
> OK, that seems to be a pragmatical approach.  (Which most configure checks
> take anyways.)  I wonder about the -d though.  (I supposed that meant
> "debug", is that wrong?)

Yup.

> Yes, I would be very much interested.  Maybe we can keep that on-list?

Sure, here it goes.  btw, I'd also be interested in better ways of doing this!

# specify where we're looking for boost
AC_ARG_WITH([boost],
	    AC_HELP_STRING([--with-boost],
		           [prefix where boost is installed]),
	    BOOST_PREFIX="$withval",
	    BOOST_PREFIX="/usr/local /usr")

# find newest installed boost-version
AC_MSG_CHECKING([for most recent boost headers])
BOOST_HEADER_PREFIX=`for dir in $BOOST_PREFIX; do echo "$dir/include"; done`
BOOST_HEADER_PATH=`find $BOOST_HEADER_PREFIX -name 'boost-*' -type d
-maxdepth 1 2>/dev/null | sort -n | head -1`
BOOST_VERSION=`echo $BOOST_HEADER_PATH | sed 's/.*boost-//'`
if test ${BOOST_VERSION}x = 'x'
then
	AC_MSG_ERROR([
	*** Couldn't find any 'boost-*' directory in
	***      $BOOST_HEADER_PREFIX
	*** Try the --with-boost option to specify the boost prefix.
	***   E.g., if your boost-installation is in /opt/boost/,
	***   with headers in /opt/boost/include/boost-1_33_1,
	***   specify '--with-boost=/opt/boost'!
	])
else
	AC_MSG_RESULT($BOOST_VERSION)
fi
CPPFLAGS=-I$BOOST_HEADER_PATH $CPPFLAGS
AC_SUBST(BOOST_PREFIX)
AC_SUBST(BOOST_VERSION)

# require Python 2.4 or higher for threading API
AM_PATH_PYTHON(2.4,,:)
AC_MSG_CHECKING([for Python distutils])
if $PYTHON -c 'from distutils import sysconfig' 2>/dev/null
then
        AC_MSG_RESULT([yes])
else
        AC_MSG_ERROR([
   *** Python distutils module not found.
   *** Did you install python-dev?])
fi


# allow user to specify which boost-python-version to link with
AC_ARG_WITH(boost_python-type,
AC_HELP_STRING([--with-boost_python-type=gcc-mt-d],[which subtype of
the boost::python library to link with; use "none" to disable]),
bp_type="$withval", bp_type="gcc-mt-d")

AC_MSG_CHECKING([boost::python type])
if test "$bp_type" != "none"; then
  AC_MSG_RESULT($bp_type)

PYTHON_PREFIX=`dirname $PYTHON`/..
PYTHON_INCLUDE=-I`$PYTHON -c 'from distutils import sysconfig; print
sysconfig.get_python_inc()'`
PYTHON_LIBDIR=`$PYTHON -c 'from distutils import sysconfig; print
sysconfig.get_config_var("LIBPL")'`
#PYTHON_LDFLAGS="-L$PYTHON_LIBDIR `$PYTHON -c 'from distutils import
sysconfig; print sysconfig.get_config_var("LINKFORSHARED")'` "
PYTHON_LIBS="-lpython$PYTHON_VERSION -ldl -lutil -lpthread"

# note: boost-headers are already included in main CPPFLAGS
BOOST_PYTHON_CPPFLAGS="$PYTHON_INCLUDE"
BOOST_PYTHON_LDFLAGS=" $PYTHON_LDFLAGS -L$BOOST_PREFIX/lib
-Wl,-rpath,$BOOST_PREFIX/lib"
BOOST_PYTHON_LIBS="-lboost_python-$bp_type $PYTHON_LIBS"

# temporarily augment flags so that CHECK_HEADER finds python headers
ac_save_CPPFLAGS="$CPPFLAGS"
CPPFLAGS="$BOOST_PYTHON_CPPFLAGS $CPPFLAGS"
ac_save_LDFLAGS="$LDFLAGS"
LDFLAGS="$BOOST_PYTHON_LDFLAGS $LDFLAGS"
ac_save_LIBS="$LIBS"
LIBS="$BOOST_PYTHON_LIBS $LIBS"
AC_CHECK_HEADER([Python.h],,
  AC_MSG_ERROR([
   *** Python development support missing Python.h.
   *** Did you install python-dev?]))

AC_CHECK_HEADER([boost/python/module.hpp],HAVE_BOOST_PYTHON=1,HAVE_BOOST_PYTHON=0)

AC_MSG_CHECKING([for boost::python])
AC_TRY_LINK([#include <boost/python/class.hpp>
   #include <string>],
  [boost::python::class_<std::string>("string");],
  AC_MSG_RESULT(yes)
  HAVE_PYTHON=1,
  AC_MSG_WARN([
        *** boost::python not found.  At least version 1.31 is required.
        *** You can download it from http://www.boost.org/]))
  CPPFLAGS="$ac_save_CPPFLAGS"
  LDFLAGS="$ac_save_LDFLAGS"
  LIBS="$ac_save_LIBS"
else
        AC_MSG_RESULT(none)
fi

-- 
Dipl.-Inform. Ingo L?tkebohle
Faculty of Technology
Bielefeld University
http://www.techfak.uni-bielefeld.de/~iluetkeb/

Fingerprint 3187 4DEC 47E6 1B1E 6F4F  57D4 CD90 C164 34AD CE5B


From iluetkeb at gmail.com  Mon Aug 28 18:43:24 2006
From: iluetkeb at gmail.com (=?ISO-8859-1?Q?Ingo_L=FCtkebohle?=)
Date: Mon, 28 Aug 2006 18:43:24 +0200
Subject: [C++-sig] Writing configure checks for libboost?
In-Reply-To: <fdd727f80608280941s7e0046fbxcd56d2dc2c19f954@mail.gmail.com>
References: <200608231531.04767.meine@kogs.informatik.uni-hamburg.de>
	<fdd727f80608260849h2ef0af64ma4cef8e21f6ee109@mail.gmail.com>
	<200608281045.43209.meine@kogs.informatik.uni-hamburg.de>
	<fdd727f80608280941s7e0046fbxcd56d2dc2c19f954@mail.gmail.com>
Message-ID: <fdd727f80608280943h2ca4bbffw46978a59470bf9ed@mail.gmail.com>

2006/8/28, Ingo L?tkebohle <iluetkeb at gmail.com>:
> else
>         AC_MSG_RESULT(none)
> fi

Sorry, remove this part, it was a left-over from a surrounding check
whether to use Python support at all.

-- 
Dipl.-Inform. Ingo L?tkebohle
Faculty of Technology
Bielefeld University
http://www.techfak.uni-bielefeld.de/~iluetkeb/

Fingerprint 3187 4DEC 47E6 1B1E 6F4F  57D4 CD90 C164 34AD CE5B


From grafikrobot at gmail.com  Mon Aug 28 18:45:28 2006
From: grafikrobot at gmail.com (Rene Rivera)
Date: Mon, 28 Aug 2006 11:45:28 -0500
Subject: [C++-sig] Writing configure checks for libboost?
In-Reply-To: <44F31794.8040701@gmail.com>
References: <200608231531.04767.meine@kogs.informatik.uni-hamburg.de>	<fdd727f80608260849h2ef0af64ma4cef8e21f6ee109@mail.gmail.com>	<200608281045.43209.meine@kogs.informatik.uni-hamburg.de>
	<44F31794.8040701@gmail.com>
Message-ID: <44F31DA8.2010101@gmail.com>

Rene Rivera wrote:
> Hans Meine wrote:
>> I wonder about the -d though.  (I supposed that meant 
>> "debug", is that wrong?)
> 
> No that's not wrong. All the variant tags are documented 
> <http://boost.org/more/getting_started.html#Results>.

Well I guess my answer might be misinterpreted. I meant "correct -d is 
debug" :-)


-- 
-- Grafik - Don't Assume Anything
-- Redshift Software, Inc. - http://redshift-software.com
-- rrivera/acm.org - grafik/redshift-software.com
-- 102708583/icq - grafikrobot/aim - grafikrobot/yahoo


From jorge.vargas at gmail.com  Mon Aug 28 21:58:42 2006
From: jorge.vargas at gmail.com (Jorge Vargas)
Date: Mon, 28 Aug 2006 15:58:42 -0400
Subject: [C++-sig] boost python mingw binaries
Message-ID: <32822fe60608281258g669fb64fj23bcd6f46133b89c@mail.gmail.com>

hi does anyone plans on distributing binaries for boost_python under windows?

I have try to compile it myself many times and fail.

Boost and Boost Jam is something new for me, the docs for boost_python
are inclined to someone that knows boost and the ones of boost itself
have no leads to what to do if something goes wrong.

I'm not a windows dev I just want to give an alternative here so all I
need is the binary since my only goal here is to make Democracy player
compile with free tools on windows.

so does anyone has boost python as a binary that works under mingw?


From achim-bpl at mol-net.com  Tue Aug 29 10:18:53 2006
From: achim-bpl at mol-net.com (Achim H.)
Date: Tue, 29 Aug 2006 10:18:53 +0200
Subject: [C++-sig] boost python mingw binaries
In-Reply-To: <32822fe60608281258g669fb64fj23bcd6f46133b89c@mail.gmail.com>
References: <32822fe60608281258g669fb64fj23bcd6f46133b89c@mail.gmail.com>
Message-ID: <200608291018.53778.achim-bpl@mol-net.com>

Am Montag, 28. August 2006 21:58 schrieb Jorge Vargas:
> so does anyone has boost python as a binary that works under mingw?

Yes, I have binaries of 1.33.1 and I can even upload them if you tell me where 
to.

Achim.


