October 2009 - Boost-users - lists.preview.boost.org

Not able to use multiple cores with boost::asio::ssl
by Ramashish Baranwal 11 Oct '09

11 Oct '09

Hi, I have an application where I am planning to use boost::asio's ssl streams. There are two parts of the app, the initial connection establishment and ssl handshake is done by boost::asio::ssl and the connection is then transferred to the core application which does the the rest. I have written a prototype using multiple threads and io_services similar to io_service_pool example. I am using a dual quad core machine, however I am not able to get the cpu usage beyond one core (i.e. it never exceeds 12-14% in the top) no matter how much the load. On doing some profiling I found that the time taken from socket accept to ssl handshake completion deteriorates linearly with the number of connections without using the other cores. I am not using any explicit locks so expected the application to scale well with the number of cores available. Is is not possible to use multiple cores with boost::asio::ssl libraries or am I not doing it right? A pseudo-code of the app- start connection acceptors handle_accept() { new_conn = new Connection(io_service_pool.get_io_service()); new_conn->start(); } Connection::start() { ssl_socket_.async_handshake(handle_handshake); } Connection::handle_handshake(error) { if (!error) // transfer connection to the application core } Thanks in advance, Ramashish

2 2

[mpl] Syntax error with BOOST_MPL_ASSERT
by Jean-Louis Leroy 11 Oct '09

11 Oct '09

What's wrong with this ? template<class Expected, class T> void check_placeholders(const T&) { BOOST_MPL_ASSERT((is_same<typename T::placeholders, Expected>)); // * } struct Test { typedef int placeholders; }; void f() { check_placeholders<int>(Test()); } msvc9 complains about : error C2059: syntax error : ')' at the line marked with a star. I have put much more complex things inside BOOST_MPL_ASSERT though :-/ I use Boost 1.40. J-L

2 1

[fusion] C arrays in fusion::vector
by Jean-Louis Leroy 11 Oct '09

11 Oct '09

msvc9+fusion 1.40 reject this: fusion::vector<int, char[20]> v; 1>C:\Users\JLL\Documents\Visual Studio 2008\Projects\boost_1_40_0\boost/fusion/container/vector/detail/vector_n.hpp(43) : error C2440: '<function-style-cast>' : cannot convert from 'int' to 'char [20]' 1> There are no conversions to array types, although there are conversions to references or pointers to arrays 1> C:\Users\JLL\Documents\Visual Studio 2008\Projects\boost_1_40_0\boost/fusion/container/vector/detail/vector_n.hpp(42) : while compiling class template member function 'boost::fusion::vector_data2<Derived,T0,T1>::vector_data2(void)' 1> with 1> [ 1> Derived=boost::fusion::vector2<int,char [20]>, 1> T0=int, 1> T1=char [20] 1> ] Am I doing something wrong ? Anyway I can wrap the C array in a class (template)... Thanks, J-L

