From vivh.29 at hotmail.fr  Sun Nov  1 12:25:01 2015
From: vivh.29 at hotmail.fr (Vivien Henry)
Date: Sun, 1 Nov 2015 18:25:01 +0100
Subject: [C++-sig] Accessing pointer properties with boost-python ?
Message-ID: <DUB119-W43AC71BA804F2591E7CC669E2D0@phx.gbl>







Hello everyone,
I am new on this mailing-list, so; first, 
My name is Vivien Henry, working as Electrical Engineer in France

I am currently trying to wrap a CAD software, gerbv (http://gerbv.geda-project.org/), which is designed to handle Gerber files. (They also provide it as a library)

the file wrapper.cpp is generated through a personal script

The script is still a work in progress.
The generated code compiles, creating objects looks ok, accessing their attributes (numbers.. enums) as well.

BUT, the problem is when i want to access a property which is actually a pointer...

For instance, with this script:

>>> from gerbv import *
>>> 
>>> 
>>> 
>>> i = gerbv_instruction_t()
>>> i
<gerbv.gerbv_instruction_t object at 0x7f655abe6788>
>>> i.opcode
gerbv.gerbv_opcodes_t.GERBV_OPCODE_NOP
>>> a = gerbv_amacro_t()
>>> a
<gerbv.gerbv_amacro_t object at 0x7f655abe67e0>
>>> a.program
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
Boost.Python.ArgumentError: Python argument types in
    None.None(gerbv_amacro_t)
did not match C++ signature:
    None(amacro {lvalue})
>>> 


I can't figure out what is happening here !

(you ll see that i am creating a proxy to access string and char array...)

Any help is welcome :D
Thanks a lot, 


Vivien Henry


here the specific code for this error, which is in the file wrapper.cpp

class gerbv_amacro_t_proxy: public gerbv_amacro_t{
    public:



        std::string get_name(){
            std::string s(name);
            return s;
        }

        void set_name(std::string s){
            name = (gchar*)s.c_str();
        }


        gerbv_amacro_t_proxy() {

        }


};




// in boost_python macro... l655

class_<gerbv_amacro_t_proxy, boost::shared_ptr<gerbv_amacro_t_proxy> >("gerbv_amacro_t")
        //Methods
        //Properties
        .add_property("name", &gerbv_amacro_t_proxy::get_name, &gerbv_amacro_t_proxy::set_name)
        .add_property("program", make_getter(&gerbv_amacro_t_proxy::program, return_internal_reference<>()), make_setter(&gerbv_amacro_t_proxy::program, return_internal_reference<>()))
        .def_readwrite("nuf_push", &gerbv_amacro_t_proxy::nuf_push)
        .add_property("next", make_getter(&gerbv_amacro_t_proxy::next, return_internal_reference<>()), make_setter(&gerbv_amacro_t_proxy::next, return_internal_reference<>()))
        ;


// in boost python macro l918
class_<gerbv_instruction_t, boost::shared_ptr<gerbv_instruction_t> >("gerbv_instruction_t")
        //Methods
        //Properties
        .def_readwrite("opcode", &gerbv_instruction_t::opcode)
        .def_readwrite("data", &gerbv_instruction_t::data)
        .add_property("next", make_getter(&gerbv_instruction_t::next, return_internal_reference<>()), make_setter(&gerbv_instruction_t::next, return_internal_reference<>()))
        ;



 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20151101/08b88495/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: wrapper.cpp
Type: text/x-c
Size: 46604 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20151101/08b88495/attachment-0001.bin>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: gerbv.h
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20151101/08b88495/attachment-0001.h>

From zhanli1986 at gmail.com  Tue Nov  3 17:13:19 2015
From: zhanli1986 at gmail.com (Zhan Li)
Date: Tue, 3 Nov 2015 17:13:19 -0500
Subject: [C++-sig] Expose a C++ function with an argument of template vector
 to Python: Boost.Python.ArgumentError
Message-ID: <CAAMT6JM4wfu8i0gWPJMfGgTsPwv19EetdWczqaNU-dgQrq384g@mail.gmail.com>

Greetings,

I'm new to Boost.Python. I have a C++ library (written and shipped by
another party) that I want to wrap into Python modules so that I can use
several functions in Python. One function takes an argument that is a C++
template vector. How can I export this function to a python function?