From jorge.vargas at gmail.com  Tue Aug 29 15:22:59 2006
From: jorge.vargas at gmail.com (Jorge Vargas)
Date: Tue, 29 Aug 2006 09:22:59 -0400
Subject: [C++-sig] boost python mingw binaries
In-Reply-To: <200608291018.53778.achim-bpl@mol-net.com>
References: <32822fe60608281258g669fb64fj23bcd6f46133b89c@mail.gmail.com>
	<200608291018.53778.achim-bpl@mol-net.com>
Message-ID: <32822fe60608290622l7c0382ffs87219f4532196fdf@mail.gmail.com>

On 8/29/06, Achim H. <achim-bpl at mol-net.com> wrote:
> Am Montag, 28. August 2006 21:58 schrieb Jorge Vargas:
> > so does anyone has boost python as a binary that works under mingw?
Great thanks
> Yes, I have binaries of 1.33.1 and I can even upload them if you tell me where
> to.
I guess the best place for that is the SF repository, since is the
official site. So I guess we could wait for someone with access there
to reply.

if for any reason that can't be done I could set the download file on
my server which currently is idle.

or I could talk to the democracy team about hosting the file in their
server, since they will eventually have to when my "port" is complete.

>
> Achim.
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
>


From roman.yakovenko at gmail.com  Tue Aug 29 19:48:57 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Tue, 29 Aug 2006 20:48:57 +0300
Subject: [C++-sig] custom smart pointers
Message-ID: <7465b6170608291048y6e2010d4ofa84fbd5760da4d3@mail.gmail.com>

Hi. I am working on some project that defines smart pointer and uses
it allover the place.
I have a problem for some reason I am not able to call functions that takes the
smart pointer instantiation as argument. The error I get is ArgumentError:

ArgumentError: Python argument types in
    custom_smart_ptr_classes.ref_get_value(add_x_t)
did not match C++ signature:
    ref_get_value(my_smart_ptr_t<controllers::controller_i> {lvalue} a)

Code high level:

template<T>
class my_smart_ptr_t{...}

struct controller_i{
  virtual int get_value() const = 0;
};

struct add_x_t : controller_i{
  int m_value;
  virtual int   get_value( ){ return m_value + get_add_value(); }
  virtual int get_add_value(){ return 0; }
};

int get_value( my_smar_ptr_t<controller_i> c ){ return c->get_value(); }

In order to export this code, I wrote next code:

namespace boost{ namespace python{

  controller_i* get_pointer( my_smart_ptr_t< controller_i > const& p )
  { return p.get(); }

  template <>
  struct pointee< my_smart_ptr_t< controller_i > >
  { typedef controller_i type; };

  //same code for add_x_t class
  ...

}}

Because both classes contains virtual function I create class-wrapper
for every class.
Classes and function registration is pretty simple. After this I add
next registration code:

register_ptr_to_python< my_smart_ptr_t< controller_i > >();
register_ptr_to_python< my_smart_ptr_t< add_x_t > >();
implicitly_convertible< my_smart_ptr_t< add_x_t >, my_smart_ptr_t<
controller_i > >();

Python code:

import my_module
add_x = my_module.add_x_t()
my_module.get_value( add_x ) # raises the error I described earlier.

I think, I exposed my classes "by the book". May be I am missing something?

Thanks in advance.

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


From quncao at gmail.com  Wed Aug 30 10:09:43 2006
From: quncao at gmail.com (Qun Cao)
Date: Wed, 30 Aug 2006 01:09:43 -0700
Subject: [C++-sig] RuntimeError: unidentifiable C++ exception
Message-ID: <1f527df80608300109q5f9df417sf8d1a23023a0a3b7@mail.gmail.com>

Hi Everyone,

I created a simple class wrapper like this

BOOST_PYTHON_MODULE(PythonTest) {
	class_<MyClass,boost::noncopyable>("pMyClass", init<char const*>());
}

When I try to initialize pMyClass in python, I got this error:

Traceback (most recent call last):
  File "<stdin>", line 1, in ?
RuntimeError: unidentifiable C++ exception

MyClass is part of a third party GameEngine, I can successfully use it in C++.
Can anyone suggest what might went wrong here?

Thanks a lot,
Qun


From roman.yakovenko at gmail.com  Wed Aug 30 13:10:36 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Wed, 30 Aug 2006 14:10:36 +0300
Subject: [C++-sig] RuntimeError: unidentifiable C++ exception
In-Reply-To: <1f527df80608300109q5f9df417sf8d1a23023a0a3b7@mail.gmail.com>
References: <1f527df80608300109q5f9df417sf8d1a23023a0a3b7@mail.gmail.com>
Message-ID: <7465b6170608300410k76c0aae4nee42f431af131c9d@mail.gmail.com>

On 8/30/06, Qun Cao <quncao at gmail.com> wrote:
> Hi Everyone,
>
> I created a simple class wrapper like this
>
> BOOST_PYTHON_MODULE(PythonTest) {
>         class_<MyClass,boost::noncopyable>("pMyClass", init<char const*>());
> }
>
> When I try to initialize pMyClass in python, I got this error:
>
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
> RuntimeError: unidentifiable C++ exception

The code generated by Py++ just works, may be you missed something?
You can try to use make_consructor function:

http://boost.org/libs/python/doc/v2/make_function.html#make_constructor-spec

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


From roman.yakovenko at gmail.com  Wed Aug 30 13:29:30 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Wed, 30 Aug 2006 14:29:30 +0300
Subject: [C++-sig] [tutorial] Auto-Overloading example bug?
Message-ID: <7465b6170608300429n564516bfi22c665cd6cbfca32@mail.gmail.com>

Hi. I could be wrong, but it seems to me that the Boost.Python tutorials
contains bug.
(
http://boost.org/libs/python/doc/tutorial/doc/html/python/functions.html#python.auto_overloading)

The compiler ( I use msvc 7.1 ) is not able to compile the code. The main
reason is that it can not decide
what "f" to select. I mean, it does not know what address it should pass to
"def" function. The same thing
occurs for BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS. It seems to me that it is
right.

Some thoughts. I think, that in this case( auto-overloading example )
function "def"
should not take reference to function. It should be enough to pass an
instance of
the constructed class. Same logic could be applied on class_::def method
too.

-- 
Roman Yakovenko
C++ Python language binding
http://www.language-binding.net/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060830/6de88e5f/attachment.htm>

From abierbaum at gmail.com  Wed Aug 30 16:07:40 2006
From: abierbaum at gmail.com (Allen Bierbaum)
Date: Wed, 30 Aug 2006 09:07:40 -0500
Subject: [C++-sig] Is there a way to automatically convert a smart_ptr to a
	held_type (try 2)
Message-ID: <bea9869d0608300707r3009cdb2u101ab96b88ec9966@mail.gmail.com>

Note: I posted this message a very long time ago on this list and
still haven't found a solution.  Several months ago David suggested
that I repost the message again to see if anyone can help out.  I am
finally getting back to this project so here is the repost.

I am wrapping a C++ library with boost.python that has multiple types
of smart ptrs for it's objects.  It has a ptr<> type that holds the
object and provides a ref counting interface (through and
addRef()/subRef()  interface).  It then has a ref_ptr<> type that acts
like a "standard"  smart pointer by holding a ptr<> internally and
calling the ref handling  operations upon construction and
destruction.

To correctly handle the memory for the objects, I want my boost.python
 class wrappers for a class T to be held using ref_ptr<T>.  I can do
this  with:

 class_<T, bases<...>, ref_ptr<T> >("T",...);

This all works fine.

I run into a problem though when I want to use wrapped method that
return ptr<>'s to objects instead of a full ref_ptr<>.

For example if I wrap a method like this:

ptr<Base> createObject()
{ ... }

then when I call this from boost.python it needs to figure out that it
should convert the ptr<Base> to a ref_ptr<Base>.  I know that I can
use register_ptr_to_python<ptr<Base> > to register a to_python
converter for the type, but as I understand it this will cause the
python wrapper to store the C++ object using a ptr<> type internally.
This won't work correctly because it will not do any memory reference
counting.

What I need is a way to tell boost python that anytime it needs to
convert a ptr<Base> to python it should first convert it to a
ref_ptr<Base>  (using a standard copy constructor in ref_ptr<>) and
then use this ref_ptr as the held type like normal.

At first I thought that implicitly_convertible<> would help:

implicitly_convertible<ptr<Base>, ref_ptr<Base> >();

But this doesn't seem to do anything for me. (this really isn't too
surprising since this isn't what implicitly_convertible is for, but it
seemed worth a try).

Has anyone else ran into anything similar?  Any ideas?

Thanks,
Allen


From amohr at pixar.com  Wed Aug 30 19:38:24 2006
From: amohr at pixar.com (Alex Mohr)
Date: Wed, 30 Aug 2006 10:38:24 -0700
Subject: [C++-sig] Is there a way to automatically convert a smart_ptr
 to a	held_type (try 2)
In-Reply-To: <bea9869d0608300707r3009cdb2u101ab96b88ec9966@mail.gmail.com>
References: <bea9869d0608300707r3009cdb2u101ab96b88ec9966@mail.gmail.com>
Message-ID: <44F5CD10.5010103@pixar.com>

> What I need is a way to tell boost python that anytime it needs to
> convert a ptr<Base> to python it should first convert it to a
> ref_ptr<Base>  (using a standard copy constructor in ref_ptr<>) and
> then use this ref_ptr as the held type like normal.
> 
> At first I thought that implicitly_convertible<> would help:
> 
> implicitly_convertible<ptr<Base>, ref_ptr<Base> >();
> 
> But this doesn't seem to do anything for me. (this really isn't too
> surprising since this isn't what implicitly_convertible is for, but it
> seemed worth a try).
> 
> Has anyone else ran into anything similar?  Any ideas?