3 2

[python] Function objects in place of member functions
by Ravi 10 Oct '09

10 Oct '09

Hello, If a free function 'func' has X* as its first argument, then, boost.python allows it to be bound to a member function on the python side, i.e., the following is legal: void func( X* x, arg1_t arg ) { ... } class_<X>( "X" ).def( "func", &func ); In order to use a function object in place of a free function, one must specialize/overload boost::python::detail::get_signature which, for some reason, does not account for function objects. Here's a very simple example that works: ---------------------------------------------- #include <boost/mpl/vector.hpp> struct X { int y; }; // Function object struct Z { int operator()( X *x, int z ) { return z + x->y; } }; namespace boost { namespace python { namespace detail { mpl::vector<int, X*, int> get_signature( Z&, X* ) {return mpl::vector<int, X*, int>();} }}} #include <boost/python/class.hpp> #include <boost/python/module.hpp> BOOST_PYTHON_MODULE( mft ) { boost::python::class_<X>( "X" ) .def( "z", Z() ).def_readwrite( "y", &X::y ); } ------------------------------------------------- However, note that the overload of get_signature precedes the inclusion of the boost.python headers, which is extremely inconvenient. However, if the headers are moved to their proper location as in the following, ---------------------------------------------- #include <boost/mpl/vector.hpp> #include <boost/python/class.hpp> #include <boost/python/module.hpp> struct X { int y; }; // Function object struct Z { int operator()( X *x, int z ) { return z + x->y; } }; namespace boost { namespace python { namespace detail { boost::mpl::vector<int, X*, int> get_signature( Z&, X* ) {return boost::mpl::vector<int, X*, int>();} }}} BOOST_PYTHON_MODULE( mft ) { boost::python::class_<X>( "X" ) .def( "z", Z() ).def_readwrite( "y", &X::y ); } ------------------------------------------------- the compilation fails with the following error message (gcc 4.4.1 with boost 1.37 or 1.39): $ g++ -Wall -shared -o mft.so memft.cc -lboost_python-mt - I/usr/include/python2.6 -fPIC In file included from memft.cc:1: /usr/include/boost/python/class.hpp: In member function ‘void boost::python::class_<T, X1, X2, X3>::def_impl(T*, const char*, Fn, const Helper&, ...) [with T = X, Fn = Z, Helper = boost::python::detail::def_helper<const char*, boost::python::detail::not_specified, boost::python::detail::not_specified, boost::python::detail::not_specified>, W = X, X1 = boost::python::detail::not_specified, X2 = boost::python::detail::not_specified, X3 = boost::python::detail::not_specified]’: /usr/include/boost/python/class.hpp:235: instantiated from ‘boost::python::class_<T, X1, X2, X3>& boost::python::class_<T, X1, X2, X3>::def(const char*, F) [with F = Z, W = X, X1 = boost::python::detail::not_specified, X2 = boost::python::detail::not_specified, X3 = boost::python::detail::not_specified]’ memft.cc:16: instantiated from here /usr/include/boost/python/class.hpp:536: error: no matching function for call to ‘get_signature(Z&, X*)’ Why is the overloaded get_signature not picked up when it is declared *after* the inclusion of the headers? The reason for using function objects is to change the argument types of a member function of X as seen by python; for example, the functor would have an argument type 'double' and would then internally convert the received 'double' argument from the python side into some custom type prior to passing it to the member function of X. In other words, the functor would act as a C++ equivalent of a python decorator. Regards, Ravi

1 0

asio tutorial question on
by Gavin Rogers 09 Oct '09

09 Oct '09

I am going through the tutorial on asio at http://www.boost.org/doc/libs/1_37_0/doc/html/boost_asio/tutorial/tutdaytim… and so far so good. I have a question, the end iterator doesn't seem to be assigned a value, in the examples. I was wondering how does it work.. e.g. tcp::resolver::iterator endpoint_iterator = resolver.resolve(query); tcp::resolver::iterator end; tcp::socket socket(io_service); boost::system::error_code error = boost::asio::error::host_not_found; while (error && endpoint_iterator != end) // <<< Question:here { socket.close(); socket.connect(*endpoint_iterator++, error); } thanks.

2 1

error: expected unqualified id before '=' token
by John Robertson 09 Oct '09

09 Oct '09

The following compiles OK typedef string Colour; struct edge_properties { vector < set< Colour > > eColours; } until I get to template < typename Graph > void populate_arrow_maps(const Graph & g) { typename graph_traits<Graph>::edge_iterator edge_iter, edges_end; typedef vector <set< string > > VCS; typedef vector <set< string > >::iterator vcs_iter, vcs_end; for (tie(edge_iter, edges_end) = edges(g); edge_iter != edges_end; ++edge_iter) { VCS & vcolour_set = get(m_eColours_map, *edge_iter); for(vcs_iter = vcolour_set.begin(); vcs_iter != vcolour_set.end(); ++vcs_iter) { ; } } } Then compiler says error: expected unqualified id before '=' token Referring to the line which is underlined above. So its referring to vcs_iter. Whats wrong? Thanks

2 2

[asio] "bad_descriptor" on async_send_to + socket.close
by Tacheon the Nucleid 09 Oct '09