In a nutshell, here is what I'm doing now but without luck. I wrap the
template vector into a Python class using
Boost.Python's vector_indexing_suite. Then in the Python script, I first
declare this Python class and then pass it to the function. But I got no
luck and the following error message,

"
Boost.Python.ArgumentError: Python argument types in
    canupo.read_msc_header(MSCFile, FloatVec, int)
did not match C++ signature:
    read_msc_header(MSCFile {lvalue}, std::vector<float,
std::allocator<float> > {lvalue}, int {lvalue})
"

More specifically, here are the C++ code and Python script I have.

In the C++ library code file "helper.hpp", it has a struct MSCFile and a
function read_msc_header.

struct MSCFile
{
  // definition of MSCFile
};

int read_msc_header(MSCFile& mscfile, std::vector<FloatType>& scales, int&
ptn)
{
  // implementation of read_msc_header
}

The "FloatType" is just a alias for float.

In the C++ wrapper code file "wrapper.cpp"
#include "helpers.hpp"
#include <boost/python.hpp>
#include <boost/python/module.hpp>
#include <boost/python/def.hpp>
#include <boost/python/implicit.hpp>
#include <boost/python/suite/indexing/vector_indexing_suite.hpp>

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

  class_<std::vector<FloatType> >("FloatVec")
    .def(vector_indexing_suite<std::vector<FloatType> >())
    ;

  class_<MSCFile>("MSCFile", init<const char*>())
    ;

  def("read_msc_header", read_msc_header);
}

In the python script, I tried to use the function as following but failed.

import canupo
mscfile = canupo.MSCFile("msc_file_name")
scales = canupo.FloatVec()
ptnparams = 0
print canupo.read_msc_header(mscfile, scales, ptnparams)

And the error goes like this:

Boost.Python.ArgumentError: Python argument types in
    canupo.read_msc_header(MSCFile, FloatVec, int)
did not match C++ signature:
    read_msc_header(MSCFile {lvalue}, std::vector<float,
std::allocator<float> > {lvalue}, int {lvalue})

I googled around but didn't find a good solution or a solution that I can
understand well. Most requires rewrite the function with Boost.Python. But
I don't want to change the C++ library code itself. I think there may be a
good way to expose such a function to Python. Thanks a lot!

Zhan.

-- 
Zhan Li, ??;
Earth & Environment, Boston University, USA;

2008 - 2011, in Chinese Academy of Sciences, China;
2004 - 2008, in Nanjing University, China;
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20151103/3adc49d8/attachment.html>

From herron at ELLINGTON.com  Wed Nov  4 10:40:25 2015
From: herron at ELLINGTON.com (Liam Herron)
Date: Wed, 4 Nov 2015 15:40:25 +0000
Subject: [C++-sig] Expose a C++ function with an argument of template
 vector to Python: Boost.Python.ArgumentError
In-Reply-To: <CAAMT6JM4wfu8i0gWPJMfGgTsPwv19EetdWczqaNU-dgQrq384g@mail.gmail.com>
References: <CAAMT6JM4wfu8i0gWPJMfGgTsPwv19EetdWczqaNU-dgQrq384g@mail.gmail.com>
Message-ID: <96426055F003C542B3FB8A0928736DF9023A627EBA@ex0.ELLINGTON.com>

What happens if you do:

print canupo.read_msc_header(mscfile, scales, float(ptnparams))

?

From: Cplusplus-sig [mailto:cplusplus-sig-bounces+herron=ellington.com at python.org] On Behalf Of Zhan Li
Sent: Tuesday, November 03, 2015 5:13 PM
To: cplusplus-sig at python.org
Subject: [C++-sig] Expose a C++ function with an argument of template vector to Python: Boost.Python.ArgumentError

Greetings,

I'm new to Boost.Python. I have a C++ library (written and shipped by another party) that I want to wrap into Python modules so that I can use several functions in Python. One function takes an argument that is a C++ template vector. How can I export this function to a python function?