At first glance, it seems like you want to register a to-python 
conversion for all your ptr<T> types.  Note that all code I write in 
this message is untested:

template <class T>
struct ptr_to_python {
     static PyObject *convert(ptr<T> const &p) {
         return Py_INCREF(object(ref_ptr<T>(p)).ptr());
     }
};

Then you need to do this for each class you register:

to_python_converter<ptr<T>, ptr_to_python<T> >();

You could automate this in a custom def visitor:

struct ref_ptr_stuff : def_visitor<ref_ptr_stuff> {
     friend class def_visitor_access;
     template <typename Class>
     void visit(Class &c) const {
         typedef typename Class::wrapped_type T;
         to_python_converter<ptr<T>, ptr_to_python<T> >();
     }
};

Then when you wrap your classes you can say:

class<T, bases<...>, ref_ptr<T> >("T",...)
     .def(ref_ptr_stuff())

Note that this gives you a convenient way to put common methods on all 
ref_ptr types.  For example, you could .def() a method that returns the 
ref count in the ref_ptr_stuff visitor.

Also note that when you get to doing polymorphic wrappers, you're going 
to want to be able to distinguish the polymorphic wrapper type from the 
"logical" wrapped type T.  To do that, you'll need to either roll your 
own or be willing to use non-public boost.python api. 
(detail::unwrap_wrapper).

Hope that helps,

Alex


From wz_ at gmx.net  Wed Aug 30 20:20:59 2006
From: wz_ at gmx.net (Nicolas Tessore)
Date: Wed, 30 Aug 2006 20:20:59 +0200
Subject: [C++-sig] Py++ issues, plus switching from Swig
Message-ID: <20060830182059.289130@gmx.net>

Hello, I hope someone can help me out:

I am relatively new to Python, and not too experienced in C++.
My situation is the following: I am trying to add Boost.Python
bindings to a library consisting of about a hundred includes,
weighting 60Megs. It is heavily relying on inheritance, in fact
all classes derive from one base.

Now the tricky bit: The lib uses templates a _lot_. The small
part that is currently included in a Swig wrapper already
contains 800 specific template conversions.

What I now want to do is switch to Boost.Python for a change.
Swig creates a huge number of problems when using the lib, and
I trust Boost to perform better. Still there is one issue: Due
to the number of classes, templates and the frequent changes of
the codebase, I cannot generate the Boost.Python wrapper manually.
I tried using Pyste, but it is still too much work to add all
classes in an interface file. The only possibility is automation, as
presented in Py++.

The reason for posting is that I cannot find the documentation
explaining how to do templates in Py++. I can (and do) use typedefs +
declaration for templated types, but I need support for template
functions as well, since I need to access a lot of Dynamic and Static
casts, in fact, each single class needs it.

Could anyone give me a hint in the right direction, or at least tell me
that it is impossible in general, so I can hack the Swig interface to
at least stop segfault'ing?

Thanks in advance,
    N.
-- 


"Feel free" ? 10 GB Mailbox, 100 FreeSMS/Monat ...
Jetzt GMX TopMail testen: http://www.gmx.net/de/go/topmail


From roman.yakovenko at gmail.com  Wed Aug 30 20:42:53 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Wed, 30 Aug 2006 21:42:53 +0300
Subject: [C++-sig] Py++ issues, plus switching from Swig
In-Reply-To: <20060830182059.289130@gmx.net>
References: <20060830182059.289130@gmx.net>
Message-ID: <7465b6170608301142y56d154b7s5734bcb539772bf3@mail.gmail.com>

On 8/30/06, Nicolas Tessore <wz_ at gmx.net> wrote:
> Hello, I hope someone can help me out:
>
> I am relatively new to Python, and not too experienced in C++.
> My situation is the following: I am trying to add Boost.Python
> bindings to a library consisting of about a hundred includes,
> weighting 60Megs. It is heavily relying on inheritance, in fact
> all classes derive from one base.
>
> Now the tricky bit: The lib uses templates a _lot_. The small
> part that is currently included in a Swig wrapper already
> contains 800 specific template conversions.

Py++ was born to solve such problems :-)

> Could anyone give me a hint in the right direction

I will try.

1. Documentation:
    http://language-binding.net/pyplusplus/documentation/how_to.html#how-to-automatically-export-template-functions-class

2. Examples:
    Using Py++ I exposed next libraries: date_time, boost.random, crc
and rational
    including unit tests.  Basically all these libraries are "one big
template".
    I used different approaches to export them.
    Here you will find more information and examples:
    http://language-binding.net/pyplusplus/examples/boost/boost.html

You can find scripts that exports these libraries here:
http://svn.sourceforge.net/viewvc/pygccxml/pyplusplus_dev/examples/pyboost_dev/dev/

Did these help? If not, you can take a look on BGL-Python package(
Python bindings for Boost Graph Library using Boost.Python ). BGL-Python
was not created with Py++, but still it could be a good example, how to expose
highly template library using Boost.Python.

P.S. Small tip: expose your code using 2 phase approach:
1. Generate C++ code that instantiate your templates.
2. Feed generated code to Py++.

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


From quncao at alumni.usc.edu  Wed Aug 30 21:38:17 2006
From: quncao at alumni.usc.edu (Qun Cao)
Date: Wed, 30 Aug 2006 12:38:17 -0700
Subject: [C++-sig] RuntimeError: unidentifiable C++ exception
Message-ID: <1f527df80608301238t1e7506c0qce27a2d420c1ca31@mail.gmail.com>

>* Hi Everyone,
*>*
*>* I created a simple class wrapper like this
*>*
*>* BOOST_PYTHON_MODULE(PythonTest) {
*>*         class_<MyClass,boost::noncopyable>("pMyClass", init<char const*>());
*>* }
*>*
*>* When I try to initialize pMyClass in python, I got this error:
*>*
*>* Traceback (most recent call last):
*>*   File "<stdin>", line 1, in ?
*>* RuntimeError: unidentifiable C++ exception
*
The code generated by Py++ just works, may be you missed something?
You can try to use make_consructor function:

http://boost.org/libs/python/doc/v2/make_function.html#make_constructor-spec

Hi Roman,
What do you mean by "the code generated by Py++ just works"?  Do you mean
you created a class
MyClass {
public:
   MyClass(char const*) ;
}
and you had successfully initialized it in Python?  I tried a similar
example class and had no problem with it.
The real problem  I had was with this specific class NiString in Gamebryo
game engine, is it possible that something
going on inside Gamebryo might cause the problem?  Or the game engine is
irrelevant to this problem?

And still, what the "unidentifiable C++ exception" is really saying?
Boost-Python called the C++ constructor code and
an exception was thrown?

I am taking a look at Py++ and make_constructor, Thanks so much for your
help!

Qun
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060830/a33e2e69/attachment.htm>

From roman.yakovenko at gmail.com  Wed Aug 30 21:44:31 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Wed, 30 Aug 2006 22:44:31 +0300
Subject: [C++-sig] RuntimeError: unidentifiable C++ exception
In-Reply-To: <1f527df80608301238t1e7506c0qce27a2d420c1ca31@mail.gmail.com>
References: <1f527df80608301238t1e7506c0qce27a2d420c1ca31@mail.gmail.com>
Message-ID: <7465b6170608301244l71d7cca9mf59a4cea8ca9b917@mail.gmail.com>

On 8/30/06, Qun Cao <quncao at alumni.usc.edu> wrote:
> Hi Roman,
> What do you mean by "the code generated by Py++ just works"?  Do you mean
> you created a class
> MyClass {
> public:
>    MyClass(char const*) ;
> }
> and you had successfully initialized it in Python?  I tried a similar
> example class and had no problem with it.

That what you described :-).

> The real problem  I had was with this specific class NiString in Gamebryo
> game engine, is it possible that something
> going on inside Gamebryo might cause the problem?  Or the game engine is
> irrelevant to this problem?

I think that NiString is relevant. Take a look on this FAQ:
http://www.boost.org/libs/python/doc/v2/faq.html#custom_string

> And still, what the "unidentifiable C++ exception" is really saying?

Boost.Python has next code:

try{
    run you C++ code
}
catch( registered exceptions ){
     convert them to Python exception
}
catch( ... ){
    throw Python RuntimeError( "unidentifiable C++ exception" )
}

So or you forgot to register your exception or ... something really
bad happened.

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


From n.tessore at gmail.com  Wed Aug 30 23:12:08 2006
From: n.tessore at gmail.com (Nicolas Tessore)
Date: Wed, 30 Aug 2006 23:12:08 +0200
Subject: [C++-sig]  Whether and how to use Py++ a template-loaded case
Message-ID: <6080bbe20608301412m1f6cdc9qc9daf2bf29471c54@mail.gmail.com>

Hello, I hope someone can help me out:

I am relatively new to Python, and not too experienced in C++.
My situation is the following: I am trying to add Boost.Python
bindings to a library consisting of about a hundred includes,
weighting 60Megs. It is heavily relying on inheritance, in fact
all classes derive from one base.

Now the tricky bit: The lib uses templates a _lot_. The small
part that is currently included in a Swig wrapper already
contains 800 specific template conversions.

What I now want to do is switch to Boost.Python for a change.
Swig creates a huge number of problems when using the lib, and
I trust Boost to perform better. Still there is one issue: Due
to the number of classes, templates and the frequent changes of
the codebase, I cannot generate the Boost.Python wrapper manually.
I tried using Pyste, but it is still too much work to add all
classes in an interface file. The only possibility is automation, as
presented in Py++.

The reason for posting is that I cannot find the documentation
explaining how to do templates in Py++. I can (and do) use typedefs +
declaration for templated types, but I need support for template
functions as well, since I need to access a lot of Dynamic and Static
casts, in fact, each single class needs it.

Could anyone give me a hint in the right direction, or at least tell me
that it is impossible in general, so I can hack the Swig interface to
at least stop segfault'ing?

