November 2012 - Boost-users - lists.preview.boost.org

Re: [Boost-users] [Thread] Win32 TLS problem with dynamic libraries
by Sarah Tappon 02 Nov '12

02 Nov '12

On Mon, Oct 29, 2012 at 12:31 AM, Сергей Захарченко wrote: > * Rename the Boost.Thread library into 'original_boost_thread'. > > * Add a custom target named 'patched_boost_thread' which uses dumpbin > and lib utilities to delete just the tss_pe.obj object from Boost.Thread. > > * Add a library called boost_thread which links to 'patched_boost_thread' and > contains a file named tss_pe.cpp which contains the aforementioned code > (essentially it's the 'null tss cleanup implementation'). I've been researching a similar issue and it's useful to hear from someone who has successfully implemented a workaround. A couple of follow-up questions, if you don't mind: Once the solution is in place, what (if anything) do you have to do to ensure that resources are cleaned up appropriately? Is it sufficient to call on_thread_exit() whenever a thread exits? > Assuming that I can guarantee that no Boost.Thread functions > are called from non-Boost.Thread threads, it shouldn't be > needed. Could you give me a concrete example of a call that would violate this guarantee? What happens if you violate it? Thanks, Sarah

1 0

[boost::python] exposing an abstract base class through boost::ptr_vector and the vector indexing suite
by Andy Falanga (afalanga) 01 Nov '12

01 Nov '12

Hi everyone, I have need to expose a couple of different classes which inherit from a single, abstract, C++ class. This abstract class is what I need to expose to python. Using the python wiki, I've figure out the correct way to expose these classes. However, the next dimension to this problem is that my users aren't supposed to be able to instantiate any of the derived classes directly. Instead, they are to use a function which will return a list of these objects. I've been working on the correct implementation for use with boost::ptr_vector. Now, I want to use the vector indexing suite in boost::python to expose these things. I've worked through properly making an assignment overload operator appropriate to this for use with the vector indexing suite. However, I'm stumped at this error: Error 2 error C2440: 'return' : cannot convert from 'Base' to 'Base *' c:\program files\boost\boost_1_47\boost\python\suite\indexing\vector_indexing_suite.hpp 77 Because of this error, I'm wondering if it's not possible, or at least not recommended, to use the vector indexing suite to expose boost::ptr_vectors. Is this true? I'm using Visual Studio 2010 and Boost 1.47. Thanks, Andy

1 0

[serialization] support for C++-11 types
by Marco Craveiro 01 Nov '12

01 Nov '12

Hi boost-users, Apologies if this question has been asked before - I'm sure it has, but my googling-fu is not finding decent matches. Anyways, the question is: with C++11 we have a number of new container types such as std::array, std::underdered_map/set, std::forward_list, etc. Are those supported by the latest release of the serialization library? If not, is there a trac ticket for this work? Many thanks for your time, Marco -- So young, and already so unknown -- Pauli blog: http://mcraveiro.blogspot.com

6 8

bug of boost algorithm library or boost asio?
by kpzhang 01 Nov '12

01 Nov '12

Hi, Apologize if this is not the right place for this mail. This program (it has been narrowed down from a larger program) always crashes after compiled in vs2008 Release(Win32) mode under windows 7. I am not familiar with assembly code and don't know it's a bug of compiler or boost::ends_with or boost::asio::buffers_iterator. It can be compiled and executed with g++ in Ubuntu without any problem. People said it's very unlikely to be compiler's bug, but when compiled in debug moded(or disable optimization), the problem does disappear. I have been stuck with this problem for quite a few hours. Any help is appreciated. Thanks in advance. #include <iostream>#include <string>#include <boost/asio.hpp>#include <boost/algorithm/string.hpp> typedef boost::asio::buffers_iterator<boost::asio::const_buffers_1> iterator_t;typedef boost::iterator_range<iterator_t> range_t;static const std::string LINE_END_MARK = "\r\n"; int main(int argc, char* argv[]){ boost::asio::streambuf _buf; std::ostream os(&_buf); os<<"END\r\n"; iterator_t cursor = boost::asio::buffers_begin(_buf.data()); iterator_t end = boost::asio::buffers_end(_buf.data()); std::ostream_iterator<char> it(std::cout," "); std::copy(LINE_END_MARK.begin(), LINE_END_MARK.end(), it); range_t r(cursor, end); if(!boost::ends_with(r, LINE_END_MARK)) return 0; return 1;}