In a nutshell, here is what I'm doing now but without luck. I wrap the template vector into a Python class using Boost.Python's vector_indexing_suite. Then in the Python script, I first declare this Python class and then pass it to the function. But I got no luck and the following error message,

"
Boost.Python.ArgumentError: Python argument types in
    canupo.read_msc_header(MSCFile, FloatVec, int)
did not match C++ signature:
    read_msc_header(MSCFile {lvalue}, std::vector<float, std::allocator<float> > {lvalue}, int {lvalue})
"

More specifically, here are the C++ code and Python script I have.

In the C++ library code file "helper.hpp", it has a struct MSCFile and a function read_msc_header.

struct MSCFile
{
  // definition of MSCFile
};

int read_msc_header(MSCFile& mscfile, std::vector<FloatType>& scales, int& ptn)
{
  // implementation of read_msc_header
}

The "FloatType" is just a alias for float.

In the C++ wrapper code file "wrapper.cpp"
#include "helpers.hpp"
#include <boost/python.hpp>
#include <boost/python/module.hpp>
#include <boost/python/def.hpp>
#include <boost/python/implicit.hpp>
#include <boost/python/suite/indexing/vector_indexing_suite.hpp>

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

  class_<std::vector<FloatType> >("FloatVec")
    .def(vector_indexing_suite<std::vector<FloatType> >())
    ;

  class_<MSCFile>("MSCFile", init<const char*>())
    ;

  def("read_msc_header", read_msc_header);
}

In the python script, I tried to use the function as following but failed.

import canupo
mscfile = canupo.MSCFile("msc_file_name")
scales = canupo.FloatVec()
ptnparams = 0
print canupo.read_msc_header(mscfile, scales, ptnparams)

And the error goes like this:

Boost.Python.ArgumentError: Python argument types in
    canupo.read_msc_header(MSCFile, FloatVec, int)
did not match C++ signature:
    read_msc_header(MSCFile {lvalue}, std::vector<float, std::allocator<float> > {lvalue}, int {lvalue})

I googled around but didn't find a good solution or a solution that I can understand well. Most requires rewrite the function with Boost.Python. But I don't want to change the C++ library code itself. I think there may be a good way to expose such a function to Python. Thanks a lot!

Zhan.

--
Zhan Li, ??;
Earth & Environment, Boston University, USA;

2008 - 2011, in Chinese Academy of Sciences, China;
2004 - 2008, in Nanjing University, China;

=============================================================================================
Email transmissions can not be guaranteed to be secure or error-free, as information 
could be intercepted, corrupted, lost, destroyed, arrive late or incomplete, or contain 
viruses.  The sender therefore does not accept liability for any errors or omissions in 
the contents of this message which arise as a result of email transmission.  In addition,
the information contained in this email message is intended only for use of the 
individual or entity named above.  If the reader of this message is not the intended
recipient, or the employee or agent responsible to deliver it to the intended recipient,
you are hereby notified that any dissemination, distribution,or copying of this communication,
disclosure of the parties to it, or any action taken or omitted to be taken in reliance on it,
is strictly prohibited, and may be unlawful.  If you are not the intended recipient please
delete this email message. 
==============================================================================================
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20151104/75113c55/attachment-0001.html>

From afalanga at micron.com  Tue Nov 17 16:17:47 2015
From: afalanga at micron.com (Andy Falanga (afalanga))
Date: Tue, 17 Nov 2015 21:17:47 +0000
Subject: [C++-sig] C++ comments to docstrings
Message-ID: <564B9958.9030809@micron.com>

Hello,

My team produces a C++ library exported to python via Boost.Python. 
We've documented our functions and classes with C#-ish style comments.  
These are parse-able by Visual Studio and XML files are built from 
them.  When our shared object file is imported in python, a function 
opens these files and makes the association of the appropriate 
"docstring" in the XML file with the __doc__ property at runtime.

