From kalin.eh at gmail.com  Sun Apr  3 10:52:15 2005
From: kalin.eh at gmail.com (kalin)
Date: Sun, 03 Apr 2005 18:52:15 +1000
Subject: [C++-sig] Re: what happened with this std::string parameter in
	boost.python
In-Reply-To: <cpt4b7$sn7$1@sea.gmane.org>
References: <20041216072015.4E0ED1E4008@bag.python.org>
	<cpt4b7$sn7$1@sea.gmane.org>
Message-ID: <d2oast$6ov$1@sea.gmane.org>

David Abrahams wrote:
> Donnie Leen wrote:
> 
>>I made a program embedding boost.python, writting a module as following:
>>
>>void write( std::string s )
>>{
>>	// do nothing
>>}
>>BOOST_PYTHON_MODULE(mym)
>>{
>>	def( "write", &write );
>>}
>>
>>
>>I run the python code:
>>import mym
>>s = 'hello'
>>mym.write(s)
>>
>>it runs ok, but if i run the python code as following:
>>import mym
>>s = int.__doc__
>>mym.write(s)
>>
>>it crashed
> 
> 
> What do you mean by "crashed?"  Are you sure it didn't throw an
> exception that was never caught?
> 
> 

I had this problem earlier. It would just access violate, as though the
string had a bad buffer pointer. I didn't look into it too much and
changed it to const char*'s so I could just get things working.

Although now, when I change it back to test, it doesn't crash anymore. :(

FWIW

kalin



From H.Duerer at gmx.net  Mon Apr  4 18:07:57 2005
From: H.Duerer at gmx.net (Holger Duerer)
Date: Mon, 04 Apr 2005 17:07:57 +0100
Subject: [C++-sig] Re: what happened with this std::string parameter in
	boost.python
References: <20041216072015.4E0ED1E4008@bag.python.org>
	<cpt4b7$sn7$1@sea.gmane.org> <d2oast$6ov$1@sea.gmane.org>
Message-ID: <87u0mmy17m.fsf@ronaldann.demon.co.uk>

>>>>> "kalin" == kalin  <kalin.eh at gmail.com> writes:

    kalin> David Abrahams wrote:
    >> Donnie Leen wrote:
    >> 
    >>> I made a program embedding boost.python, writting a module as following:
    >>> 
    >>> void write( std::string s )
    >>> {
    >>> // do nothing
    >>> }
    >>> BOOST_PYTHON_MODULE(mym)
    >>> {
    >>> def( "write", &write );
    >>> }
    >>> 
    >>> 
    >>> I run the python code:
    >>> import mym
    >>> s = 'hello'
    >>> mym.write(s)
    >>> 
    >>> it runs ok, but if i run the python code as following:
    >>> import mym
    >>> s = int.__doc__
    >>> mym.write(s)
    >>> 
    >>> it crashed
    >> What do you mean by "crashed?"  Are you sure it didn't throw an
    >> exception that was never caught?
    >> 

[...]

Is this under Windows?  We had problems like that -- short string (<16
bytes?) work, long strings cause a crash.  We traced this to multiple
versions of the C++ runtime library.  Are you linking the runtime
library statically?

        Holger




From paniq at paniq.org  Tue Apr  5 05:39:09 2005
From: paniq at paniq.org (Leonard "paniq" Ritter)
Date: Tue, 05 Apr 2005 05:39:09 +0200
Subject: [C++-sig] doing iterators another way...
Message-ID: <4252085D.6010304@paniq.org>

the example that is given in the tutorial for iterators is giving a 
standard way of solving the problem for c++ iterators, but i'm also 
using enumerators which provide a set of current/next/reset commands. 
for those interested, this is how i turned my class into an iterator:

        template<typename T>
        struct Enumerator : Mu::Interface::Unknown
        {           
            typedef typename Ptrfy<T>::Type ElementReturnType;

            virtual Mu::Result Mu_InterfaceCall MoveNext(Bool& 
succeeded) = 0;
            virtual Mu::Result Mu_InterfaceCall Current(T& current) = 0;
            virtual Mu::Result Mu_InterfaceCall Reset() = 0;

            Mu_WrapGetMethod(MoveNext,Bool);
            Mu_WrapGetMethod(Current,ElementReturnType);
        };

        template<typename T>
        typename T::ElementReturnType iterator_next(T* enumerator)
        {
            if (!enumerator->MoveNext())
            {
                PyErr_SetString(PyExc_StopIteration,"");
                boost::python::throw_error_already_set();
            }
            return enumerator->Current();
        }

        template<typename T>
        T* MakePythonIteratorFromEnumerator(T* enumerator)
        {           
            enumerator->Reset();
            return enumerator;
        }

    and here's some pseudo-boost.python code:

    class_<EnumeratorClass, bases<Mu::Interface::Unknown> 
 >("EnumeratorClass")
        .def("__iter__", 
&MakePythonIteratorFromEnumerator<EnumeratorClass>, 
return_internal_reference<>())
        .def("next", &iterator_next<EnumeratorClass>)
   ;



From paniq at paniq.org  Tue Apr  5 10:16:28 2005
From: paniq at paniq.org (Leonard "paniq" Ritter)
Date: Tue, 05 Apr 2005 10:16:28 +0200
Subject: [C++-sig] doing iterators another way...
In-Reply-To: <4252085D.6010304@paniq.org>
References: <4252085D.6010304@paniq.org>
Message-ID: <4252495C.9020604@paniq.org>

the approach below has the flaw that recursive iterations will fail 
since the enumerator is not cloned. for that purpose i added a Clone() 
method to the enumerator interface which is called in 
MakePythonIteratorFromEnumerator, where a Ptr<T> (a custom smart pointer 
class) is returned, hence suspending the need for a return value call 
policy.

Leonard "paniq" Ritter wrote:

> the example that is given in the tutorial for iterators is giving a 
> standard way of solving the problem for c++ iterators, but i'm also 
> using enumerators which provide a set of current/next/reset commands. 
> for those interested, this is how i turned my class into an iterator:
>
>        template<typename T>
>        struct Enumerator : Mu::Interface::Unknown
>        {                      typedef typename Ptrfy<T>::Type 
> ElementReturnType;
>
>            virtual Mu::Result Mu_InterfaceCall MoveNext(Bool& 
> succeeded) = 0;
>            virtual Mu::Result Mu_InterfaceCall Current(T& current) = 0;
>            virtual Mu::Result Mu_InterfaceCall Reset() = 0;
>
>            Mu_WrapGetMethod(MoveNext,Bool);
>            Mu_WrapGetMethod(Current,ElementReturnType);
>        };
>
>        template<typename T>
>        typename T::ElementReturnType iterator_next(T* enumerator)
>        {
>            if (!enumerator->MoveNext())
>            {
>                PyErr_SetString(PyExc_StopIteration,"");
>                boost::python::throw_error_already_set();
>            }
>            return enumerator->Current();
>        }
>
>        template<typename T>
>        T* MakePythonIteratorFromEnumerator(T* enumerator)
>        {                      enumerator->Reset();
>            return enumerator;
>        }
>
>    and here's some pseudo-boost.python code:
>
>    class_<EnumeratorClass, bases<Mu::Interface::Unknown> 
> >("EnumeratorClass")
>        .def("__iter__", 
> &MakePythonIteratorFromEnumerator<EnumeratorClass>, 
> return_internal_reference<>())
>        .def("next", &iterator_next<EnumeratorClass>)
>   ;
>
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
>
>



From thomas.schipolowski at tu-berlin.de  Tue Apr  5 13:28:55 2005
From: thomas.schipolowski at tu-berlin.de (Thomas Schipolowski)
Date: Tue, 5 Apr 2005 13:28:55 +0200
Subject: [C++-sig] crashes on functions with std::string arguments
Message-ID: <001101c539d2$a6d7c0c0$89689582@dynamik.local>

 Hello,
 I am using Python 2.3 and Boost_1_32_0 with VC6 on Windows 2000. I
experience crashes due to heap corruption when a function or method with a
std::string argument is called. The same problem is also triggered by the
extract<std::string> facility.
 The following test code reproduces both conditions:

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

int test1(std::string s) {
            return s.end() - s.begin();
}

int test2(object o) {
            extract<std::string> extr(o);
            if (extr.check()) {
                        std::string s(extr());
                        return s.end() - s.begin();
            } else {
                        return -1;
            }
}

BOOST_PYTHON_MODULE(test)
{
            def("test1", test1);
            def("test2", test2);
}

The error will not be immediately apparent until the extension is compiled
in debug mode. Calling any of the above functions from python will then
(always) pop up a message saying  "Debug Assertion _CrtIsValidHeapPtr()
failed". When compiled without debug, the function calls may succeed many
times before an "unidentifyable C++ exception" is reported. The interpreter
will segfault at this point or later on, depending on memory usage.

Now I saw this message
http://mail.python.org/pipermail/c++-sig/2005-April/008760.html on the list
which seems related. With regard to the explanation given there (multiple
versions of the C++ runtime library on windows) I agree that this is a
possible reason in my case, as the problem mysteriously went away when I
switched from my Win 2000 box at work to a Win XP Laptop. But yesterday I
tried on a third computer (also XP) and there it crashed the same as on Win
2000. However, I think the standard library is statically linked in my case.
(But I am not sure how to check and control this in the VC-IDE.)

I tried to dig into the source with a debugger, but couldn't find a solution
so far. Details are below. I would be glad about any advice on how to
correct this or work around it.

Many thanks,
Thomas Schipolowski


The debugger shows me that the error happens at the point where the
temporary std::string object stored in
rvalue_from_python_data<>::storage.bytes is destroyed. (In case of the
test1() function this happens when
detail::caller_arity<1>::impl<>::operator() finishes. In the test2()
function it is the destruction of the extr variable when test2() returns.)
Further inside the std::string destructor, the problem is apparently caused
by the destructor trying to deallocate the internal character buffer. At
this point I thought there might be something wrong with the copy-on-write
implementation of std::string. In fact, the string s in test2() is destroyed
first, deleting its buffer, and then it crashes when the string contained in
extr is destroyed, again deleting its buffer. But I am not sure if there is
actually anything shared between both string objects. From what I saw in the
debugger, it appears that the constructor of s initializes its pointer to
the character buffer of a temporary string which is created as return value
of extract_rvalue<T>::operator(). The buffer address of s is different from
the buffer of the string in extract_rvalue<>.m_data.storage.bytes. I have no
idea what can be wrong here.






From dave at boost-consulting.com  Tue Apr  5 14:11:54 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Tue, 05 Apr 2005 08:11:54 -0400
Subject: [C++-sig] Re: crashes on functions with std::string arguments
References: <001101c539d2$a6d7c0c0$89689582@dynamik.local>
Message-ID: <uzmwdpgmt.fsf@boost-consulting.com>

"Thomas Schipolowski" <thomas.schipolowski at tu-berlin.de> writes:

>  Hello,
>  I am using Python 2.3 and Boost_1_32_0 with VC6 on Windows 2000. I
> experience crashes due to heap corruption when a function or method with a
> std::string argument is called. The same problem is also triggered by the
> extract<std::string> facility.
>  The following test code reproduces both conditions:

Not by itself, it doesn't!  What Python code do you need to execute to
generate the crash?

> #include <string>
> #include <boost/python.hpp>
> using namespace boost::python;
>
> int test1(std::string s) {
>             return s.end() - s.begin();
> }
>
> int test2(object o) {
>             extract<std::string> extr(o);
>             if (extr.check()) {
>                         std::string s(extr());
>                         return s.end() - s.begin();
>             } else {
>                         return -1;
>             }
> }
>
> BOOST_PYTHON_MODULE(test)
> {
>             def("test1", test1);
>             def("test2", test2);
> }

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



From thomas.schipolowski at tu-berlin.de  Tue Apr  5 14:56:06 2005
From: thomas.schipolowski at tu-berlin.de (Thomas Schipolowski)
Date: Tue, 5 Apr 2005 14:56:06 +0200
Subject: [C++-sig] Re: crashes on functions with std::string arguments
References: <001101c539d2$a6d7c0c0$89689582@dynamik.local>
	<uzmwdpgmt.fsf@boost-consulting.com>
Message-ID: <000701c539de$d4a3e180$89689582@dynamik.local>

Thanks for the quick reply Dave!

> >  The following test code reproduces both conditions:
>
> Not by itself, it doesn't!  What Python code do you need to execute to
> generate the crash?

I simple invoke python.exe in a console window (that's also what the
debugger in the VC-IDE does) and say
>>> import test
>>> test.test1("Otto")
or
>>> test.test2("Otto")

In debug mode, the message box with the _CrtIsValidHeapPointer Assertion
Failure opens instantly before the python prompt returns. I can answer the
message box with "Retry" and get an "unidentifyable C++ exception".
When not compiled with debug, I do something like
>>> for i in range(100):
          print i, test.test2("Otto")

This prints "4" between 10 and 30 times, then prints the above mentioned
exception. If I invoke the loop statement again I usually get only one "4"
or even no "4", followed by the exception. Repeating the loop statement
further, I get varying numbers of "4"s, but never again 20... At some point,
the interpreter will crash with a memory write access violation (I dont know
the English wording of the dialog.) This can happen at any time - I have
seen it sometimes as early as the tenth invokation of test2() and sometimes
only when I terminated the interpreter via Crtl-Z after many repetitions of
the above loop command.
However, calling test2() with something that is not accepted by
extract<std::string>, e.g. test2(None) or test2(10), will never crash (I
tried the loop with range(100000)). It will print -1 as it should.


Thanks again,
Thomas.






From dave at boost-consulting.com  Tue Apr  5 16:19:51 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Tue, 05 Apr 2005 10:19:51 -0400
Subject: [C++-sig] Re: crashes on functions with std::string arguments
References: <001101c539d2$a6d7c0c0$89689582@dynamik.local>
	<uzmwdpgmt.fsf@boost-consulting.com>
	<000701c539de$d4a3e180$89689582@dynamik.local>
Message-ID: <uekdpi9vc.fsf@boost-consulting.com>

"Thomas Schipolowski" <thomas.schipolowski at tu-berlin.de> writes:

> Thanks for the quick reply Dave!
>
>> >  The following test code reproduces both conditions:
>>
>> Not by itself, it doesn't!  What Python code do you need to execute to
>> generate the crash?
>
> I simple invoke python.exe in a console window (that's also what the
> debugger in the VC-IDE does) and say
>>>> import test
>>>> test.test1("Otto")
> or
>>>> test.test2("Otto")

Works fine for me here, building and testing with Boost.Build.  This
is a build problem on your end (see the tip at the top of
http://www.boost.org/libs/python/doc/tutorial/doc/html/python/hello.html).
If you examine the build command-lines it should be easy to verify
that you have multiple runtime libraries in use.

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



From thomas.schipolowski at tu-berlin.de  Wed Apr  6 09:18:40 2005
From: thomas.schipolowski at tu-berlin.de (Thomas Schipolowski)
Date: Wed, 6 Apr 2005 09:18:40 +0200
Subject: [C++-sig] Re: crashes on functions with std::string arguments
References: <001101c539d2$a6d7c0c0$89689582@dynamik.local><uzmwdpgmt.fsf@boost-consulting.com><000701c539de$d4a3e180$89689582@dynamik.local>
	<uekdpi9vc.fsf@boost-consulting.com>
Message-ID: <000501c53a78$dd1bcb80$89689582@dynamik.local>

> Works fine for me here, building and testing with Boost.Build.  This
> is a build problem on your end


Yes, I compiled the test case with BJam and now it works flawlessly. Great!
One complication is left: I wonder now how to explain to BJam that some of
my files need to be fed to a fortran compiler, before they are linked with
the c++ objects. Would you say that using BJams 'prebuilt binaries' feature
with an external fortran build is the best way to achieve that, or is it
reasonably easy to tell BJam about the fortran toolchain? Is the latter
documented somewhere?

Again, thanks very much for the quick help.
Thomas.






From dave at boost-consulting.com  Wed Apr  6 15:05:18 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Wed, 06 Apr 2005 09:05:18 -0400
Subject: [C++-sig] Re: crashes on functions with std::string arguments
References: <001101c539d2$a6d7c0c0$89689582@dynamik.local>
	<uzmwdpgmt.fsf@boost-consulting.com>
	<000701c539de$d4a3e180$89689582@dynamik.local>
	<uekdpi9vc.fsf@boost-consulting.com>
	<000501c53a78$dd1bcb80$89689582@dynamik.local>
Message-ID: <uk6ng2gz5.fsf@boost-consulting.com>

"Thomas Schipolowski" <thomas.schipolowski at tu-berlin.de> writes:

>> Works fine for me here, building and testing with Boost.Build.  This
>> is a build problem on your end
>
>
> Yes, I compiled the test case with BJam and now it works flawlessly. Great!
> One complication is left: I wonder now how to explain to BJam that some of
> my files need to be fed to a fortran compiler, before they are linked with
> the c++ objects. Would you say that using BJams 'prebuilt binaries' feature
> with an external fortran build is the best way to achieve that, or is it
> reasonably easy to tell BJam about the fortran toolchain? Is the latter
> documented somewhere?

Well, Boost.Build v2 has a fortran toolset, but it seems to be under-documented

  $ bjam --v2 --help fortran

  Module 'fortran':

    This file contains common settings for all fortran tools

  Module 'fortran' classes:

    Use --help fortran.<class-name> to get more information.

  * fortran-compiling-generator:

I'm not sure what you need to do to configure it.  Maybe its author
can help?  I suggest following up to the jamboost list
(http://www.boost.org/more/mailing_lists.htm#jamboost).

> Again, thanks very much for the quick help.
> Thomas.

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



From jbrandmeyer at earthlink.net  Thu Apr  7 05:25:31 2005
From: jbrandmeyer at earthlink.net (Jonathan Brandmeyer)
Date: Wed, 06 Apr 2005 23:25:31 -0400
Subject: [C++-sig] build system feature requests
Message-ID: <1112844331.17274.34.camel@localhost.localdomain>

Could you add two features to the build scripts for Boost.Python?
First, when running bjam from libs/python/build, the libraries are not
name mangled as described here:
http://www.boost.org/more/getting_started.html#Results

Second, It is fairly common (for me, at least), to have more than one
version of Python installed on my system at any given time.  Is it
possible to add -py23, -py24, etc. to the standard mangling scheme and
the build directory layout?

Thanks,
-Jonathan Brandmeyer



From grafik.list at redshift-software.com  Thu Apr  7 05:46:55 2005
From: grafik.list at redshift-software.com (Rene Rivera)
Date: Wed, 06 Apr 2005 22:46:55 -0500
Subject: [C++-sig] build system feature requests
In-Reply-To: <1112844331.17274.34.camel@localhost.localdomain>
References: <1112844331.17274.34.camel@localhost.localdomain>
Message-ID: <4254AD2F.4050808@redshift-software.com>

Jonathan Brandmeyer wrote:
> Could you add two features to the build scripts for Boost.Python?
> First, when running bjam from libs/python/build, the libraries are not
> name mangled as described here:
> http://www.boost.org/more/getting_started.html#Results

Ok that's easy enough. And other libraries already do that.

> Second, It is fairly common (for me, at least), to have more than one
> version of Python installed on my system at any given time.  Is it
> possible to add -py23, -py24, etc. to the standard mangling scheme and
> the build directory layout?

Also easy, but this one I would ask Dave and others to comment as I'm 
not an expert on BPL usage. I would think that this only applies to BPL 
extensions and the BPL library itself.


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


From dave at boost-consulting.com  Thu Apr  7 06:58:41 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 07 Apr 2005 00:58:41 -0400
Subject: [C++-sig] Re: build system feature requests
References: <1112844331.17274.34.camel@localhost.localdomain>
	<4254AD2F.4050808@redshift-software.com>
Message-ID: <ur7hn41z2.fsf@boost-consulting.com>

Rene Rivera <grafik.list at redshift-software.com> writes:

> Jonathan Brandmeyer wrote:
>> Could you add two features to the build scripts for Boost.Python?
>> First, when running bjam from libs/python/build, the libraries are not
>> name mangled as described here:
>> http://www.boost.org/more/getting_started.html#Results
>
> Ok that's easy enough. And other libraries already do that.
>
>> Second, It is fairly common (for me, at least), to have more than one
>> version of Python installed on my system at any given time.  Is it
>> possible to add -py23, -py24, etc. to the standard mangling scheme and
>> the build directory layout?
>
> Also easy, but this one I would ask Dave and others to comment as I'm 
> not an expert on BPL usage. 

I don't have much to say about it.  Sounds OK to me.

