align="left" dir="ltr">
/> class="539543310-28032008"> References: <1JfAyJ-0taLxI0@fwd35.aul.t-online.de><0130AB6AEBCC2045B3F203151123DA651CBB7B@SXLM-101.fg.rbc.com> <1JfCek-14PaL20@fwd27.aul.t-online.de> <0130AB6AEBCC2045B3F203151123DA651CBB7E@SXLM-101.fg.rbc.com> Message-ID: <1JfDuc-1z7EcS0@fwd26.aul.t-online.de> > >>> import TTBridge > Traceback (most recent call last): > File "", line 1, in ? > ImportError: No module named TTBridge > Are you sure your module is on the Python25 PYTHONPATH ? Yes, I tested it also. However, we start considering another approach to this problem. We can use Python 2.4 until we found a workaraound. That is a pity but unavoidable in these conditions (I don't have the time). Cheers Stefan P.S.: Sorry about the HTML junk, hope this message is plain text only! From nat at lindenlab.com Fri Mar 28 16:03:52 2008 From: nat at lindenlab.com (Nat Goodspeed) Date: Fri, 28 Mar 2008 11:03:52 -0400 Subject: [C++-sig] ActiveState Python 2.5.1.1, Boost 1.34.1 and MSVC 8 SP1 problem In-Reply-To: <1JfAyJ-0taLxI0@fwd35.aul.t-online.de> References: <1JfAyJ-0taLxI0@fwd35.aul.t-online.de> Message-ID: <47ED08D8.1030007@lindenlab.com> Stefan.Borovac at t-online.de wrote: > I'am currently working on small project which involves > a python programm which calls a C++ DLL using Boost Python. > > Everything worked fine using ActiveState Python 2.3.2-232 and ActiveState Python > 2.4.3.12. > But after switching to ActiveState Python 2.5.1.1, I can not import the DLL > anymore (without > any detailed error message). It simply doesn't work. Try renaming your .dll to .pyd? See this section in the "What's New in Python 2.5" documentation: http://docs.python.org/whatsnew/ports.html#SECTION0001510000000000000000 "Windows: .dll is no longer supported as a filename extension for extension modules. .pyd is now the only filename extension that will be searched for." From venkat83 at gmail.com Fri Mar 28 17:58:40 2008 From: venkat83 at gmail.com (Venkatraman S) Date: Fri, 28 Mar 2008 22:28:40 +0530 Subject: [C++-sig] boost::python PyUnicodeUCS4_FromEncodedObject In-Reply-To: <47EBB5FA.2040209@sympatico.ca> References: <47EBB5FA.2040209@sympatico.ca> Message-ID: On Thu, Mar 27, 2008 at 2:58 PM, Stefan Seefeld wrote: > Can you print out the compile and link commands this generates ? I'm > using distutils myself, but found its extension handling limited and > broken, and so have reimplemented the 'build_ext' command. I wouldn't be > surprized if distutils was causing the error you are experiencing. > > Well, i compiled Python again with unicode support(UCS4) and it worked. But still, would like to know as to how you reimplemented build_ext - can you share some information. -- Venkat Blog @ http://blizzardzblogs.blogspot.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From seefeld at sympatico.ca Fri Mar 28 18:08:19 2008 From: seefeld at sympatico.ca (Stefan Seefeld) Date: Fri, 28 Mar 2008 13:08:19 -0400 Subject: [C++-sig] boost::python PyUnicodeUCS4_FromEncodedObject In-Reply-To: References: <47EBB5FA.2040209@sympatico.ca> Message-ID: <47ED2603.3000600@sympatico.ca> Venkatraman S wrote: > > On Thu, Mar 27, 2008 at 2:58 PM, Stefan Seefeld > wrote: > > Can you print out the compile and link commands this generates ? I'm > using distutils myself, but found its extension handling limited and > broken, and so have reimplemented the 'build_ext' command. I wouldn't be > surprized if distutils was causing the error you are experiencing. > > > Well, i compiled Python again with unicode support(UCS4) and it worked. > But still, would like to know as to how you reimplemented build_ext - > can you share some information. It's brute-force, really: I made it a wrapper around other build systems (GNU make / autotools based, in my case). The result is that I'm dealing with a hybrid system: make/autotools for individual extension modules, and distutils for the overall project (especially the packaging). This seems to be a good compromize between effort and features. HTH, Stefan -- ...ich hab' noch einen Koffer in Berlin... From rschlup at highspeedcrow.ca Fri Mar 28 20:20:57 2008 From: rschlup at highspeedcrow.ca (Rolf Schlup) Date: Fri, 28 Mar 2008 14:20:57 -0500 Subject: [C++-sig] Py++ filter based on return type Message-ID: Hi Roman, I'm trying to