Thanks in advance,
    N.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060830/5f4031db/attachment.htm>

From wz_ at gmx.net  Wed Aug 30 23:44:22 2006
From: wz_ at gmx.net (wz)
Date: Wed, 30 Aug 2006 23:44:22 +0200
Subject: [C++-sig]  Py++ issues, plus switching from Swig
Message-ID: <000101c6cc7d$ce216070$1700a8c0@HeikesComputer>

Those links were indeed helpful, especially the sizeof() method once again 
shows
that beauty is in simplicity.

Still, my concerns are not completely vanished.

1) If I understood it right, in order to make my instanciated vector<float>
    named vector_Float in python, do I have to use this construct:
        typedef vector<float> vector_Float; sizeof(vector_Float);
    ? Is there a better way that would not let me write out all 1000s of 
templates
    possibilities from over 250 included headers? A simple
        mb.collect_templates('vector', ['float', 'bool', 'Point3', 
'Quaternion'])
    would be a godsend.

2) How to deal with templated functions? As there are at least one 
DynamicCast<T>
    and one StaticCast<T> for each class, how can I automate their use in 
Py++?

As said, I cannot manually make C++ instantiations of the classes. They 
change too
frequently, and I could not catch up to their rate of change.

Thanks again, N. :) 



From quncao at gmail.com  Thu Aug 31 02:13:29 2006
From: quncao at gmail.com (Qun Cao)
Date: Wed, 30 Aug 2006 17:13:29 -0700
Subject: [C++-sig] RuntimeError: unidentifiable C++ exception
Message-ID: <1f527df80608301713q4687b9c7i552a3f50b4ffd573@mail.gmail.com>

>
> >* Hi Roman,
> *>* What do you mean by "the code generated by Py++ just works"?  Do you mean
> *>* you created a class
> *>* MyClass {
> *>* public:
> *>*    MyClass(char const*) ;
> *>* }
> *>* and you had successfully initialized it in Python?  I tried a similar
> *>* example class and had no problem with it.
> *
> That what you described :-).
>
> >* The real problem  I had was with this specific class NiString in Gamebryo
> *>* game engine, is it possible that something
> *>* going on inside Gamebryo might cause the problem?  Or the game engine is
> *>* irrelevant to this problem?
> *
> I think that NiString is relevant. Take a look on this FAQ:
> http://www.boost.org/libs/python/doc/v2/faq.html#custom_string
>
> Hi Roman,
Thanks for you suggestions.
I looked into the custom_string code, it is a nice example but different
with my problem.  I don't want python automatically convert my NiFixedString
to python string, all I want is a way to use the constructor of
NiFixedString class to create a python object.
 a = NiFixedString("test")

Theoretically this is no different than wrapping any other class with a
constructor that takes a string/char const*, and this code should do it,
right?
BOOST_PYTHON_MODULE(NiStringTest)
{
    class_<NiFixedString,boost::noncopyable>("pNiFixedString", init<char
const*>());
}

Btw, NiFixedString encapsulates a memory efficient implementation of
reference-counted ASCII strings, here is the header file, is there any thing
that potentially cause the exception?  Also, from my understanding, PY++
simplifies the process of generating the wrapper code, could adopting PY++
possibly help with this problem?

#ifndef NIFIXEDSTRINGTABLE_H
#define NIFIXEDSTRINGTABLE_H

#include "NiRTLib.h"
#include "NiGlobalStringTable.h"
#include "NiPath.h"

class NIMAIN_ENTRY NiFixedString : public NiMemObject
{
public:
    NiFixedString();
    NiFixedString(const char* pcString);
    NiFixedString(const NiFixedString& kString);
    ~NiFixedString();

    operator const char*() const;

    bool Exists() const;

    NiFixedString& operator=(const NiFixedString& kString);
    NiFixedString& operator=(const char* pcString);

    size_t GetLength() const;
    unsigned int GetRefCount() const;

    bool Equals(const char* pcStr) const;
    bool EqualsNoCase(const char* pcStr) const;

    bool Contains(const char* pcStr) const;
    bool ContainsNoCase(const char* pcStr) const;

    friend bool operator==(const NiFixedString& s1, const NiFixedString&
s2);
    friend bool operator!=(const NiFixedString& s1, const NiFixedString&
s2);

    friend bool operator==(const NiFixedString& s1, const char* s2);
    friend bool operator!=(const NiFixedString& s1, const char* s2);

    friend bool operator==(const char* s1, const NiFixedString& s2);
    friend bool operator!=(const char* s1, const NiFixedString& s2);

    // Begin Emergent internal use only
    // End Emergent internal use only
protected:
    NiGlobalStringTable::GlobalStringHandle m_kHandle;
};

NIMAIN_ENTRY void NiStandardizeFilePath(NiFixedString& kString);

#include "NiFixedString.inl"

#endif

Excerpt from NifixedString.inl

inline NiFixedString::NiFixedString(const char* pcString)
{
    if (pcString == NULL)
    {
        m_kHandle = NiGlobalStringTable::NULL_STRING;
    }
    else
    {
        m_kHandle = NiGlobalStringTable::AddString(pcString);
    }
}











>* And still, what the "unidentifiable C++ exception" is really saying?
> *
> Boost.Python has next code:
>
> try{
>     run you C++ code
> }
> catch( registered exceptions ){
>      convert them to Python exception
> }
> catch( ... ){
>     throw Python RuntimeError( "unidentifiable C++ exception" )
> }
>
> So or you forgot to register your exception or ... something really
> bad happened.
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060830/a5a68249/attachment.htm>

From dave at boost-consulting.com  Thu Aug 31 04:28:01 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Wed, 30 Aug 2006 22:28:01 -0400
Subject: [C++-sig] Py++ 0.8.1
References: <7465b6170608241156o5119227ds30b3b215c937e781@mail.gmail.com>
Message-ID: <upsehvdu6.fsf@boost-consulting.com>

"Roman Yakovenko" <roman.yakovenko at gmail.com> writes:

> I'm glad to announce the new version of Py++.
>
> Download page: http://language-binding.net/pyplusplus/download.html
>
> What is it?
> Py++ is an object-oriented framework for creating a code generator for
> Boost.Python library.
>
> Project home page: http://language-binding.net/pyplusplus/pyplusplus.html

We need to get Py++ referenced from Boost Python, and we need to get
its home page off of that horrible (sorry) advertisement-riddled place
where it lives now!

> Changes:
> 1. Support for Boost.Python indexing suite version 2 was implemented. Indexing
>     suite 2, provides support for almost all STD containers.
> 2. Py++ generates "stable" code. If header files were not changed,
> Py++ will not
>     change any file.
> 3. Support for huge classes was added. Py++ is able to split
> registration code for
>     the class to multiple cpp files.
> 4. User code could be added almost anywhere, without use of low level API.
> 5. Py++ will provide user with feedback: warnings about exported declarations.
> 6. Py++ now supports generation of documentation strings.
> 7. A lot of bugs were fixed.
> 8. Documentation was written.

Excellent!

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From dave at boost-consulting.com  Thu Aug 31 04:41:09 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Wed, 30 Aug 2006 22:41:09 -0400
Subject: [C++-sig] boost python doc problem? (return_self)
References: <ecnj0q$8e7$1@sea.gmane.org>
Message-ID: <uhcztvd8a.fsf@boost-consulting.com>

Neal Becker <ndbecker2 at gmail.com> writes:

> The example in ref man for return_self, shows:
>    class_<widget>("Widget")
>       .def("sensitive", &Widget::get_sensitive)
>       .def("sensitive", &Widget::set_sensitive, return_self<>())
>       ;
>
> But Widget::set_sensitive is void, it doesn't return anything.  What am I
> missing here?

That's the whole point of return_self: to give the wrapped function a
return value other than None.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From dave at boost-consulting.com  Thu Aug 31 04:46:10 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Wed, 30 Aug 2006 22:46:10 -0400
Subject: [C++-sig] custom smart pointers
References: <7465b6170608291048y6e2010d4ofa84fbd5760da4d3@mail.gmail.com>
Message-ID: <uac5lvczx.fsf@boost-consulting.com>

"Roman Yakovenko" <roman.yakovenko at gmail.com> writes:

> Code high level:
>
> template<T>
> class my_smart_ptr_t{...}
>
> struct controller_i{
>   virtual int get_value() const = 0;
> };
>
> struct add_x_t : controller_i{
>   int m_value;
>   virtual int   get_value( ){ return m_value + get_add_value(); }
>   virtual int get_add_value(){ return 0; }
> };
>
> int get_value( my_smar_ptr_t<controller_i> c ){ return c->get_value(); }
>
> In order to export this code, I wrote next code:
>
> namespace boost{ namespace python{
>
>   controller_i* get_pointer( my_smart_ptr_t< controller_i > const& p )
>   { return p.get(); }
>
>   template <>
>   struct pointee< my_smart_ptr_t< controller_i > >
>   { typedef controller_i type; };
>
>   //same code for add_x_t class
>   ...
>
> }}
>
> Because both classes contains virtual function I create class-wrapper
> for every class.
> Classes and function registration is pretty simple. After this I add
> next registration code:
>
> register_ptr_to_python< my_smart_ptr_t< controller_i > >();
> register_ptr_to_python< my_smart_ptr_t< add_x_t > >();
> implicitly_convertible< my_smart_ptr_t< add_x_t >, my_smart_ptr_t<
> controller_i > >();
>
> Python code:
>
> import my_module
> add_x = my_module.add_x_t()
> my_module.get_value( add_x ) # raises the error I described earlier.
>
> I think, I exposed my classes "by the book". May be I am missing something?