Now that you know how we do it currently, I'd like to know are there 
better methods in use by, or known to, those in this list?  The main 
hurdle to overcome at this point is that our library is built for both 
Windows and Linux.  Because the documentation method has been to use 
Visual Studio to build the XML, the Windows build must be done first and 
the XML left in some place that the Linux build can reach.  I'd like to 
eliminate this dependency for Linux.

Andy

From amohr at pixar.com  Thu Nov 19 13:06:40 2015
From: amohr at pixar.com (Alex Mohr)
Date: Thu, 19 Nov 2015 10:06:40 -0800
Subject: [C++-sig] C++ comments to docstrings
In-Reply-To: <564B9958.9030809@micron.com>
References: <564B9958.9030809@micron.com>
Message-ID: <564E0FB0.7090404@pixar.com>

Hi Andy,

We do something very similar to what you describe, except we use doxygen 
(www.doxygen.org) to output XML files that we use to generate __doc__ 
strings for our python bindings.  We're primarily a linux shop.

I don't know how hard it would be to switch to doxygen, maybe it can 
deal with C#-ish style comments?  But I believe it runs on Linux and 
Windows so perhaps it might help?

Alex

On 11/17/2015 1:17 PM, Andy Falanga (afalanga) wrote:
> Hello,
>
> My team produces a C++ library exported to python via Boost.Python.
> We've documented our functions and classes with C#-ish style comments.
> These are parse-able by Visual Studio and XML files are built from
> them.  When our shared object file is imported in python, a function
> opens these files and makes the association of the appropriate
> "docstring" in the XML file with the __doc__ property at runtime.
>
> Now that you know how we do it currently, I'd like to know are there
> better methods in use by, or known to, those in this list?  The main
> hurdle to overcome at this point is that our library is built for both
> Windows and Linux.  Because the documentation method has been to use
> Visual Studio to build the XML, the Windows build must be done first and
> the XML left in some place that the Linux build can reach.  I'd like to
> eliminate this dependency for Linux.
>
> Andy
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> https://mail.python.org/mailman/listinfo/cplusplus-sig
>


From afalanga at micron.com  Thu Nov 19 19:33:58 2015
From: afalanga at micron.com (Andy Falanga (afalanga))
Date: Fri, 20 Nov 2015 00:33:58 +0000
Subject: [C++-sig] C++ comments to docstrings
In-Reply-To: <564E0FB0.7090404@pixar.com>
References: <564B9958.9030809@micron.com> <564E0FB0.7090404@pixar.com>
Message-ID: <564E6A33.3090001@micron.com>

On 11/19/2015 11:06 AM, Alex Mohr wrote:
> Hi Andy,
>
> We do something very similar to what you describe, except we use 
> doxygen (www.doxygen.org) to output XML files that we use to generate 
> __doc__ strings for our python bindings.  We're primarily a linux shop.
>
> I don't know how hard it would be to switch to doxygen, maybe it can 
> deal with C#-ish style comments?  But I believe it runs on Linux and 
> Windows so perhaps it might help?
>
> Alex

Alex,

Knowing better how our system works, using doxygen may be a viable 
alternative.  I'll investigate this.  I've used it before on our 
codebase and it worked fairly well.  However, the output wasn't the same 
as the XML produced by VS.  At the time, I thought this was a 
showstopper.  Knowing more now, however, it might be possible to alter 
our parsing routines to evaluate the doxygen XML.  This would be grand 
since doxygen is pretty simple to install on Linux.