bind a relatively complex interface class of an IG control system and it has a lot of methods that return references to member which are classes in themselves. I'm also working on a script that will automatically create a Py++ script from Rational Rose model. I have become a little stuck in creating a Py++ that will create a module class that will build. Hopefully, you'll be able to answer the following questions so that I can get back to it 1) Since there are a whole lot of methods, what is the best way to work with the call policies - especially with the member functions that return references to other classes using & and not pointers 2) Minimize the need for 20 or more include declarations. Including only those files that are absolutely necessary for the class being bound 3) I've been looking for examples for what I'm trying to do, but they are very scarce. I think I'm past the whole hello_world thing. Is there a place on the web that has what I'm looking for? Thank you in advance, Rolf Schlup -------------- next part -------------- An HTML attachment was scrubbed... URL: From srossross at gmail.com Fri Mar 28 22:59:56 2008 From: srossross at gmail.com (Sean Ross-Ross) Date: Fri, 28 Mar 2008 14:59:56 -0700 Subject: [C++-sig] c++ exception handleing Message-ID: <69018DC9-9930-435E-A80D-11D777A6F189@gmail.com> Hi, had coded a c++ algorithm that I have just wrapped with boost, inside the algorithm I have some python callback functions when an error is raised in one of the python callbacks, the c++ code (file and line number) is obviously not included in the stack trace. Is there an elegant way to provide the user with more information of the exception? perhaps insert the c++ file and line number that the python function was called from? ~Sean From seefeld at sympatico.ca Fri Mar 28 23:26:07 2008 From: seefeld at sympatico.ca (Stefan Seefeld) Date: Fri, 28 Mar 2008 18:26:07 -0400 Subject: [C++-sig] c++ exception handleing In-Reply-To: <69018DC9-9930-435E-A80D-11D777A6F189@gmail.com> References: <69018DC9-9930-435E-A80D-11D777A6F189@gmail.com> Message-ID: <47ED707F.8060604@sympatico.ca> Sean Ross-Ross wrote: > Hi, had coded a c++ algorithm that I have just wrapped with boost, > inside the algorithm I have some python callback functions > when an error is raised in one of the python callbacks, the c++ code > (file and line number) is obviously not included in the stack trace. > Is there an elegant way to provide the user with more information of > the exception? perhaps insert the c++ file and line number that the > python function was called from? Certainly. However, the respective Python API isn't (yet) nicely wrapped into boost.python, so you'll have to do that manually, i.e. by calling PyErr_SetObject() yourself. The type and value you set there can be anything you like, and so you are free to inject useful information about the location the exception happened in. (Of course, you can't synthesize a call stack reaching into C++ code, though...) HTH, Stefan -- ...ich hab' noch einen Koffer in Berlin... From venkat83 at gmail.com Sat Mar 29 02:36:10 2008 From: venkat83 at gmail.com (Venkatraman S) Date: Sat, 29 Mar 2008 07:06:10 +0530 Subject: [C++-sig] Python/C PyRun_String Message-ID: Hi All, Am using Pythong/C API for embedding. I have written a small-simple wrapper that does the low level functions and exposes the functionalities in a c++ style(i.,e the application layer doensnt use 'PyObject'). I am reading the contents of a file(Python script) and am running the same using PyRun_String; the script setups some variables in the Main Dict which i access it using the PyArg_Parse...etc etc(the standard way). This keeps running in a loop - i.,e it needs to be processed for every transaction - hence the Python wrapper is a singleton which preserves the main dict. Am stumbling on a problem wherein when i try to access a variable in Python from C as a string - this variable in python was constructed by concatenating a float with a string(using 'str'), and hence the resulting python variable is also a string. If i comment out the variable , then the program runs fine; else it(C program) throws out a segmentation fault. A clueless as to how to debug this and suitable hints would be immense help. (If required , i can share the code for this - i just didnt want to send it here) Secondly, can anyone point me to the Boost::python way of embedding - basically, i would like to know how to build/compile the program without using bjam. Regards, -- Venkat Blog @ http://blizzardzblogs.blogspot.