>From this I can't understand what you're doing or what errors you're
getting.  Please do what I tell everyone to do: make a _complete_,
_minimal_ example that reproduces your problem and post that.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From dave at boost-consulting.com  Thu Aug 31 04:53:28 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Wed, 30 Aug 2006 22:53:28 -0400
Subject: [C++-sig] Is there a way to automatically convert a smart_ptr
	to a	held_type (try 2)
References: <bea9869d0608300707r3009cdb2u101ab96b88ec9966@mail.gmail.com>
	<44F5CD10.5010103@pixar.com>
Message-ID: <uwt8pty3b.fsf@boost-consulting.com>

Alex Mohr <amohr at pixar.com> writes:

>> What I need is a way to tell boost python that anytime it needs to
>> convert a ptr<Base> to python it should first convert it to a
>> ref_ptr<Base>  (using a standard copy constructor in ref_ptr<>) and
>> then use this ref_ptr as the held type like normal.
>> 
>> At first I thought that implicitly_convertible<> would help:
>> 
>> implicitly_convertible<ptr<Base>, ref_ptr<Base> >();
>> 
>> But this doesn't seem to do anything for me. (this really isn't too
>> surprising since this isn't what implicitly_convertible is for, but it
>> seemed worth a try).
>> 
>> Has anyone else ran into anything similar?  Any ideas?
>
> At first glance, it seems like you want to register a to-python 
> conversion for all your ptr<T> types.  

Good idea.

> Note that all code I write in 
> this message is untested:
>
> template <class T>
> struct ptr_to_python {
>      static PyObject *convert(ptr<T> const &p) {
>          return Py_INCREF(object(ref_ptr<T>(p)).ptr());
>      }
> };
>
> Then you need to do this for each class you register:
>
> to_python_converter<ptr<T>, ptr_to_python<T> >();

Looks good so far.

> You could automate this in a custom def visitor:
>
> struct ref_ptr_stuff : def_visitor<ref_ptr_stuff> {
>      friend class def_visitor_access;
>      template <typename Class>
>      void visit(Class &c) const {
>          typedef typename Class::wrapped_type T;
>          to_python_converter<ptr<T>, ptr_to_python<T> >();
>      }
> };

That is a very clever and elegant abuse of def_visitor, I must say!

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From djowel at gmail.com  Thu Aug 31 06:34:00 2006
From: djowel at gmail.com (Joel de Guzman)
Date: Thu, 31 Aug 2006 12:34:00 +0800
Subject: [C++-sig] [tutorial] Auto-Overloading example bug?
In-Reply-To: <7465b6170608300429n564516bfi22c665cd6cbfca32@mail.gmail.com>
References: <7465b6170608300429n564516bfi22c665cd6cbfca32@mail.gmail.com>
Message-ID: <44F666B8.9080609@boost-consulting.com>

Roman Yakovenko wrote:
> Hi. I could be wrong, but it seems to me that the Boost.Python tutorials 
> contains bug.
> ( 
> http://boost.org/libs/python/doc/tutorial/doc/html/python/functions.html#python.auto_overloading 
> <http://boost.org/libs/python/doc/tutorial/doc/html/python/functions.html#python.auto_overloading> 
> )
> 
> The compiler ( I use msvc 7.1 ) is not able to compile the code. The 
> main reason is that it can not decide
> what "f" to select. I mean, it does not know what address it should pass 
> to "def" function. The same thing
> occurs for BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS. It seems to me that 
> it is right.
> 
> Some thoughts. I think, that in this case( auto-overloading example ) 
> function "def"
> should not take reference to function. It should be enough to pass an 
> instance of
> the constructed class. Same logic could be applied on class_::def method 
> too.

I am not quite sure what you mean. Please see defaults.cpp in the
libs/python/test directory. The example compiles just fine with
VC7.1. If you are not convinced, you might want to provide a minimal
sample code that we can try.

Regards,
-- 
Joel de Guzman
http://www.boost-consulting.com
http://spirit.sf.net


From djowel at gmail.com  Thu Aug 31 08:03:11 2006
From: djowel at gmail.com (Joel de Guzman)
Date: Thu, 31 Aug 2006 14:03:11 +0800
Subject: [C++-sig] [tutorial] Auto-Overloading example bug?
In-Reply-To: <44F666B8.9080609@boost-consulting.com>
References: <7465b6170608300429n564516bfi22c665cd6cbfca32@mail.gmail.com>
	<44F666B8.9080609@boost-consulting.com>
Message-ID: <44F67B9F.5060802@boost-consulting.com>

Joel de Guzman wrote:
> Roman Yakovenko wrote:
>> Hi. I could be wrong, but it seems to me that the Boost.Python tutorials 
>> contains bug.
>> ( 
>> http://boost.org/libs/python/doc/tutorial/doc/html/python/functions.html#python.auto_overloading 
>> <http://boost.org/libs/python/doc/tutorial/doc/html/python/functions.html#python.auto_overloading> 
>> )
>>
>> The compiler ( I use msvc 7.1 ) is not able to compile the code. The 
>> main reason is that it can not decide
>> what "f" to select. I mean, it does not know what address it should pass 
>> to "def" function. The same thing
>> occurs for BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS. It seems to me that 
>> it is right.
>>
>> Some thoughts. I think, that in this case( auto-overloading example ) 
>> function "def"
>> should not take reference to function. It should be enough to pass an 
>> instance of
>> the constructed class. Same logic could be applied on class_::def method 
>> too.
> 
> I am not quite sure what you mean. Please see defaults.cpp in the
> libs/python/test directory. The example compiles just fine with
> VC7.1. If you are not convinced, you might want to provide a minimal
> sample code that we can try.

Ok, disregard that. I did check it out and the error is the def it
should be:

     .def("foo", (void(*)(bool, int, char))0, foo_overloads());

I fixed the docs and updated to CVS.

Regards,
-- 
Joel de Guzman
http://www.boost-consulting.com
http://spirit.sf.net


From roman.yakovenko at gmail.com  Thu Aug 31 12:17:29 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 31 Aug 2006 13:17:29 +0300
Subject: [C++-sig] Py++ issues, plus switching from Swig
In-Reply-To: <000101c6cc7d$ce216070$1700a8c0@HeikesComputer>
References: <000101c6cc7d$ce216070$1700a8c0@HeikesComputer>
Message-ID: <7465b6170608310317n7cf1ce18s18004c9dd32d20fc@mail.gmail.com>

On 8/31/06, wz <wz_ at gmx.net> wrote:
> Those links were indeed helpful, especially the sizeof() method once again
> shows
> that beauty is in simplicity.
>
> Still, my concerns are not completely vanished.
>
> 1) If I understood it right, in order to make my instanciated vector<float>
>     named vector_Float in python, do I have to use this construct:
>         typedef vector<float> vector_Float; sizeof(vector_Float);
>     ? Is there a better way that would not let me write out all 1000s of
> templates
>     possibilities from over 250 included headers? A simple
>         mb.collect_templates('vector', ['float', 'bool', 'Point3',
> 'Quaternion'])
>     would be a godsend.

Take a look on contrib package. It contains class that may be does what you
want.  Before you go with this solution consider to read "best practices"
http://language-binding.net/pyplusplus/documentation/best_practices.html

You may have performance problems, Py++ will have to work on a huge amount
of data ( in your case ).

> 2) How to deal with templated functions? As there are at least one
> DynamicCast<T>
>     and one StaticCast<T> for each class, how can I automate their use in
> Py++?

I don't understand this part of question

> As said, I cannot manually make C++ instantiations of the classes. They
> change too
> frequently, and I could not catch up to their rate of change.

GCCXML does not dump template classes, but thier instantiation.
So, somewhere you will have to keep a matrix of template classes and thier
possible ( arguments )instantiations. I don't see how you can avoid this.

In some cases it is possible to reuse coding conventions, physical and logical
designs. For example directory A contains  template classes, while directory
B contains class-arguments to classes defined in A.  Something like this.

May be you can run Py++ on user code, so you will have all instantiations that
users use?

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


From roman.yakovenko at gmail.com  Thu Aug 31 12:26:48 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 31 Aug 2006 13:26:48 +0300
Subject: [C++-sig] RuntimeError: unidentifiable C++ exception
In-Reply-To: <1f527df80608301713q4687b9c7i552a3f50b4ffd573@mail.gmail.com>
References: <1f527df80608301713q4687b9c7i552a3f50b4ffd573@mail.gmail.com>
Message-ID: <7465b6170608310326x652c025bvcb57f770e4e351c0@mail.gmail.com>

On 8/31/06, Qun Cao <quncao at gmail.com> wrote:
> Hi Roman,
> Thanks for you suggestions.
> I looked into the custom_string code, it is a nice example but different
> with my problem.  I don't want python automatically convert my NiFixedString
> to python string, all I want is a way to use the constructor of
> NiFixedString class to create a python object.
>  a = NiFixedString("test")
>
> Theoretically this is no different than wrapping any other class with a
> constructor that takes a string/char const*, and this code should do it,
> right?

Right

> BOOST_PYTHON_MODULE(NiStringTest)
> {
>
> class_<NiFixedString,boost::noncopyable>("pNiFixedString",
> init<char const*>());
> }
>
> Btw, NiFixedString encapsulates a memory efficient implementation of
> reference-counted ASCII strings, here is the header file, is there any thing
> that potentially cause the exception?  Also, from my understanding, PY++
> simplifies the process of generating the wrapper code, could adopting PY++
> possibly help with this problem?

No, but it could help you with other problems

> inline NiFixedString::NiFixedString(const char* pcString)
> {
>     if (pcString == NULL)
>     {
>         m_kHandle = NiGlobalStringTable::NULL_STRING;
>     }
>     else
>     {
>         m_kHandle =
> NiGlobalStringTable::AddString(pcString);
^^^^^^^^^^^^^^^^^^^^^^^^^^^
My guess here is your problem - NiGlobalStringTable may be have to be
initialized earlier.

