- Boost-users - lists.boost.org

M68K-Coldfire: Boost 1.44 seems to have a problem with low level locking
by Andreas Wehrmann 15 Sep '10

15 Sep '10

Hello, I recently upgraded boost on my embedded linux platform (m68k coldfire architecture) from version 1.41 (cmake release) to 1.44. Boost itself compiles flawlessly but when I tried to link my application (which makes use of the thread, regex, system libraries) it fails with (note the last line): squ_loopback_timer.o: In function `boost::asio::detail::wait_handler<boost::_bi::bind_t<void, boost::_mfi::mf1<void, rapp::squ_loopback_timer_t, boost::system::error_code const&>, boost::_bi::list2<boost::_bi::value<rapp::squ_loopback_timer_t*>, boost::arg<1> > > >::do_complete(boost::asio::detail::task_io_service*, boost::asio::detail::task_io_service_operation*, boost::system::error_code, unsigned int)': squ_loopback_timer.cpp:(.text._ZN5boost4asio6detail12wait_handlerINS_3_bi6bind_tIvNS_4_mfi3mf1IvN4rapp20squ_loopback_timer_tERKNS_6system10error_codeEEENS3_5list2INS3_5valueIPS8_EENS_3argILi1EEEEEEEE11do_completeEPNS1_15task_io_serviceEPNS1_25task_io_service_operationESA_j[boost::asio::detail::wait_handler<boost::_bi::bind_t<void, boost::_mfi::mf1<void, rapp::squ_loopback_timer_t, boost::system::error_code const&>, boost::_bi::list2<boost::_bi::value<rapp::squ_loopback_timer_t*>, boost::arg<1> > > >::do_complete(boost::asio::detail::task_io_service*, boost::asio::detail::task_io_service_operation*, boost::system::error_code, unsigned int)]+0x72): undefined reference to `__sync_lock_test_and_set_4' What changed about the locking mechanism? Btw. I'm also using boost version 1.38 in another project (but on the same platform). My compiler is g++ version 4.2.0 20070318 (prerelease) (Sourcery G++ Lite 4.2-35). Any hints or help is appreciated! Best regards -- Dipl.-Ing. (FH) Andreas Wehrmann Software Development -------------------------------------------------------------- Center Communication Systems GmbH A-1210 Wien, Ignaz-Köck-Straße 19 Sitz in Wien FN 796 88p, Firmenbuchgericht Wien www.centersystems.com Tel.: +43 (0) 190 199 - 3616 Mobile: +43 (0) 664 884 75916 Fax: +43 (0) 190 199 - 2110 E-Mail: a.wehrmann(a)centersystems.com ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.

3 5

[asio] io_service::work and cpu load
by Daniele Barzotti 15 Sep '10

15 Sep '10

Hi, in my DLL I create an endless io_service::work object to assign to a serial_port. I noted that, when I load the application that uses this DLL my CPU load raise up to 50% only when the io_service is created. Any suggestions? regards, Daniele.

1 0

Building PBGL
by David Doria 15 Sep '10

15 Sep '10

I downloaded PBGL from here: http://osl.iu.edu/research/pbgl/ <http://osl.iu.edu/research/pbgl/>When running CMake, it said that it couldn't find boost_mpi. I read that I need to do ./bjam --with-mpi I did that and it seemed to be successful (though very very quick): -------------------- Component configuration: ... - mpi : building ... ...found 1 target... The Boost C++ Libraries were successfully built! -------------------- So my questions are: 1) Does PBGL not come with the normal boost package from sourceforge ( http://sourceforge.net/projects/boost/files/boost/1.44.0/)? 2) How can I tell if boost_mpi built properly? Thanks, David

2 1

Re: [Boost-users] Why doesn't boost::function use get_pointer() ?
by Peter Soetens 15 Sep '10