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From roman.yakovenko at gmail.com Sat Mar 29 07:46:59 2008 From: roman.yakovenko at gmail.com (Roman Yakovenko) Date: Sat, 29 Mar 2008 09:46:59 +0300 Subject: [C++-sig] Py++ filter based on return type In-Reply-To: References: Message-ID: <7465b6170803282346p207dd07dvdf52f1f356c694ea@mail.gmail.com> 2008/3/28 Rolf Schlup : > Hi Roman, Good morning > I'm trying to bind a relatively complex interface class of an IG control > system and it has a lot of methods that return references to member which > are classes in themselves. I'm also working on a script that will > automatically create a Py++ script from Rational Rose model. Coool :-) > I have become > a little stuck in creating a Py++ that will create a module class that will > build. > > Hopefully, you'll be able to answer the following questions so that I can > get back to it > > 1) Since there are a whole lot of methods, what is the best way to > work with the call policies ? especially with the member functions that > return references to other classes using & and not pointers Py++ & call policies documentation: http://language-binding.net/pyplusplus/documentation/functions/call_policies.html type_traits ( similar to Boost.TypeTraits ) functionality: http://www.language-binding.net/pygccxml/apidocs/pygccxml.declarations.type_traits-module.html The usage example for your use case from pygccxml import declarations from pyplusplus.module_builder import call_policies mb = module_builder_t(...) def has_return_internal_reference_policy( f ): if not f.return_type: #constructor or destructor return False r_type = f.return_type if declarations.is_reference(r_type) and declarations.is_class( declarations.remove_reference( r_type ): return True if declarations.is_pointer(r_type) and declarations.is_class( declarations.remove_pointer( r_type ): return True #TODO: add const treatment return False scope = mb.[namespace|class_](....) #select the appropriate scope scope.calldefs( has_return_internal_reference_policy ).call_policies = call_policies.return_internal_reference() > 2) Minimize the need for 20 or more include declarations. Including > only those files that are absolutely necessary for the class being bound What version do you use? I worked hard to minimize the number of included files. Can you create small example? > 3) I've been looking for examples for what I'm trying to do, but they > are very scarce. I think I'm past the whole hello_world thing. Is there a > place on the web that has what I'm looking for? I don't exactly understand what you are trying to do, but here few links that I hope will help you: * few boost libraries wrapped by me: http://pygccxml.svn.sourceforge.net/viewvc/pygccxml/pyplusplus_dev/examples/pyboost_dev/dev/ * python-ogre - huge repository of modules wrapped with Py++ - http://python-ogre.svn.sourceforge.net/viewvc/python-ogre/trunk/python-ogre/code_generators/ take a look on "common_utils" directory HTH, if not come back and ask more specific question :-) -- Roman Yakovenko C++ Python language binding http://www.language-binding.net/ From rwgk at yahoo.com Sun Mar 30 07:39:53 2008 From: rwgk at yahoo.com (Ralf W. Grosse-Kunstleve) Date: Sat, 29 Mar 2008 22:39:53 -0700 (PDT) Subject: [C++-sig] c++ exception handleing Message-ID: <850525.65895.qm@web31107.mail.mud.yahoo.com> > (Of course, you can't > synthesize a call stack reaching into C++ code, though...) Actually, under Linux and Mac OS X you can: http://cctbx.svn.sourceforge.net/viewvc/cctbx/trunk/boost_adaptbx/meta_ext.cpp?view=markup Ralf From rschlup at highspeedcrow.ca Sun Mar 30 16:36:30 2008 From: rschlup at highspeedcrow.ca (Rolf Schlup) Date: Sun, 30 Mar 2008 09:36:30 -0500 Subject: [C++-sig] New to Py++ Message-ID: Hi all, I'm new to Py++ and I need to create bindings for a Visual Simulation control library for the company I work for. The library consists of over 200 classes of which 50 or need to be "exposed" to be usable. I've tried the Py++ GUI, but the C++ code I get from it doesn't