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

shared_ptr race condition
by Sharon Paisner 22 Oct '04

22 Oct '04

I'm having a problem wherein the block of code below hangs at line 16. A race condition is being created with the shared_ptrs. The Call class contains an std::map of DialogPtrs (where a DialogPtr is just a boost::shared_ptr< Dialog >) - is it legal to use a boost::shared_ptr to wrap a class that contains a dynamically allocated container like a map? note: the createDialog() method is write-locked, so that the map cannot be accessed by more than one thread at a time. 1 int main( int argc, char *argv[] ) 2 { 3 AAFramework aa; 4 CallPtr call( new Call( aa ) ); 5 6 UriPtr callee( new Uri( "sip:name@domain" ) ); 7 call->setInitialCallee( callee ); 8 9 assert( call->getInitialCallee() == callee ); 10 11 TagPtr tag( new Tag() ); 12 CallIdPtr callId( new CallId() ); 13 UriPtr contact( new Uri( "sip:contact@domain" ) ); 14 UriPtr uriComplete( new Uri( "sip:complete@domain" ) ); 15 16 DialogPtr newDialog = call->createDialog( DT_CALLEE, tag, callId, contact, uriComplete ); 17 assert( call->getDialog( DT_CALLEE ) == newDialog ); 18 19 return EXIT_SUCCESS; 20 } one of the strangest aspects of this problem is that it occurs on some machines, but not others, and I have yet to determine the difference between the systems on which it works and the systems on which it doesn't work. Boost's use of so much inline code makes it very difficult to debug the program. in case it helps any, I'm using boost-1.31 and gcc-3.4.2. thanks, Sharon

2 1

[sandbox/socket]1.32/vc7.1: Who's in charge???
by Francis ANDRE 22 Oct '04

22 Oct '04

Hi I am trying to get the boost-sandbox/socket running but it seems that this socket library is out of sync, at least for vc7.1. Who is in charge of this library since Hugo DUNCAN has moved its code base out of the boost foundry?? Cheers Francis

5 13

boost::serialization problems
by Andrzej Cieslar 22 Oct '04

22 Oct '04

Hello, I'm having some problems trying to use the boost::serialization library. Here's why: I have a class hierarchy (~80 classes). I was trying to add the serialization abilities to the entire hierarchy (intrusive template-based version). Because I need to serialize things through pointers to base classes, I do the register_type on each class. But the compilation fails on MSVC 7.1 with "internal structure overflow" and "out of keys" errors. So I googled a little bit and found out that this is a known issue with MSVC and that the workaround is to split the .cpp file into smaller parts. So I did. But now the linker fails due to excessive .lib file size (I had to put the register_type function for each class in separate .cpp file) Ok so I switched to the polymorphic archives. It compiles fine, but fails at runtime when saving a null pointer. Here's a short example: --- #include <fstream> #include <boost/archive/polymorphic_text_oarchive.hpp> struct A { void serialize( boost::archive::polymorphic_oarchive & ar, const unsigned int version) {} }; int main(int argc, char* argv[]) { std::ofstream ofs("serialized.txt"); boost::archive::polymorphic_text_oarchive oa(ofs); A* a = 0; oa << a; } --- The failure is because of trying to dynamic_cast from interface_oarchive& to basic_oarchive&, which I believe are not related: oserializer.hpp, l.422: if(NULL == t){ basic_oarchive & boa = boost::smart_cast_reference<basic_oarchive &>(ar); Are these known issues or am I missing something? I'd be grateful for your help All best, Andrzej Cieslar ---------------------------------------------------- Pistolety wiatrówki bez zezwolenia w Militaria.pl Teraz do 30% taniej! http://klik.wp.pl/?adr=http%3A%2F%2Fadv.reklama.wp.pl%2Fas%2Fsklep1.html&si…

1 0

Newbie Question regarding \algorithm\string
by Ghorashi, Ali 22 Oct '04

22 Oct '04

Hello All, Just a quick question: Why isn't the \algorithm\string library not included in the standard boost download bundle? Thanks, A.G.

3 2

Using template in property types
by Stephen torri 22 Oct '04

22 Oct '04

I have a need to store an instance of a template class in each vertex of a adjacency_list. I was wondering if my idea below will do: template <class House_Type, class Number_Type> class Home { public: uint32_t id; House_Type type; Number_Type digits; }; The following classes are specific instances of the Home class: class Ranch_Home : public Home <std::string, uint8_t> {} class Tudor_Home : public Home <std::string, float> {} I was thinking that the property to store this would have to be something allows for either Home type to be stored at a vertex. struct house_obj{}; typedef property < house_obj, boost::shared_ptr< Home< House_Type_Base, Digit_Type_Base> > > VertexProperty; In this case the House_Type_Base and Digit_Type_Base what I am not sure about here. I want to use a Factory to create the appropriate Home based on a string I read from the file as was seen above. For example, for each line of the file I plan on using a boost spirit parser to parse the file since I need all three elements before calling the factory. 1 "Ranch" 20 2 "Tudor" 20509.0 So how can I use the property when the template arguments are objects that have no common base class between them? Stephen -- Email: storri(a)torri.org