15 Sep '10

On Tuesday 14 September 2010 23:47:41 Steven Watanabe wrote: > AMDG > > Peter Soetens wrote: > > When using boost::bind, I can write this: > > > > struct X > > { > > void foo(int) {} > > }; > > > > shared_ptr<X> x( new X ); > > // bind takes shared_ptr: > > bind(&X::foo, x, 3)(); > > > > But I can't write this: > > > > function<void(X*,int)> foo = &X::foo; > > // function refuses shared_ptr: > > foo(x,1); > > > > Why doesn't boost::function have the same get_pointer() trick ? > > Overload resolution for calling a boost::function is exactly the > same as for calling a function with that signature. shared_ptr<X> > is not convertible to X*, so the call fails. So this is by design I presume. > > > Or how can I > > rewrite the function signature such that it does work with shared > > pointers ? > > function<void(const shared_ptr<X>&, int)> foo; Thank you. The problem was that I didn't know if the first argument was a regular pointer or a shared pointer (generic code!). I have also found another solution, that is, by 'always' writing the call like this: foo(get_pointer(x),1); Which will work for all cases of x, and the signature of foo can remain void(X*,int). Peter

1 0

[BGL] changes in graph => changes in filtered_graph?
by Adam Spargo 15 Sep '10

15 Sep '10

Hi, Just a simple question - if I have a filtered_graph fg, based on adjacency_list g, do changes to g automagically turn up in fg? Or do I need to make a new fg, once I have finished changing g? I know I could write a simple test to find out, which I am doing, but I'd be grateful if someone with more experience of BGL could let me know of any exceptions or potential pitfalls. Thanks in advance, Adam. -- Dr Adam Spargo High Performance Assembly Group email: aws(a)sanger.ac.uk Wellcome Trust Sanger Institute Tel: +44 (0)1223 834244 x7728 Hinxton, Cambridge CB10 1SA Fax: +44 (0)1223 494919 -- The Wellcome Trust Sanger Institute is operated by Genome Research Limited, a charity registered in England with number 1021457 and a company registered in England with number 2742969, whose registered office is 215 Euston Road, London, NW1 2BE.

3 2

[Filesystem] v3: preferred() vs. make_preferred()
by PB 15 Sep '10

15 Sep '10

Hello, I have built up a path in which the underlying string has a mixture of slashes and backslashes. I'm passing my_path.string().c_str() to a Windows API function which is getting confused because the string is being provided verbatim. I'm therefore looking to convert the string. According to the Filesystem tutorial (http://www.boost.org/doc/libs/1_44_0/libs/filesystem/v3/doc/tutorial.html) I can use the preferred() observer method to get the string that is suitable for the current (Windows) platform. This sounds like quite a nice design. However, it seems the preferred() method doesn't exist (i.e. the tutorial is wrong). There is a make_preferred() method which is non-const because it *changes the underlying representation* to the format preferred by the operating system. >From my application developer point of view, the preferred() observer is neater/better. I don't care and don't need to know what the underlying path format is. I have no desire to change it. All I am bothered about is that I can store it to a string (perhaps in an external file), and reconstruct the path from the same string. There are a few points in the application however where I need to present the path in a particular way, and it'd be nice to just throw a call to preferred() there. e.g. listBox_.AddString(my_path.preferred().c_str()); I've now got to do: bfs::path pref = my_path; pref.make_preferred(); listBox_.AddString(pref.string().c_str()); because the function that does the rendering of the string to the list box accepts the path as a const boost::filesystem::path& and therefore cannot change it. Wanted to point out the mismatch between the tutorial and the source code, and am wondering on the rationale behind make_preferred() over preferred()? Always possible I'm missing something! Thanks, Pete