Thanks again,
Andy

From machang at princeton.edu  Mon Nov 23 17:45:21 2015
From: machang at princeton.edu (Michael Chang)
Date: Mon, 23 Nov 2015 17:45:21 -0500
Subject: [C++-sig] Compilation (?) Problem with boost.python
Message-ID: <CABFku2Gc1ZtXVHdF2zvAEi-O-VsXhSQnHqtLY0sbc9fGhbpo6Q@mail.gmail.com>

Hello,

I am running boost_1_59_0 on Ubuntu, and I bootstrapped/installed (using
b2) boost.python without any errors. When I run the quick start example, I
get the following error:

chachang at chachang-Lenovo-Y50-70:~/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/libs/python/example/quickstart$
bjam toolset=gcc --verbose-test test
warning: mismatched versions of Boost.Build engine and core
warning: Boost.Build engine (bjam) is 2011.12.01
warning: Boost.Build core (at
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src)
is 2015.07-git
Performing configuration checks

    - 32-bit                   : no
    - 64-bit                   : yes
    - arm                      : no
    - mips1                    : no
    - power                    : no
    - sparc                    : no
    - x86                      : yes
    - symlinks supported       : yes
*/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/tools/link.jam:398:
in link-recursively*
*ERROR: rule "READLINK" unknown in module "link".*
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/tools/link.jam:472:
in link.do-link-recursively
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/kernel/modules.jam:107:
in modules.call-in
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/util/indirect.jam:98:
in indirect.call
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:798:
in class at action.actualize
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:311:
in actualize-action
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:497:
in actualize-no-scanner
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:134:
in class at virtual-target.actualize
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:855:
in actualize-sources
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:780:
in class at action.actualize
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:311:
in actualize-action
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:497:
in actualize-no-scanner
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:134:
in class at virtual-target.actualize
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:819:
in actualize-source-type
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:840:
in actualize-sources
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:780:
in class at action.actualize
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:311:
in actualize-action
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:497:
in actualize-no-scanner
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:134:
in class at virtual-target.actualize
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:819:
in actualize-source-type
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:840:
in actualize-sources
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:780:
in class at action.actualize
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:311:
in actualize-action
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:497:
in actualize-no-scanner
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:134:
in class at virtual-target.actualize
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:819:
in actualize-source-type
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:840:
in actualize-sources
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:780:
in class at action.actualize
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:311:
in actualize-action
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:497:
in actualize-no-scanner
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build/virtual-target.jam:134:
in class at virtual-target.actualize
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/tools/build/src/build-system.jam:720:
in load
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/libs/python/example/quickstart/../../../../tools/build/src/kernel/modules.jam:295:
in import
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/libs/python/example/quickstart/../../../../tools/build/src/kernel/bootstrap.jam:139:
in boost-build
/home/chachang/Desktop/thesis/VeriFlow-v0.3/boost_1_59_0/libs/python/example/quickstart/boost-build.jam:7:
in module scope


Is there any insight to what could be causing this problem? I have
installed Boost at local prefix (not /usr/local/); where can I specify the
non-default prefix when I run/compile the code?

Thanks!

Michael Alan Chang
Princeton University '16
Computer Science
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20151123/e111e862/attachment.html>

From slide.o.mix at gmail.com  Tue Nov 24 18:30:05 2015
From: slide.o.mix at gmail.com (Slide)
Date: Tue, 24 Nov 2015 23:30:05 +0000
Subject: [C++-sig] Boost Python in dynamic libraries
Message-ID: <CAPiUgVfgHtXKEAHewF2fPFTza-y=APRCwNFQ22ib6-Ve1n5N6Q@mail.gmail.com>

I have a set of plugins that I am loading from shared libraries on Linux. I
thought I would want to call the initModule function when loading the
library, but I am getting the following error when I do so:

Fatal Python error: PyThreadState_Get: no current thread
Aborted (core dumped)

Can anyone point me to an example of something like this? I am using
BOOST_PYTHON_MODULE in my shared libraries and then loading them at
runtime. I would like to import that module into my main namespace where
most of my code is being run.