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



From beyer at imb-jena.de  Thu Apr  7 23:42:01 2005
From: beyer at imb-jena.de (Andreas Beyer)
Date: Thu, 07 Apr 2005 14:42:01 -0700
Subject: [C++-sig] "_POSIX_C_SOURCE" redefined
Message-ID: <4255A929.9030200@imb-jena.de>

Hi,

I keep getting the "_POSIX_C_SOURCE" redefined warning when compiling my 
own project with BJAM.
The problem was described in:
http://mail.python.org/pipermail/c++-sig/2005-February/008564.html
In my case the examples get compiled without a warning, but my own 
project causes the warning:

gcc-C++-action 
bin/csrc/network.so/gcc/release/shared-linkable-true/network.o
In file included from 
/cellar/users/abeyer/boost/boost_current/boost/python/detail/wrap_python.hpp:30,
                 from 
/cellar/users/abeyer/boost/boost_current/boost/python/detail/prefix.hpp:13,
                 from 
/cellar/users/abeyer/boost/boost_current/boost/python/args.hpp:8,
                 from 
/cellar/users/abeyer/boost/boost_current/boost/python.hpp:11,
                 from property_handle.hpp:16,
                 from node.hpp:26,
                 from connection.hpp:18,
                 from network.cpp:13:
/cellar/users/abeyer/include/python2.4/pyconfig.h:835:1: warning: 
"_POSIX_C_SOURCE" redefined
In file included from 
/usr/include/c++/3.2.3/i386-redhat-linux/bits/os_defines.h:39,
                 from 
/usr/include/c++/3.2.3/i386-redhat-linux/bits/c++config.h:34,
                 from /usr/include/c++/3.2.3/string:45,
                 from connection.hpp:15,
                 from network.cpp:13:
/usr/include/features.h:131:1: warning: this is the location of the 
previous definition

Since examples work, I guess I am doing something wrong, but I can't 
figure out what. I tried different orderings of the includes in 
property_handle.hpp. I tried #include <boost/python.hpp> first, before 
the standard library includes (like <string> etc.) and I tried including 
first the std-lib includes and including boost/python.hpp afterwards. 
Neither worked.
So which ordering is correct? What do I have to look at?


Thanks!!
   Andreas





From rwgk at yahoo.com  Fri Apr  8 00:29:42 2005
From: rwgk at yahoo.com (Ralf W. Grosse-Kunstleve)
Date: Thu, 7 Apr 2005 15:29:42 -0700 (PDT)
Subject: [C++-sig] "_POSIX_C_SOURCE" redefined
In-Reply-To: 6667
Message-ID: <20050407222942.2709.qmail@web31510.mail.mud.yahoo.com>

--- Andreas Beyer <beyer at imb-jena.de> wrote:
> Since examples work, I guess I am doing something wrong, but I can't 
> figure out what. I tried different orderings of the includes in 
> property_handle.hpp. I tried #include <boost/python.hpp> first, before 
> the standard library includes (like <string> etc.) and I tried including 
> first the std-lib includes and including boost/python.hpp afterwards. 
> Neither worked.
> So which ordering is correct? What do I have to look at?

#include <boost/python.hpp>

should be at the very top of your wrapper code.

Cheers,
        Ralf



		
__________________________________ 
Do you Yahoo!? 
Yahoo! Personals - Better first dates. More second dates. 
http://personals.yahoo.com



From lothar at xcerla.com  Fri Apr  8 00:58:06 2005
From: lothar at xcerla.com (Lothar Werzinger)
Date: Thu, 7 Apr 2005 15:58:06 -0700
Subject: [C++-sig] "_POSIX_C_SOURCE" redefined
In-Reply-To: <20050407222942.2709.qmail@web31510.mail.mud.yahoo.com>
References: <20050407222942.2709.qmail@web31510.mail.mud.yahoo.com>
Message-ID: <200504071558.07154.lothar@xcerla.com>

On Thursday 07 April 2005 15:29, Ralf W. Grosse-Kunstleve wrote:
> --- Andreas Beyer <beyer at imb-jena.de> wrote:
> > Since examples work, I guess I am doing something wrong, but I can't
> > figure out what. I tried different orderings of the includes in
> > property_handle.hpp. I tried #include <boost/python.hpp> first, before
> > the standard library includes (like <string> etc.) and I tried including
> > first the std-lib includes and including boost/python.hpp afterwards.
> > Neither worked.
> > So which ordering is correct? What do I have to look at?
>
> #include <boost/python.hpp>
>
> should be at the very top of your wrapper code.

And even that may not be enough. We use boost::python and log4cplus and 
unfortunately log4cplus also defines some macros that conflict with macros 
previously defined in pythons headers. The only sane way for the probme is 
that the developers of Python (and log4cplus) fix their headers.

/opt2/log4cplus/include/log4cplus/config.h:79:1: warning: "HAVE_STAT" 
redefined
In file included from /opt2/python-2.4/include/python2.4/Python.h:55,
                 
from /opt2/boost/include/boost-1_32/boost/python/detail/wrap_python.hpp:121,
                 
from /opt2/boost/include/boost-1_32/boost/python/detail/prefix.hpp:13,
                 from /opt2/boost/include/boost-1_32/boost/python/args.hpp:8,
                 from /opt2/boost/include/boost-1_32/boost/python.hpp:11,
                 from src/utility/PythonInterpreterGuard.h:10,
                 from src/rating/test/PythonRatingStepTest.cpp:10:
/opt2/python-2.4/include/python2.4/pyport.h:139:1: warning: this is the 
location of the previous definition

Lothar
-- 
Lothar Werzinger Dipl.-Ing. Univ.
framework & platform architect
Xcerla Corporation
275 Tennant Avenue, Suite 202
Morgan Hill, Ca 95037
email: lothar at xcerla.com
phone: +1-408-776-9018


From beyer at imb-jena.de  Fri Apr  8 01:09:38 2005
From: beyer at imb-jena.de (Andreas Beyer)
Date: Thu, 07 Apr 2005 16:09:38 -0700
Subject: [C++-sig] "_POSIX_C_SOURCE" redefined
In-Reply-To: <200504071558.07154.lothar@xcerla.com>
References: <20050407222942.2709.qmail@web31510.mail.mud.yahoo.com>
	<200504071558.07154.lothar@xcerla.com>
Message-ID: <4255BDB2.8030006@imb-jena.de>

Lothar Werzinger wrote:

>On Thursday 07 April 2005 15:29, Ralf W. Grosse-Kunstleve wrote:
>  
>
>>--- Andreas Beyer <beyer at imb-jena.de> wrote:
>>    
>>
>>>Since examples work, I guess I am doing something wrong, but I can't
>>>figure out what. I tried different orderings of the includes in
>>>property_handle.hpp. I tried #include <boost/python.hpp> first, before
>>>the standard library includes (like <string> etc.) and I tried including
>>>first the std-lib includes and including boost/python.hpp afterwards.
>>>Neither worked.
>>>So which ordering is correct? What do I have to look at?
>>>      
>>>
>>#include <boost/python.hpp>
>>
>>should be at the very top of your wrapper code.
>>    
>>
>
>And even that may not be enough. We use boost::python and log4cplus and 
>unfortunately log4cplus also defines some macros that conflict with macros 
>previously defined in pythons headers. The only sane way for the probme is 
>that the developers of Python (and log4cplus) fix their headers.
>
>/opt2/log4cplus/include/log4cplus/config.h:79:1: warning: "HAVE_STAT" 
>redefined
>In file included from /opt2/python-2.4/include/python2.4/Python.h:55,
>                 
>from /opt2/boost/include/boost-1_32/boost/python/detail/wrap_python.hpp:121,
>                 
>from /opt2/boost/include/boost-1_32/boost/python/detail/prefix.hpp:13,
>                 from /opt2/boost/include/boost-1_32/boost/python/args.hpp:8,
>                 from /opt2/boost/include/boost-1_32/boost/python.hpp:11,
>                 from src/utility/PythonInterpreterGuard.h:10,
>                 from src/rating/test/PythonRatingStepTest.cpp:10:
>/opt2/python-2.4/include/python2.4/pyport.h:139:1: warning: this is the 
>location of the previous definition
>
>Lothar
>  
>
Can you post that to the Python developers? I thought it was common 
sense, that include order shouldn't matter.

Andreas



From Marc.Kwiatkowski at veritas.com  Fri Apr  8 01:13:36 2005
From: Marc.Kwiatkowski at veritas.com (Marc Kwiatkowski)
Date: Thu, 7 Apr 2005 16:13:36 -0700
Subject: [C++-sig] "_POSIX_C_SOURCE" redefined
Message-ID: <7AA63464B644CD40A3BAD0AFEF26FA5307ABE8@svlxchcln3.enterprise.veritas.com>

This seems to be a bug introduced in the 2.4 configure.in/pyconfig.h.in logic.  If _XOPEN_SOURCE is defined configure.in will also define _POSIX_C_SOURCE will also be defined and, for most platforms, configure.in does define _XOPEN_SOURCE.  I try to avoid the hell that is autoconf.  As a workaround you might try wrapping the _POSIX_C_SOURCE definition in pyconfig.h with a "#if !defined(_POSIX_C_SOURCE)"  I tried coaxing configure to generate the right settings for RHEL AS3 but no luck.

> -----Original Message-----
> From: c++-sig-bounces at python.org [mailto:c++-sig-bounces at python.org]On
> Behalf Of Lothar Werzinger
> Sent: Thursday, April 07, 2005 15:58
> To: c++-sig at python.org
> Subject: Re: [C++-sig] "_POSIX_C_SOURCE" redefined
> 
> 
> On Thursday 07 April 2005 15:29, Ralf W. Grosse-Kunstleve wrote:
> > --- Andreas Beyer <beyer at imb-jena.de> wrote:
> > > Since examples work, I guess I am doing something wrong, 
> but I can't
> > > figure out what. I tried different orderings of the includes in
> > > property_handle.hpp. I tried #include <boost/python.hpp> 
> first, before
> > > the standard library includes (like <string> etc.) and I 
> tried including
> > > first the std-lib includes and including boost/python.hpp 
> afterwards.
> > > Neither worked.
> > > So which ordering is correct? What do I have to look at?
> >
> > #include <boost/python.hpp>
> >
> > should be at the very top of your wrapper code.
> 
> And even that may not be enough. We use boost::python and 
> log4cplus and 
> unfortunately log4cplus also defines some macros that 
> conflict with macros 
> previously defined in pythons headers. The only sane way for 
> the probme is 
> that the developers of Python (and log4cplus) fix their headers.
> 
> /opt2/log4cplus/include/log4cplus/config.h:79:1: warning: "HAVE_STAT" 
> redefined
> In file included from /opt2/python-2.4/include/python2.4/Python.h:55,
>                  
> from 
> /opt2/boost/include/boost-1_32/boost/python/detail/wrap_python
> .hpp:121,
>                  
> from /opt2/boost/include/boost-1_32/boost/python/detail/prefix.hpp:13,
>                  from 
> /opt2/boost/include/boost-1_32/boost/python/args.hpp:8,
>                  from 
> /opt2/boost/include/boost-1_32/boost/python.hpp:11,
>                  from src/utility/PythonInterpreterGuard.h:10,
>                  from src/rating/test/PythonRatingStepTest.cpp:10:
> /opt2/python-2.4/include/python2.4/pyport.h:139:1: warning: 
> this is the 
> location of the previous definition
> 
> Lothar
> -- 
> Lothar Werzinger Dipl.-Ing. Univ.
> framework & platform architect
> Xcerla Corporation
> 275 Tennant Avenue, Suite 202
> Morgan Hill, Ca 95037
> email: lothar at xcerla.com
> phone: +1-408-776-9018
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
> 


From jgresula at seznam.cz  Fri Apr  8 16:20:06 2005
From: jgresula at seznam.cz (Jaroslav Gresula)
Date: Fri, 08 Apr 2005 16:20:06 +0200
Subject: [C++-sig] [Pyste] template instantiation with a user defined type
Message-ID: <42569316.5060109@seznam.cz>

I have a problem with the following scenario:

// -- file type.h
class Type {};

// -- file template.h
template<class T>
class Templ
{
   T m_data;
};

// -- test.pyste
Type = Class( "Type", "type.h" )
Templ = Template( "Templ", "template.h" )
Templ( "Type" )


When I run Pyste, it internally creates the following temporary file:
// -- ~temporary
#include "c:\absolute\path\to\template.h"

typedef Templ< Type > Templ_Type;
void __instantiate_Templ_Type()
{ sizeof(Templ_Type); }

that is passed to gccxml which issues an error saying 'Type was not 
declared in this scope'. Obviously, type.h header is missing.

Is Pyste capable of handling a template specialized with a user defined 
type?

Thanks,
Jarda


From dave at boost-consulting.com  Fri Apr  8 16:54:14 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Fri, 08 Apr 2005 10:54:14 -0400
Subject: [C++-sig] Re: [Pyste] template instantiation with a user defined
	type
References: <42569316.5060109@seznam.cz>
Message-ID: <u7jjdz5d5.fsf@boost-consulting.com>

Jaroslav Gresula <jgresula at seznam.cz> writes:

> I have a problem with the following scenario:
>
> // -- file type.h
> class Type {};
>
> // -- file template.h
> template<class T>
> class Templ
> {
>   T m_data;
> };
>
> // -- test.pyste
> Type = Class( "Type", "type.h" )
> Templ = Template( "Templ", "template.h" )
> Templ( "Type" )
>
>
> When I run Pyste, it internally creates the following temporary file:
> // -- ~temporary
> #include "c:\absolute\path\to\template.h"
>
> typedef Templ< Type > Templ_Type;
> void __instantiate_Templ_Type()
       ^^

I know this has no relevance to your problem, but all identifiers
containing double underscores are reserved to the C++ implementation.
Pyste should never be generating such names.

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



From mvc at di.fct.unl.pt  Fri Apr  8 17:10:00 2005
From: mvc at di.fct.unl.pt (Marco Correia)
Date: Fri, 8 Apr 2005 16:10:00 +0100
Subject: [C++-sig] inheriting and overloading
Message-ID: <200504081610.00461.mvc@di.fct.unl.pt>

hi,

I'm using an overload to a base class member:

class Base 
{
	void something() {}
};

class Derived : public Base
{
	void something(int i) {}
};

I don't get error by building this with boost::bython, but in python I'm not 
able to use the Base::something member.

Is this a known issue? Should I post here a test case?

thanks
Marco
-- 
Marco Correia <mvc at di.fct.unl.pt>


From paniq at paniq.org  Sun Apr 10 05:00:38 2005
From: paniq at paniq.org (Leonard "paniq" Ritter)
Date: Sun, 10 Apr 2005 05:00:38 +0200
Subject: [C++-sig] wrapper.cpp not in 1.32.0 dsp/vcproj file?
In-Reply-To: <200504081610.00461.mvc@di.fct.unl.pt>
References: <200504081610.00461.mvc@di.fct.unl.pt>
Message-ID: <425896D6.30808@paniq.org>

it seems that wrapper.cpp has been forgotten in the 
boost_1_32_0\libs\python\build\VisualStudio boost_python.dsp and .vcproj 
file... can someone verify this?

i was wondering about a linker error concerning a missing symbol and 
found this to be the cause.



From mike at thewilsonfamily.freeserve.co.uk  Tue Apr 12 05:30:02 2005
From: mike at thewilsonfamily.freeserve.co.uk (Mike)
Date: Tue, 12 Apr 2005 04:30:02 +0100
Subject: [C++-sig] wrapper.cpp not in 1.32.0 dsp/vcproj file?
References: <200504081610.00461.mvc@di.fct.unl.pt> <425896D6.30808@paniq.org>
Message-ID: <001401c53f0f$ea41fe60$597c4e51@MIKESPC01>

I had the same problem (with the .dsp file - but no .vcproj file in 1_32 
as far as I'm aware(?)).  
As you imply, adding wrapper.cpp seems to solve the problem.

Mike
----- Original Message ----- 
From: "Leonard "paniq" Ritter" <paniq at paniq.org>
To: "Development of Python/C++ integration" <c++-sig at python.org>
Sent: Sunday, April 10, 2005 4:00 AM
Subject: [C++-sig] wrapper.cpp not in 1.32.0 dsp/vcproj file?


> it seems that wrapper.cpp has been forgotten in the 
> boost_1_32_0\libs\python\build\VisualStudio boost_python.dsp and .vcproj 
> file... can someone verify this?
> 
> i was wondering about a linker error concerning a missing symbol and 
> found this to be the cause.
> 
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
> 
> 



From paniq at paniq.org  Sun Apr 10 15:55:15 2005
From: paniq at paniq.org (Leonard "paniq" Ritter)
Date: Sun, 10 Apr 2005 15:55:15 +0200
Subject: [C++-sig] wrapper.cpp not in 1.32.0 dsp/vcproj file?
In-Reply-To: <001401c53f0f$ea41fe60$597c4e51@MIKESPC01>
References: <200504081610.00461.mvc@di.fct.unl.pt> <425896D6.30808@paniq.org>
	<001401c53f0f$ea41fe60$597c4e51@MIKESPC01>
Message-ID: <42593043.7010501@paniq.org>

ah yes, i guess i generated the sln/vcproj files in 7.1. so its the dsp 
that misses the file.

Mike wrote:

>I had the same problem (with the .dsp file - but no .vcproj file in 1_32 
>as far as I'm aware(?)).  
>As you imply, adding wrapper.cpp seems to solve the problem.
>
>Mike
>----- Original Message ----- 
>From: "Leonard "paniq" Ritter" <paniq at paniq.org>
>To: "Development of Python/C++ integration" <c++-sig at python.org>
>Sent: Sunday, April 10, 2005 4:00 AM
>Subject: [C++-sig] wrapper.cpp not in 1.32.0 dsp/vcproj file?
>
>
>  
>
>>it seems that wrapper.cpp has been forgotten in the 
>>boost_1_32_0\libs\python\build\VisualStudio boost_python.dsp and .vcproj 
>>file... can someone verify this?
>>
>>i was wondering about a linker error concerning a missing symbol and 
>>found this to be the cause.
>>
>>_______________________________________________
>>C++-sig mailing list
>>C++-sig at python.org
>>http://mail.python.org/mailman/listinfo/c++-sig
>>
>>
>>    
>>
>
>_______________________________________________
>C++-sig mailing list
>C++-sig at python.org
>http://mail.python.org/mailman/listinfo/c++-sig
>
>
>  
>



From dave at boost-consulting.com  Sun Apr 10 15:57:34 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Sun, 10 Apr 2005 09:57:34 -0400
Subject: [C++-sig] Re: wrapper.cpp not in 1.32.0 dsp/vcproj file?
References: <200504081610.00461.mvc@di.fct.unl.pt> <425896D6.30808@paniq.org>
Message-ID: <u7jjavinl.fsf@boost-consulting.com>

"Leonard \"paniq\" Ritter" <paniq at paniq.org> writes:

> it seems that wrapper.cpp has been forgotten in the 
> boost_1_32_0\libs\python\build\VisualStudio boost_python.dsp and .vcproj 
> file... can someone verify this?
>
> i was wondering about a linker error concerning a missing symbol and 
> found this to be the cause.

The maintainer of that VC++ project is listed on the "building" page
of the Boost.Python docs.  You should contact him about this problem
if you want it fixed.

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



From paniq at paniq.org  Mon Apr 11 20:47:17 2005
From: paniq at paniq.org (Leonard "paniq" Ritter)
Date: Mon, 11 Apr 2005 20:47:17 +0200
Subject: [C++-sig] Re: wrapper.cpp not in 1.32.0 dsp/vcproj file?
In-Reply-To: <u7jjavinl.fsf@boost-consulting.com>
References: <200504081610.00461.mvc@di.fct.unl.pt> <425896D6.30808@paniq.org>
	<u7jjavinl.fsf@boost-consulting.com>
Message-ID: <425AC635.4050200@paniq.org>

no i dont want it fixed, i'm perfectly fine with a broken project file.

besides for strengthening the elite and all, there should really be some 
things putting off the newbies.

*ahem*

seriously, i was only trying to verify if it was my or someone elses 
mistake. i contacted brett calcott for you ;)

David Abrahams wrote:

>"Leonard \"paniq\" Ritter" <paniq at paniq.org> writes:
>
>  
>
>>it seems that wrapper.cpp has been forgotten in the 
>>boost_1_32_0\libs\python\build\VisualStudio boost_python.dsp and .vcproj 
>>file... can someone verify this?
>>
>>i was wondering about a linker error concerning a missing symbol and 
>>found this to be the cause.
>>    
>>
>
>The maintainer of that VC++ project is listed on the "building" page
>of the Boost.Python docs.  You should contact him about this problem
>if you want it fixed.
>
>  
>



From nicodemus at esss.com.br  Tue Apr 12 02:19:52 2005
From: nicodemus at esss.com.br (Nicodemus)
Date: Mon, 11 Apr 2005 22:19:52 -0200
Subject: [C++-sig] [Pyste] template instantiation with a user defined type
In-Reply-To: <42569316.5060109@seznam.cz>
References: <42569316.5060109@seznam.cz>
Message-ID: <425B1428.2080908@esss.com.br>

Hi,

Jaroslav Gresula wrote:

> I have a problem with the following scenario:
>
> // -- file type.h
> class Type {};
>
> // -- file template.h
> template<class T>
> class Templ
> {
>   T m_data;
> };
>
> // -- test.pyste
> Type = Class( "Type", "type.h" )
> Templ = Template( "Templ", "template.h" )
> Templ( "Type" )
>
>
> When I run Pyste, it internally creates the following temporary file:
> // -- ~temporary
> #include "c:\absolute\path\to\template.h"
>
> typedef Templ< Type > Templ_Type;
> void __instantiate_Templ_Type()
> { sizeof(Templ_Type); }
>
> that is passed to gccxml which issues an error saying 'Type was not 
> declared in this scope'. Obviously, type.h header is missing.
>
> Is Pyste capable of handling a template specialized with a user 
> defined type?


Not right now, but I don't see why we can't add support for it. I belive 
this would cover all cases:

Templ( "Type", "type.h" )

and

Templ( Type )

I will implement this shortly in CVS.

Best Regards,
Bruno.