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


From abierbaum at gmail.com  Thu Aug 31 15:14:40 2006
From: abierbaum at gmail.com (Allen Bierbaum)
Date: Thu, 31 Aug 2006 08:14:40 -0500
Subject: [C++-sig] Is there a way to automatically convert a smart_ptr
	to a held_type (try 2)
In-Reply-To: <44F5CD10.5010103@pixar.com>
References: <bea9869d0608300707r3009cdb2u101ab96b88ec9966@mail.gmail.com>
	<44F5CD10.5010103@pixar.com>
Message-ID: <bea9869d0608310614w16e953ferc75692ce47883297@mail.gmail.com>

Thanks Alex, this looks very good.  I will try it later today.

On 8/30/06, Alex Mohr <amohr at pixar.com> wrote:
[snip]
>
> Also note that when you get to doing polymorphic wrappers, you're going
> to want to be able to distinguish the polymorphic wrapper type from the
> "logical" wrapped type T.  To do that, you'll need to either roll your
> own or be willing to use non-public boost.python api.
> (detail::unwrap_wrapper).

I think I am a little lost here or missing something.  I am definitely
going to be using this code in in the presence of polymorphism.

Can you explain this part further or do you have a code example where
you have had to do this?

-Allen

> Hope that helps,
>
> Alex
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
>


From wz_ at gmx.net  Thu Aug 31 16:06:19 2006
From: wz_ at gmx.net (Nicolas Tessore)
Date: Thu, 31 Aug 2006 16:06:19 +0200
Subject: [C++-sig] Lost Again! :)
Message-ID: <44F6ECDB.3070104@gmx.net>

Hello, thanks to your helpful replies, I think I got one step closer to
understanding how Py++ should work :)

I built a python script that dynamically creates a header file including
all the libraries from one folder. Fortunately, each class is contained
in one single header called <classname>.h, so that creating the template
typedefs and instantiations work fine.

Now my understanding is still lacking with respect to template functions:
I followed the howto-advise and call every function once for
instantiation. This works, however I need to control the name that
function later has in Python. At the moment I use function pointers (I
could not find something like typedef), but they don't seem to make
their way into the Boost.Python binding file.

---

Now something from my tests: I noticed I often get a error like

     super(type, obj): obj must be instance or subtype of type.

from a condition in decl_wrappers/calldef_wrapper.py (line 309).
However, running your GUI utility for a test, this error never occurs.
What is my mistake?

---

Another problem arises when I try to build an example I created for my
lib: I always run into

     'boost::python::init' : too many template arguments

errors. I had no time to dive deeper into Boost.Python, anyone up for a
quick heads-up?

---

This is all :D I really appreciate your efforts in helping out, and I am
confident Py++ results will soon overtake Swig and pay off the time
invested.

Regards, N.



From dave at boost-consulting.com  Thu Aug 31 17:12:08 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 31 Aug 2006 11:12:08 -0400
Subject: [C++-sig] boost:python::arg and default value
References: <7465b6170607041320j2f745b89xff3f8d44e475b1c0@mail.gmail.com>
Message-ID: <u64g9rlbr.fsf@boost-consulting.com>

"Roman Yakovenko" <roman.yakovenko at gmail.com> writes:

> Hi.
> I want to resolve the issue discussed in this thread:
> http://thread.gmane.org/gmane.comp.python.c++/9878/focus=9881
>
> The short version of the discussion is:
>
> class Foo;
>
> class Foo2
> {
>   friend class Foo;
>   Foo2(Foo *f=0);
> };
>
> class Foo
> {
>   Foo(Foo2 *f=0);
> };
>
> It is not possible to expose both Foo and Foo2 constructors with
> default values,
> without usage of some tricks. The main problem is that, when Boost.Python
> initializes boost::python::arg, it requires type of default value to
> be registered.
>
> I had some free time and took a look on the code. I think, I understand what is
> going on and how to fix the situation, but still need some help here and there.
>
> I see only one possible solution - to add to Boost.Python
> functionality, that will
> allow user to control argument default value conversion time. 

I disagree.  It should be lazy and do the conversion on demand.  Extra
control is not a significant benefit in this instance.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From dave at boost-consulting.com  Thu Aug 31 17:23:33 2006
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 31 Aug 2006 11:23:33 -0400
Subject: [C++-sig] [Python] null pointer access violation when calling
	Python function
References: <ecupjn$e6u$1@sea.gmane.org> <871wqym2gp.fsf@pereiro.peloton>
	<44F6B88F.2010001@spaceapplications.com>
Message-ID: <uirk9q68a.fsf@boost-consulting.com>

Tanguy Fautre <tanguy.fautre at spaceapplications.com> writes:

>>And please bring
>> your Boost.Python questions to the C++-sig
>> <http://boost.org/more/mailing_lists.htm#cplussig>.
>
>
> After doing a minimal example reproducing this error, I think I found 
> the problem.
>
> According to the C++ Programming Language (3rd ed), Section 14.7: "It is 
> implementation-defined whether destructors are invoked when a program is 
> terminated because of an uncaught exception.".

Yes.

> That, I did not know. I thought destructor were always called.
>
> GCC 4.x and Visual C++ 2005 are indeed calling std::terminate without 
> calling the destructors. Except... Except when debugging with VC++ 2005. 
> When debugging, you have the choice to continue after a uncaught 
> exception is thrown as if nothing happened. This effectively leads to a 
> segfault as described as above.

I think changing the semantics of "throw x" from "unwind the stack and
maybe terminate" to "continue as though nothing has happened" seems
like it could break any program.

> So it's not a problem with my application nor Boost.Python. Just my 
> misunderstanding of uncaught exceptions and of Visual C++.
>
> Sorry I bothered you with this false alert.

'S ok.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From eduardo.picado at gmail.com  Thu Aug 31 18:59:38 2006
From: eduardo.picado at gmail.com (Eduardo Picado)
Date: Thu, 31 Aug 2006 18:59:38 +0200
Subject: [C++-sig] Trouble choosing boost python policy
Message-ID: <6c75ff8f0608310959m1a4de5pedc4dcd24bf6d5fd@mail.gmail.com>

Hi,

I'm wrapping the following function:



 /**
  * System retrieving.
  * \throw std::bad_alloc memory allocation error.
  * \return a reference to the system associated to the calling process.
  */
  static TASRSystem & getASRSystem() throw (std::bad_alloc);



Which policy should I choose?

I read the individual doc for all of them, but it's a bit hard to get it.
I imagine it might be return_internal_reference but it is not working?



>>> system = TASRSystem.getASRSystem()
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
IndexError: boost::python::with_custodian_and_ward_postcall: argument
index out of range


Thanks,

Eduardo
-- 
Eduardo PICADO
MSN eduardo.picado at gmail.com
portable +33 6 32 47 10 10 (NEW)
fix +33 1 45 03 52 54
|??<?????>??|

From roman.yakovenko at gmail.com  Thu Aug 31 19:08:07 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 31 Aug 2006 20:08:07 +0300
Subject: [C++-sig] boost:python::arg and default value
In-Reply-To: <u64g9rlbr.fsf@boost-consulting.com>
References: <7465b6170607041320j2f745b89xff3f8d44e475b1c0@mail.gmail.com>
	<u64g9rlbr.fsf@boost-consulting.com>
Message-ID: <7465b6170608311008n5b7ec413p535eeb27e92b3249@mail.gmail.com>

On 8/31/06, David Abrahams <dave at boost-consulting.com> wrote:
> > I see only one possible solution - to add to Boost.Python
> > functionality, that will
> > allow user to control argument default value conversion time.
>
> I disagree.  It should be lazy and do the conversion on demand.  Extra
> control is not a significant benefit in this instance.

Does "lazy and on demand" mean first time a user calls function?
If so, I will try to work on the patch, but I think I still will need help with
boost::python::objects::function class.

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


From roman.yakovenko at gmail.com  Thu Aug 31 19:21:25 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 31 Aug 2006 20:21:25 +0300
Subject: [C++-sig] [tutorial] Auto-Overloading example bug?
In-Reply-To: <44F67B9F.5060802@boost-consulting.com>
References: <7465b6170608300429n564516bfi22c665cd6cbfca32@mail.gmail.com>
	<44F666B8.9080609@boost-consulting.com>
	<44F67B9F.5060802@boost-consulting.com>
Message-ID: <7465b6170608311021y2f5bebecr90ab1766c5b84ad2@mail.gmail.com>

On 8/31/06, Joel de Guzman <djowel at gmail.com> wrote:
> Ok, disregard that. I did check it out and the error is the def it
> should be:
>
>      .def("foo", (void(*)(bool, int, char))0, foo_overloads());
>
> I fixed the docs and updated to CVS.

Do you want to consider to fix "def"?

.def("foo", foo_overloads() );

It is not clear why user should pass second argument.

Thanks.

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


From roman.yakovenko at gmail.com  Thu Aug 31 19:33:02 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 31 Aug 2006 20:33:02 +0300
Subject: [C++-sig] Lost Again! :)
In-Reply-To: <44F6ECDB.3070104@gmx.net>
References: <44F6ECDB.3070104@gmx.net>
Message-ID: <7465b6170608311033j4c678819o3d91c804f9f2a156@mail.gmail.com>

On 8/31/06, Nicolas Tessore <wz_ at gmx.net> wrote:
> I built a python script that dynamically creates a header file including
> all the libraries from one folder. Fortunately, each class is contained
> in one single header called <classname>.h, so that creating the template
> typedefs and instantiations work fine.

:-)

> Now my understanding is still lacking with respect to template functions:
> I followed the howto-advise and call every function once for
> instantiation. This works, however I need to control the name that
> function later has in Python. At the moment I use function pointers (I
> could not find something like typedef), but they don't seem to make
> their way into the Boost.Python binding file.

Please read these tutorials:
http://language-binding.net/pyplusplus/documentation/tutorials/module_builder/module_builder.html

> ---
>
> Now something from my tests: I noticed I often get a error like
>
>      super(type, obj): obj must be instance or subtype of type.
>
> from a condition in decl_wrappers/calldef_wrapper.py (line 309).
> However, running your GUI utility for a test, this error never occurs.
> What is my mistake?