Thanks!

slide
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20151124/8e7f033a/attachment.html>

From giuseppe.corbelli at copangroup.com  Wed Nov 25 03:51:56 2015
From: giuseppe.corbelli at copangroup.com (Giuseppe Corbelli)
Date: Wed, 25 Nov 2015 09:51:56 +0100
Subject: [C++-sig] Boost Python in dynamic libraries
In-Reply-To: <CAPiUgVfgHtXKEAHewF2fPFTza-y=APRCwNFQ22ib6-Ve1n5N6Q@mail.gmail.com>
References: <CAPiUgVfgHtXKEAHewF2fPFTza-y=APRCwNFQ22ib6-Ve1n5N6Q@mail.gmail.com>
Message-ID: <565576AC.3020709@copangroup.com>

On 25/11/2015 00:30, Slide wrote:
> I have a set of plugins that I am loading from shared libraries on Linux. I
> thought I would want to call the initModule function when loading the library,
> but I am getting the following error when I do so:
>
> Fatal Python error: PyThreadState_Get: no current thread
> Aborted (core dumped)
>
> Can anyone point me to an example of something like this? I am using
> BOOST_PYTHON_MODULE in my shared libraries and then loading them at runtime. I
> would like to import that module into my main namespace where most of my code
> is being run.

If you have a look at the interpreter code (assuming 2.7 series), you'll see that

PyThreadState *
PyThreadState_Get(void)
{
     if (_PyThreadState_Current == NULL)
         Py_FatalError("PyThreadState_Get: no current thread");

     return _PyThreadState_Current;
}

PyThreadState *_PyThreadState_Current is defined in Python/pystate.c and 
initialized as NULL.

In Python/pythonrun.c within Py_InitializeEx() call a new thread state is 
initialized and PyThreadState_Swap() sets it.

In short: I'd say you didn't call Py_Initialize() before dlopen()ing the 
shared objects.

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

From slide.o.mix at gmail.com  Wed Nov 25 11:01:28 2015
From: slide.o.mix at gmail.com (Slide)
Date: Wed, 25 Nov 2015 16:01:28 +0000
Subject: [C++-sig] Boost Python in dynamic libraries
In-Reply-To: <565576AC.3020709@copangroup.com>
References: <CAPiUgVfgHtXKEAHewF2fPFTza-y=APRCwNFQ22ib6-Ve1n5N6Q@mail.gmail.com>
 <565576AC.3020709@copangroup.com>
Message-ID: <CAPiUgVf3ZBtoyVunyEHyH4oEnJGuc3CgLdSh6wwBTh8H7qTXdA@mail.gmail.com>

Wow, that was great! I appreciate your help. I indeed was not calling
Py_Initialize before loading the libraries. Much appreciated!

On Wed, Nov 25, 2015 at 7:40 AM Giuseppe Corbelli <
giuseppe.corbelli at copangroup.com> wrote:

> On 25/11/2015 00:30, Slide wrote:
> > I have a set of plugins that I am loading from shared libraries on
> Linux. I
> > thought I would want to call the initModule function when loading the
> library,
> > but I am getting the following error when I do so:
> >
> > Fatal Python error: PyThreadState_Get: no current thread
> > Aborted (core dumped)
> >
> > Can anyone point me to an example of something like this? I am using
> > BOOST_PYTHON_MODULE in my shared libraries and then loading them at
> runtime. I
> > would like to import that module into my main namespace where most of my
> code
> > is being run.
>
> If you have a look at the interpreter code (assuming 2.7 series), you'll
> see that
>
> PyThreadState *
> PyThreadState_Get(void)
> {
>      if (_PyThreadState_Current == NULL)
>          Py_FatalError("PyThreadState_Get: no current thread");
>
>      return _PyThreadState_Current;
> }
>
> PyThreadState *_PyThreadState_Current is defined in Python/pystate.c and
> initialized as NULL.
>
> In Python/pythonrun.c within Py_InitializeEx() call a new thread state is
> initialized and PyThreadState_Swap() sets it.
>
> In short: I'd say you didn't call Py_Initialize() before dlopen()ing the
> shared objects.
>
> --
>              Giuseppe Corbelli
> WASP Software Engineer, Copan Italia S.p.A
> Phone: +390303666318  Fax: +390302659932
> E-mail: giuseppe.corbelli at copangroup.com
> _______________________________________________
> Cplusplus-sig mailing list
> Cplusplus-sig at python.org
> https://mail.python.org/mailman/listinfo/cplusplus-sig
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cplusplus-sig/attachments/20151125/de0e9748/attachment.html>