From nicodemus at esss.com.br  Tue Apr 12 02:20:48 2005
From: nicodemus at esss.com.br (Nicodemus)
Date: Mon, 11 Apr 2005 22:20:48 -0200
Subject: [C++-sig] Pyste bug? Using Wrapper with member functions and
	--pyste-ns
In-Reply-To: <5ab8740205032002161499b77d@mail.gmail.com>
References: <5ab8740205032002161499b77d@mail.gmail.com>
Message-ID: <425B1460.3090104@esss.com.br>

Hi Dan,

Dan Haffey wrote:

>Hello,
>
>I just noticed that when using the Pyste Wrapper call to create a
>member function wrapper, it places the wrapper definition code outside
>of the specified (--pyste-ns=) namespace, but the generated module
>code looks for it in the namespace. This doesn't seem to happen with
>normal functions, just members.
>  
>

Really sorry for not responding earlier. I somehow missed your message. 
I will fix this bug shortly in CVS.

Best Regards,
Bruno.



From beyer at imb-jena.de  Tue Apr 12 04:45:18 2005
From: beyer at imb-jena.de (Andreas Beyer)
Date: Mon, 11 Apr 2005 19:45:18 -0700
Subject: [C++-sig] Pointer to existing Python object
Message-ID: <425B363E.8090905@imb-jena.de>

Hi:

I know that there has been some discussion on this subject on c++ sig. 
However, I could not find a solution for my specific problem and maybe 
there is no solution.

Here we go:
I have a container class B that creates instances of class A - both in 
C++. In Python I can get instances of A from B with a get() method. I 
would like to get an existing Python object whenever such object existis 
and I would like to get a new Python object if no Python object for A 
exists.

An example:


using namespace boost;

class A : public enable_shared_from_this<A> {
  public:
     A() : val(0) {};
     int val;
     typedef shared_ptr<A> A_ptr;
     A_ptr self() {
        A_ptr self;
        self = shared_from_this();
        return self;
      }

};


class B {
    public:
      B() {
           a = A::A_ptr(new A());
      }
      A::A_ptr a;
      void set(A::A_ptr a) {
          this->a = a;
      }

      A::A_ptr get() {
          return a->self();
      }
};


BOOST_PYTHON_MODULE(ptr_test) {
  python::class_<A> ("A")
    .def_readwrite("val", &A::val)
  ;
  python::register_ptr_to_python< A::A_ptr >();
 
  python::class_<B>("B")
     .def("set", &B::set)
     .def("get", &B::get)
  ;
}


In Python I would like to do this:

 >>> b = B()
 >>> a1 = b.get() # obtain instance of class A
 >>> a2 = b.get() # get the *same Python instance* again
 >>> assert(a1==a2)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
AssertionError
 >>>

The internal representations of instances of A are the same! Whenever I 
access a C++ member of A I am actually accessing the same object.  So I 
can do this:
 >>> a1.val = 42 # change internal reference
 >>> print a2.val
42
 >>>
However, I cannot do this:
 >>> a1.foo = 42
 >>> print a2.foo
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
AttributeError: 'A' object has no attribute 'foo'
 >>>

I think instances of B have to check if there is already an existing 
Python object. In that case  B (or a wrapper of B) should return the 
existing Python object (i.e. instances of python::object). Is there a 
way to find such an existing instance?

Andreas




From colin_irwin at hotmail.com  Tue Apr 12 06:16:09 2005
From: colin_irwin at hotmail.com (Colin Irwin)
Date: Tue, 12 Apr 2005 14:16:09 +1000
Subject: [C++-sig] Re: Wrapper for exception translation
In-Reply-To: <d1r4u1$n3q$1@sea.gmane.org>
References: <2040C0A1CA23D51181A30050BAAC9902011046D8@berexch.ber.haufemg.com>	<d1ofeg$p6$1@sea.gmane.org>	<u1xa77n4g.fsf@boost-consulting.com>
	<d1r4u1$n3q$1@sea.gmane.org>
Message-ID: <d3fhsl$33m$1@sea.gmane.org>

Dave - In the previous post you stated that Daniel Wallin was doing a 
bit of cleaning in the code base for the named parameter library.  Did 
this happen?  Is it now possible to add unnamed parameter support?

Anyway, I think I'm still keen to do whatever on this front to help with 
a wrapper to translate exceptions through into Python.

Colin



From paniq at paniq.org  Wed Apr 13 00:38:33 2005
From: paniq at paniq.org (Leonard "paniq" Ritter)
Date: Wed, 13 Apr 2005 00:38:33 +0200
Subject: [C++-sig] Re: wrapper.cpp not in 1.32.0 dsp/vcproj file?
In-Reply-To: <425AC635.4050200@paniq.org>
References: <200504081610.00461.mvc@di.fct.unl.pt>
	<425896D6.30808@paniq.org>	<u7jjavinl.fsf@boost-consulting.com>
	<425AC635.4050200@paniq.org>
Message-ID: <425C4DE9.4020205@paniq.org>

fyi, the mail to brett calcott bounced.

Leonard "paniq" Ritter wrote:

> no i dont want it fixed, i'm perfectly fine with a broken project file.
>
> besides for strengthening the elite and all, there should really be 
> some things putting off the newbies.
>
> *ahem*
>
> seriously, i was only trying to verify if it was my or someone elses 
> mistake. i contacted brett calcott for you ;)
>
> David Abrahams wrote:
>
>> "Leonard \"paniq\" Ritter" <paniq at paniq.org> writes:
>>
>>  
>>
>>> it seems that wrapper.cpp has been forgotten in the 
>>> boost_1_32_0\libs\python\build\VisualStudio boost_python.dsp and 
>>> .vcproj file... can someone verify this?
>>>
>>> i was wondering about a linker error concerning a missing symbol and 
>>> found this to be the cause.
>>>   
>>
>>
>> The maintainer of that VC++ project is listed on the "building" page
>> of the Boost.Python docs.  You should contact him about this problem
>> if you want it fixed.
>>
>>  
>>
>
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
>
>



From leejd80 at gmail.com  Wed Apr 13 15:40:26 2005
From: leejd80 at gmail.com (Gerald Lee)
Date: Wed, 13 Apr 2005 21:40:26 +0800
Subject: [C++-sig] Linker Error
Message-ID: <7dbf822f05041306403f3e728d@mail.gmail.com>

I will Embedded Python to my Application.The C++ compiler is BCB6's.
I compiled the a simple code, OK, but error occured when link.
The link error message followed:
Build
[Linker Error] Unresolved external '__Py_RefTotal' referenced from 
E:\PROJECT\GRIDCAM\SOURCE\IOSERVICES\NEWMODULE.OBJ
[Linker Error] Unresolved external '__Py_NegativeRefcount' referenced from 
E:\PROJECT\GRIDCAM\SOURCE\IOSERVICES\NEWMODULE.OBJ
[Linker Error] Unresolved external '__Py_Dealloc' referenced from 
E:\PROJECT\GRIDCAM\SOURCE\IOSERVICES\NEWMODULE.OBJ

I already convert python23.lib to omf format. How to pass it?

Best regard,
Gerald Lee

-- 
My Blog >> http://leejd.cndev.org
My QQ >> 9847243
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20050413/985832d3/attachment.htm>

From domma at procoders.net  Wed Apr 13 15:52:19 2005
From: domma at procoders.net (Achim Domma (Procoders))
Date: Wed, 13 Apr 2005 15:52:19 +0200
Subject: [C++-sig] Initialize SparseVector
Message-ID: <425D2413.40300@procoders.net>

Hi,

I have wrapped a sparse vector class, which currently can be used like this:

v=svm.vector()
v[3]=7
v[9]=-3.2
v[5]=0.3
...

Everything works finde so far, but I want to implement a more pythonic 
way to initialize the vector. Something like this would be nice:

v=svm.vector( (3,7), (9,-3.2), (5,0.3), ... )

What's the easiest way to get something like this with boost.python? I 
have control over the complete code, so small modifications on the 
sparse vector class would be no problem.

regards,
Achim


From nicodemus at esss.com.br  Wed Apr 13 16:27:39 2005
From: nicodemus at esss.com.br (Nicodemus)
Date: Wed, 13 Apr 2005 11:27:39 -0300
Subject: [C++-sig] Initialize SparseVector
In-Reply-To: <425D2413.40300@procoders.net>
References: <425D2413.40300@procoders.net>
Message-ID: <425D2C5B.3010702@esss.com.br>

Hi Achim,

Take a look at make_constructor. I don't have time right now to provide 
an example, but this should get you going.

Regards,
Bruno.

Achim Domma (Procoders) wrote:

> Hi,
>
> I have wrapped a sparse vector class, which currently can be used like 
> this:
>
> v=svm.vector()
> v[3]=7
> v[9]=-3.2
> v[5]=0.3
> ...
>
> Everything works finde so far, but I want to implement a more pythonic 
> way to initialize the vector. Something like this would be nice:
>
> v=svm.vector( (3,7), (9,-3.2), (5,0.3), ... )
>
> What's the easiest way to get something like this with boost.python? I 
> have control over the complete code, so small modifications on the 
> sparse vector class would be no problem.
>
> regards,
> Achim
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
>
>



From jgresula at seznam.cz  Wed Apr 13 23:11:59 2005
From: jgresula at seznam.cz (Jaroslav Gresula)
Date: Wed, 13 Apr 2005 23:11:59 +0200
Subject: [C++-sig] Pyste enhancements - mini patch
Message-ID: <425D8B1F.7090709@seznam.cz>

Hi,

I extended my local copy of Pyste by the following two enhancements:

i) a class can be declared as final:
  C = Class( 'C', 'c.h' )
  final( C )

ii) registration of converters for user-defined smart pointers:
  C = Class( 'C', 'c.h' )
  Include( "my_smart_pointers.h" )
  use_smart_ptr( C, "my_shared_ptr" )
  use_smart_ptr( C, "my_auto_ptr" )


I thought that also someone else might find it useful. If anyone is 
interested, the patch is attached.

Cheers,
Jarda
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pyste-minipatch.zip
Type: application/x-zip-compressed
Size: 1398 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20050413/034c5c72/attachment.bin>

From beyer at imb-jena.de  Wed Apr 13 23:14:44 2005
From: beyer at imb-jena.de (Andreas Beyer)
Date: Wed, 13 Apr 2005 14:14:44 -0700
Subject: [C++-sig] shared_ptr -> weak_ptr -> shared_ptr
Message-ID: <425D8BC4.50905@imb-jena.de>

Hi.

I would like to transform a shared pointer originating from a python 
object to a weak pointer. Later I would like to transform it back to a 
shared pointer. For some strange reason I keep getting seg-faults when 
trying to lock the weak pointer.
Is there any danger in storing shared pointers from python objects as 
weak pointers?
According to the weak pointer's docu, lock() returns an empty shared_ptr 
if the refered object expired. However, in my example this shouldn't be 
the case anyway.

The test case:

using namespace boost;

struct A { };

typedef shared_ptr<A> A_ptr_shared;
typedef weak_ptr<A> A_ptr_weak;

void transform(A_ptr_shared a) {
   std::cerr << a.get() << std::endl;
   A_ptr_weak aw(a);
   std::cerr << "have weak" << std::endl;
   A_ptr_shared as = aw.lock(); // boom!
   std::cerr << "lock is ok" << std::endl;
   std::cerr << as.get() << std::endl;
}

BOOST_PYTHON_MODULE(ptr_test) {
  python::class_<A> ("A");
  python::register_ptr_to_python< A_ptr_shared >();

  python::def("transform", &transform);
}


The python session:

 >>> from ptr_test import *
 >>> a=A()
 >>> transform(a)
0x8190a58
have weak
Segmentation fault (core dumped)




From cppsig-eric at sneakemail.com  Thu Apr 14 03:11:33 2005
From: cppsig-eric at sneakemail.com (Eric)
Date: Wed, 13 Apr 2005 18:11:33 -0700
Subject: [C++-sig] serializing threaded access to Boost.Python-wrapped class
In-Reply-To: <425D8B1F.7090709@seznam.cz>
Message-ID: <10212-54350@sneakemail.com>

I have a C++ class, accessed from python via Boost.Python, that I am trying
to prevent from being reentered in the context of a multithreaded program. I
can do this relatively easily with the RLock class in the threading module.
However this requires me to track down every caller of this class and wrap
all calls in an acquire()/release() pair. I'm worried that it would be
pretty easy to miss one of these and then cause hard-to-find bugs.

Is there any way to do this automatically with Boost.Python?



From gritsch at iue.tuwien.ac.at  Thu Apr 14 09:27:07 2005
From: gritsch at iue.tuwien.ac.at (gritsch at iue.tuwien.ac.at)
Date: Thu, 14 Apr 2005 09:27:07 +0200
Subject: [C++-sig] Re: Wrapper for exception translation
Message-ID: <1113463627.425e1b4b2b242@l01.iue.tuwien.ac.at>

Hi,

I am also looking forward to some nice mechanism to integrate custom exceptions
nicely with Boost.Python.  However, in the meantime I use the following
approach, and I wonder, if this is currently the correct way:

--- 8< ---

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

struct MyException {
   std::string what() const {
        return "MyException occurred";
    }
};

void throw_it() {
    throw MyException();
}

static PyObject * my_exception = NULL;

void translator( MyException const & e ) {
    PyErr_SetString( my_exception, e.what().c_str() );
}

BOOST_PYTHON_MODULE( extest ) {
    my_exception = PyErr_NewException( "extest.MyException", NULL, NULL );
    PyModule_AddObject( boost::python::scope().ptr(), "MyException",
my_exception );
    boost::python::register_exception_translator<MyException>( translator );

    boost::python::def( "throw_it", throw_it );
}

--- 8< ---

Do I really have to use the Python C-API funcions
  PyErr_NewException and
 PyModule_AddObject



----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.


From gritsch at iue.tuwien.ac.at  Thu Apr 14 09:33:22 2005
From: gritsch at iue.tuwien.ac.at (gritsch at iue.tuwien.ac.at)
Date: Thu, 14 Apr 2005 09:33:22 +0200
Subject: [C++-sig] Re: Wrapper for exception translation
Message-ID: <1113464002.425e1cc257eb1@l01.iue.tuwien.ac.at>

Hi,

I am also looking forward to some mechanism to integrate custom exceptions
nicely with Boost.Python.  However, in the meantime I use the following
approach, and I wonder, if this is currently the correct way:

--- 8< ---

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

struct MyException {
   std::string what() const {
        return "MyException occurred";
    }
};

void throw_it() {
    throw MyException();
}

static PyObject * my_exception = NULL;

void translator( MyException const & e ) {
    PyErr_SetString( my_exception, e.what().c_str() );
}

BOOST_PYTHON_MODULE( extest ) {
    my_exception = PyErr_NewException( "extest.MyException", NULL, NULL );
    PyModule_AddObject( boost::python::scope().ptr(), "MyException",
my_exception );
    boost::python::register_exception_translator<MyException>( translator );

    boost::python::def( "throw_it", throw_it );
}

--- 8< ---

Do I really have to use the Python C-API funcions PyErr_NewException and
PyModule_AddObject?

The module can be used from Python as

--- 8< ---

import extest

try:
    extest.throw_it()
except extest.MyException, msg:
    print msg

--- 8< ---

Thank you for reading, and please excuse the accidentally sended incomplete
email,
Markus


----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.


From dave at boost-consulting.com  Thu Apr 14 09:48:35 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 14 Apr 2005 09:48:35 +0200
Subject: [C++-sig] Re: wrapper.cpp not in 1.32.0 dsp/vcproj file?
References: <200504081610.00461.mvc@di.fct.unl.pt> <425896D6.30808@paniq.org>
	<u7jjavinl.fsf@boost-consulting.com> <425AC635.4050200@paniq.org>
	<425C4DE9.4020205@paniq.org>
Message-ID: <ur7hdpzn0.fsf@boost-consulting.com>

"Leonard \"paniq\" Ritter" <paniq at paniq.org> writes:

> fyi, the mail to brett calcott bounced.

What a shame.

> Leonard "paniq" Ritter wrote:
>
>> no i dont want it fixed, i'm perfectly fine with a broken project file.
>>
>> besides for strengthening the elite and all, there should really be
>> some things putting off the newbies.
>>
>> *ahem*
>>
>> seriously, i was only trying to verify if it was my or someone elses
>> mistake. i contacted brett calcott for you ;)

It's not "for me."  I told Brett I'd accept the project file if he
would maintain it ongoingly, because I can't afford to test and
maintain it.  Unless Brett shows up or someone else volunteers I'm
going to remove it from the CVS for the next release.

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



From dave at boost-consulting.com  Thu Apr 14 09:51:54 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 14 Apr 2005 09:51:54 +0200
Subject: [C++-sig] Re: Wrapper for exception translation
References: <2040C0A1CA23D51181A30050BAAC9902011046D8@berexch.ber.haufemg.com>
	<d1ofeg$p6$1@sea.gmane.org> <u1xa77n4g.fsf@boost-consulting.com>
	<d1r4u1$n3q$1@sea.gmane.org> <d3fhsl$33m$1@sea.gmane.org>
Message-ID: <ull7lpzhh.fsf@boost-consulting.com>

Colin Irwin <colin_irwin at hotmail.com> writes:

> Dave - In the previous post you stated that Daniel Wallin was doing a 
> bit of cleaning in the code base for the named parameter library.  Did 
> this happen?  

Yes, Daniel and I have done that cleanup.

> Is it now possible to add unnamed parameter support?

Yep, we're starting on it now.

> Anyway, I think I'm still keen to do whatever on this front to help
> with a wrapper to translate exceptions through into Python.

Thanks; I'll keep it in mind.

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



From dave at boost-consulting.com  Thu Apr 14 09:55:59 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 14 Apr 2005 09:55:59 +0200
Subject: [C++-sig] Re: shared_ptr -> weak_ptr -> shared_ptr
References: <425D8BC4.50905@imb-jena.de>
Message-ID: <ud5sxpzao.fsf@boost-consulting.com>

Andreas Beyer <beyer at imb-jena.de> writes:

> Hi.
>
> I would like to transform a shared pointer originating from a python 
> object to a weak pointer. Later I would like to transform it back to a 
> shared pointer. For some strange reason I keep getting seg-faults when 
> trying to lock the weak pointer.
> Is there any danger in storing shared pointers from python objects as 
> weak pointers?

Shouldn't be, especially in your case.  This behavior really surprises
me.  Peter, do you have anything to say about this?

> According to the weak pointer's docu, lock() returns an empty shared_ptr 
> if the refered object expired. However, in my example this shouldn't be 
> the case anyway.
>
> The test case:
>
> using namespace boost;
>
> struct A { };
>
> typedef shared_ptr<A> A_ptr_shared;
> typedef weak_ptr<A> A_ptr_weak;
>
> void transform(A_ptr_shared a) {
>   std::cerr << a.get() << std::endl;
>   A_ptr_weak aw(a);
>   std::cerr << "have weak" << std::endl;
>   A_ptr_shared as = aw.lock(); // boom!
>   std::cerr << "lock is ok" << std::endl;
>   std::cerr << as.get() << std::endl;
> }
>
> BOOST_PYTHON_MODULE(ptr_test) {
>  python::class_<A> ("A");
>  python::register_ptr_to_python< A_ptr_shared >();
>
>  python::def("transform", &transform);
> }
>
>
> The python session:
>
>>>> from ptr_test import *
>>>> a=A()
>>>> transform(a)
> 0x8190a58
> have weak
> Segmentation fault (core dumped)

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



From ghost at cs.msu.su  Thu Apr 14 11:32:26 2005
From: ghost at cs.msu.su (Vladimir Prus)
Date: Thu, 14 Apr 2005 13:32:26 +0400
Subject: [C++-sig] Re: serializing threaded access to Boost.Python-wrapped
	class
References: <425D8B1F.7090709@seznam.cz> <10212-54350@sneakemail.com>
Message-ID: <d3ld51$rt6$1@sea.gmane.org>

Eric wrote:

> I have a C++ class, accessed from python via Boost.Python, that I am
> trying to prevent from being reentered in the context of a multithreaded
> program. I can do this relatively easily with the RLock class in the
> threading module. However this requires me to track down every caller of
> this class and wrap all calls in an acquire()/release() pair. I'm worried
> that it would be pretty easy to miss one of these and then cause
> hard-to-find bugs.
> 
> Is there any way to do this automatically with Boost.Python?

There's a Python-only solution -- after importing a class, you can walk over
all its methods, and "decorate" them, adding any pre/post processing you
like. The attached example does this with the 'hello' class from
'example/getting_started2' example.

This is much easier that tracking down every caller -- this only requires to
to call a function for each imported *class*.

- Volodya


-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: pre_post.py
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20050414/7fb0cb8f/attachment.txt>

From pdimov at mmltd.net  Thu Apr 14 11:42:11 2005
From: pdimov at mmltd.net (Peter Dimov)
Date: Thu, 14 Apr 2005 12:42:11 +0300
Subject: [C++-sig] Re: shared_ptr -> weak_ptr -> shared_ptr
References: <425D8BC4.50905@imb-jena.de> <ud5sxpzao.fsf@boost-consulting.com>
Message-ID: <d3ldn8$ug1$1@sea.gmane.org>

David Abrahams wrote:
> Andreas Beyer <beyer at imb-jena.de> writes:
>
>> Hi.
>>
>> I would like to transform a shared pointer originating from a python
>> object to a weak pointer. Later I would like to transform it back to
>> a shared pointer. For some strange reason I keep getting seg-faults
>> when trying to lock the weak pointer.
>> Is there any danger in storing shared pointers from python objects as
>> weak pointers?
>
> Shouldn't be, especially in your case.  This behavior really surprises
> me.  Peter, do you have anything to say about this?