4 3

__MINGW32_MAJOR_VERSION confusing, and causing problems with boost thread
by Florian Burkart 01 Nov '12

01 Nov '12

Hi, __MINGW32_MAJOR_VERSION is confusing me at the moment. Where is it coming from? What does it stand for? Even though my compiler at the moment informs me of PS C:\Sys\boost\boost_1_50_0> g++ --version g++.exe (GCC) 4.6.3 20111208 (prerelease) is is defined as 3 when boost compiles. __MINGW32_MINOR_VERSION is defined as 11. I had to hard check that, no idea where those numbers are coming from. This leads to this statement from boost_1_50_0\libs\thread\src\win32\tss_pe.cpp #if defined(__MINGW64__) || (__MINGW32_MAJOR_VERSION >3) || \ ((__MINGW32_MAJOR_VERSION==3) && (__MINGW32_MINOR_VERSION>=18)) evaluating to false and this extern "C" const IMAGE_TLS_DIRECTORY32 _tls_used __attribute__ ((section(".rdata$T"))) = to compile in, which in turn produces c:/mingw-4.6.1/bin/../lib/gcc/mingw32/4.6.1/../../../libmingw32.a(tlssup.o):tlssup.c:(.tls+0x0): multiple definition of `_tls_used' /Sys/boost/MinGW32/boost_1_50_0/lib/libboost_thread-mgw46-mt-d-1_50.a(tss_pe.o):tss_pe.cpp:(.rdata$T+0x0): first defined here when trying to link to the static library. I fixed that by hacking the 18 in the source down to a 11 and recompiling boost with PS C:\Sys\boost\boost_1_50_0> .\b2.exe --prefix=c:\Sys\boost\MinGW32\boost_1_50_0 toolset=gcc address-model=32 link=static -s BZIP2_SOURCE=C:\Sys\bzip2-1.0.6 -s ZLIB_SOURCE=C:\Sys\zlib-1.2.7 install That seems to have solved my issue for now, but I don't know if my major or minor versions are set incorrectly or whether this might a bug in the code base? Thoughts? Florian

2 2

Re: [Boost-users] FW: bug of boost algorithm library or boost asio?
by kpzhang 01 Nov '12

01 Nov '12

Hi Alex, Thanks for your reply. To Igor: The exception type is 'Access Violation' raised in boost::ends_with function. Regards, Ivan 2012/10/31 Alex Perry <Alex.Perry(a)smartlogic.com> > Sorry forgot to add you to cc - it speeds up conversation if you are in > digest mode for boost list! > > -----Original Message----- > From: Alex Perry > Sent: 31 October 2012 15:04 > To: 'boost-users(a)lists.boost.org' > Subject: RE: [Boost-users] bug of boost algorithm library or boost asio? > > On 31 October 2012 09:05 kpzhang [mailto:sail4dream@gmail.com] wrote :- > > ...snip... > >People said it's very unlikely to be compiler's bug, but when compiled in > debug moded(or disable optimization), the problem does disappear. > >I have been stuck with this problem for quite a few hours. Any help is > appreciated. Thanks in advance. > ...snip... > > Not sure if its any help but just tried compiling this with VS2010 and > VS2012 > > From VS command prompt (with boost version 1.51 pathed into include and > lib) > > cl test.cpp /EHsc /O2 > > and resulting test.exe is fine (with or without /O2) > > So does look like a problem with vs2008 specifically (sorry don't have one > available to try any longer) > > Alex > >

1 0

C++ Developers - ADVANCED AI PLATFORM (Artificial General Intelligence)
by B Hart 01 Nov '12

01 Nov '12

I am putting together a Google Plus Group for C++ developers who are (or would be) passionately interested in building a platform for strong AI (and eventual commercialization). If you want to be in this Google Plus Group please e-mail me at bhartsb at gmail dot com. Subject line "AI NOW" Please state a few details about yourself. Below is a link to a very abbreviated presentation. More details will be forthcoming if you join. Regards, Brian Hart Presentation: Building A Platform For Strong AI<https://docs.google.com/presentation/d/1OhPurShl7CP4qHDBRyW1cn2Aj2rYGKX3-eR…>

1 0