compile. There is one main interface class - VSControl - that has methods for returning references to members ( mostly of other classes ). Some of the classes have forward referencing of their dependencies which causes the C++ to completely haywire and I'm thinking it has something to do with how Py++ generates the wrapper code. I'm also having a lot of issues with Call Policies. Mainly, the ones that have with return of reference types are giving me warnings when I generate the C++ code. I've been using the followingcode, but the resulting C++ code only compiles when I manually add the necessary header files to the generated code. But, then when I try to import it, I get unresolved reference errors. If I try to add all necessary files to the builder lines, the warnings and compiler errors increase exponentially. Again, any help or examples that can yield usable binding code will be appreciated. I've included the VSControl header file at the end of this message. Thank you in advance Rolf Schlup **************************************************************************** **************************************************************************** ******************* #!/usr/bin/python import os import sys sys.path.append( '../../../..' ) from pyplusplus import module_builder from pyplusplus.module_builder import call_policies from pygccxml import declarations from pygccxml.declarations import custom_matcher_t as c_matcher from pygccxml.declarations import regex_matcher_t as re_matcher #Creating an instance of class that will help you to expose your declarations mb = module_builder.module_builder_t( files=[ 'VSControl.hpp'] , gccxml_path=r"/usr/local/bin/gccxml" , working_directory=r"/home/rolf/source/ws_stes/BER/cobusiness/test/VisualSyst emTest/VSPython/" , include_paths=['/home/rolf/source/ws_stes/BER/cobusiness/VisualSystem/VSCore ' ,'/home/rolf/source/ws_stes/BER/cobusiness/VisualSystem/VSCommBase' ,'/home/rolf/source/ws_stes/BER/cobusiness/VisualSystem/VSException' ,'/home/rolf/source/ws_stes/UNT/real/Types' ,'/home/rolf/source/ws_stes/BER/cobusiness/VisualSystem/VSCoordConverter' ,'/home/rolf/source/ws_stes/BER/cobusiness/VisualSystem/VSEnumAndTypes' ,'/home/rolf/source/ws_stes/BER/cobusiness/VisualSystem/VSDefFileReader' ,'/usr/include/python/' ,'/home/rolf/source/ws_stes/BER/real/' ,'/home/rolf/source/ws_stes/BER/real/Global' ,'/home/rolf/source/ws_stes/BER/reglobal' ,'/home/rolf/source/ws_stes/BER/cobusiness/' ,'/home/rolf/source/ws_stes/UNT/real/' ,'/home/rolf/source/ws_stes/BER/cobusiness/VisualSystem/VSCommManager' ,'/home/rolf/source/ws_stes/BER/cobusiness/VisualSystem/VSComm' ] , define_symbols=[] ) # The following 2 lines do not have any effect mb.build_code_creator('VSControl') mb.code_creator.add_include( "path.hpp" ) mem_funs = mb.calldefs() mem_funs.create_with_signature = True for mem_fun in mem_funs: if mem_fun.call_policies: continue if not mem_fun.call_policies and (declarations.is_reference(mem_fun.return_type) or declarations.is_pointer (mem_fun.return_type) ): mem_fun.call_policies = call_policies.return_value_policy(call_policies.reference_existing_object ) #Well, don't you want to see what's going on? mb.print_declarations() #Give a name to the module. After this step you should not modify/customize declarations. mb.build_code_creator( module_name='vscontrol' ) #I don't want absolute includes within code mb.code_creator.user_defined_directories.append( os.path.abspath('.') ) #Write code to file. mb.write_module( os.path.join( os.path.abspath('.'),'VSControl_Bindings.cpp' ) ) **************************************************************************** *********************************************************************** #ifndef VSControl_h #define VSControl_h 1 #include "Global/REALGlobal.hpp" #include "VisualSystem/VSCommBase/VSAllPacketDefinitions.hpp" namespace N_VisualSystem { class VSBasePacket; class VSCommManager; class VSConverterBase; class VSMunitionMan; class VSIsectorMan; class VSCreatureMan; class VSSensorMan; class VSCelestial; class VSEntityMan; class VSObjectAllocater; class VSSystem; class VSEffectMan; class VSTerrainMan; class VSHudMan; class VSTextMan; class VSWeatherMan; class VSSightMan; } // namespace N_VisualSystem namespace N_VisualSystem { class VSControl { public: VSControl (); virtual ~VSControl(); virtual bool init (const std::string& a_configFileName, const std::string& a_interfaceName); VSWeatherMan& getWeatherMan (); VSCelestial& getCelestial () const; VSCreatureMan& getCreatureMan () const; VSEffectMan& getEffectMan (); VSEntityMan& getEntityMan () const; VSHudMan& getHudMan (); VSIsectorMan& getIsectorMan () const; VSMunitionMan& getMunitionMan (); VSSensorMan& getSensorMan (); VSSightMan& getSightMan () const; VSSystem& getSystem (); VSTerrainMan& getTerrainMan (); VSTextMan& getTextMan (); virtual bool isUpdateRequested (); void startUpdate (); void endUpdate (); bool getConnectionState () const; void preUpdate (); protected: void createPacketRefList (); void evalIncommingPackets (const std::vector& a_packetList); void handleSystemResponse (const VSSystemPacket& a_systemPacket); void handleCollisionResponse (); void handleTerrainResponse (const VSTerrainResponsePacket& a_terrainPacket); void handleEffectResponse (const VSEffectResponsePacket& a_effectResponsePacket); VSCommManager& getCommManager (); void handleIsectorResponse (const VSBasePacket& a_isectorPacket); bool m_Initialized; std::vector m_PacketRefList; ShVal m_CommManager; private: VSControl(const VSControl &right); VSControl & operator=(const VSControl &right); bool operator==(const VSControl &right) const; bool operator!=(const VSControl &right) const; void installDefaultCoordConverter (); ShVal m_EntityManager; ShVal m_TerrainManager; ShVal m_System; ShVal m_SightManager; ShVal m_WeatherManager; ShVal m_TextManager; ShVal m_HudManager; ShVal m_EffectManager; ShVal m_Celestial; ShVal m_SensorManager; ShVal m_CreatureManager; ShVal m_IsectorManager; ShVal m_MunitionManager; VSConverterBase* m_DefaultConverter; }; } // namespace N_VisualSystem #endif -------------- next part -------------- An HTML attachment was scrubbed... URL: From srossross at gmail.com Sun Mar 30 17:23:21 2008 From: srossross at gmail.com (Sean Ross-Ross) Date: Sun, 30 Mar 2008 08:23:21 -0700 Subject: [C++-sig] c++ exception handleing In-Reply-To: <850525.65895.qm@web31107.mail.mud.yahoo.com> References: <850525.65895.qm@web31107.mail.mud.yahoo.com> Message-ID: <6AC2E71C-E739-4125-AAD7-B965D5374324@gmail.com> This meta_ext.cpp sounds cool. Are there any demos or tutorials on how to use the stack synth? On a second note, when my python callback from c++ raises an exception I get an "Abort trap" after the function exits. This just started recently and I don't know what I changed to make it do that. Does anyone have any idea where to start looking for the problem? Thanks for the help On 29-Mar-08, at 10:39 PM, Ralf W. Grosse-Kunstleve wrote: >> (Of course, you can't >> synthesize a call stack reaching into C++ code, though...) > > Actually, under Linux and Mac OS X you can: > > http://cctbx.svn.sourceforge.net/viewvc/cctbx/trunk/boost_adaptbx/ > meta_ext.cpp?view=markup > > Ralf > -------------- next part -------------- An HTML attachment was scrubbed... URL: From mathem at tician.de Sun Mar 30 19:03:00 2008 From: mathem at tician.de (Andreas =?iso-8859-1?q?Kl=F6ckner?=) Date: Sun, 30 Mar 2008 13:03:00 -0400 Subject: [C++-sig] A vector of objects with private == operators In-Reply-To: <3dc9bcf00803270912w3772b29dq778c92cac4eab1ec@mail.gmail.com> References: <3dc9bcf00803261135h5f4feb80jb980e1497be20f81@mail.gmail.com> <200803261442.56918.lists@informa.tiker.net> <3dc9bcf00803270912w3772b29dq778c92cac4eab1ec@mail.gmail.com> Message-ID: <200803301303.02279.mathem@tician.de> Argh, sorry. My bad. #define PYTHON_ERROR(TYPE, REASON) \ { \ PyErr_SetString(PyExc_##TYPE, REASON); \ throw boost::python::error_already_set(); \ } Andreas On Donnerstag 27 M?rz 2008, Matthew Scouten wrote: > Thank you, > This looks like it will fix my problems. > However, the PYTHON_ERROR line will not compile because there is apparently > no such function. Is this a macro specific to your project? if so, what is > behind it? > > 2008/3/26 Andreas Kl?ckner : > > On Mittwoch 26 M?rz 2008, Matthew Scouten wrote: > > > I have noticed that the vector indexing suite only compiles for a > > > vector if foo has an == operator. Some of the objects that I need > > > > to > > > > > wrap in a library have to wrap have no == operator. That's fine, I just > > > create a dummy operator as a global function. The dummy