It should work. Is it possible that Boost.Python is single-threaded but 
ptr_test is mutlithreaded? Which platform? 





From dave at boost-consulting.com  Thu Apr 14 09:48:35 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 14 Apr 2005 09:48:35 +0200
Subject: [C++-sig] Re: wrapper.cpp not in 1.32.0 dsp/vcproj file?
In-Reply-To: <425C4DE9.4020205@paniq.org> (Leonard Ritter's message of "Wed,
	13 Apr 2005 00:38:33 +0200")
References: <200504081610.00461.mvc@di.fct.unl.pt> <425896D6.30808@paniq.org>
	<u7jjavinl.fsf@boost-consulting.com> <425AC635.4050200@paniq.org>
	<425C4DE9.4020205@paniq.org>
Message-ID: <ur7hdpzn0.fsf@boost-consulting.com>

"Leonard \"paniq\" Ritter" <paniq at paniq.org> writes:

> fyi, the mail to brett calcott bounced.

What a shame.

> Leonard "paniq" Ritter wrote:
>
>> no i dont want it fixed, i'm perfectly fine with a broken project file.
>>
>> besides for strengthening the elite and all, there should really be
>> some things putting off the newbies.
>>
>> *ahem*
>>
>> seriously, i was only trying to verify if it was my or someone elses
>> mistake. i contacted brett calcott for you ;)

It's not "for me."  I told Brett I'd accept the project file if he
would maintain it ongoingly, because I can't afford to test and
maintain it.  Unless Brett shows up or someone else volunteers I'm
going to remove it from the CVS for the next release.

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



From beyer at imb-jena.de  Thu Apr 14 18:16:12 2005
From: beyer at imb-jena.de (Andreas Beyer)
Date: Thu, 14 Apr 2005 09:16:12 -0700
Subject: [C++-sig] Re: shared_ptr -> weak_ptr -> shared_ptr
In-Reply-To: <d3ldn8$ug1$1@sea.gmane.org>
References: <425D8BC4.50905@imb-jena.de> <ud5sxpzao.fsf@boost-consulting.com>
	<d3ldn8$ug1$1@sea.gmane.org>
Message-ID: <425E974C.20103@imb-jena.de>

My platform is
  Python 2.4.1 (#1, Apr  8 2005, 15:53:09)
  gcc (GCC) 3.2.3 20030502 (Red Hat Linux 3.2.3-20)
  boost cvs download from Apr. 5th

The program runs single-threaded. There is nothing else as what I gave 
in my example. I didn't even use Idle for the python part, but just a 
plain python shell.

Can someone else try to reproduce this behavior on another Linux platform?
David, are there other ways to obtain a weak pointer from a python object?

Thanks!
    Andreas

Peter Dimov wrote:

>David Abrahams wrote:
>  
>
>>Andreas Beyer <beyer at imb-jena.de> writes:
>>
>>    
>>
>>>Hi.
>>>
>>>I would like to transform a shared pointer originating from a python
>>>object to a weak pointer. Later I would like to transform it back to
>>>a shared pointer. For some strange reason I keep getting seg-faults
>>>when trying to lock the weak pointer.
>>>Is there any danger in storing shared pointers from python objects as
>>>weak pointers?
>>>      
>>>
>>Shouldn't be, especially in your case.  This behavior really surprises
>>me.  Peter, do you have anything to say about this?
>>    
>>
>
>It should work. Is it possible that Boost.Python is single-threaded but 
>ptr_test is mutlithreaded? Which platform? 
>
>
>
>_______________________________________________
>C++-sig mailing list
>C++-sig at python.org
>http://mail.python.org/mailman/listinfo/c++-sig
>
>  
>



From beyer at imb-jena.de  Thu Apr 14 19:02:05 2005
From: beyer at imb-jena.de (Andreas Beyer)
Date: Thu, 14 Apr 2005 10:02:05 -0700
Subject: [C++-sig] Pointer to existing Python object
Message-ID: <425EA20D.5050700@imb-jena.de>

Hi,

I just want to bring this issue back on the table. I still didn't get a 
brilliant idea how to solve it.

So far, I see only this option:
- Wrap B, in particular wrap B.get().
- In get() create the actual python object, like:
    python::object py_a(B::get());
- Store those objects in a map.
- When B_wrap.get() gets called again, check if a python object already 
exists and return it.

However, in my real-world application I had to wrap quite a lot of 
methods returning As. I also have methods returning vectors of A_ptr, 
which makes it even more enoying. (I would have to wrap the 
vector<A_ptr>, too.)

What is this has_back_reference doing? Is that somehow related to my 
problem? (I regret that I don't really understand what it is doing, as I 
don't know what a "predicate metafunction" is.)
Just adopting the example at 
http://www.boost.org/libs/python/doc/v2/has_back_reference.html is not 
really a solution, because my instances of A are created inside the C++ 
code (by factory methods in class B). I would have to be very intrusive 
and make the actual class B aware of Python/boost::python for adopting 
the above solution.

I know that boost::python is collecting the shared pointers somewhere. 
Is it possible to use that mechanism somewhow for solving my problem? 
Any ideas?

Thanks!!
    Andreas



From cludwig at cdc.informatik.tu-darmstadt.de  Thu Apr 14 19:18:16 2005
From: cludwig at cdc.informatik.tu-darmstadt.de (Christoph Ludwig)
Date: Thu, 14 Apr 2005 19:18:16 +0200
Subject: [C++-sig] Re: shared_ptr -> weak_ptr -> shared_ptr
In-Reply-To: <425E974C.20103@imb-jena.de>
References: <425D8BC4.50905@imb-jena.de> <ud5sxpzao.fsf@boost-consulting.com>
	<d3ldn8$ug1$1@sea.gmane.org> <425E974C.20103@imb-jena.de>
Message-ID: <20050414171816.GE2729@lap200.cdc.informatik.tu-darmstadt.de>

On Thu, Apr 14, 2005 at 09:16:12AM -0700, Andreas Beyer wrote:
> My platform is
>  Python 2.4.1 (#1, Apr  8 2005, 15:53:09)
>  gcc (GCC) 3.2.3 20030502 (Red Hat Linux 3.2.3-20)
>  boost cvs download from Apr. 5th
> 
> The program runs single-threaded. There is nothing else as what I gave 
> in my example. I didn't even use Idle for the python part, but just a 
> plain python shell.

Have you BOOST_DISABLE_THREADS defined in boost/config/user.hpp or on the
compiler's command line? 

The Boost config system cannot reliable detect under Linux whether the
compiler was called with the option -pthread. IIRC it assumes the code is MT
unless BOOST_DISABLE_THREADS was defined. So your python interpreter may be
single-threaded and you compiled and linked your test program without
-pthread, but the shared_ptr code still refers to pthread symbols unless
you defined BOOST_DISABLE_THREADS. (For some reason such mismatches are not
caught by the linker.)

I experienced the same problem some time ago and spent many hours with the
debugger before I found the problem's cause. The symptoms were similar to what
you report.

> Can someone else try to reproduce this behavior on another Linux platform?

I can't reproduce it, but then I have no single-threaded python interpreter
installed anymore. So even if the shared_ptr code refers to pthread symbols
they can be resolved when the extension module is loaded.


HTH, regards

Christoph

> Peter Dimov wrote:
> 
> >David Abrahams wrote:
> > 
> >
> >>Andreas Beyer <beyer at imb-jena.de> writes:
> >>
> >>   
> >>
> >>>Hi.
> >>>
> >>>I would like to transform a shared pointer originating from a python
> >>>object to a weak pointer. Later I would like to transform it back to
> >>>a shared pointer. For some strange reason I keep getting seg-faults
> >>>when trying to lock the weak pointer.
> >>>Is there any danger in storing shared pointers from python objects as
> >>>weak pointers?
> >>>     
> >>>
> >>Shouldn't be, especially in your case.  This behavior really surprises
> >>me.  Peter, do you have anything to say about this?
> >>   
> >>
> >
> >It should work. Is it possible that Boost.Python is single-threaded but 
> >ptr_test is mutlithreaded? Which platform? 
> >
> >
> >
> >_______________________________________________
> >C++-sig mailing list
> >C++-sig at python.org
> >http://mail.python.org/mailman/listinfo/c++-sig
> >
> > 
> >
> 
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
> 
> 

-- 
http://www.informatik.tu-darmstadt.de/TI/Mitarbeiter/cludwig.html
LiDIA: http://www.informatik.tu-darmstadt.de/TI/LiDIA/Welcome.html



From beyer at imb-jena.de  Thu Apr 14 23:52:59 2005
From: beyer at imb-jena.de (Andreas Beyer)
Date: Thu, 14 Apr 2005 14:52:59 -0700
Subject: [C++-sig] Re: shared_ptr -> weak_ptr -> shared_ptr
In-Reply-To: <20050414171816.GE2729@lap200.cdc.informatik.tu-darmstadt.de>
References: <425D8BC4.50905@imb-jena.de>
	<ud5sxpzao.fsf@boost-consulting.com>	<d3ldn8$ug1$1@sea.gmane.org>
	<425E974C.20103@imb-jena.de>
	<20050414171816.GE2729@lap200.cdc.informatik.tu-darmstadt.de>
Message-ID: <425EE63B.4010705@imb-jena.de>

Unfortunately that doesn't help. I called bjam like this:
bjam -sBUILD=release -sBOOST_DISABLE_THREADS  -q
and I get the same problem.
I then tried to compile the debug version and the story gets even worse: 
I get an internal comiler error:

bjam -sBUILD=debug -sBOOST_DISABLE_THREADS  -q
[...]
gcc-C++-action 
bin/csrc/ptr_test.so/gcc/debug/shared-linkable-true/ptr_test.o
/cellar/users/abeyer/boost/boost_current/boost/detail/sp_counted_base_gcc_x86.hpp: 
In
   function `long int boost::detail::atomic_conditional_increment(long 
int*)':
/cellar/users/abeyer/boost/boost_current/boost/detail/sp_counted_base_gcc_x86.hpp:94: 
Internal
   compiler error in instantiate_virtual_regs_1, at function.c:3974
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://bugzilla.redhat.com/bugzilla/> for instructions.
 
    set -e
    "g++"   -c -Wall -ftemplate-depth-255  -DBOOST_PYTHON_DYNAMIC_LIB  
-g -O0 -fno-inline -fPIC   -I"bin/csrc"   -I 
"/cellar/users/abeyer//include/python2.4" -I 
"/cellar/users/abeyer/boost/boost_current"  -o 
"bin/csrc/ptr_test.so/gcc/debug/shared-linkable-true/ptr_test.o"  
"ptr_test.cpp"
    "/usr/bin/objcopy" --set-section-flags .debug_str=contents,debug 
"bin/csrc/ptr_test.so/gcc/debug/shared-linkable-true/ptr_test.o"
 
This error is caused by the same line of code that causes the seg-fault 
in the release version. If I comment out the following line of code, the 
compiler works fine:
   A_ptr_shared as = aw.lock(); // boom!

I would assume that many people work with Red Hat's gcc 3.2.3. I tried 
gcc 3.2.3-20 and 3.2.3-49; both crash in the respective line of code.

Christoph Ludwig wrote:

>On Thu, Apr 14, 2005 at 09:16:12AM -0700, Andreas Beyer wrote:
>  
>
>>My platform is
>> Python 2.4.1 (#1, Apr  8 2005, 15:53:09)
>> gcc (GCC) 3.2.3 20030502 (Red Hat Linux 3.2.3-20)
>> boost cvs download from Apr. 5th
>>
>>The program runs single-threaded. There is nothing else as what I gave 
>>in my example. I didn't even use Idle for the python part, but just a 
>>plain python shell.
>>    
>>
>
>Have you BOOST_DISABLE_THREADS defined in boost/config/user.hpp or on the
>compiler's command line? 
>
>The Boost config system cannot reliable detect under Linux whether the
>compiler was called with the option -pthread. IIRC it assumes the code is MT
>unless BOOST_DISABLE_THREADS was defined. So your python interpreter may be
>single-threaded and you compiled and linked your test program without
>-pthread, but the shared_ptr code still refers to pthread symbols unless
>you defined BOOST_DISABLE_THREADS. (For some reason such mismatches are not
>caught by the linker.)
>
>I experienced the same problem some time ago and spent many hours with the
>debugger before I found the problem's cause. The symptoms were similar to what
>you report.
>
>  
>
>>Can someone else try to reproduce this behavior on another Linux platform?
>>    
>>
>
>I can't reproduce it, but then I have no single-threaded python interpreter
>installed anymore. So even if the shared_ptr code refers to pthread symbols
>they can be resolved when the extension module is loaded.
>
>
>HTH, regards
>
>Christoph
>
>  
>
>>Peter Dimov wrote:
>>
>>    
>>
>>>David Abrahams wrote:
>>>
>>>
>>>      
>>>
>>>>Andreas Beyer <beyer at imb-jena.de> writes:
>>>>
>>>>  
>>>>
>>>>        
>>>>
>>>>>Hi.
>>>>>
>>>>>I would like to transform a shared pointer originating from a python
>>>>>object to a weak pointer. Later I would like to transform it back to
>>>>>a shared pointer. For some strange reason I keep getting seg-faults
>>>>>when trying to lock the weak pointer.
>>>>>Is there any danger in storing shared pointers from python objects as
>>>>>weak pointers?
>>>>>    
>>>>>
>>>>>          
>>>>>
>>>>Shouldn't be, especially in your case.  This behavior really surprises
>>>>me.  Peter, do you have anything to say about this?
>>>>  
>>>>
>>>>        
>>>>
>>>It should work. Is it possible that Boost.Python is single-threaded but 
>>>ptr_test is mutlithreaded? Which platform? 
>>>
>>>
>>>
>>>
>>>      
>>>



From pdimov at mmltd.net  Fri Apr 15 00:10:37 2005
From: pdimov at mmltd.net (Peter Dimov)
Date: Fri, 15 Apr 2005 01:10:37 +0300
Subject: [C++-sig] Re: Re: shared_ptr -> weak_ptr -> shared_ptr
References: <425D8BC4.50905@imb-jena.de><ud5sxpzao.fsf@boost-consulting.com>	<d3ldn8$ug1$1@sea.gmane.org><425E974C.20103@imb-jena.de><20050414171816.GE2729@lap200.cdc.informatik.tu-darmstadt.de>
	<425EE63B.4010705@imb-jena.de>
Message-ID: <d3mpi5$r6o$1@sea.gmane.org>

Andreas Beyer wrote:
> Unfortunately that doesn't help. I called bjam like this:
> bjam -sBUILD=release -sBOOST_DISABLE_THREADS  -q
> and I get the same problem.
> I then tried to compile the debug version and the story gets even
> worse: I get an internal comiler error:
>
> bjam -sBUILD=debug -sBOOST_DISABLE_THREADS  -q
> [...]
> gcc-C++-action
> bin/csrc/ptr_test.so/gcc/debug/shared-linkable-true/ptr_test.o
> /cellar/users/abeyer/boost/boost_current/boost/detail/sp_counted_base_gcc_x86.hpp:
> In
>   function `long int boost::detail::atomic_conditional_increment(long
> int*)':
> /cellar/users/abeyer/boost/boost_current/boost/detail/sp_counted_base_gcc_x86.hpp:94:
> Internal
>   compiler error in instantiate_virtual_regs_1, at function.c:3974
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See <URL:http://bugzilla.redhat.com/bugzilla/> for instructions.

Please try the latest CVS version, where this problem has been fixed. 
Hopefully it will take care of your other issue as well. Please let us know 
whether this solves the problem. 





From beyer at imb-jena.de  Fri Apr 15 02:20:22 2005
From: beyer at imb-jena.de (Andreas Beyer)
Date: Thu, 14 Apr 2005 17:20:22 -0700
Subject: [C++-sig] Re: Re: shared_ptr -> weak_ptr -> shared_ptr
In-Reply-To: <d3mpi5$r6o$1@sea.gmane.org>
References: <425D8BC4.50905@imb-jena.de><ud5sxpzao.fsf@boost-consulting.com>	<d3ldn8$ug1$1@sea.gmane.org><425E974C.20103@imb-jena.de><20050414171816.GE2729@lap200.cdc.informatik.tu-darmstadt.de>	<425EE63B.4010705@imb-jena.de>
	<d3mpi5$r6o$1@sea.gmane.org>
Message-ID: <425F08C6.6060503@imb-jena.de>

Peter Dimov wrote:

>Andreas Beyer wrote:
>  
>
>>Unfortunately that doesn't help. I called bjam like this:
>>bjam -sBUILD=release -sBOOST_DISABLE_THREADS  -q
>>and I get the same problem.
>>I then tried to compile the debug version and the story gets even
>>worse: I get an internal comiler error:
>>
>>bjam -sBUILD=debug -sBOOST_DISABLE_THREADS  -q
>>[...]
>>gcc-C++-action
>>bin/csrc/ptr_test.so/gcc/debug/shared-linkable-true/ptr_test.o
>>/cellar/users/abeyer/boost/boost_current/boost/detail/sp_counted_base_gcc_x86.hpp:
>>In
>>  function `long int boost::detail::atomic_conditional_increment(long
>>int*)':
>>/cellar/users/abeyer/boost/boost_current/boost/detail/sp_counted_base_gcc_x86.hpp:94:
>>Internal
>>  compiler error in instantiate_virtual_regs_1, at function.c:3974
>>Please submit a full bug report,
>>with preprocessed source if appropriate.
>>See <URL:http://bugzilla.redhat.com/bugzilla/> for instructions.
>>    
>>
>
>Please try the latest CVS version, where this problem has been fixed. 
>Hopefully it will take care of your other issue as well. Please let us know 
>whether this solves the problem. 
>
>  
>
That's it! Now it works even without the -sBOOST_DISABLE_THREADS flag.
Both problems are gone.

Peter, thanks a lot!!

Andreas




From paniq at paniq.org  Sun Apr 17 15:01:40 2005
From: paniq at paniq.org (Leonard "paniq" Ritter)
Date: Sun, 17 Apr 2005 15:01:40 +0200
Subject: [C++-sig] virtual overrides and custom smart pointers
Message-ID: <42625E34.7000408@paniq.org>

hi,

i'm running into a problem i'm trying to fix since several days without 
much luck. it seems my custom smart pointer class becomes worthless when 
wrapper<> is used to construct an object whose methods can be overridden 
in python.

wrapper export code:

            class_<WindowEventsWrap, bases<Mu::Interface::Unknown>, 
Ptr<WindowEventsWrap>, boost::noncopyable>("WindowEvents")
                
.def("OnWidgetEvent",pure_virtual(&WindowEvents::OnWidgetEvent))
                
.def("WidgetEventHandled",pure_virtual(&WindowEvents::WidgetEventHandled))
            ;

wrapper declaration:

        struct WindowEventsWrap
            :    InterfaceImplementation<WindowEventsWrap, 
Implements<Mu::Sandoz::Interface::WindowEvents> >,
                wrapper< Mu::Sandoz::Interface::WindowEvents >

python test code:

class MyWindowEvents(Sandoz.WindowEvents):
    def OnWidgetEvent(self, event):
        print event.type
        return Sandoz.Result.NotImplemented
   
    def WidgetEventHandled(self, event, widget):
        print event.type, widget
        return Sandoz.Result.NotImplemented

def Main():
    mwe = MyWindowEvents()
    view = Sandoz.ViewFrame("?.Sandoz.ViewFrame")
    view.Connect(mwe)
    del mwe

connect adds a reference to the mwe callback object and view calls that 
object later on. this is known to work. however when i delete mwe right 
after Connect(), the object itself persists, but its overridden methods 
(and i guess the underlying python object as well) are gone. no Release 
or AddReference on my object is called from wrapper_base, so i assume it 
is not using my custom smart pointer class which deals with this stuff.

can someone suggest a solution to this problem?

best regards,
leonard



From blast2001 at 21cn.com  Mon Apr 18 04:11:32 2005
From: blast2001 at 21cn.com (Blade)
Date: Mon, 18 Apr 2005 10:11:32 +0800
Subject: [C++-sig] Ask  for  help .
Message-ID: <000001c543be$9e3974e0$68c810ac@TEOMartin>

Hi everyone,

 I am a new boost.python user. I am trying to extend and embed python with my C++ application.
 now I encounter the following problem:
 My do steps:
 1. Write a simple python extend module(library).
 2. Write some python scripts according to extend module.
 3. In my application, call PyRun_SimpleString function to execute the scripts. It works well.
 
My question: how to exchange info or access variable(defined in C++ application or Python script) between C++ and scripts?
   That is to say: in scripts, how to access global variable which has been defined in C++ main application ?


Could someone show me a example ?

Thanks in advance,
Martin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20050418/e8db223f/attachment.htm>

From seefeld at sympatico.ca  Mon Apr 18 04:41:53 2005
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Sun, 17 Apr 2005 22:41:53 -0400
Subject: [C++-sig] Ask  for  help .
In-Reply-To: <000001c543be$9e3974e0$68c810ac@TEOMartin>
References: <000001c543be$9e3974e0$68c810ac@TEOMartin>
Message-ID: <42631E71.1040905@sympatico.ca>

Blade wrote:

> My question: how to exchange info or access variable(defined in C++ 
> application or Python script) between C++ and scripts?
>    That is to say: in scripts, how to access global variable which has 
> been defined in C++ main application ?