I am not sure that this is your mistake. Can you create small script that
reproduce the error?

> ---
>
> Another problem arises when I try to build an example I created for my
> lib: I always run into
>
>      'boost::python::init' : too many template arguments
>
> errors. I had no time to dive deeper into Boost.Python, anyone up for a
> quick heads-up?

You don't read Py++ warnings, right?

  The function has more than %d arguments ( %d ).
  You should adjust BOOST_PYTHON_MAX_ARITY macro.
  For more information see:
http://mail.python.org/pipermail/c++-sig/2002-June/001554.html

You should find this message in your output script, where first %d
will be 10, and second %d will
contain actual number of arguments.

Please read them, you will save a lot of time.

> ---
>
> This is all :D I really appreciate your efforts in helping out, and I am
> confident Py++ results will soon overtake Swig and pay off the time
> invested.

We will see.


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


From roman.yakovenko at gmail.com  Thu Aug 31 19:37:20 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 31 Aug 2006 20:37:20 +0300
Subject: [C++-sig] Trouble choosing boost python policy
In-Reply-To: <6c75ff8f0608310959m1a4de5pedc4dcd24bf6d5fd@mail.gmail.com>
References: <6c75ff8f0608310959m1a4de5pedc4dcd24bf6d5fd@mail.gmail.com>
Message-ID: <7465b6170608311037waa7f25drfa3d65a795ac2df7@mail.gmail.com>

On 8/31/06, Eduardo Picado <eduardo.picado at gmail.com> wrote:
> Hi,
>
> I'm wrapping the following function:
>
>
>
>  /**
>   * System retrieving.
>   * \throw std::bad_alloc memory allocation error.
>   * \return a reference to the system associated to the calling process.
>   */
>   static TASRSystem & getASRSystem() throw (std::bad_alloc);
>
>
>
> Which policy should I choose?

The only way to answer this question is to see function implementation.
If getASRSystem is reference to some kind of single tone, than
return_value_policy<reference_existing_object> should work.



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


From wz_ at gmx.net  Thu Aug 31 20:13:03 2006
From: wz_ at gmx.net (Nicolas Tessore)
Date: Thu, 31 Aug 2006 20:13:03 +0200
Subject: [C++-sig] Lost Again! :)
In-Reply-To: <7465b6170608311033j4c678819o3d91c804f9f2a156@mail.gmail.com>
References: <44F6ECDB.3070104@gmx.net>
	<7465b6170608311033j4c678819o3d91c804f9f2a156@mail.gmail.com>
Message-ID: <44F726AF.4010001@gmx.net>

Roman Yakovenko wrote:
> Please read these tutorials:
> http://language-binding.net/pyplusplus/documentation/tutorials/module_builder/module_builder.html

I did, though I fear I am too underexperienced to see any direct 
connection to my problem. Or did I misspell my problem? The situation is 
the following:

I have a base class MyClass. From it, I derive HisClass and HerClass.
Then I have a function
   DynamicCast<T>(MyClass *)
and another one called(MyClass *)
   StaticCast<T>
which are used to cast one of MyClass's derivatives into another, ie.

   HisClass h1;
   HerClass h2;
   h2 = DynamicCast<HisClass>(h2);

Now sorry if I am turning around here, and you answered this already. If 
so, please tell me and I will stop asking until I get the knack myself. :)

I then used a helper to write my special Py++ header:

   01:  /* The cast object */
   02:  MyClass *CastObj;
   03:
   04:  /* Cast every object once to make it known to Py++ */
   05:  DynamicCast<HisClass>(CastObj);
   06:  DynamicCast<HerClass>(CastObj);
   07:  ...

This produces a result in my binding source file. I understood I then 
have to rename the functions, but is there a way to do this in a 
pattern-like way so I can rename all DynamicCast<T> to DynamicCastToT in 
one step? Though I just checked that using a loop for renaming them 
"manually" should work as well.

> I am not sure that this is your mistake. Can you create small script that
> reproduce the error?

Unfortunately, I cannot. It does not happen on smaller test projects, 
and only occasionally on the library now.

> You don't read Py++ warnings, right?

I try to, but it is a bit hard to spot this one in 11237 lines of 
output. Especially if you are unsure what to look for.

Still I am happy to get more and more functionality to work :)

Best regards,

N.


From amohr at pixar.com  Thu Aug 31 20:24:43 2006
From: amohr at pixar.com (Alex Mohr)
Date: Thu, 31 Aug 2006 11:24:43 -0700
Subject: [C++-sig] Is there a way to automatically convert a smart_ptr
 to a held_type (try 2)
In-Reply-To: <bea9869d0608310614w16e953ferc75692ce47883297@mail.gmail.com>
References: <bea9869d0608300707r3009cdb2u101ab96b88ec9966@mail.gmail.com>
	<44F5CD10.5010103@pixar.com>
	<bea9869d0608310614w16e953ferc75692ce47883297@mail.gmail.com>
Message-ID: <44F7296B.4060708@pixar.com>

>> Also note that when you get to doing polymorphic wrappers, you're going
>> to want to be able to distinguish the polymorphic wrapper type from the
>> "logical" wrapped type T.  To do that, you'll need to either roll your
>> own or be willing to use non-public boost.python api.
>> (detail::unwrap_wrapper).
> 
> I think I am a little lost here or missing something.  I am definitely
> going to be using this code in in the presence of polymorphism.
> 
> Can you explain this part further or do you have a code example where
> you have had to do this?

The basic problem is that if you use the def_visitor approach I posted, 
then for classes exposed as polymorphic on the python side, 
Class::wrapped_type will be the polymorphic wrapper type (a class 
derived from boost::python::wrapper<T>) rather than T itself.

This is important because when you register your to-python converter, 
you want it to be for ptr<T> not ptr<wrapper<T> >, so you need to get 
the unwrapped T.  Doing this means either using detail::unwrap_wrapper, 
or rolling your own version of the same.

If you don't use the def_visitor, then you won't have this problem 
because you'll use the right T explicitly for each wrapped class. 
However, I would recommend considering the def_visitor approach if you 
have a large system because it gives you a centralized place to add 
stuff to whole categories of classes.

Does this make more sense?  Here's what my def visitor could potentially 
look like, if you were comfortable with using non-public api:

struct ref_ptr_stuff : def_visitor<ref_ptr_stuff> {
      friend class def_visitor_access;
      template <class Class, class Wrapper, class T>
      void register_conversion(Class &c, Wrapper *, T *) {
          to_python_converter<ptr<T>, ptr_to_python<T> >();
      }
      template <class Class>
      void visit(Class &c) const {
          typedef typename Class::wrapped_type T;
          register_conversion(c, (T *)0,
                              detail::unwrap_wrapper((T *)0));
      }
};

Of course, in this example I don't need Wrapper in register_conversion, 
but I thought I'd leave it to be illustrative.  Also, let me reemphasize 
that this is non-public api.

If you're able to do partial specialization, I believe you can roll your 
own unwrap_wrapper that works in a nicer way.

Alex


From abierbaum at gmail.com  Thu Aug 31 20:32:37 2006
From: abierbaum at gmail.com (Allen Bierbaum)
Date: Thu, 31 Aug 2006 13:32:37 -0500
Subject: [C++-sig] Describing how boost.python works
Message-ID: <bea9869d0608311132t512e7a20j399e303686282abf@mail.gmail.com>

Hello all,

I don't know about the rest of you, but I have been using boost.python
for a number of years now and there is still a lot about how it works
behind the scenes that seems like magic to me.  Normally this isn't a
bad thing, but when trying to track down bugs it becomes a problem
because I just don't know what the system is trying to do.

Earlier this year I became interested enough to spend several evenings
looking through old mailing list questions and through the code to try
to start documenting it for myself.  I have put up the results of this
effort on a wiki at:

https://realityforge.vrsource.org/view/PyOpenSG/HowBoostPythonWorks

There is still much to be documented (especially the conversion of
data across the language barrier), but it is a start and hopefully it
will be helpful to someone.

Does anyone know of any more complete documentation of how
boost.python works?  Are there any other people willing to collaborate
on documenting the inner workings of the system?

-Allen


From roman.yakovenko at gmail.com  Thu Aug 31 21:00:49 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 31 Aug 2006 22:00:49 +0300
Subject: [C++-sig] Lost Again! :)
In-Reply-To: <44F726AF.4010001@gmx.net>
References: <44F6ECDB.3070104@gmx.net>
	<7465b6170608311033j4c678819o3d91c804f9f2a156@mail.gmail.com>
	<44F726AF.4010001@gmx.net>
Message-ID: <7465b6170608311200j41b6f088tb725d832253338b6@mail.gmail.com>

On 8/31/06, Nicolas Tessore <wz_ at gmx.net> wrote:
> I did, though I fear I am too underexperienced to see any direct
> connection to my problem. Or did I misspell my problem? The situation is
> the following:
>
> I have a base class MyClass. From it, I derive HisClass and HerClass.
>
> Then I have a function
>    DynamicCast<T>(MyClass *)
> and another one called(MyClass *)
>    StaticCast<T>
> which are used to cast one of MyClass's derivatives into another, ie.
>
>    HisClass h1;
>    HerClass h2;
>    h2 = DynamicCast<HisClass>(h2);
>
> Now sorry if I am turning around here, and you answered this already. If
> so, please tell me and I will stop asking until I get the knack myself. :)
>
> I then used a helper to write my special Py++ header:
>
>    01:  /* The cast object */
>    02:  MyClass *CastObj;
>    03:
>    04:  /* Cast every object once to make it known to Py++ */
>    05:  DynamicCast<HisClass>(CastObj);
>    06:  DynamicCast<HerClass>(CastObj);
>    07:  ...