1 0

Swap allocator problem in multi_array.hpp
by Ewan Campbell 20 Oct '04

20 Oct '04

When resizing a multi-dimensional array I receive the warning: d:\microsoft visual studio\vc98\include\xutility(100) : warning C4101: '_Tmp' : unreferenced local variable d:\boost\include\boost-1_31\boost\multi_array.hpp(246) : see reference to function template instantiation 'void __cdecl std::swap(class std::allocator<double> &,class std::allocator<double> &)' being compiled An unreference local variable is not exactly the end of the world, but it is churned out every time a file is compiled that include this function and I'd rather not disable this warning. I can find no mention of anyone else having this problem for such a widely used function (dynamically resizing arrays). The specific line that appears to be causing the problem is: swap(this->allocator_,new_array.allocator_); in "multi_array.hpp" I am resizing the array like so header file: typedef boost::multi_array<double, 2> array2d_type; typedef array2d_type::index index; Class CArrayConfiguration{ ... protected: ... array2d_type m_pReceiveArray; ... }; implementation file: m_pReceiveArray.resize(boost::extents[M][5]); // where M = 10 -> 1000 and it is this line that prompts the warning. I am using MSVC++ 6.0 on WinXP. If I initalising the array to shape and then resize it there is no problem, but I require it to be a class member variable. Thanks in advance, ec

4 6

Constructing vertex_iterator from vertex_descriptor
by Martin Magnusson 20 Oct '04

20 Oct '04

I'm using BGL, and I noticed that add_vertex returns a vertex_descriptor to the new vertex. However, I would like to have a vertex_iterator pointing to the same element. Is there a way to construct a vertex_iterator when given a vertex_descriptor? / martin

3 2

compiling bgl
by Emre Turkay 19 Oct '04

19 Oct '04

Hi all, I've compiled boost on windows with vc71. However, I don't see a generated binary file for the boost graph library, though it was generating in linux. Do I need to do some manual tricks? Thanks, Emre Turkay

2 1

[BGL]DIJKSTRA - ospf-example
by Khaled Saleh 18 Oct '04

18 Oct '04

I'm trying to grasp the BGL using the user guide and reference manual book by J.Seik et al. Trying the ospf-example (\boost_1_31_0\libs\graph\example) using Borland C++ 5.6.4 for Win32 raise the follwing error: "dijkstra_shortest_paths.hpp": E2285 Could not find a match for 'dijkstra_shortest_paths<VertexListGraph,Param,Tag,Rest>(const adjacency_list<vecS,vecS,directedS,no_property,property<edge_weight_t,int,no _property>,no_property,listS>,unsigned int,unsigned int * const,iterator_property_map<int *,vec_adj_list_vertex_id_map<no_property,unsigned int>,int,int &>,adj_list_edge_property_map<directed_tag,int,const int &,unsigned int,const property<edge_weight_t,int,no_property>,edge_weight_t>,vec_adj_list_vertex_i d_map<no_property,unsigned int>,undefined,const closed_plus<int>,undefined,undefined,const default_dijkstra_visitor)' in function void dijkstra_dispatch2<adjacency_list<vecS,vecS,directedS,no_property,property<e dge_weight_t,int,no_property>,no_property,listS>,const iterator_property_map<int *,vec_adj_list_vertex_id_map<no_property,unsigned int>,int,int &>,adj_list_edge_property_map<directed_tag,int,const int &,unsigned int,const property<edge_weight_t,int,no_property>,edge_weight_t>,vec_adj_list_vertex_i d_map<no_property,unsigned int>,bgl_named_params<unsigned int *,vertex_predecessor_t,no_property> >(const adjacency_list<vecS,vecS,directedS,no_property,property<edge_weight_t,int,no _property>,no_property,listS> &,unsigned int,iterator_property_map<int *,vec_adj_list_vertex_id_map<no_property,unsigned int>,int,int &>,adj_list_edge_property_map<directed_tag,int,const int &,unsigned int,const property<edge_weight_t,int,no_property>,edge_weight_t>,vec_adj_list_vertex_i d_map<no_property,unsigned int>,const bgl_named_params<unsigned int *,vertex_predecessor_t,no_property> &) at line 247 The offending line is: choose_param(get_param(params, graph_visitor), make_dijkstra_visitor(null_visitor()))); Please advise in order to get myself going. Best Regards

1 0

Vertex descriptor invalidation
by Greg Emerson 18 Oct '04

18 Oct '04

Re: Boost Graph Library I have a feeling I am having some problems due to add_edge and vertex invalidation. According to the table on http://www.boost.org/libs/graph/doc/adjacency_list.html decriptors are stable, but http://groups.yahoo.com/group/boost/message/5435 claims otherwise. Could you clear this up for me? Thanks much, Greg Emerson

3 3