To reflect C++ variables into your python environment you have to bind
them to names in the dictionary that gets passed to PyRun_SimpleString().

The same works in the other direction, too: once PyRun_SimpleString() has
finished, you can query the same dictionary for variables that your script
defined.

HTH,
		Stefan


From tom.denniston at gmail.com  Mon Apr 18 04:54:15 2005
From: tom.denniston at gmail.com (Tom Denniston)
Date: Sun, 17 Apr 2005 22:54:15 -0400
Subject: [C++-sig] boost.python on mac os x
Message-ID: <d6f2d3dd050417195441ce38a@mail.gmail.com>

When I look at the regression test results for boost I notice that the
Darwin platform does not have a result for Boost.Python.  Does this
mean that Boost.Python does not support mac os x?  If it does not is
this because of something about the mac that makes it particularly
difficult to support or is there just a lack of interest in that
platform?

Has anyone succeeded in compiling boost.python for mac os x.  When I
try I get the following GCC error.  From googling it seems that there
is not much you can do about it.

The error I get is below:

...failed gcc-Link-action
bin/boost/libs/regex/build/libboost_regex.dylib/gcc/debug/shared-linkable-true/libboost_regex-gcc-d-1_32.dylib...
gcc-Link-action
bin/boost/libs/regex/build/libboost_regex.dylib/gcc/release/shared-linkable-true/libboost_regex-gcc-1_32.dylib
c++: unrecognized option `-shared'


-Tom


From blast2001 at 21cn.com  Mon Apr 18 05:06:56 2005
From: blast2001 at 21cn.com (Blade)
Date: Mon, 18 Apr 2005 11:06:56 +0800
Subject: [C++-sig] Ask  for  help .
References: <000001c543be$9e3974e0$68c810ac@TEOMartin>
	<42631E71.1040905@sympatico.ca>
Message-ID: <003401c543c3$b07a83b0$68c810ac@TEOMartin>

Thanks for your answer!

Further:  how to  bind the C++ variables into the dictionary with  boost.python  not  using  Python/C API  ?

I have not found the boost.python Doc mentioned these.
Could  you  give me some hints or  some relative resource ?
Are there any  open source projects using python for script language  for  reference ?

Thanks again !

Best Regards,
Blade

Ps: My english is not good enough, Maybe I have described something unclearly. :)


----- Original Message ----- 
From: "Stefan Seefeld" <seefeld at sympatico.ca>
To: "Blade" <blast2001 at 21cn.com>; "Development of Python/C++ integration" <c++-sig at python.org>
Sent: Monday, April 18, 2005 10:41 AM
Subject: Re: [C++-sig] Ask for help .


> Blade wrote:
> 
> > My question: how to exchange info or access variable(defined in C++ 
> > application or Python script) between C++ and scripts?
> >    That is to say: in scripts, how to access global variable which has 
> > been defined in C++ main application ?
> 
> To reflect C++ variables into your python environment you have to bind
> them to names in the dictionary that gets passed to PyRun_SimpleString().
> 
> The same works in the other direction, too: once PyRun_SimpleString() has
> finished, you can query the same dictionary for variables that your script
> defined.
> 
> HTH,
> Stefan
> 

From seefeld at sympatico.ca  Mon Apr 18 05:31:56 2005
From: seefeld at sympatico.ca (Stefan Seefeld)
Date: Sun, 17 Apr 2005 23:31:56 -0400
Subject: [C++-sig] Ask  for  help .
In-Reply-To: <003401c543c3$b07a83b0$68c810ac@TEOMartin>
References: <000001c543be$9e3974e0$68c810ac@TEOMartin>	<42631E71.1040905@sympatico.ca>
	<003401c543c3$b07a83b0$68c810ac@TEOMartin>
Message-ID: <42632A2C.7040309@sympatico.ca>

Blade wrote:
> Thanks for your answer!
> 
> Further:  how to  bind the C++ variables into the dictionary with  boost.python  not  using  Python/C API  ?
> 
> I have not found the boost.python Doc mentioned these.
> Could  you  give me some hints or  some relative resource ?
> Are there any  open source projects using python for script language  for  reference ?

You may want to look into the 'embedded.cpp' test that ships with the boost sources, i.e.
'boost/libs/python/test/embedding.cpp'. It shows how the 'main_namespace' dictionary
is constructed before being passed to a call to 'PyRun_String()'. Try some modifications,
i.e. add new variables to it before calling PyRun_String(), and then try accessing them
from inside the executed python script.

Regards,
		Stefan


From patrick.hartling at gmail.com  Mon Apr 18 14:40:23 2005
From: patrick.hartling at gmail.com (Patrick Hartling)
Date: Mon, 18 Apr 2005 07:40:23 -0500
Subject: [C++-sig] boost.python on mac os x
In-Reply-To: <d6f2d3dd050417195441ce38a@mail.gmail.com>
References: <d6f2d3dd050417195441ce38a@mail.gmail.com>
Message-ID: <944f7d8e050418054024acc20a@mail.gmail.com>

On 4/17/05, Tom Denniston <tom.denniston at gmail.com> wrote:
> When I look at the regression test results for boost I notice that the
> Darwin platform does not have a result for Boost.Python.  Does this
> mean that Boost.Python does not support mac os x?  If it does not is
> this because of something about the mac that makes it particularly
> difficult to support or is there just a lack of interest in that
> platform?

Boost.Python works fine for me on Mac OS X.  I have used Boost.Python
from the 1.31.0 and 1.32.0 releases of Boost.  From looking through
many messages on the Boost developer list, I believe that support for
OS X was improved between 1.31.0 and 1.32.0.  Furthermore, the hard
work of the Boost developers to get support for OS X solidified in
Boost in general helped me improve the usability of my own software on
Mac OS X.

> Has anyone succeeded in compiling boost.python for mac os x.  When I
> try I get the following GCC error.  From googling it seems that there
> is not much you can do about it.
> 
> The error I get is below:
> 
> ...failed gcc-Link-action
> bin/boost/libs/regex/build/libboost_regex.dylib/gcc/debug/shared-linkable-true/libboost_regex-gcc-d-1_32.dylib...
> gcc-Link-action
> bin/boost/libs/regex/build/libboost_regex.dylib/gcc/release/shared-linkable-true/libboost_regex-gcc-1_32.dylib
> c++: unrecognized option `-shared'

It looks as though you are using the 'gcc' toolset, which is the
default on OS X.   Try the 'darwin' toolset instead.  I am not sure
why the 'darwin' toolset is not the default.  There is one caveat,
however: there is a file in the boost_wserialization library that
requires more RAM to compile than my PowerBook has.  As such, I
usually restrict the build so that it either excludes that library or
so that it only builds the Boost libraries that I am going to use.

 -Patrick


-- 
Patrick L. Hartling
http://www.137.org/patrick/


From beyer at imb-jena.de  Mon Apr 18 21:02:33 2005
From: beyer at imb-jena.de (Andreas Beyer)
Date: Mon, 18 Apr 2005 12:02:33 -0700
Subject: [C++-sig] Pointer to existing Python object
In-Reply-To: <425EA20D.5050700@imb-jena.de>
References: <425EA20D.5050700@imb-jena.de>
Message-ID: <42640449.1030602@imb-jena.de>

Well, still no response, maybe the issue is really tough. I keep 
thinking about it, this is my current state:

It is already possible to obtain smart pointers for objects that are 
created in the Python environment. Coming back to my initial example:
http://mail.python.org/pipermail/c++-sig/2005-April/008787.html
I can create 'A's in Python, pass them to instances of B and B::get() 
will return the same *Python* object. I think this feature is enabled by
python::register_ptr_to_python< A::A_ptr >();

I am looking for the reverse, objects A created by instances of B (i.e. 
inside C++) need to be recognized by boost::python whenever the Python 
environment asks for them. That means, whenever instances of B return a 
smart pointer to an object for which a Python object exists already, 
that existing Python object should be returned. For objects created in 
the C++ environment this means, the first call to B::get() should return 
a new Python wrapper, subsequent calls should return the existing 
object. This could become tricky. E.g., this concept may not work for 
all kinds of return policies. However, I am looking for a specific 
solution for my code, where it may be possible to circumvent these 
difficulties.
It should be possible to implement the idea by defining a 
to_python_converter for A_ptr. That to_python_converter had to check for 
existing Pyhon objects. I tried a few things, but so-far I was 
unsuccessfull. Could somebody give me advice on how to proceed?

1. Are my thoughts on the right track?
2. I never wrote a to_python_converter. I am not even sure I have to: 
converting A* to Python with the default mechanism is no problem. In 
fact I want to wrap the A_ptr -> Python conversion. Can I somehow use 
the existing boost::python tools for converting A* inside my custom 
converter? How?
3. Is there an existing solution for a similar problem on which I could 
expand on?

Thanks!!
  Andreas

PS: Another difficulty: The wrapper that I am thinking of would manage 
some kind of global map, which maps pointers A* to their respective 
python wrappers. Pointers not in the map would get new python wrappers. 
But what about objects that are deleted? How gets my code informed about 
this? If I don't delete those entries from the map, the map will keep 
growing and growing. Is there a solution without changing ~A()?


Andreas Beyer wrote:

> Hi,
>
> I just want to bring this issue back on the table. I still didn't get 
> a brilliant idea how to solve it.
>
> So far, I see only this option:
> - Wrap B, in particular wrap B.get().
> - In get() create the actual python object, like:
>    python::object py_a(B::get());
> - Store those objects in a map.
> - When B_wrap.get() gets called again, check if a python object 
> already exists and return it.
>
> However, in my real-world application I had to wrap quite a lot of 
> methods returning As. I also have methods returning vectors of A_ptr, 
> which makes it even more enoying. (I would have to wrap the 
> vector<A_ptr>, too.)
>
> What is this has_back_reference doing? Is that somehow related to my 
> problem? (I regret that I don't really understand what it is doing, as 
> I don't know what a "predicate metafunction" is.)
> Just adopting the example at 
> http://www.boost.org/libs/python/doc/v2/has_back_reference.html is not 
> really a solution, because my instances of A are created inside the 
> C++ code (by factory methods in class B). I would have to be very 
> intrusive and make the actual class B aware of Python/boost::python 
> for adopting the above solution.
>
> I know that boost::python is collecting the shared pointers somewhere. 
> Is it possible to use that mechanism somewhow for solving my problem? 
> Any ideas?
>
> Thanks!!
>    Andreas
>
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
>



From tom.denniston at gmail.com  Tue Apr 19 01:15:32 2005
From: tom.denniston at gmail.com (Tom Denniston)
Date: Mon, 18 Apr 2005 18:15:32 -0500
Subject: [C++-sig] boost.python on mac os x
In-Reply-To: <944f7d8e050418054024acc20a@mail.gmail.com>
References: <d6f2d3dd050417195441ce38a@mail.gmail.com>
	<944f7d8e050418054024acc20a@mail.gmail.com>
Message-ID: <d6f2d3dd05041816155dac6766@mail.gmail.com>

how do I use darwin rather than gcc?

On 4/18/05, Patrick Hartling <patrick.hartling at gmail.com> wrote:
> On 4/17/05, Tom Denniston <tom.denniston at gmail.com> wrote:
> > When I look at the regression test results for boost I notice that the
> > Darwin platform does not have a result for Boost.Python.  Does this
> > mean that Boost.Python does not support mac os x?  If it does not is
> > this because of something about the mac that makes it particularly
> > difficult to support or is there just a lack of interest in that
> > platform?
> 
> Boost.Python works fine for me on Mac OS X.  I have used Boost.Python
> from the 1.31.0 and 1.32.0 releases of Boost.  From looking through
> many messages on the Boost developer list, I believe that support for
> OS X was improved between 1.31.0 and 1.32.0.  Furthermore, the hard
> work of the Boost developers to get support for OS X solidified in
> Boost in general helped me improve the usability of my own software on
> Mac OS X.
> 
> > Has anyone succeeded in compiling boost.python for mac os x.  When I
> > try I get the following GCC error.  From googling it seems that there
> > is not much you can do about it.
> >
> > The error I get is below:
> >
> > ...failed gcc-Link-action
> > bin/boost/libs/regex/build/libboost_regex.dylib/gcc/debug/shared-linkable-true/libboost_regex-gcc-d-1_32.dylib...
> > gcc-Link-action
> > bin/boost/libs/regex/build/libboost_regex.dylib/gcc/release/shared-linkable-true/libboost_regex-gcc-1_32.dylib
> > c++: unrecognized option `-shared'
> 
> It looks as though you are using the 'gcc' toolset, which is the
> default on OS X.   Try the 'darwin' toolset instead.  I am not sure
> why the 'darwin' toolset is not the default.  There is one caveat,
> however: there is a file in the boost_wserialization library that
> requires more RAM to compile than my PowerBook has.  As such, I
> usually restrict the build so that it either excludes that library or
> so that it only builds the Boost libraries that I am going to use.
> 
> -Patrick
> 
> --
> Patrick L. Hartling
> http://www.137.org/patrick/
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
>


From patrick.hartling at gmail.com  Tue Apr 19 02:10:07 2005
From: patrick.hartling at gmail.com (Patrick Hartling)
Date: Mon, 18 Apr 2005 19:10:07 -0500
Subject: [C++-sig] boost.python on mac os x
In-Reply-To: <d6f2d3dd05041816155dac6766@mail.gmail.com>
References: <d6f2d3dd050417195441ce38a@mail.gmail.com>
	<944f7d8e050418054024acc20a@mail.gmail.com>
	<d6f2d3dd05041816155dac6766@mail.gmail.com>
Message-ID: <944f7d8e050418171059edbbdc@mail.gmail.com>

On 4/18/05, Tom Denniston <tom.denniston at gmail.com> wrote:
> how do I use darwin rather than gcc?

Just as described in the documentation for building Boost:

   http://www.boost.org/more/getting_started.html#step5

Just specify the "darwin" toolset.  A brief descrption of  all the
toolsets can also be found on that page, along with links to more
detailed information.

 -Patrick

> On 4/18/05, Patrick Hartling <patrick.hartling at gmail.com> wrote:
> > On 4/17/05, Tom Denniston <tom.denniston at gmail.com> wrote:
> > > When I look at the regression test results for boost I notice that the
> > > Darwin platform does not have a result for Boost.Python.  Does this
> > > mean that Boost.Python does not support mac os x?  If it does not is
> > > this because of something about the mac that makes it particularly
> > > difficult to support or is there just a lack of interest in that
> > > platform?
> >
> > Boost.Python works fine for me on Mac OS X.  I have used Boost.Python
> > from the 1.31.0 and 1.32.0 releases of Boost.  From looking through
> > many messages on the Boost developer list, I believe that support for
> > OS X was improved between 1.31.0 and 1.32.0.  Furthermore, the hard
> > work of the Boost developers to get support for OS X solidified in
> > Boost in general helped me improve the usability of my own software on
> > Mac OS X.
> >
> > > Has anyone succeeded in compiling boost.python for mac os x.  When I
> > > try I get the following GCC error.  From googling it seems that there
> > > is not much you can do about it.
> > >
> > > The error I get is below:
> > >
> > > ...failed gcc-Link-action
> > > bin/boost/libs/regex/build/libboost_regex.dylib/gcc/debug/shared-linkable-true/libboost_regex-gcc-d-1_32.dylib...
> > > gcc-Link-action
> > > bin/boost/libs/regex/build/libboost_regex.dylib/gcc/release/shared-linkable-true/libboost_regex-gcc-1_32.dylib
> > > c++: unrecognized option `-shared'
> >
> > It looks as though you are using the 'gcc' toolset, which is the
> > default on OS X.   Try the 'darwin' toolset instead.  I am not sure
> > why the 'darwin' toolset is not the default.  There is one caveat,
> > however: there is a file in the boost_wserialization library that
> > requires more RAM to compile than my PowerBook has.  As such, I
> > usually restrict the build so that it either excludes that library or
> > so that it only builds the Boost libraries that I am going to use.
> >
> > -Patrick
> >
> > --
> > Patrick L. Hartling
> > http://www.137.org/patrick/
> > _______________________________________________
> > C++-sig mailing list
> > C++-sig at python.org
> > http://mail.python.org/mailman/listinfo/c++-sig
> >
> 


-- 
Patrick L. Hartling
http://www.137.org/patrick/


From jbrandmeyer at earthlink.net  Tue Apr 19 02:34:15 2005
From: jbrandmeyer at earthlink.net (Jonathan Brandmeyer)
Date: Mon, 18 Apr 2005 20:34:15 -0400
Subject: [C++-sig] boost.python on mac os x
In-Reply-To: <d6f2d3dd050417195441ce38a@mail.gmail.com>
References: <d6f2d3dd050417195441ce38a@mail.gmail.com>
Message-ID: <1113870856.4652.7.camel@localhost.localdomain>

On Sun, 2005-04-17 at 22:54 -0400, Tom Denniston wrote:
> When I look at the regression test results for boost I notice that the
> Darwin platform does not have a result for Boost.Python.  Does this
> mean that Boost.Python does not support mac os x?  If it does not is
> this because of something about the mac that makes it particularly
> difficult to support or is there just a lack of interest in that
> platform?
> 
> Has anyone succeeded in compiling boost.python for mac os x.

There is a prebuilt package in Fink.
http://fink.sourceforge.net/pdb/search.php?summary=boost

HTH,
-Jonathan



From tom.denniston at gmail.com  Tue Apr 19 04:36:44 2005
From: tom.denniston at gmail.com (Tom Denniston)
Date: Mon, 18 Apr 2005 22:36:44 -0400
Subject: [C++-sig] boost.python on mac os x
In-Reply-To: <944f7d8e050418171059edbbdc@mail.gmail.com>
References: <d6f2d3dd050417195441ce38a@mail.gmail.com>
	<944f7d8e050418054024acc20a@mail.gmail.com>
	<d6f2d3dd05041816155dac6766@mail.gmail.com>
	<944f7d8e050418171059edbbdc@mail.gmail.com>
Message-ID: <d6f2d3dd05041819366305f01d@mail.gmail.com>

That works, all the targets seem to compile except one (see below). 
It seems to just hang.  Is this the one you were having trouble with. 
what surprises me is that I have 1 GB of ram and still have the
problem.

The ouput I am seeing is below:


Tom-Dennistons-Computer:/development/libraries/boost_1_32_0
tomdenniston$ /development/bjam-3.1.10-1/bjam -sTOOLS=darwin
-sPYTHON_VERSION=2.3
...found 8239 targets...
...updating 261 targets...
darwin-C++-action
bin/boost/libs/serialization/build/libboost_serialization.a/darwin/release/xml_grammar.o




On 4/18/05, Patrick Hartling <patrick.hartling at gmail.com> wrote:
> On 4/18/05, Tom Denniston <tom.denniston at gmail.com> wrote:
> > how do I use darwin rather than gcc?
> 
> Just as described in the documentation for building Boost:
> 
>    http://www.boost.org/more/getting_started.html#step5
> 
> Just specify the "darwin" toolset.  A brief descrption of  all the
> toolsets can also be found on that page, along with links to more
> detailed information.
> 
>  -Patrick
> 
> > On 4/18/05, Patrick Hartling <patrick.hartling at gmail.com> wrote:
> > > On 4/17/05, Tom Denniston <tom.denniston at gmail.com> wrote:
> > > > When I look at the regression test results for boost I notice that the
> > > > Darwin platform does not have a result for Boost.Python.  Does this
> > > > mean that Boost.Python does not support mac os x?  If it does not is
> > > > this because of something about the mac that makes it particularly
> > > > difficult to support or is there just a lack of interest in that
> > > > platform?
> > >
> > > Boost.Python works fine for me on Mac OS X.  I have used Boost.Python
> > > from the 1.31.0 and 1.32.0 releases of Boost.  From looking through
> > > many messages on the Boost developer list, I believe that support for
> > > OS X was improved between 1.31.0 and 1.32.0.  Furthermore, the hard
> > > work of the Boost developers to get support for OS X solidified in
> > > Boost in general helped me improve the usability of my own software on
> > > Mac OS X.
> > >
> > > > Has anyone succeeded in compiling boost.python for mac os x.  When I
> > > > try I get the following GCC error.  From googling it seems that there
> > > > is not much you can do about it.
> > > >
> > > > The error I get is below:
> > > >
> > > > ...failed gcc-Link-action
> > > > bin/boost/libs/regex/build/libboost_regex.dylib/gcc/debug/shared-linkable-true/libboost_regex-gcc-d-1_32.dylib...
> > > > gcc-Link-action
> > > > bin/boost/libs/regex/build/libboost_regex.dylib/gcc/release/shared-linkable-true/libboost_regex-gcc-1_32.dylib
> > > > c++: unrecognized option `-shared'
> > >
> > > It looks as though you are using the 'gcc' toolset, which is the
> > > default on OS X.   Try the 'darwin' toolset instead.  I am not sure
> > > why the 'darwin' toolset is not the default.  There is one caveat,
> > > however: there is a file in the boost_wserialization library that
> > > requires more RAM to compile than my PowerBook has.  As such, I
> > > usually restrict the build so that it either excludes that library or
> > > so that it only builds the Boost libraries that I am going to use.
> > >
> > > -Patrick
> > >
> > > --
> > > Patrick L. Hartling
> > > http://www.137.org/patrick/
> > > _______________________________________________
> > > C++-sig mailing list
> > > C++-sig at python.org
> > > http://mail.python.org/mailman/listinfo/c++-sig
> > >
> >
> 
> --
> Patrick L. Hartling
> http://www.137.org/patrick/
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
>


From blast2001 at 21cn.com  Tue Apr 19 09:50:30 2005
From: blast2001 at 21cn.com (Blade)
Date: Tue, 19 Apr 2005 15:50:30 +0800
Subject: Fw: [C++-sig] Ask  for  help .
Message-ID: <00c801c544b4$799c0fc0$68c810ac@TEOMartin>

Sorry to interrupt you again !