2 1

boost::bind correct way
by Kraus Philipp 15 Sep '10

15 Sep '10

Hi, I'm using boost::bind for creating a thread group. My code shows this way: boost::thread_group l_threadgroup; for(std::size_t i=0; i < boost::thread::hardware_concurrency(); ++i) l_threadgroup .create_thread( boost::bind( &ncd<T>::multithread_deflate_symmetric, this, i ) ); The code runs within the class method "symmetric" and should be called from the self object the multithread_deflate_symmetric. My class "ncd" is a template class. I debug the code on a amd64 Gentoo Linux with valgrind 3.5 and get a "data race condition". Valgrind shows this message: ==6914== Possible data race during read of size 8 at 0x7fefff6d0 by thread #1 ==6914== at 0x4C2B138: ??? (in /usr/lib64/valgrind/ vgpreload_helgrind-amd64-linux.so) ==6914== by 0x6152B10: boost::thread::start_thread() (in /opt/ library/Boost/1.44/lib/libboost_thread.so.1.44.0) ==6914== by 0x414B50: boost::thread::thread<boost::_bi::bind_t<void, boost::_mfi::mf1<void, ncd<double>, unsigned long const&>, boost::_bi::list2<boost::_bi::value<ncd<double>*>, boost::_bi::value<unsigned long> > > >(boost::_bi::bind_t<void, boost::_mfi::mf1<void, ncd<double>, unsigned long const&>, boost::_bi::list2<boost::_bi::value<ncd<double>*>, boost::_bi::value<unsigned long> > >, boost::disable_if<boost::is_convertible<boost::_bi::bind_t<void, boost::_mfi::mf1<void, ncd<double>, unsigned long const&>, boost::_bi::list2<boost::_bi::value<ncd<double>*>, boost::_bi::value<unsigned long> > >&, boost::detail::thread_move_t<boost::_bi::bind_t<void, boost::_mfi::mf1<void, ncd<double>, unsigned long const&>, boost::_bi::list2<boost::_bi::value<ncd<double>*>, boost::_bi::value<unsigned long> > > > >, boost::thread::dummy*>::type) (in main) ==6914== by 0x419F4D: ncd<double>::symmetric(std::vector<std::string, std::allocator<std::string> > const&, bool const&) (in main) ==6914== by 0x40B10F: main (in main) ==6914== This conflicts with a previous write of size 8 by thread #2 ==6914== at 0x4C2B263: ??? (in /usr/lib64/valgrind/ vgpreload_helgrind-amd64-linux.so) Do I use boost::bind in the correct way? Thanks Phil

2 1

Property Tree problem with equal range.
by Andrew Maclean 14 Sep '10

14 Sep '10

I am using Boost 1.43 and last line in the following code causes this error: /usr/local/include/boost/property_tree/detail/ptree_implementation.hpp:471: error: ‘class boost::multi_index::detail::bidir_node_iterator<boost::multi_index::detail::ordered_index_node<boost::multi_index::detail::index_node_base<std::pair<const std::basic_string<char, std::char_traits<char>, std::allocator<char> >, boost::property_tree::basic_ptree<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > > > >, std::allocator<std::pair<const std::basic_string<char, std::char_traits<char>, std::allocator<char> >, boost::property_tree::basic_ptree<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > > > > > > > >’ has no member named ‘base’ Here is the code fragment. pt is the property tree. std::string key = "program.query"; std::size_t num = pt->count(key); std::cout << "Number = " << num << std::endl; typedef boost::property_tree::ptree::assoc_iterator assocIt; typedef std::pair< assocIt, assocIt > TRangeItPair; TRangeItPair er = pt->equal_range(key); The only fix I could see was here: http://lists.boost.org/boost-users/2010/02/55806.php But I cannot see how to apply it to the above code fragment. Any help is appreciated. Regards Andrew -- ___________________________________________ Andrew J. P. Maclean Centre for Autonomous Systems The Rose Street Building J04 The University of Sydney 2006 NSW AUSTRALIA Ph: +61 2 9351 3283 Fax: +61 2 9351 7474 URL: http://www.acfr.usyd.edu.au/ ___________________________________________

1 0

Why doesn't boost::function use get_pointer() ?
by Peter Soetens 14 Sep '10

14 Sep '10

When using boost::bind, I can write this: struct X { void foo(int) {} }; shared_ptr<X> x( new X ); // bind takes shared_ptr: bind(&X::foo, x, 3)(); But I can't write this: function<void(X*,int)> foo = &X::foo; // function refuses shared_ptr: foo(x,1); Why doesn't boost::function have the same get_pointer() trick ? Or how can I rewrite the function signature such that it does work with shared pointers ? Any chance it gets into C++0x ;-) ? Thanks, Peter

2 1

[Filesystem] query
by Sean Farrow 14 Sep '10

14 Sep '10

Hi: I have a path of the form: Drive\directory1\directory2, how would I go about retrieving the second directory only i.e in this example directory2? Is this the same code in filesystem 2 and 3? Regards Sean.

2 1