This is what I understand until now:

struct MyClass{...};
struct HisClass : MyClass
{};
struct HerClass : MyClass
{};

And you want to cast from HisClass to HerClass, right?
( I don't want to know why you have to do this, really :) )

If HisClass derives from HerClass, than you don't need to introduce
casting.
http://www.boost.org/libs/python/doc/tutorial/doc/html/python/exposing.html#python.inheritance

> This produces a result in my binding source file. I understood I then
> have to rename the functions, but is there a way to do this in a
> pattern-like way so I can rename all DynamicCast<T> to DynamicCastToT in
> one step? Though I just checked that using a loop for renaming them
> "manually" should work as well.

from pygccxml import declarations

def rename_cast( cast ):
    cls_to = declarations.class_traits.get_declaration( cast.return_type )
    cast.rename( 'DynamicCastTo' + cls_to.name )

casts = mb.free_functions( lambda decl: decl.name.startswith( 'DynamicCast<'
) )
map( rename_cast, casts )

> > I am not sure that this is your mistake. Can you create small script
that
> > reproduce the error?
>
> Unfortunately, I cannot. It does not happen on smaller test projects,
> and only occasionally on the library now.

May be you can debug the code? I can do it too, but you will have to send me
GCCXML generated xml file and script that reporoduce the error. C++ source
code is not requiered.

> > You don't read Py++ warnings, right?
>
> I try to, but it is a bit hard to spot this one in 11237 lines of
> output. Especially if you are unsure what to look for.

By default Py++ prints all messages, formatted to feat your console
width. You can redefine this functionality and print a message into
file and all message could be written within one line. After this
you can use grep to find and monitor interesting messages.

Read this section of docs:
http://language-binding.net/pyplusplus/documentation/feedback.html

-- 
Roman Yakovenko
C++ Python language binding
http://www.language-binding.net/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060831/0ac1a620/attachment.htm>

From abierbaum at gmail.com  Thu Aug 31 22:06:26 2006
From: abierbaum at gmail.com (Allen Bierbaum)
Date: Thu, 31 Aug 2006 15:06:26 -0500
Subject: [C++-sig] Is there a way to automatically convert a smart_ptr
	to a held_type (try 2)
In-Reply-To: <uwt8pty3b.fsf@boost-consulting.com>
References: <bea9869d0608300707r3009cdb2u101ab96b88ec9966@mail.gmail.com>
	<44F5CD10.5010103@pixar.com> <uwt8pty3b.fsf@boost-consulting.com>
Message-ID: <bea9869d0608311306v43136b0aof87bd634c508650a@mail.gmail.com>

Is there a way while access a wrapped object in python to ask it the
type of the currently in use held_type?  In other words, is there a
way I can write a test case in python that will make sure things are
behaving as I want them and that I am not using a ptr<Base> to hold
the C++ object behind the scenes?

-Allen

On 8/30/06, David Abrahams <dave at boost-consulting.com> wrote:
> Alex Mohr <amohr at pixar.com> writes:
>
> >> What I need is a way to tell boost python that anytime it needs to
> >> convert a ptr<Base> to python it should first convert it to a
> >> ref_ptr<Base>  (using a standard copy constructor in ref_ptr<>) and
> >> then use this ref_ptr as the held type like normal.
> >>
> >> At first I thought that implicitly_convertible<> would help:
> >>
> >> implicitly_convertible<ptr<Base>, ref_ptr<Base> >();
> >>
> >> But this doesn't seem to do anything for me. (this really isn't too
> >> surprising since this isn't what implicitly_convertible is for, but it
> >> seemed worth a try).
> >>
> >> Has anyone else ran into anything similar?  Any ideas?
> >
> > At first glance, it seems like you want to register a to-python
> > conversion for all your ptr<T> types.
>
> Good idea.
>
> > Note that all code I write in
> > this message is untested:
> >
> > template <class T>
> > struct ptr_to_python {
> >      static PyObject *convert(ptr<T> const &p) {
> >          return Py_INCREF(object(ref_ptr<T>(p)).ptr());
> >      }
> > };
> >
> > Then you need to do this for each class you register:
> >
> > to_python_converter<ptr<T>, ptr_to_python<T> >();
>
> Looks good so far.
>
> > You could automate this in a custom def visitor:
> >
> > struct ref_ptr_stuff : def_visitor<ref_ptr_stuff> {
> >      friend class def_visitor_access;
> >      template <typename Class>
> >      void visit(Class &c) const {
> >          typedef typename Class::wrapped_type T;
> >          to_python_converter<ptr<T>, ptr_to_python<T> >();
> >      }
> > };
>
> That is a very clever and elegant abuse of def_visitor, I must say!
>
> --
> Dave Abrahams
> Boost Consulting
> www.boost-consulting.com
>
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
>


From quncao at gmail.com  Thu Aug 31 22:37:47 2006
From: quncao at gmail.com (Qun Cao)
Date: Thu, 31 Aug 2006 13:37:47 -0700
Subject: [C++-sig] RuntimeError: unidentifiable C++ exception
In-Reply-To: <7465b6170608310326x652c025bvcb57f770e4e351c0@mail.gmail.com>
References: <1f527df80608301713q4687b9c7i552a3f50b4ffd573@mail.gmail.com>
	<7465b6170608310326x652c025bvcb57f770e4e351c0@mail.gmail.com>
Message-ID: <1f527df80608311337v1e1ba4b4n9fc07b89a962c37@mail.gmail.com>

Hi Roman,
NiGlobalStringTable here is another class that manages the global NiString
table.  There is a line:
#include "NiGlobalStringTable.h"
in NiFixedString.h

In this case, Boost_Python should be able to look up
NiGlobalStringTable.hand the corresponding lib file to execute the C++
code, right?  Or do you
suggest that I should manually wrap
NiGlobalStringTable class too?

Thanks,
Qun

On 8/31/06, Roman Yakovenko <roman.yakovenko at gmail.com> wrote:
>
>
> > inline NiFixedString::NiFixedString(const char* pcString)
> > {
> >     if (pcString == NULL)
> >     {
> >         m_kHandle = NiGlobalStringTable::NULL_STRING;
> >     }
> >     else
> >     {
> >         m_kHandle =
> > NiGlobalStringTable::AddString(pcString);
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^
> My guess here is your problem - NiGlobalStringTable may be have to be
> initialized earlier.
>
> --
> Roman Yakovenko
> C++ Python language binding
> http://www.language-binding.net/
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060831/ab0d6a7a/attachment.htm>

From roman.yakovenko at gmail.com  Thu Aug 31 22:41:33 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 31 Aug 2006 23:41:33 +0300
Subject: [C++-sig] RuntimeError: unidentifiable C++ exception
In-Reply-To: <1f527df80608311337v1e1ba4b4n9fc07b89a962c37@mail.gmail.com>
References: <1f527df80608301713q4687b9c7i552a3f50b4ffd573@mail.gmail.com>
	<7465b6170608310326x652c025bvcb57f770e4e351c0@mail.gmail.com>
	<1f527df80608311337v1e1ba4b4n9fc07b89a962c37@mail.gmail.com>
Message-ID: <7465b6170608311341w23fb2e2eg72eeba6dae52b5d5@mail.gmail.com>

On 8/31/06, Qun Cao <quncao at gmail.com> wrote:
> Hi Roman,
>  NiGlobalStringTable here is another class that manages the global NiString
> table.  There is a line:
> #include "NiGlobalStringTable.h"
> in NiFixedString.h
>
> In this case, Boost_Python should be able to look up NiGlobalStringTable.h
> and the corresponding lib file to execute the C++ code, right?  Or do you
> suggest that I should manually wrap
> NiGlobalStringTable class too?

I don't understand you. All I said is that NiGlobalStringTable  could be some
kind of single tone, that should be initialized. May be you forgot to
intialize it?

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


From roman.yakovenko at gmail.com  Thu Aug 31 22:58:39 2006
From: roman.yakovenko at gmail.com (Roman Yakovenko)
Date: Thu, 31 Aug 2006 23:58:39 +0300
Subject: [C++-sig] custom smart pointers
In-Reply-To: <uac5lvczx.fsf@boost-consulting.com>
References: <7465b6170608291048y6e2010d4ofa84fbd5760da4d3@mail.gmail.com>
	<uac5lvczx.fsf@boost-consulting.com>
Message-ID: <7465b6170608311358s1c02c6c9hc26222fae8cfdef6@mail.gmail.com>

On 8/31/06, David Abrahams <dave at boost-consulting.com> wrote:
> >From this I can't understand what you're doing or what errors you're
> getting.  Please do what I tell everyone to do: make a _complete_,
> _minimal_ example that reproduces your problem and post that.

I attached 3 files. This is complete and minimal example.

custom_smart_ptr.h
  file contains smart pointer class definition.

custom_smart_ptr_classes_to_be_exported.hpp
   file contains definition of 2 classes:  "base" and "derived" classes.
   This file also defines "get_value" function. this function is
called from Python.

custom_smart_ptr_classes.cpp
   file contains Boost.Python bindings.

Python code is next:
import custom_smart_ptr_classes as cspc
d = cspc.derived( 2 )
cspc.get_value( d )

The last line raises next exception:

ArgumentError: Python argument types in
    custom_smart_ptr_classes.get_value(derived)
did not match C++ signature:
    get_value(my_smart_ptr_t<base> a)

Thank you.

-- 
Roman Yakovenko
C++ Python language binding
http://www.language-binding.net/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: custom_smart_ptr.h
Type: text/x-chdr
Size: 920 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060831/ba3138b5/attachment.h>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: custom_smart_ptr_classes_to_be_exported.hpp
Type: text/x-c++hdr
Size: 412 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060831/ba3138b5/attachment.hpp>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: custom_smart_ptr_classes.cpp
Type: text/x-c++src
Size: 2341 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20060831/ba3138b5/attachment.cpp>