I understand your standpoint, but don't know the specific steps to implement those.

add new variables to it before calling PyRun_String(), -------------  how  to  do  it ?
try accessing them from inside the executed python script. ----------  How ?

Regards,
Blade


----- Original Message ----- 
From: "Stefan Seefeld" <seefeld at sympatico.ca>
To: "Blade" <blast2001 at 21cn.com>; "Development of Python/C++ integration" <c++-sig at python.org>
Sent: Monday, April 18, 2005 11:31 AM
Subject: Re: [C++-sig] Ask for help .


> Blade wrote:
> > Thanks for your answer!
> > 
> > Further:  how to  bind the C++ variables into the dictionary with  boost.python  not  using  Python/C API  ?
> > 
> > I have not found the boost.python Doc mentioned these.
> > Could  you  give me some hints or  some relative resource ?
> > Are there any  open source projects using python for script language  for  reference ?
> 
> You may want to look into the 'embedded.cpp' test that ships with the boost sources, i.e.
> 'boost/libs/python/test/embedding.cpp'. It shows how the 'main_namespace' dictionary
> is constructed before being passed to a call to 'PyRun_String()'. Try some modifications,
> i.e. add new variables to it before calling PyRun_String(), and then try accessing them
> from inside the executed python script.
> 
> Regards,
> Stefan
> 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20050419/242f4528/attachment.htm>

From brett at coombs.anu.edu.au  Tue Apr 19 07:42:58 2005
From: brett at coombs.anu.edu.au (Brett Calcott)
Date: Tue, 19 Apr 2005 15:42:58 +1000
Subject: [C++-sig] Re: wrapper.cpp not in 1.32.0 dsp/vcproj file?
In-Reply-To: <425C4DE9.4020205@paniq.org>
References: <200504081610.00461.mvc@di.fct.unl.pt>	<425896D6.30808@paniq.org>	<u7jjavinl.fsf@boost-consulting.com>	<425AC635.4050200@paniq.org>
	<425C4DE9.4020205@paniq.org>
Message-ID: <d425h9$cou$1@sea.gmane.org>

Leonard "paniq" Ritter wrote:
> fyi, the mail to brett calcott bounced.
> 

That is me. My apologies. I haven't been using Boost Python for some 
time now. I have had my head buried in my PhD. This email should be more 
stable:
brett.calcott at gmail.com

I am still happy to maintain the vc build support. Leonard I assume that 
this just requires the addition of a file?

Also, I owe Dave a personal apology. I promised some documentation of 
the internals way back in '03, and Dave took the time to answer some of 
my questions (http://www.boost.org/libs/python/doc/internals.html). I 
never did this. I remember making some notes, and finding it to be the 
hardest code to understand I had ever looked at. But I never sent 
anything to the list.

Dave, are you still happy for me to maintain this given my track record 
here? The gmail address will be stable.


Brett




From blast2001 at 21cn.com  Tue Apr 19 11:10:59 2005
From: blast2001 at 21cn.com (Blade)
Date: Tue, 19 Apr 2005 17:10:59 +0800
Subject: [C++-sig] Ask  for  help .
Message-ID: <00e601c544bf$b4e66ed0$68c810ac@TEOMartin>

Hi  everyone,

    Now  I  have solved the following 2 issues when the variable is simple type,  but when I bind the C++ class object into the dictionary,  it compile  successfully,  it will abort once execute it. 
The below is the c++ source code:

    if (PyImport_AppendInittab("embedded_hello", initembedded_hello) == -1)
        throw std::runtime_error("Failed to add embedded_hello to the interpreter's "
                                 "builtin modules");
    Py_Initialize();
    python::object main_module((
        python::handle<>(python::borrowed(PyImport_AddModule("__main__")))));

    python::object main_namespace((main_module.attr("__dict__")));
    Base * pBase = new Base(2000);
    main_namespace["retValue"] = 5;
    main_namespace["AppBase"] = pBase;    //abort here

    python::handle<> result(
        PyRun_String(
            "from embedded_hello import *        \n"
            "print AppBase.i \n"
        Py_file_input, main_namespace.ptr(), main_namespace.ptr())
        );


Notes: Base is the class name I defined and main_namespace is the dict which is passed to PyRun_String.


Could someone figure out the cause ?
Thanks very much !


Best Regards,
Blade

  ----- Original Message ----- 
  From: Blade 
  To: Stefan Seefeld 
  Cc: c++-sig at python.org 
  Sent: Tuesday, April 19, 2005 3:50 PM
  Subject: Fw: [C++-sig] Ask for help .


  Sorry to interrupt you again !

  I understand your standpoint, but don't know the specific steps to implement those.

  add new variables to it before calling PyRun_String(), -------------  how  to  do  it ?
  try accessing them from inside the executed python script. ----------  How ?

  Regards,
  Blade


  ----- Original Message ----- 
  From: "Stefan Seefeld" <seefeld at sympatico.ca>
  To: "Blade" <blast2001 at 21cn.com>; "Development of Python/C++ integration" <c++-sig at python.org>
  Sent: Monday, April 18, 2005 11:31 AM
  Subject: Re: [C++-sig] Ask for help .


  > Blade wrote:
  > > Thanks for your answer!
  > > 
  > > Further:  how to  bind the C++ variables into the dictionary with  boost.python  not  using  Python/C API  ?
  > > 
  > > I have not found the boost.python Doc mentioned these.
  > > Could  you  give me some hints or  some relative resource ?
  > > Are there any  open source projects using python for script language  for  reference ?
  > 
  > You may want to look into the 'embedded.cpp' test that ships with the boost sources, i.e.
  > 'boost/libs/python/test/embedding.cpp'. It shows how the 'main_namespace' dictionary
  > is constructed before being passed to a call to 'PyRun_String()'. Try some modifications,
  > i.e. add new variables to it before calling PyRun_String(), and then try accessing them
  > from inside the executed python script.
  > 
  > Regards,
  > Stefan
  > 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20050419/9ac1b4e8/attachment.htm>

From dave at boost-consulting.com  Tue Apr 19 14:22:04 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Tue, 19 Apr 2005 14:22:04 +0200
Subject: [C++-sig] Re: Pointer to existing Python object
References: <425EA20D.5050700@imb-jena.de> <42640449.1030602@imb-jena.de>
Message-ID: <ud5sryn0z.fsf@boost-consulting.com>

Andreas Beyer <beyer at imb-jena.de> writes:

> Well, still no response, maybe the issue is really tough. I keep 
> thinking about it, this is my current state:

Andreas,

I have been travelling and on vacation, so I haven't had a chance to
look at your problem yet.  I'll try to give it a look before I get
home on 2005/4/25, but I can't make any promises.

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



From dave at boost-consulting.com  Tue Apr 19 18:50:51 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Tue, 19 Apr 2005 18:50:51 +0200
Subject: [C++-sig] Re: wrapper.cpp not in 1.32.0 dsp/vcproj file?
References: <200504081610.00461.mvc@di.fct.unl.pt> <425896D6.30808@paniq.org>
	<u7jjavinl.fsf@boost-consulting.com> <425AC635.4050200@paniq.org>
	<425C4DE9.4020205@paniq.org> <d425h9$cou$1@sea.gmane.org>
Message-ID: <uvf6iww0k.fsf@boost-consulting.com>

Brett Calcott <brett at coombs.anu.edu.au> writes:

> Leonard "paniq" Ritter wrote:
>> fyi, the mail to brett calcott bounced.
>> 
>
> That is me. My apologies. I haven't been using Boost Python for some 
> time now. I have had my head buried in my PhD. This email should be more 
> stable:
> brett.calcott at gmail.com
>
> I am still happy to maintain the vc build support. Leonard I assume that 
> this just requires the addition of a file?
>
> Also, I owe Dave a personal apology. I promised some documentation of 
> the internals way back in '03, and Dave took the time to answer some of 
> my questions (http://www.boost.org/libs/python/doc/internals.html). I 
> never did this. I remember making some notes, and finding it to be the 
> hardest code to understand I had ever looked at. But I never sent 
> anything to the list.
>
> Dave, are you still happy for me to maintain this given my track record 
> here? The gmail address will be stable.

If you will actually do it, I'm happy for you to do it ;-)

If you are saying that you are going to be unreliable, that's another
matter.

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



From mike at thewilsonfamily.freeserve.co.uk  Thu Apr 21 14:31:34 2005
From: mike at thewilsonfamily.freeserve.co.uk (Mike)
Date: Thu, 21 Apr 2005 13:31:34 +0100
Subject: [C++-sig] Re: wrapper.cpp not in 1.32.0 dsp/vcproj file?
References: <200504081610.00461.mvc@di.fct.unl.pt>	<425896D6.30808@paniq.org>	<u7jjavinl.fsf@boost-consulting.com>	<425AC635.4050200@paniq.org><425C4DE9.4020205@paniq.org>
	<d425h9$cou$1@sea.gmane.org>
Message-ID: <000401c5466e$0ef7e0c0$d8b44c51@MIKESPC01>

Bret Calcott wrote....

> 
> I am still happy to maintain the vc build support. Leonard I assume that 
> this just requires the addition of a file?
> 

I have just noticed that 
   slice.cpp 
also needs to be added to the project (besides wrapper.cpp).

Mike




From patrick.hartling at gmail.com  Wed Apr 20 00:08:54 2005
From: patrick.hartling at gmail.com (Patrick Hartling)
Date: Tue, 19 Apr 2005 17:08:54 -0500
Subject: [C++-sig] boost.python on mac os x
In-Reply-To: <d6f2d3dd05041819366305f01d@mail.gmail.com>
References: <d6f2d3dd050417195441ce38a@mail.gmail.com>
	<944f7d8e050418054024acc20a@mail.gmail.com>
	<d6f2d3dd05041816155dac6766@mail.gmail.com>
	<944f7d8e050418171059edbbdc@mail.gmail.com>
	<d6f2d3dd05041819366305f01d@mail.gmail.com>
Message-ID: <944f7d8e0504191508371993cf@mail.gmail.com>

On 4/18/05, Tom Denniston <tom.denniston at gmail.com> wrote:
> That works, all the targets seem to compile except one (see below).
> It seems to just hang.  Is this the one you were having trouble with.
> what surprises me is that I have 1 GB of ram and still have the
> problem.
> 
> The ouput I am seeing is below:
> 
> Tom-Dennistons-Computer:/development/libraries/boost_1_32_0
> tomdenniston$ /development/bjam-3.1.10-1/bjam -sTOOLS=darwin
> -sPYTHON_VERSION=2.3
> ...found 8239 targets...
> ...updating 261 targets...
> darwin-C++-action
> bin/boost/libs/serialization/build/libboost_serialization.a/darwin/release/xml_grammar.o

That's the one.  Perhaps it's not a memory problem but rather a
compiler bug?  I haven't ever bothered to dig into it because I am not
using the serialization library.  As I said before, I usually restrict
the build to just the libraries that I need.  It saves time and space.

 -Patrick

> On 4/18/05, Patrick Hartling <patrick.hartling at gmail.com> wrote:
> > On 4/18/05, Tom Denniston <tom.denniston at gmail.com> wrote:
> > > how do I use darwin rather than gcc?
> >
> > Just as described in the documentation for building Boost:
> >
> >    http://www.boost.org/more/getting_started.html#step5
> >
> > Just specify the "darwin" toolset.  A brief descrption of  all the
> > toolsets can also be found on that page, along with links to more
> > detailed information.
> >
> >  -Patrick
> >
> > > On 4/18/05, Patrick Hartling <patrick.hartling at gmail.com> wrote:
> > > > On 4/17/05, Tom Denniston <tom.denniston at gmail.com> wrote:
> > > > > When I look at the regression test results for boost I notice that the
> > > > > Darwin platform does not have a result for Boost.Python.  Does this
> > > > > mean that Boost.Python does not support mac os x?  If it does not is
> > > > > this because of something about the mac that makes it particularly
> > > > > difficult to support or is there just a lack of interest in that
> > > > > platform?
> > > >
> > > > Boost.Python works fine for me on Mac OS X.  I have used Boost.Python
> > > > from the 1.31.0 and 1.32.0 releases of Boost.  From looking through
> > > > many messages on the Boost developer list, I believe that support for
> > > > OS X was improved between 1.31.0 and 1.32.0.  Furthermore, the hard
> > > > work of the Boost developers to get support for OS X solidified in
> > > > Boost in general helped me improve the usability of my own software on
> > > > Mac OS X.
> > > >
> > > > > Has anyone succeeded in compiling boost.python for mac os x.  When I
> > > > > try I get the following GCC error.  From googling it seems that there
> > > > > is not much you can do about it.
> > > > >
> > > > > The error I get is below:
> > > > >
> > > > > ...failed gcc-Link-action
> > > > > bin/boost/libs/regex/build/libboost_regex.dylib/gcc/debug/shared-linkable-true/libboost_regex-gcc-d-1_32.dylib...
> > > > > gcc-Link-action
> > > > > bin/boost/libs/regex/build/libboost_regex.dylib/gcc/release/shared-linkable-true/libboost_regex-gcc-1_32.dylib
> > > > > c++: unrecognized option `-shared'
> > > >
> > > > It looks as though you are using the 'gcc' toolset, which is the
> > > > default on OS X.   Try the 'darwin' toolset instead.  I am not sure
> > > > why the 'darwin' toolset is not the default.  There is one caveat,
> > > > however: there is a file in the boost_wserialization library that
> > > > requires more RAM to compile than my PowerBook has.  As such, I
> > > > usually restrict the build so that it either excludes that library or
> > > > so that it only builds the Boost libraries that I am going to use.
> > > >
> > > > -Patrick
> > > >
> > > > --
> > > > Patrick L. Hartling
> > > > http://www.137.org/patrick/
> > > > _______________________________________________
> > > > C++-sig mailing list
> > > > C++-sig at python.org
> > > > http://mail.python.org/mailman/listinfo/c++-sig
> > > >
> > >
> >
> > --
> > Patrick L. Hartling
> > http://www.137.org/patrick/
> > _______________________________________________
> > C++-sig mailing list
> > C++-sig at python.org
> > http://mail.python.org/mailman/listinfo/c++-sig
> >
> 


-- 
Patrick L. Hartling
http://www.137.org/patrick/


From mikeowen at llnl.gov  Thu Apr 21 02:50:14 2005
From: mikeowen at llnl.gov (J. Michael Owen)
Date: Wed, 20 Apr 2005 17:50:14 -0700
Subject: [C++-sig] RTLD_GLOBAL cures my g++ + Boost.Python + cross module
	blues?
Message-ID: <200504201750.14065.mikeowen@llnl.gov>

I've recently come across a method to get my g++ compiled Boost.Python classes 
wrapped in different shared libraries to work (cross module class 
inheritance, dynamic_cast, etc.), and I thought I'd share this to see if it's 
useful to anyone else here, as well as see if anyone knows of any problems 
with this approach.

For completeness I'm using Python 2.3.4, g++ 3.4.1, and the last Boost release 
(1.32.0) on Linux x86.  When compiling with g++ I've been plagued by the 
cross library type_info problems described elsewhere 
(http://www.python.org/moin/boost.python/CrossExtensionModuleDependencies), 
but recently I discovered that if I force python's dlopen to use RTLD_GLOBAL 
then upon importing my extensions the different libraries resolve these 
type_info problems and everything starts working.  All I've done is issue the 
following python commands before importing any Boost.Python wrapped code:

import sys, dl
sys.setdlopenflags(dl.RTLD_NOW|dl.RTLD_GLOBAL)

While this seems to have solved all of my current cross-library problems, I'm 
suspicious that it could be this simple.  Does anyone else here have any 
experience with these g++/Boost.Python/cross module issues?  Is there a 
reason the above might be a bad idea?

Mike.

-- 
 "Hey...where are the sunflower seeds?" |       J. Michael Owen         
        o_o /                           |       
        (")                             |       Email:  mikeowen at llnl.gov
       \/'\/                            |
____(__(,_,)_______________________________________________________________



From brett at coombs.anu.edu.au  Thu Apr 21 05:48:13 2005
From: brett at coombs.anu.edu.au (Brett Calcott)
Date: Thu, 21 Apr 2005 13:48:13 +1000
Subject: [C++-sig] Re: wrapper.cpp not in 1.32.0 dsp/vcproj file?
In-Reply-To: <uvf6iww0k.fsf@boost-consulting.com>
References: <200504081610.00461.mvc@di.fct.unl.pt>
	<425896D6.30808@paniq.org>	<u7jjavinl.fsf@boost-consulting.com>
	<425AC635.4050200@paniq.org>	<425C4DE9.4020205@paniq.org>
	<d425h9$cou$1@sea.gmane.org> <uvf6iww0k.fsf@boost-consulting.com>
Message-ID: <d477hl$l6r$1@sea.gmane.org>

David Abrahams wrote:
> 
> If you will actually do it, I'm happy for you to do it ;-)

Yes. I will do it. I have VC6 and 7.1, so should be able to sort this in 
the next week.

> 
> If you are saying that you are going to be unreliable, that's another
> matter.
> 

Looking after the vc build is doable. I can be relied on for this. I bit 
off more than I could chew with documenting internals :(

Cheers,
Brett



From news4vovan at mail.ru  Fri Apr 22 17:24:21 2005
From: news4vovan at mail.ru (Vladimir Sukhoy)
Date: Fri, 22 Apr 2005 18:24:21 +0300
Subject: [C++-sig] pyste, shared_ptr and constructor override
Message-ID: <d4b4mo$cad$1@sea.gmane.org>

I use boost.python to export class which is enclosed in
smart pointer and has protected constructor to force
smart pointer usage, like that:

class MyClass;
typedef boost::shared_ptr<MyClass> MyClassPtr;

class MyClass
{
public:
    static MyClassPtr Create();
    virtual ~MyClass();
protected:
    MyClass();
<skip>
};

So, after pyste has processed the code for MyClass I have something like
that:

 class_< MyClass, bases< boost::noncopyable_::noncopyable > ,
boost::noncopyable >("MyClass", no_init)
        .def("Create", &MyClass::Create)
....
 register_ptr_to_python< boost::shared_ptr< space::http::SimpleGet > >();

So after compiling extension module I can use MyClass from python like that

mc = MyClass.Create()

But that's not exactly what I want, because I really meant that shared_ptr 
IS MyClass,
so I want to write

mc = MyClass()

and have C++ function MyClass::Create being called instead.

Of course, I can override this with Python (which rox), even procedurally, 
like that:

def override_constructor(cls, cons, d):
    def override_func(cls, *args, **kwargs):
        return cons(*args, **kwargs)
    dc = {'__new__' : override_func}
    from new import classobj
    d[cls.__name__] = classobj(cls.__name__, (cls,), dc)

override_constructor(MyClass, MyClass.Create, globals())

Is there a way to do the same thing with pyste at wrapper generation step?






From tom.denniston at gmail.com  Sat Apr 23 20:36:49 2005
From: tom.denniston at gmail.com (Tom Denniston)
Date: Sat, 23 Apr 2005 14:36:49 -0400
Subject: [C++-sig] Running python using Boost.Python modules
Message-ID: <d6f2d3dd0504231136372879d2@mail.gmail.com>

Is there a way to run python so that it picks up the C++ modules
compiled using Boost.Python and bjam.  I saw in the examples that you
could run python from a bjam target such as the following:

# Declare a test for the extension module
boost-python-runtest test1
    :  # Python test driver
    src/python/test_getting_started1.py
    # extension modules to use
    <pyd>getting_started1 ;
    
    

But I didn't see anything on how to run a python process independent
of a bjam target.  one possibility is to copy the modules into the
python installation but i would prefer to use the libraries from
python without having and side efffect on my python installation.

Would appreciate any thoughts or direction toward the appropriate
documentation.  I might just be missing something really obvious here.

-Tom


From jbrandmeyer at earthlink.net  Sat Apr 23 22:36:10 2005
From: jbrandmeyer at earthlink.net (Jonathan Brandmeyer)
Date: Sat, 23 Apr 2005 16:36:10 -0400
Subject: [C++-sig] Running python using Boost.Python modules
In-Reply-To: <d6f2d3dd0504231136372879d2@mail.gmail.com>
References: <d6f2d3dd0504231136372879d2@mail.gmail.com>
Message-ID: <1114288570.20396.11.camel@localhost.localdomain>

On Sat, 2005-04-23 at 14:36 -0400, Tom Denniston wrote:
> Is there a way to run python so that it picks up the C++ modules
> compiled using Boost.Python and bjam.  I saw in the examples that you
> could run python from a bjam target such as the following:
> 
> # Declare a test for the extension module
> boost-python-runtest test1
>     :  # Python test driver
>     src/python/test_getting_started1.py
>     # extension modules to use
>     <pyd>getting_started1 ;
>     
>     
> 
> But I didn't see anything on how to run a python process independent
> of a bjam target.  one possibility is to copy the modules into the
> python installation but i would prefer to use the libraries from
> python without having and side efffect on my python installation.
> 
> Would appreciate any thoughts or direction toward the appropriate
> documentation.  I might just be missing something really obvious here.