09 Oct '09

[asio] "bad_descriptor" error on async_send_to + socket.close Hi, I am trying to close a UDP socket on which an async_send_to operation is running. The handler is like this: void Sender::handleSend(error_code const & e, size_t bt) { this_thread::sleep(posix_time::seconds(1)); socket.async_send_to(buffer, remoteEndpoint, bind(&Sender::handleSend, this, placeholders::error, placeholders::bytes_transferred)); } The sender itself is fine. But on some special occasions I have to close the socket. I do it like this: void Sender::do_close() { socket.close(); } void close() { socket.get_io_service().post( bind(&Sender::do_close, this)); } When close() is called, the async handler yields a "bad_descriptor" error. I would expect an "operation_aborted" error as stated in the documentation. For async_receive, everything is fine and works as expected. What am i getting wrong? Greetings, B. Haeberlein -- GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT! Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01

1 0

[BGL] reset of the random number generator in the sorted_erdos_renyi_iterator
by Marti Cuquet 09 Oct '09

09 Oct '09

Hi, I noticed that when generating several graphs with the sorted_erdos_renyi_iterator with the same random generator all of them happen to be the same. Looking at erdos_renyi_generator.hpp I see that there is a reset in the generator. This does not happens in the normal (unsorted) erdos_renyi_iterator since the generator is passed by reference in the constructor. I'm still not confortable with the BGL and the random libraries, but I might say that it is better to pass the random generator by reference. Is there a special reason for not doing so? Thank you, Martí

2 2

[BGL] Modern benchmarks available?
by Nathan Cooprider 08 Oct '09

08 Oct '09

I'm evaluating the Boost Graph Library for a project would like to present my manager with some benchmark numbers. The only ones I have been able to find are from the GGCL OOPSLA paper, and those are pretty out of date. Does anybody know of any modern work at benchmarking BGL performance? I am looking at adjacency lists with sets in particular, but anything is better than nothing. I realize that I may just have to give big-O estimates in the end (because it is all dependent on the graph structure), but any estimates would help me for my initial evaluation. I have generally steered clear of Parallel BGL because it does not seem as mature as BGL. Is this a false assumption? Should I look further into PBGL? Thank you very much. Nathan

3 2

[serialization?] converting utf8 string to unicode wstring
by Igor R 08 Oct '09

08 Oct '09

Hello, I try to accomplish the subj with help of boost's utf8_codecvt_facet. I based my code on this example: http://www.boost.org/doc/libs/1_40_0/libs/serialization/doc/codecvt.html . The only difference is that my utf8 text resides in std::string: #include <sstream> #include <iostream> #include "boost/archive/detail/utf8_codecvt_facet.hpp" // link with boost/libs/serialization/src/utf8_codecvt_facet.cpp int main() { std::string utf; utf.resize(11); // hardcode some utf8 text utf[0] = 0xd7; utf[1] = 0x90; utf[2] = 0xd7; utf[3] = 0x99; utf[4] = 0xd7; utf[5] = 0x92; utf[6] = 0xd7; utf[7] = 0x95; utf[8] = 0xd7; utf[9] = 0xa8; utf[10] = 0x0; std::locale old_locale; std::locale utf8_locale(old_locale, new boost::archive::detail::utf8_codecvt_facet()); std::locale::global(utf8_locale); std::stringstream in; in.imbue(utf8_locale); in.str(utf); std::wstringstream out; out << in; std::wcout << out.str() << std::endl; } The above code doesn't work: "out" buffer doesn't contain correct unicode interpretation of the string. Actually, all i want is a c++ equivalent to the following WinAPI: #include "windows.h" int main() { std::string utf; utf.resize(11); // hardcode some utf8 text utf[0] = 0xd7; utf[1] = 0x90; utf[2] = 0xd7; utf[3] = 0x99; utf[4] = 0xd7; utf[5] = 0x92; utf[6] = 0xd7; utf[7] = 0x95; utf[8] = 0xd7; utf[9] = 0xa8; utf[10] = 0x0; wchar_t outBuff[11]; MultiByteToWideChar(CP_UTF8, 0, utf.c_str(), -1, outBuff, 10); } ...which works well. Any idea would be greatly appreciated! Thanks.

3 3