September 2010 - Boost-users - lists.preview.boost.org

boost interprocess library and concurrency
by Shane Miller 04 Sep '10

04 Sep '10

Hi, After reading through the BOOST interprocess library documentation I cannot workout whether - its STL set/hash/vector containers are multi-threaded safe. (Seeming, no, is the answer) - if a program cores in the middle of write operation to a STL container in shared memory then clearly the container is in an undefined state. Locks might be left locked (if this BOOST library uses locks say in a STL set or hash map insert/update). Now what? We just throw away the shared memory segment and start again? Has anybody solved these issues using CAS, transactions, or some other lock free solution? - aside from these issues, suppose one allocates a 2GB shared memory segment and all 2GB is used by a STL set. Does this BOOST interprocess library support adding more shared memory segments so this STL set can grow larger? Does doing so require relocating it which is a slow operation? What are the details here. I am looking for a crash recoverable, concurrent shared memory implementation of STL. Mainly I have a cache of items that will be read-intensive with some periodic updates and insertions throughout the day. Key-value pairs won't work; it must operate on pointers to C++ structs. Shane

2 1

Run all Boost.Test cases simulteniously
by Gaurav N Shah 04 Sep '10

04 Sep '10

Hi, As far as I know, boost::test behaior is to execute each test case by the sequence those are added. However, I want to execute all test cases simulteniously. Is there any way for it? This communication is for informational purposes only. It is not intended as an offer or solicitation for the purchase or sale of any financial instrument or as an official confirmation of any transaction. All market prices, data and other information are not warranted as to completeness or accuracy and are subject to change without notice. Any comments or statements made herein do not necessarily reflect those of JPMorgan Chase & Co., its subsidiaries and affiliates. This transmission may contain information that is privileged, confidential, legally privileged, and/or exempt from disclosure under applicable law. If you are not the intended recipient, you are hereby notified that any disclosure, copying, distribution, or use of the information contained herein (including any reliance thereon) is STRICTLY PROHIBITED. Although this transmission and any attachments are believed to be free of any virus or other defect that might affect any computer system into which it is received and opened, it is the responsibility of the recipient to ensure that it is virus free and no responsibility is accepted by JPMorgan Chase & Co., its subsidiaries and affiliates, as applicable, for any loss or damage arising in any way from its use. If you received this transmission in error, please immediately contact the sender and destroy the material in its entirety, whether in electronic or hard copy format. Thank you. Please refer to http://www.jpmorgan.com/pages/disclosures for disclosures relating to European legal entities.

3 2

[MSM] stack overflow happens when there is a false-guard anonymous transition
by Tang Jiang Jun 04 Sep '10

04 Sep '10

Hello guys, I met a stack overflow crash in the following MSM sample (boost 1.44). struct Some_Event {}; struct State0 : public msm::front::state<> {}; struct State1 : public msm::front::state<> {}; struct State2 : public msm::front::state<> {}; struct State_Machine_ : public msm::front::state_machine_def< State_Machine_ > { typedef State0 initial_state; template< typename Event > bool is_possible( const Event& ) { return false; } struct transition_table : public mpl::vector< _row< State0, Some_Event, State1 >, g_row< State0, msm::front::none, State2, &State_Machine_::is_possible > > {}; }; struct State_Machine : public msm::back::state_machine< State_Machine_ > {}; void main() { State_Machine state_machine; state_machine.start(); state_machine.process_event( Some_Event() ); } When start() got called in main(), it will enter a endless reentering between state_machine::process_event() and state_machine::handle_eventless_transitions_helper::process_completion_event(). Finally it caused a stack overflow. Could somebody please help me to fix this problem? Many thanks! Regards, Tang

1 0

Using subscript operator with interprocess map
by Michi Mutsuzaki 04 Sep '10

04 Sep '10

Hello, I'm using a map from basic_string to int on managed_mapped_file (boost 1.35.0). Everything works except that I get a compilation error if I use subscript operator. I'm trying something like this: #include <boost/interprocess/managed_mapped_file.hpp> #include <boost/interprocess/containers/map.hpp> #include <boost/interprocess/containers/string.hpp> #include <boost/interprocess/allocators/allocator.hpp> namespace bi = boost::interprocess; typedef bi::allocator<char, bi::managed_mapped_file::segment_manager> char_alloc; typedef bi::basic_string<char, std::char_traits<char>, char_alloc> char_string; typedef std::pair<const char_string, int> ValueType; typedef bi::allocator<ValueType, bi::managed_mapped_file::segment_manager> map_alloc; typedef bi::map<char_string, int, std::less<char_string>, map_alloc> MyMap; int main () { bi::managed_mapped_file segment(bi::open_or_create, "/tmp/test", 65536); map_alloc alloc_inst (segment.get_segment_manager()); MyMap *mymap = segment.find_or_construct<MyMap>("MyMap") (std::less<bi::string>(), alloc_inst); (*mymap)[char_string("key")] = 1; return 0; } What am I doing wrong? Thanks! --Michi

2 1

[BGL] Can't set non-const edge_descriptor = const edge_descriptor
by John Robertson 04 Sep '10