See sys.path in the Python Library Reference, and Installing Python
Modules, section 4.1.

HTH,
-Jonathan



From nicodemus at esss.com.br  Sun Apr 24 17:49:16 2005
From: nicodemus at esss.com.br (Nicodemus)
Date: Sun, 24 Apr 2005 13:49:16 -0200
Subject: [C++-sig] pyste, shared_ptr and constructor override
In-Reply-To: <d4b4mo$cad$1@sea.gmane.org>
References: <d4b4mo$cad$1@sea.gmane.org>
Message-ID: <426BBFFC.5020900@esss.com.br>

Hi Vladmir,

Vladimir Sukhoy wrote:

>But that's not exactly what I want, because I really meant that shared_ptr 
>IS MyClass,
>so I want to write
>
>mc = MyClass()
>
>and have C++ function MyClass::Create being called instead.
>
>Of course, I can override this with Python (which rox), even procedurally, 
>like that:
>
>def override_constructor(cls, cons, d):
>    def override_func(cls, *args, **kwargs):
>        return cons(*args, **kwargs)
>    dc = {'__new__' : override_func}
>    from new import classobj
>    d[cls.__name__] = classobj(cls.__name__, (cls,), dc)
>
>override_constructor(MyClass, MyClass.Create, globals())
>
>Is there a way to do the same thing with pyste at wrapper generation step?
>  
>

Not right now, and I don't have any definite plans for it either. :/ I 
suggest you keep inserting the constructor like you are doing.

Regards,
Bruno.


From itamar at itamarst.org  Mon Apr 25 17:54:58 2005
From: itamar at itamarst.org (Itamar Shtull-Trauring)
Date: Mon, 25 Apr 2005 11:54:58 -0400
Subject: [C++-sig] RTLD_GLOBAL cures my g++ + Boost.Python + cross
	module blues?
In-Reply-To: <200504201750.14065.mikeowen@llnl.gov>
References: <200504201750.14065.mikeowen@llnl.gov>
Message-ID: <1114444498.8213.29.camel@localhost.localdomain>

On Wed, 2005-04-20 at 17:50 -0700, J. Michael Owen wrote:

> import sys, dl
> sys.setdlopenflags(dl.RTLD_NOW|dl.RTLD_GLOBAL)
> 
> While this seems to have solved all of my current cross-library problems, I'm 
> suspicious that it could be this simple.  Does anyone else here have any 
> experience with these g++/Boost.Python/cross module issues?  Is there a 
> reason the above might be a bad idea?

This is what I do. The only possible problem, I think, is that the
symbols might conflict with other C modules you will load later. Using
namespaces should make this unlikely, and I typically reset dlopenflags
back to normal (i.e. no RTLD_GLOBAL) once I've imported my boost module.

And of course this won't work on Windows :)




From nicodemus at esss.com.br  Tue Apr 26 00:11:32 2005
From: nicodemus at esss.com.br (Nicodemus)
Date: Mon, 25 Apr 2005 20:11:32 -0200
Subject: [C++-sig] Getting the pointer for a member operator
Message-ID: <426D6B14.9040309@esss.com.br>

Hello all,

Currently Pyste does this:

// C++ code
struct C
{
     operator const char*() { return "C"; }
};

// generated:

BOOST_PYTHON_MODULE(_operators)
{
     class_< operators::C >("C", init<  >())
         .def("__str__", &operators::C::operator const char*)
     ;
}

Problem is that this is broken under MSVC.NET 2003. The compiler 
complains with:

"""
_operators.cpp(24) : error C2833: 'operator const' is not a recognized 
operator or type
"""

ie, the compiler doesn't realize that its the operator const char* that 
it should get the address to. I tried changing it to:

     .def("__str__", &operators::C::operator (const char*))

But without success. Does anybody know how I can fix this?

Best Regards,
Nicodemus.



From nicodemus at esss.com.br  Tue Apr 26 00:40:36 2005
From: nicodemus at esss.com.br (Nicodemus)
Date: Mon, 25 Apr 2005 20:40:36 -0200
Subject: [C++-sig] Pyste bug? Using Wrapper with member functions and
	--pyste-ns
In-Reply-To: <5ab8740205032002161499b77d@mail.gmail.com>
References: <5ab8740205032002161499b77d@mail.gmail.com>
Message-ID: <426D71E4.3040605@esss.com.br>

Dan Haffey wrote:

>Hello,
>
>I just noticed that when using the Pyste Wrapper call to create a
>member function wrapper, it places the wrapper definition code outside
>of the specified (--pyste-ns=) namespace, but the generated module
>code looks for it in the namespace. This doesn't seem to happen with
>normal functions, just members.
>  
>

Fixed. Thanks Dan!

Regards,
Nicodemus.



From news4vovan at mail.ru  Tue Apr 26 16:01:49 2005
From: news4vovan at mail.ru (Vladimir Sukhoy)
Date: Tue, 26 Apr 2005 17:01:49 +0300
Subject: [C++-sig] Re: Getting the pointer for a member operator
References: <426D6B14.9040309@esss.com.br>
Message-ID: <d4lhal$a9q$1@sea.gmane.org>

> Problem is that this is broken under MSVC.NET 2003. The compiler complains 
> with:
I tried similar with .NET 2003 compiler - it works, and I cannot reproduce 
this error! The full module code is:

#include <boost/python.hpp>
using namespace boost::python;
namespace operators
{
struct C
{
operator const char*() { return "C"; }
};
}
BOOST_PYTHON_MODULE(_ace)
{
class_< operators::C >("C", init< >())
.def("__str__", &operators::C::operator const char*)
;
}






From blast2001 at 21cn.com  Wed Apr 27 05:44:25 2005
From: blast2001 at 21cn.com (Martin Dai)
Date: Wed, 27 Apr 2005 11:44:25 +0800
Subject: [C++-sig] Pass C++ object to python  issue
Message-ID: <d4n1j3$qpq$1@sea.gmane.org>

Hi  everyone,

    Now  I  have solved the pass variable when the variable is simple type,
but when I bind the C++ class object into the dictionary,  it compile
successfully,  it will abort once execute it.
The below is the c++ source code:

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

python::handle<>(python::borrowed(PyImport_AddModule("__main__")))));

    python::object main_namespace((main_module.attr("__dict__")));
    Base * pBase = new Base(2000);
    main_namespace["retValue"] = 5;
    main_namespace["AppBase"] = python::object(python::ptr(pBase));
//abort here

    python::handle<> result(
        PyRun_String(
            "from embedded_hello import *        \n"
            "print AppBase.i \n"
        Py_file_input, main_namespace.ptr(), main_namespace.ptr())
        );


Notes: Base is the class name I defined and main_namespace is the dict which
is passed to PyRun_String.


Could someone figure out the cause ?
Thanks very much !


Best Regards,
Blade





From ghost at cs.msu.su  Wed Apr 27 09:32:42 2005
From: ghost at cs.msu.su (Vladimir Prus)
Date: Wed, 27 Apr 2005 11:32:42 +0400
Subject: [C++-sig] Re: Pass C++ object to python  issue
References: <d4n1j3$qpq$1@sea.gmane.org>
Message-ID: <d4neu4$qas$1@sea.gmane.org>

Martin Dai wrote:

> Hi  everyone,
> 
>     Now  I  have solved the pass variable when the variable is simple
>     type,
> but when I bind the C++ class object into the dictionary,  it compile
> successfully,  it will abort once execute it.

With what error message?

> The below is the c++ source code:
> 
>     if (PyImport_AppendInittab("embedded_hello", initembedded_hello) ==
>     -1)

Please always provide a complete program.

>     main_namespace["AppBase"] = python::object(python::ptr(pBase));
> //abort here
> 
>     python::handle<> result(
>         PyRun_String(
>             "from embedded_hello import *        \n"
>             "print AppBase.i \n"
>         Py_file_input, main_namespace.ptr(), main_namespace.ptr())
>         );

I've tried to reproduce this, and got:

    http://zigzag.cs.msu.su/~ghost/e3.cpp

that works fine. However, when I move the 

    main_namespace["context"] = ptr(b);

above call to PyRun_String (as is done in your example), 

    http://zigzag.cs.msu.su/~ghost/e3f.cpp

I get this:

   TypeError: No Python class registered for C++ class B

which looks quite reasonable. Python has no idea that class 'B' even exists,
but you're trying to add instance of that class. I wonder if it's possible
to automatically register Python class in this case, though.

- Volodya




From Ben.Cain at us.army.mil  Wed Apr 27 16:09:52 2005
From: Ben.Cain at us.army.mil (Cain, Ben (Contractor-Summit))
Date: Wed, 27 Apr 2005 09:09:52 -0500
Subject: [C++-sig] Newbie ... trouble compiling Boost.Python
Message-ID: <6DB8567B5D54F443982465B984BE4AA99B2B18@amr-ex5.ds.amrdec.army.mil>

Hello,
 
I'm getting an error when trying to compile Boost.Python (using GCC on
Windows).  Most of Boost seemed to compile successfully.
 
I used the following command to compile from C:\Boost_1_32_0
   bjam "-sTOOLS=gcc" install
 
However, I get the following error during the Boost.Python build phase:
   Python.h: No such file or directory
 
I have the following environment variables set with Python 2.3.5
installed:
   PYTHON_ROOT       C:\Python23
   PYTHON_VERSION    2.3
   PYTHON_INCLUDES   C:\Python23\include
   PYTHON_LIB_PATH   C:\Python23\libs
 
I then tried to build Boost.Python directly from
C:\Boost_1_32_0\libs\python\build.  However, bjam didn't know how to
build from this directory.
 
Any help will be greatly appreciated.
 
Thanks,
  Ben
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20050427/5275b1b4/attachment.htm>

From dave at boost-consulting.com  Wed Apr 27 20:34:39 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Wed, 27 Apr 2005 14:34:39 -0400
Subject: [C++-sig] Re: wrapper.cpp not in 1.32.0 dsp/vcproj file?
References: <200504081610.00461.mvc@di.fct.unl.pt> <425896D6.30808@paniq.org>
	<u7jjavinl.fsf@boost-consulting.com> <425AC635.4050200@paniq.org>
	<425C4DE9.4020205@paniq.org> <d425h9$cou$1@sea.gmane.org>
	<uvf6iww0k.fsf@boost-consulting.com> <d477hl$l6r$1@sea.gmane.org>
Message-ID: <uis28ukzk.fsf@boost-consulting.com>

Brett Calcott <brett at coombs.anu.edu.au> writes:

> David Abrahams wrote:
>> If you will actually do it, I'm happy for you to do it ;-)
>
> Yes. I will do it. I have VC6 and 7.1, so should be able to sort this in 
> the next week.

Thanks!

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



From dave at boost-consulting.com  Wed Apr 27 20:51:53 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Wed, 27 Apr 2005 14:51:53 -0400
Subject: [C++-sig] Re: Getting the pointer for a member operator
References: <426D6B14.9040309@esss.com.br>
Message-ID: <ud5sguk6u.fsf@boost-consulting.com>

Nicodemus <nicodemus at esss.com.br> writes:

> Problem is that this is broken under MSVC.NET 2003. The compiler 
> complains with:
>
> """
> _operators.cpp(24) : error C2833: 'operator const' is not a recognized 
> operator or type
> """
>
> ie, the compiler doesn't realize that its the operator const char* that 
> it should get the address to. I tried changing it to:
>
>     .def("__str__", &operators::C::operator (const char*))
>
> But without success. Does anybody know how I can fix this?

Did you try a typedef?

    typedef char const* string;

    ...

         .def("__str__", &operators::C::operator string)

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



From dave at boost-consulting.com  Wed Apr 27 20:56:36 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Wed, 27 Apr 2005 14:56:36 -0400
Subject: [C++-sig] Re: Newbie ... trouble compiling Boost.Python
References: <6DB8567B5D54F443982465B984BE4AA99B2B18@amr-ex5.ds.amrdec.army.mil>
Message-ID: <u7jioujyz.fsf@boost-consulting.com>

"Cain, Ben (Contractor-Summit)" <Ben.Cain at us.army.mil> writes:

> Hello,
>  
> I'm getting an error when trying to compile Boost.Python (using GCC on
> Windows).  Most of Boost seemed to compile successfully.
>  
> I used the following command to compile from C:\Boost_1_32_0
>    bjam "-sTOOLS=gcc" install
>  
> However, I get the following error during the Boost.Python build phase:
>    Python.h: No such file or directory
>  
> I have the following environment variables set with Python 2.3.5
> installed:
>    PYTHON_ROOT       C:\Python23
>    PYTHON_VERSION    2.3
>    PYTHON_INCLUDES   C:\Python23\include
>    PYTHON_LIB_PATH   C:\Python23\libs
>  
> I then tried to build Boost.Python directly from
> C:\Boost_1_32_0\libs\python\build.  However, bjam didn't know how to
> build from this directory.


The directory is not the issue.  My guess is that you are trying to
use a Cygwin GCC from a Windows command-prompt with a Windows (not
Cygwin) installation of Python?

You either need to use a MinGW compiler (or gcc-nocygwin) or use a
Cygwin-based python.
See
http://www.boost.org/libs/python/doc/building.html#cygwin_configuration


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



From Ben.Cain at us.army.mil  Thu Apr 28 00:45:23 2005
From: Ben.Cain at us.army.mil (Cain, Ben (Contractor-Summit))
Date: Wed, 27 Apr 2005 17:45:23 -0500
Subject: [C++-sig] Re: Newbie ... trouble compiling Boost.Python
Message-ID: <6DB8567B5D54F443982465B984BE4AA99B2B1F@amr-ex5.ds.amrdec.army.mil>

David,

I am trying to build from Windows command-prompt.  However, I think I'm
using MinGW's gcc.

gcc -v yields

Reading specs from C:/MinGW/bin/../lib/gcc-lib/mingw32/3.2.3/specs
Configured with: ../gcc/configure --with-gcc --with-gnu-ld --with-gnu-as
--host=mingw32 --target=mingw32 --prefix=/mingw --enable-threads
--disable-nls --enable-languages=c++,f77,objc --disable-win32-registry
--disable-shared --enable-sjlj-exceptions
Thread model: win32
gcc version 3.2.3 (mingw special 20030504-1)

I downloaded Python from ...

http://www.python.org/ftp/python/2.3.5/Python-2.3.5.exe

I downloaded Boost  from ...

http://prdownloads.sourceforge.net/boost/boost_1_32_0.exe?download

Thanks,
  Ben

David Abrahams wrote:
> 
> My guess is that you are trying to
> use a Cygwin GCC from a Windows command-prompt with a Windows (not
> Cygwin) installation of Python?
> 
> You either need to use a MinGW compiler (or gcc-nocygwin) or use a
> Cygwin-based python.


From nyamatongwe at gmail.com  Thu Apr 28 01:19:31 2005
From: nyamatongwe at gmail.com (Neil Hodgson)
Date: Thu, 28 Apr 2005 09:19:31 +1000
Subject: [C++-sig] Visual C++ 2005
Message-ID: <50862ebd050427161953cfc3ab@mail.gmail.com>

   Does anyone know if compilation of Boost.Python based code works
better with Visual C++ 2005 compared to the 2003 version? Visual C++
2003 is OK but requires breaking wrappers up into multiple files and
compilation can be very slow and memory-intensive.

   How can Boost.Jam be instructed to look for Visual C++ 2005? Both
versions of the compiler are installed here and Boost.Jam always finds
2003. A beta of the express edition of Visual C++ 2005 is a free
download, IIRC about 60 Megs.

   Neil


From blast2001 at 21cn.com  Thu Apr 28 03:47:35 2005
From: blast2001 at 21cn.com (Martin Dai)
Date: Thu, 28 Apr 2005 09:47:35 +0800
Subject: [C++-sig] Re: Pass C++ object to python  issue
References: <d4n1j3$qpq$1@sea.gmane.org> <d4neu4$qas$1@sea.gmane.org>
Message-ID: <d4pf3t$k8p$1@sea.gmane.org>

Hi  Volodya,

    Thanks for your great help.  I have tried those two files(those are very
visual  :).
But  just as you mentioned:      is  it  possible  to  register  the
corresponding  Python class  for  custom C++  class   by   an  explicit
way
before  Invoke  PyRun_String?

Best Regards,
Martin


"Vladimir Prus" <ghost at cs.msu.su> ???? news:d4neu4$qas$1 at sea.gmane.org...
> Martin Dai wrote:
>
> > Hi  everyone,
> >
> >     Now  I  have solved the pass variable when the variable is simple
> >     type,
> > but when I bind the C++ class object into the dictionary,  it compile
> > successfully,  it will abort once execute it.
>
> With what error message?
>
> > The below is the c++ source code:
> >
> >     if (PyImport_AppendInittab("embedded_hello", initembedded_hello) ==
> >     -1)
>
> Please always provide a complete program.
>
> >     main_namespace["AppBase"] = python::object(python::ptr(pBase));
> > //abort here
> >
> >     python::handle<> result(
> >         PyRun_String(
> >             "from embedded_hello import *        \n"
> >             "print AppBase.i \n"
> >         Py_file_input, main_namespace.ptr(), main_namespace.ptr())
> >         );
>
> I've tried to reproduce this, and got:
>
>     http://zigzag.cs.msu.su/~ghost/e3.cpp
>
> that works fine. However, when I move the
>
>     main_namespace["context"] = ptr(b);
>
> above call to PyRun_String (as is done in your example),
>
>     http://zigzag.cs.msu.su/~ghost/e3f.cpp
>
> I get this:
>
>    TypeError: No Python class registered for C++ class B
>
> which looks quite reasonable. Python has no idea that class 'B' even
exists,
> but you're trying to add instance of that class. I wonder if it's possible
> to automatically register Python class in this case, though.
>
> - Volodya





From paniq at paniq.org  Thu Apr 28 19:54:15 2005
From: paniq at paniq.org (Leonard "paniq" Ritter)
Date: Thu, 28 Apr 2005 19:54:15 +0200
Subject: [C++-sig] virtual overrides and custom smart pointers
In-Reply-To: <42625E34.7000408@paniq.org>
References: <42625E34.7000408@paniq.org>
Message-ID: <42712347.6090902@paniq.org>

the problem still persists. does anyone have a solution?

Leonard "paniq" Ritter wrote:

> hi,
>
> i'm running into a problem i'm trying to fix since several days 
> without much luck. it seems my custom smart pointer class becomes 
> worthless when wrapper<> is used to construct an object whose methods 
> can be overridden in python.
>
> wrapper export code:
>
>            class_<WindowEventsWrap, bases<Mu::Interface::Unknown>, 
> Ptr<WindowEventsWrap>, boost::noncopyable>("WindowEvents")
>                
> .def("OnWidgetEvent",pure_virtual(&WindowEvents::OnWidgetEvent))
>                
> .def("WidgetEventHandled",pure_virtual(&WindowEvents::WidgetEventHandled)) 
>
>            ;
>
> wrapper declaration:
>
>        struct WindowEventsWrap
>            :    InterfaceImplementation<WindowEventsWrap, 
> Implements<Mu::Sandoz::Interface::WindowEvents> >,
>                wrapper< Mu::Sandoz::Interface::WindowEvents >
>
> python test code:
>
> class MyWindowEvents(Sandoz.WindowEvents):
>    def OnWidgetEvent(self, event):
>        print event.type
>        return Sandoz.Result.NotImplemented
>      def WidgetEventHandled(self, event, widget):
>        print event.type, widget
>        return Sandoz.Result.NotImplemented
>
> def Main():
>    mwe = MyWindowEvents()
>    view = Sandoz.ViewFrame("?.Sandoz.ViewFrame")
>    view.Connect(mwe)
>    del mwe
>
> connect adds a reference to the mwe callback object and view calls 
> that object later on. this is known to work. however when i delete mwe 
> right after Connect(), the object itself persists, but its overridden 
> methods (and i guess the underlying python object as well) are gone. 
> no Release or AddReference on my object is called from wrapper_base, 
> so i assume it is not using my custom smart pointer class which deals 
> with this stuff.
>
> can someone suggest a solution to this problem?
>
> best regards,
> leonard
>
> _______________________________________________
> C++-sig mailing list
> C++-sig at python.org
> http://mail.python.org/mailman/listinfo/c++-sig
>
>



From dave at boost-consulting.com  Thu Apr 28 23:09:21 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 28 Apr 2005 17:09:21 -0400
Subject: [C++-sig] Re: Pointer to existing Python object
References: <425B363E.8090905@imb-jena.de>
Message-ID: <uwtqmobge.fsf@boost-consulting.com>

Andreas Beyer <beyer at imb-jena.de> writes:

> Hi:
>
> I know that there has been some discussion on this subject on c++ sig. 
> However, I could not find a solution for my specific problem and maybe 
> there is no solution.
>
> Here we go:
>
> I have a container class B that creates instances of class A - both
> in C++. In Python I can get instances of A from B with a get()
> method. I would like to get an existing Python object whenever such
> object existis and I would like to get a new Python object if no
> Python object for A exists.

This is hard to do.  I see that you're using shared_ptr and that's a
good start.  What you need to do, the moment a shared_ptr is about to
be released to Python, is replace it:

   template <class T>
   hold_python(shared_ptr<T>& x)
   {
       x = extract<shared_ptr<T> >( python::object(x) );
   }

> An example:
>
>
> using namespace boost;
>
> class A : public enable_shared_from_this<A> {
>  public:
>     A() : val(0) {};
>     int val;
>     typedef shared_ptr<A> A_ptr;
>     A_ptr self() {
>        A_ptr self;
>        self = shared_from_this();
>        return self;
>      }
>
> };
>
>
> class B {
>    public:
>      B() {
>           a = A::A_ptr(new A());
>      }
>      A::A_ptr a;
>      void set(A::A_ptr a) {
>          this->a = a;
>      }
>
>      A::A_ptr get() {
>          return a->self();

What was wrong with 

          return a;

here?

