- Boost-users - lists.boost.org

performance of dijkstra_shortest_paths
by Bhaskara Marthi 01 May '09

01 May '09

I recently ran into a situation where dijkstra_shortest_paths in the boost graph library was running very slowly. When I replaced the call to dijkstra_shortest_paths with the code below, things were a couple of orders magnitude faster. Any idea why? - Bhaskara Version 1: with dijkstra_shortest_paths ReachableCostVector GridGraph::singleSourceCosts (const Cell2D& source, const vector<Cell2D>& dests) { GridDistances distances; GridGraphVertex start = cellVertex(source); resetIndices(); typedef map<GridGraphVertex, GridGraphVertex> GridPreds; GridPreds predecessors; boost::dijkstra_shortest_paths (graph_, start, weight_map(get(&EdgeCost::cost, graph_)). vertex_index_map(get(&CellInfo::index, graph_)). distance_map(associative_property_map<GridDistances>(distances)). predecessor_map(associative_property_map<GridPreds>(predecessors)). // to avoid warnings visitor(GraphSearchVisitor())); ReachableCostVector v(dests.size()); transform (dests.begin(), dests.end(), v.begin(), bind(extractCost, this, distances, _1)); return v; } Version 2: without dijkstra_shortest_paths // Approximation to Dijkstra ReachableCostVector GridGraph::singleSourceCosts (const Cell2D& source, const vector<Cell2D>& dests) { GridDistances distances; GridGraphVertex start = cellVertex(source); typedef pair<GridGraphVertex, double> QueueItem; queue<QueueItem> q; set<GridGraphVertex> visited; q.push(QueueItem(start,0.0)); visited.insert(start); while (!q.empty()) { GridGraphVertex v; double d; tie(v,d) = q.front(); q.pop(); GridGraphAdjacencyIterator iter, end; for (tie(iter,end)=adjacent_vertices(v, graph_); iter!=end; iter++) { if (visited.find(*iter)==visited.end()) { double cost = d+graph_[edge(v,*iter,graph_).first].cost; q.push(QueueItem(*iter, cost)); visited.insert(*iter); distances[*iter] = cost; } } } ReachableCostVector costs; for (vector<Cell2D>::const_iterator iter=dests.begin(); iter!=dests.end(); ++iter) { GridDistances::iterator pos = distances.find(cellVertex(*iter)); if (pos==distances.end()) costs.push_back(ReachableCost(false,-42.0)); else costs.push_back(ReachableCost(true, pos->second)); } return costs; }

4 7

New declaration ambiguates old declaration mutant.hpp
by Celine.Murphy＠barclayscapital.com 01 May '09

01 May '09

/ThirdParty/include/boost-1.38.0/boost-1_38/boost/bimap/relation/detail/ mutant.hpp:68: new declaration `template<class View, class Type> boost::enable_if<boost::is_const<Reference>, const View&>::type boost::bimaps::relation::detail::mutate(Type&)' /ThirdParty/include/boost-1.38.0/boost-1_38/boost/bimap/relation/detail/ mutant.hpp:55: ambiguates old declaration `template<class View, class Type> boost::enable_if<boost::mpl::not_<boost::is_const<Reference> >, View&>::type boost::bimaps::relation::detail::mutate(Type&)' Hi, I am getting the above error when I try to compile a class that includes boost/bimap.hpp. It happens at the include stage. Has anyone seen this before? Gcc 3.2.3, boost 1.38. Thanks Celine _______________________________________________ This e-mail may contain information that is confidential, privileged or otherwise protected from disclosure. If you are not an intended recipient of this e-mail, do not duplicate or redistribute it by any means. Please delete it and any attachments and notify the sender that you have received it in error. Unless specifically indicated, this e-mail is not an offer to buy or sell or a solicitation to buy or sell any securities, investment products or other financial product or service, an official confirmation of any transaction, or an official statement of Barclays. Any views or opinions presented are solely those of the author and do not necessarily represent those of Barclays. This e-mail is subject to terms available at the following link: www.barcap.com/emaildisclaimer. By messaging with Barclays you consent to the foregoing. Barclays Capital is the investment banking division of Barclays Bank PLC, a company registered in England (number 1026167) with its registered office at 1 Churchill Place, London, E14 5HP. This email may relate to or be sent from other members of the Barclays Group. _______________________________________________

2 2

boost::posix_time::from_time_t bug ???
by Łukasz Majewski 01 May '09

01 May '09

Hello, i think that i've found bug in date time library or maybe i made something wrog. SAMPLE CODE: #include <boost/date_time/posix_time/posix_time.hpp> #include <iostream> using namespace boost::posix_time; int main( int argc, char** argv ) { time_t t = time( 0 ); std::cout << "time_t: " << ctime( &t ) << std::endl; struct tm* t2 = localtime( &t ); std::cout << "struct tm: " << asctime( t2 ) << std::endl; ptime t3 = from_time_t( t ); std::cout << "ptime from time_t: " << to_simple_string( t3 ) << std::endl; ptime t4 = ptime_from_tm( *t2 ); std::cout << "ptime from tm: " << to_simple_string( t4 ) << std::endl; return 0; } OUTPUT: time_t: Fri May 1 14:22:01 2009 struct tm: Fri May 1 14:22:01 2009 ptime from time_t: 2009-May-01 12:22:01 // incorrect result ptime from tm: 2009-May-01 14:22:01

2 2

[boost-users][StateChart] polymorphic_downcast failure
by Igor R 01 May '09

01 May '09

Hi, The above failure occurs in reaction_dispatcher::derived_with_action::react(). IIUC, state_machine tries to invoke in_state_reaction that's defined for an event, which is unrelated to the one being processed. What logical error in my state-machine might cause such a problem? Unfortunately, I don't have any self-contained code at the moment, so please treat my question as theoretical :). Thanks.