operator just > > > > tests > > > > > &a==&b. > > > > > > Now I come to a problem. the 'bar' class has a == operator, but it is > > > declared private. I tried to create a global == operator but the > > > > compiler > > > > > complained that it was ambiguous (the compiler is such a whiny little > > > brat.) > > > > > > So my questions are 3 fold, > > > Why the heck does vector_indexing_suite need an == operator? > > > Assuming I can't change the problem class, what are possible > > > > workarounds? > > > > > Does vector_indexing_suite2 still need an == operator? > > > > This is a hack I've been using: > > > > 8< -------------------------------------------------------------- > > template > > class no_compare_indexing_suite : > > public boost::python::vector_indexing_suite > no_compare_indexing_suite > > > { > > public: > > static bool contains(T &container, typename T::value_type const &key) > > { PYTHON_ERROR(NotImplementedError, "containment checking not > > supported on > > this container"); } > > }; > > 8< -------------------------------------------------------------- > > > > Use like so: > > > > cl.def(no_compare_indexing_suite()) > > > > Andreas > > > > _______________________________________________ > > C++-sig mailing list > > C++-sig at python.org > > http://mail.python.org/mailman/listinfo/c++-sig -- Andreas Kloeckner ^ mathem at tician.de 114 Camp Street Apt 2 | http://mathema.tician.de Providence, RI 02906 | cell: (401) 497-9385 Applied Mathematics, Brown | home: (401) 648-0599 ---------------------------+------------------------------------> AIM/ICQ: inducer77 | XMPP/GoogleTalk: inducer at amessage.de -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: This is a digitally signed message part. URL: From roman.yakovenko at gmail.com Sun Mar 30 19:13:47 2008 From: roman.yakovenko at gmail.com (Roman Yakovenko) Date: Sun, 30 Mar 2008 20:13:47 +0300 Subject: [C++-sig] New to Py++ In-Reply-To: References: Message-ID: <7465b6170803301013q6d3f3a99v5cb57e55f0a7924c@mail.gmail.com> 2008/3/30 Rolf Schlup : > Hi all, Good evening > I'm new to Py++ and I need to create bindings for a Visual Simulation > control library for the company I work for. The library consists of over > 200 classes of which 50 or need to be "exposed" to be usable. I've tried > the Py++ GUI, but the C++ code I get from it doesn't compile. Py++ GUI was created to suite single goal: to show what code generates Py++, without customization. You can use it almost as-is in small projects or when you just start with Boost.Python. > There is one > main interface class ? VSControl - that has methods for returning > references to members ( mostly of other classes ). Some of the classes have > forward referencing of their dependencies which causes the C++ to completely > haywire and I'm thinking it has something to do with how Py++ generates the > wrapper code. Well, you can expose forward declarated classes as opaque type. I think that Py++ doesn't support thi use case. Normally what you want is to expose that classes too. In order to do this, you need to create header file, that will include all relevant files and than to path is to Py++ > I'm also having a lot of issues with Call Policies. Mainly, > the ones that have with return of reference types are giving me warnings > when I generate the C++ code. I've been using the followingcode, but the > resulting C++ code only compiles when I manually add the necessary header > files to the generated code. I guess this errors relaed to forward declarations. Try to pass single header file, built as described above. > But, then when I try to import it, I get > unresolved reference errors. I think this error is related to your environment. Use "dependency viewer" to verify that your dll can load others. >If I try to add all necessary files to the > builder lines, the warnings and compiler errors increase exponentially. > Again, any help or examples that can yield usable binding code will be > appreciated. I need more specific error description. From my experience it is better to start from something small, just few classes. To get to the point you can import and use few simple exported classes and than start to add other classes and functions. -- Roman Yakovenko C++ Python language binding http://www.language-binding.net/