>      }
> };
>

A::A_ptr get_b_a(B& b)
{
    hold_python(b.a);
    return b.a;
}

> BOOST_PYTHON_MODULE(ptr_test) {
>  python::class_<A> ("A")
>    .def_readwrite("val", &A::val)
>  ;
>  python::register_ptr_to_python< A::A_ptr >();
>
>  python::class_<B>("B")
>     .def("set", &B::set)
>     .def("get", &B::get)

      .def("get", get_b_a) // <=== here
>  ;
> }
>
>
> In Python I would like to do this:
>
>>>> b = B()
>>>> a1 = b.get() # obtain instance of class A
>>>> a2 = b.get() # get the *same Python instance* again
>>>> assert(a1==a2)
> Traceback (most recent call last):
>  File "<stdin>", line 1, in ?
> AssertionError
>>>>
>
> The internal representations of instances of A are the same! Whenever I 
> access a C++ member of A I am actually accessing the same object.  So I 
> can do this:
>>>> a1.val = 42 # change internal reference
>>>> print a2.val
> 42
>>>>
> However, I cannot do this:
>>>> a1.foo = 42
>>>> print a2.foo
> Traceback (most recent call last):
>  File "<stdin>", line 1, in ?
> AttributeError: 'A' object has no attribute 'foo'
>>>>
>
> I think instances of B have to check if there is already an existing 
> Python object. In that case  B (or a wrapper of B) should return the 
> existing Python object (i.e. instances of python::object). Is there a 
> way to find such an existing instance?

HTH,

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



From dave at boost-consulting.com  Thu Apr 28 16:51:47 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 28 Apr 2005 10:51:47 -0400
Subject: [C++-sig] Re: Visual C++ 2005
References: <50862ebd050427161953cfc3ab@mail.gmail.com>
Message-ID: <usm1bhs3g.fsf@boost-consulting.com>

Neil Hodgson <nyamatongwe at gmail.com> writes:

>    Does anyone know if compilation of Boost.Python based code works
> better with Visual C++ 2005 compared to the 2003 version? Visual C++
> 2003 is OK but requires breaking wrappers up into multiple files and
> compilation can be very slow and memory-intensive.
>
>    How can Boost.Jam be instructed to look for Visual C++ 2005? 

With the latest Boost CVS,

   myjam -sTOOLS=vc-8_0 whatever

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



From dave at boost-consulting.com  Thu Apr 28 15:35:07 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 28 Apr 2005 09:35:07 -0400
Subject: [C++-sig] Re: Newbie ... trouble compiling Boost.Python
References: <6DB8567B5D54F443982465B984BE4AA99B2B1F@amr-ex5.ds.amrdec.army.mil>
Message-ID: <umzrjkos4.fsf@boost-consulting.com>

"Cain, Ben (Contractor-Summit)" <Ben.Cain at us.army.mil> writes:

> David,
>
> I am trying to build from Windows command-prompt.  However, I think I'm
> using MinGW's gcc.
>
> gcc -v yields
>
> Reading specs from C:/MinGW/bin/../lib/gcc-lib/mingw32/3.2.3/specs
> Configured with: ../gcc/configure --with-gcc --with-gnu-ld --with-gnu-as
> --host=mingw32 --target=mingw32 --prefix=/mingw --enable-threads
> --disable-nls --enable-languages=c++,f77,objc --disable-win32-registry
> --disable-shared --enable-sjlj-exceptions
> Thread model: win32
> gcc version 3.2.3 (mingw special 20030504-1)
>
> I downloaded Python from ...
>
> http://www.python.org/ftp/python/2.3.5/Python-2.3.5.exe
>
> I downloaded Boost  from ...
>
> http://prdownloads.sourceforge.net/boost/boost_1_32_0.exe?download

In that case you need to use -sTOOLS=mingw.
http://www.boost.org/more/getting_started.html#Tools shows that
-sTOOLS=gcc is reserved for Unix and Cygwin.

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



From ustc9305 at yahoo.com  Thu Apr 28 23:17:41 2005
From: ustc9305 at yahoo.com (jianbing wu)
Date: Thu, 28 Apr 2005 14:17:41 -0700 (PDT)
Subject: [C++-sig] for python library file
Message-ID: <20050428211741.42216.qmail@web51109.mail.yahoo.com>

Hi guys,
I am currently write a C++ program which will use the
"python24.lib", but when I compiled my program under
the debug version, the .Net compiler complains not
finding "pythod24_d.lib" (as I knew this library file
is only used for debugging purpose). But I cannot find
it in the python installation folder. Would someone
please give me this file or tell me how to create this
file?

Thank you,
Ivory

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


From Ben.Cain at us.army.mil  Fri Apr 29 04:24:22 2005
From: Ben.Cain at us.army.mil (Cain, Ben (Contractor-Summit))
Date: Thu, 28 Apr 2005 21:24:22 -0500
Subject: [C++-sig] Re: Newbie ... trouble compiling Boost.Python
Message-ID: <6DB8567B5D54F443982465B984BE4AA99B2B26@amr-ex5.ds.amrdec.army.mil>

David,

Thank you for your patience with a Python newbie.  I hope to contribute
to this mailing list as I come up to speed.

Many Thanks,
  Ben

David Abrahams wrote:
> 
> In that case you need to use -sTOOLS=mingw.
> http://www.boost.org/more/getting_started.html#Tools shows that
> -sTOOLS=gcc is reserved for Unix and Cygwin.
.org/mailman/listinfo/c++-sig
[Ben Cain wrote:] 



From sascha at askja.de  Fri Apr 29 08:36:38 2005
From: sascha at askja.de (Sascha Jensen)
Date: Fri, 29 Apr 2005 08:36:38 +0200
Subject: [C++-sig] for python library file
In-Reply-To: <20050428211741.42216.qmail@web51109.mail.yahoo.com>
References: <20050428211741.42216.qmail@web51109.mail.yahoo.com>
Message-ID: <4271D5F6.2040607@askja.de>



jianbing wu wrote:
> Hi guys,
> I am currently write a C++ program which will use the
> "python24.lib", but when I compiled my program under
> the debug version, the .Net compiler complains not
> finding "pythod24_d.lib" (as I knew this library file
> is only used for debugging purpose). But I cannot find
> it in the python installation folder. Would someone
> please give me this file or tell me how to create this
> file?

in pyconfig.h is defined to use python24_d.lib


#ifdef MS_COREDLL
#	ifndef Py_BUILD_CORE /* not building the core - must be an ext */
#		if defined(_MSC_VER)
			/* So MSVC users need not specify the .lib file in
			their Makefile (other compilers are generally
			taken care of by distutils.) */
#			ifdef _DEBUG
#				pragma comment(lib,"python24_d.lib")
#			else
#				pragma comment(lib,"python24.lib")
#			endif /* _DEBUG */
#		endif /* _MSC_VER */
#	endif /* Py_BUILD_CORE */
#endif /* MS_COREDLL */


i used this inside my header to undefine _DEBUG for the Python lib, 
because i did not need debugging

#ifdef _DEBUG
#undef _DEBUG
#include "Python.h"
#define _DEBUG
#else
#include "Python.h"
#endif

probably there is a more elegant way, but this works fine for me,
Sascha



From dave at boost-consulting.com  Fri Apr 29 16:40:25 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Fri, 29 Apr 2005 10:40:25 -0400
Subject: [C++-sig] Re: virtual overrides and custom smart pointers
References: <42625E34.7000408@paniq.org> <42712347.6090902@paniq.org>
Message-ID: <ur7gtir3a.fsf@boost-consulting.com>

"Leonard \"paniq\" Ritter" <paniq at paniq.org> writes:

>> can someone suggest a solution to this problem?

Post a minimal complete test case and I might be able to help.

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



From jaehokim1 at yahoo.com  Fri Apr 29 17:19:25 2005
From: jaehokim1 at yahoo.com (jaeho kim)
Date: Fri, 29 Apr 2005 08:19:25 -0700 (PDT)
Subject: [C++-sig] Need to understand the following Python Program- Help me..
Message-ID: <20050429151925.32787.qmail@web51409.mail.yahoo.com>

Hi,,,

I am very interested about the following Python program, but I am
very beginner on Python.. I do not understand this algorithm,,

I would appreciated if you give me this algorithm to some other
popular programming language.


filename="Karp.py"




from __future__ import nested_scopes

import bisect

class _Num:
    def __init__(self, value, index):
        self.value = value
        self.i = index

    def __lt__(self, other):
        return self.value < other.value

# This implements the Karmarkar-Karp heuristic for partitioning a set
# in two, i.e. into two disjoint subsets s.t. their sums are
# approximately equal.  It produces only one result, in O(N*log N)
# time.  A remarkable property is that it loves large sets:  in
# general, the more numbers you feed it, the better it does.

class Partition:
    def __init__(self, nums):
        self.nums = nums
        sorted = [_Num(nums[i], i) for i in range(len(nums))]
        sorted.sort()
        self.sorted = sorted

    def run(self):
        sorted = self.sorted[:]
        N = len(sorted)
        connections = [[] for i in range(N)]

        while len(sorted) > 1:
            bigger  = sorted.pop()
            smaller = sorted.pop()

            # Force these into different sets, by "drawing a
            # line" connecting them.
            i, j = bigger.i, smaller.i
            connections[i].append(j)
            connections[j].append(i)

            diff = bigger.value - smaller.value
            assert diff >= 0
            bisect.insort(sorted, _Num(diff, i))

        # Now sorted contains only 1 element x, and x.value is
        # the difference between the subsets' sums.

        # Theorem:  The connections matrix represents a spanning tree
        # on the set of index nodes, and any tree can be 2-colored.
        # 2-color this one (with "colors" 0 and 1).

        index2color = [None] * N

        def color(i, c):
            if index2color[i] is not None:
                assert index2color[i] == c
                return
            index2color[i] = c
            for j in connections[i]:
                color(j, 1-c)

        color(0, 0)

        # Partition the indices by their colors.
        subsets = [[], []]
        for i in range(N):
            subsets[index2color[i]].append(i)

        return subsets

N = 50
import math
x = [math.sqrt(i) for i in range(1, N+1)]
p = Partition(x)
s, t = p.run()
sum1 = 0L
sum2 = 0L
for i in s:
    sum1 += x[i]
for i in t:
    sum2 += x[i]
print "Set 1 sum", repr(sum1)
print "Set 2 sum", repr(sum2)
print "difference", repr(abs(sum1 - sum2))


Thanks for you help...

Larry Jaeho Kim


Larry Jaeho Kim

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


From MKhesin at liquidnet.com  Fri Apr 29 17:25:24 2005
From: MKhesin at liquidnet.com (Max Khesin)
Date: Fri, 29 Apr 2005 11:25:24 -0400
Subject: [C++-sig] Need to understand the following Python Program- He
	lp me..
Message-ID: <4B6FB7F60D37D41188C300B0D022E0C0039653FD@exchange.lnholdings.com>

Wrong forum - this one is dedicated to Python/C++ binding. Please try
comp.lang.python on Usenet, or better yet google "karp algorithm" - I am
sure you will find many implementations.

max

-----Original Message-----
From: jaeho kim [mailto:jaehokim1 at yahoo.com]
Sent: Friday, April 29, 2005 11:19 AM
To: c++-sig at python.org
Subject: [C++-sig] Need to understand the following Python Program- Help
me..

Hi,,,

I am very interested about the following Python program, but I am
very beginner on Python.. I do not understand this algorithm,,

I would appreciated if you give me this algorithm to some other
popular programming language.


filename="Karp.py"




from __future__ import nested_scopes

import bisect

class _Num:
    def __init__(self, value, index):
        self.value = value
        self.i = index

    def __lt__(self, other):
        return self.value < other.value

# This implements the Karmarkar-Karp heuristic for partitioning a set
# in two, i.e. into two disjoint subsets s.t. their sums are
# approximately equal.  It produces only one result, in O(N*log N)
# time.  A remarkable property is that it loves large sets:  in
# general, the more numbers you feed it, the better it does.

class Partition:
    def __init__(self, nums):
        self.nums = nums
        sorted = [_Num(nums[i], i) for i in range(len(nums))]
        sorted.sort()
        self.sorted = sorted

    def run(self):
        sorted = self.sorted[:]
        N = len(sorted)
        connections = [[] for i in range(N)]

        while len(sorted) > 1:
            bigger  = sorted.pop()
            smaller = sorted.pop()

            # Force these into different sets, by "drawing a
            # line" connecting them.
            i, j = bigger.i, smaller.i
            connections[i].append(j)
            connections[j].append(i)

            diff = bigger.value - smaller.value
            assert diff >= 0
            bisect.insort(sorted, _Num(diff, i))

        # Now sorted contains only 1 element x, and x.value is
        # the difference between the subsets' sums.

        # Theorem:  The connections matrix represents a spanning tree
        # on the set of index nodes, and any tree can be 2-colored.
        # 2-color this one (with "colors" 0 and 1).

        index2color = [None] * N

        def color(i, c):
            if index2color[i] is not None:
                assert index2color[i] == c
                return
            index2color[i] = c
            for j in connections[i]:
                color(j, 1-c)

        color(0, 0)

        # Partition the indices by their colors.
        subsets = [[], []]
        for i in range(N):
            subsets[index2color[i]].append(i)

        return subsets

N = 50
import math
x = [math.sqrt(i) for i in range(1, N+1)]
p = Partition(x)
s, t = p.run()
sum1 = 0L
sum2 = 0L
for i in s:
    sum1 += x[i]
for i in t:
    sum2 += x[i]
print "Set 1 sum", repr(sum1)
print "Set 2 sum", repr(sum2)
print "difference", repr(abs(sum1 - sum2))


Thanks for you help...

Larry Jaeho Kim


Larry Jaeho Kim

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
_______________________________________________
C++-sig mailing list
C++-sig at python.org
http://mail.python.org/mailman/listinfo/c++-sig
This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the system manager. This message contains confidential information and is intended only for the individual named. If you are not the named addressee you should not disseminate, distribute or copy this e-mail.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20050429/21dd6ac9/attachment.htm>

From nicodemus at esss.com.br  Fri Apr 29 23:05:43 2005
From: nicodemus at esss.com.br (Nicodemus)
Date: Fri, 29 Apr 2005 19:05:43 -0200
Subject: [C++-sig] Re: Getting the pointer for a member operator
In-Reply-To: <d4lhal$a9q$1@sea.gmane.org>
References: <426D6B14.9040309@esss.com.br> <d4lhal$a9q$1@sea.gmane.org>
Message-ID: <4272A1A7.8000401@esss.com.br>

Vladimir Sukhoy wrote:

>>Problem is that this is broken under MSVC.NET 2003. The compiler complains 
>>with:
>>
>I tried similar with .NET 2003 compiler - it works, and I cannot reproduce 
>this error! The full module code is:
>
><snip code>
>

Thanks for trying it Vladimir! Indeed the above code does work. Since 
the error message is pretty specific, I assumed the error was on that 
line. If I tried to post a reproductible test case, as I should, I would 
learn that the error only shows in the following code:

//----------------------------------------

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

namespace operators
{
    struct C
    {
        operator int() const { return 1; }
        operator const char*() { return "C"; }
    };
}
BOOST_PYTHON_MODULE(_ace)
{
    class_< operators::C >("C", init< >())
        .def("__int__", &operators::C::operator int) // comment this line
        .def("__str__", &operators::C::operator const char*)
    ;
}

//----------------------------------------

If you comment the marked line, the code compiles fine. Otherwise, it 
gives this error:

op2.cpp(16) : error C2833: 'operator const' is not a recognized operator 
or type

Does anyone know what is going on?


From nicodemus at esss.com.br  Fri Apr 29 23:09:27 2005
From: nicodemus at esss.com.br (Nicodemus)
Date: Fri, 29 Apr 2005 19:09:27 -0200
Subject: [C++-sig] Re: Getting the pointer for a member operator
In-Reply-To: <ud5sguk6u.fsf@boost-consulting.com>
References: <426D6B14.9040309@esss.com.br> <ud5sguk6u.fsf@boost-consulting.com>
Message-ID: <4272A287.1080802@esss.com.br>

David Abrahams wrote:

>Nicodemus <nicodemus at esss.com.br> writes:
>
>  
>
>>But without success. Does anybody know how I can fix this?
>>    
>>
>
>Did you try a typedef?
>
>    typedef char const* string;
>
>    ...
>
>         .def("__str__", &operators::C::operator string)
>
>  
>

Hi Dave,

This would work, but I would like to avoid that for Pyste. I believe its 
just a matter of syntax, hence easily solvable. I googled around a bit, 
but without success. Please take a look at my post for Valdimir, I 
posted a reproductible test case there. Perhaps you can shed some light 
in the matter?

Thanks,
Bruno.



From dave at boost-consulting.com  Sat Apr 30 01:53:54 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Fri, 29 Apr 2005 19:53:54 -0400
Subject: [C++-sig] Re: Need to understand the following Python Program- Help
	me..
References: <20050429151925.32787.qmail@web51409.mail.yahoo.com>
Message-ID: <uu0lpta0d.fsf@boost-consulting.com>

jaeho kim <jaehokim1 at yahoo.com> writes:

> Hi,,,
>
> I am very interested about the following Python program, but I am
> very beginner on Python.. I do not understand this algorithm,,

You're in the wrong forum for pure Python questions; this list is
about Python/C++ integration.  I suggest you try comp.lang.python
(http://groups-beta.google.com/group/comp.lang.python)

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



From dave at boost-consulting.com  Sat Apr 30 01:53:54 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Fri, 29 Apr 2005 19:53:54 -0400
Subject: [C++-sig] Re: Need to understand the following Python Program- Help
	me..
References: <20050429151925.32787.qmail@web51409.mail.yahoo.com>
Message-ID: <uu0lpta0d.fsf@boost-consulting.com>

jaeho kim <jaehokim1 at yahoo.com> writes:

> Hi,,,
>
> I am very interested about the following Python program, but I am
> very beginner on Python.. I do not understand this algorithm,,

You're in the wrong forum for pure Python questions; this list is
about Python/C++ integration.  I suggest you try comp.lang.python
(http://groups-beta.google.com/group/comp.lang.python)

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



From dave at boost-consulting.com  Sat Apr 30 03:25:37 2005
From: dave at boost-consulting.com (David Abrahams)
Date: Fri, 29 Apr 2005 21:25:37 -0400
Subject: [C++-sig] Re: Getting the pointer for a member operator
References: <426D6B14.9040309@esss.com.br>
	<ud5sguk6u.fsf@boost-consulting.com> <4272A287.1080802@esss.com.br>
Message-ID: <uekctrr72.fsf@boost-consulting.com>

Nicodemus <nicodemus at esss.com.br> writes:

> David Abrahams wrote:
>
>>Nicodemus <nicodemus at esss.com.br> writes:
>>
>>  
>>
>>>But without success. Does anybody know how I can fix this?
>>>    
>>>
>>
>>Did you try a typedef?
>>
>>    typedef char const* string;
>>
>>    ...
>>
>>         .def("__str__", &operators::C::operator string)
>>
>>  
>>
>
> Hi Dave,
>
> This would work, but I would like to avoid that for Pyste. 

I don't know why.

> I believe its just a matter of syntax, hence easily solvable. I
> googled around a bit, but without success. Please take a look at my
> post for Valdimir, I posted a reproductible test case there. Perhaps
> you can shed some light in the matter?

No, I've already supplied one workaround.  I don't know why it's not
good enough for you.

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



From news4vovan at mail.ru  Sat Apr 30 12:36:33 2005
From: news4vovan at mail.ru (Vladimir Sukhoy)
Date: Sat, 30 Apr 2005 13:36:33 +0300
Subject: [C++-sig] Re: Re: Getting the pointer for a member operator
References: <426D6B14.9040309@esss.com.br> <d4lhal$a9q$1@sea.gmane.org>
	<4272A1A7.8000401@esss.com.br>
Message-ID: <d4vmoq$2uv$1@sea.gmane.org>

> Does anyone know what is going on?

Well, I dont know, but if you do something like that:

BOOST_PYTHON_MODULE(_ace)
{
    class_< operators::C > ref("C", init< >());
    ref.def("__int__", &operators::C::operator int);
    ref.def("__str__", &operators::C::operator const char*);
}

it compiles successfully. In fact, it looks like lexer problem.






From nicodemus at esss.com.br  Sat Apr 30 21:52:16 2005
From: nicodemus at esss.com.br (Nicodemus)
Date: Sat, 30 Apr 2005 16:52:16 -0300
Subject: [C++-sig] Re: Getting the pointer for a member operator
In-Reply-To: <uekctrr72.fsf@boost-consulting.com>
References: <426D6B14.9040309@esss.com.br>	<ud5sguk6u.fsf@boost-consulting.com>
	<4272A287.1080802@esss.com.br> <uekctrr72.fsf@boost-consulting.com>
Message-ID: <4273E1F0.8030402@esss.com.br>

David Abrahams wrote:

>No, I've already supplied one workaround.  I don't know why it's not
>good enough for you.
>

Fine, thanks for the help.

Regards,
Bruno.