2 6

[boost-users][python]
by Germán Diago 01 May '09

01 May '09

I have a question about managing python objects. I have a base class implemented in c++, which I use in python. >From python I inherit that class. And the I try to use the python instance from c++, like this. //This class has a function which is registered in python. I want to //be able to register python instances with this class class AnotherClass { void registerobject(const std::string & name, MyBaseClass * a) { v.push_back(a); } }; //Interface class, which I inherit in python class MyBaseClass { virtual const std::string getData(Object * o) const = 0; }; >From python: class MyPyClass(MyBaseClass): def __init__(self): MyBaseClass.__init__(self) def getData(self, o): return "hello" My question is. In order to register that class, namely, MyPyClass(which is a python class): can I do it through registerobject? Or should I implement something like this?: class AnotherClass { void registerobject(const std::string & name, python::object & o) { MyBaseClass * m = extract<MyBaseClass>(o); v.push_back(m); } }; In all two cases I get error_already_set exception when I try to execute the registerobject function and don't know why.

1 1

[statechart] Asynch statechart in dll
by Bill Clark 01 May '09

01 May '09

I have modified the Pingpong sample (an asynchronous statechart) so that the statechart itself is in a dll, with the calling code (in PingPong.cpp) in an exe. This is using VS2005 sp1. I have used a macro in the usual way to mark the Player class as exported. The app builds and runs, but it seems that events don't get acted on. We hit Player's initiate_impl(), and we get to state_machine::process_event() with a BallReturned event passed in, but we never hit the handler for that event (Waiting::react(const BallReturned&)). Is this something that should work? Is it necessary to do something more than just export the necessary classes from the dll? Thanks, Bill **

2 4

Overloading errors
by A B 01 May '09

01 May '09

Does anyone know why the following sample fails (it was adapted from the boost overloading example) and how to do this correctly. Thanks in advance. #include <boost/python/module.hpp> #include <boost/python/def.hpp> #include <boost/python/args.hpp> #include <boost/python/tuple.hpp> #include <boost/python/class.hpp> #include <boost/python/overloads.hpp> #include <boost/python/return_internal_reference.hpp> using namespace boost::python; struct X { int f(int x, double y = 4.25, char const* z = "wow") { return x; } int f(int x, bool b = true) { return x; } }; BOOST_PYTHON_MEMBER_FUNCTION_OVERLOADS(f_member_overloads, X::f, 1, 3) BOOST_PYTHON_MODULE(args_ext) { class_<X>("X", "This is X's docstring") .def("f1", &X::f, f_member_overloads(args("x", "y", "z"), "f's docstring" )[return_internal_reference<>()] ) ; } g++ -fpic -shared -o test.so test.cpp -lpython2.4 -lboost_python -I/usr/include/python2.4 test.cpp: In function ‘void init_module_args_ext()’: test.cpp:31: error: no matching function for call to ‘boost::python::class_<X, boost::python::detail::not_specified, boost::python::detail::not_specified, boost::python::detail::not_specified>::def(const char [3], <unresolved overloaded function type>, boost::python::detail::overloads_proxy<boost::python::return_internal_reference<1ul, boost::python::default_call_policies>, f_member_overloads>)’

2 2

Boost.Date_Time building for Windows CE
by Tory Bjorklund 01 May '09

01 May '09

Hello, My application builds fine when using Boost.Date_Time with win32 but when building with a configuration for Windows CE I get the following errors: boost/date_time/c_time.hpp(83) : error C2039: 'localtime' : is not a member of 'std' boost/date_time/c_time.hpp(83) : error C2660: 'boost::date_time::c_time::localtime' : function does not take 1 arguments boost/date_time/c_time.hpp(92) : error C2039: 'gmtime' : is not a member of 'std' boost/date_time/c_time.hpp(92) : error C2660: 'boost::date_time::c_time::gmtime' : function does not take 1 arguments boost/date_time/gregorian/gregorian_io.hpp(12) : fatal error C1083: Cannot open include file: 'locale': No such file or directory Does anyone know how I may be able to build an app for CE that includes the Date_Time library? Thanks, Tory

1 0

Boost.MultiIndex question on merge
by arm2arm 01 May '09

01 May '09

Hello, Are there way to merge two multiindex objects ? suppouse we have : struct strData{ int id; float a, b; strData(int id_, float a_, float b_):a(a_),b(b_),id(id_){} }; typedef multi_index_container< strData, indexed_by< ordered_unique< tag<id>, BOOST_MULTI_INDEX_MEMBER(strData,int,id)>, ordered_non_unique< tag<snap>,BOOST_MULTI_INDEX_MEMBER(strData,float,a)> > > data_setA, data_setB; Are there way to do: data_setA.merge(data_setB) ? thank you beforehand. Arman. -- View this message in context: http://www.nabble.com/Boost.MultiIndex-question-on-merge-tp23277496p2327749… Sent from the Boost - Users mailing list archive at Nabble.com.

5 11

boost coroutine status
by Robert Ramey 30 Apr '09

30 Apr '09

This package has been in the vault or some time. At least one well known person on this list has indicated that the quality is good and that he is using it in production code. Turns out that I now have a need of this facility but I've got a couple of problems. a) When I try and build it, I get ConvertFiberToThread which looks to me that there is some bjam magic missing. b) I would like to look the documentation. But that's hard to do as it's in quick book files and I don't have a quick book reader. I don't know that I have the ability to generate html files from quickbook. Maybe I do, but I can't really figure this out. c) I don't think that this package has ever been reviews nor do I see it scheduled for review. What is the status of this library. Anyone having information on this topic is encouraged to bring me up to date. Robert Ramey

5 13