04 Sep '10

I have struct edge_properties { Traits::edge_descriptor ePredecessor; } some_function ( const Graph g ) { e = edge(vertex_a, vertex_b, g).first; //fails here g[e].ePredecessor = e; } compiler gives ./Includes/Utilities.cpp:451: error: passing ‘const boost::detail::edge_desc_impl<boost::bidirectional_tag, long unsigned int>’ as ‘this’ argument of ‘boost::detail::edge_desc_impl<boost::bidirectional_tag, long unsigned int>& boost::detail::edge_desc_impl<boost::bidirectional_tag, long unsigned int>::operator=(const boost::detail::edge_desc_impl<boost::bidirectional_tag, long unsigned int>&)’ discards qualifiers make: *** [pl] Error 1 My guess is that e is returned as a const but g[e].ePredecessor is declared a variable. How can I fix this? Thanks John

2 1

Understanding of Boost License
by Ronghui Yu 03 Sep '10

03 Sep '10

Hi, All, If I have a software developed based on Boost library, does it mean I have to open source this software? Thanks in advance -- =========================== Regards Ronghui Yu

2 1

[filesystem] v3 recursive_directory_iterator does not enumerate entries
by Timothy Madden 03 Sep '10

03 Sep '10

Hello I use recursive_directory_iterator in my project and the dereference operarator on the iterator object always returns the the base directory that I am iterating over, and not the directory entries, each one in turn. Am I doing something wrong ? int iLevel = -1; // loop over and copy each file from the source directory for ( recursive_directory_iterator it(tmpName); it != recursive_directory_iterator(); it++ ) { //if (it.level() != iLevel) if (true) { // recursion just entered a new directory, create it #ifdef _DEBUG boost::filesystem::path dir_entry(*it); cout << "Entry: " << dir_entry << endl; cout << "\tas:"; for ( boost::filesystem::path::iterator p_it = dir_entry.begin(); p_it != dir_entry.end(); p_it++ ) { if (p_it != dir_entry.begin()) cout << "/"; cout << *p_it; } cout << endl; #endif iLevel = it.level(); } } Thank you, Timothy Madden

1 0

Is there a way for a library to add symbols to the standard headers ?
by Timothy Madden 03 Sep '10

03 Sep '10

Hello To write a library to add a POSIX emulation layer on Windows, is there a way to add symbols from my library to C/C++ standard library headers, without getting into recursive inclusion ? For example, is there a way to get the 'Additions to fstream' from the boost::filesystem library, actually look like they are added to the standard <fstream> header ? I know users can use #include "cstdio" (with quotes instead of brackets) throughout a project, in order to later be able to include some local version of the header, if the need for it arise during during some porting at a later time for example. However this works for user-level code, while for a library it is not that simple any more ... Thank you, Timothy Madden

5 7

gcc refuses to compile Boost.Extension example
by René Tschirley 03 Sep '10

03 Sep '10

Dear Boost experts, I struggle to get Jeremy Pack's Boost.Extension going. I want to use the platform independent way to explicitly load shared libraries. Starting with the tutorial, it took me some time to find this out: boost::extensions::shared_library lib("mylib.dll"); boost::function<void (int)> fun1(lib.get<void, int>("DoSomething")); works, but boost::function<void (void)> fun2(lib.get<void, void>("DoSomething")); does not work as the compiler (Eclipse CDT Helios using gcc 4.4.0 on Windows with MinGW) complains: somefile.cpp:74: error: no matching function for call to 'boost::extensions::shared_library::get(const char [12])' I'm a bit confused. Am I blind? Is this a bug? Weird feature? User-too-dumb-error? Sincerely, Rene

2 2

BOOST_IS_MPI_DATATYPE macro and std::vector
by Prevrhal, Sven 03 Sep '10

03 Sep '10

I have a class whose data I want to send/recv and which are of type std::vector. However, once it's allocated the size of this vector is fixed during runtime. Can I use BOOST_IS_MPI_DATATYPE to speed up data transfer and how would I do it? Thanks, Sven ________________________________ The information contained in this message may be confidential and legally protected under applicable law. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, forwarding, dissemination, or reproduction of this message is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message.

2 1

boost::shared_ptr vs std::shared_ptr
by Raindog 03 Sep '10

03 Sep '10

I was just watching the video here: http://channel9.msdn.com/shows/Going+Deep/C9-Lectures-Stephan-T-Lavavej-Sta… And was wondering if anyone had written up a brief guide about the differences between boost::shared_ptr and std::shared_ptr.

1 0

treat a range of ranges as a single range?
by Nathan Ridge 03 Sep '10

03 Sep '10

Hello, Is it possible, using the facilities in Boost.Range, to treat a range whose elements are themselves ranges, as a range whose elements are the elements of the component ranges? For example, I would like to treat a range of strings as a range of characters. A detokenizer in essence. Thanks, Nate.

1 0

replace_range already declared in this scope?
by Nathan Ridge 03 Sep '10

03 Sep '10

Hello, It seems the following two headers do not play nicely with each other: #include <boost/algorithm/string.hpp> #include <boost/range/adaptors.hpp> When both are included in a source file, I get the error: boost/range/adaptor/replaced.hpp:102: error: ‘replace_range’ is already declared in this scope Removing either of the headers fixes the problem, but I cannot do that since I need functionality from both. What can I do to fix this? I am using Boost 1.43. Thanks, Nate.

1 0

Re: [Boost-users] Problem using serialization::xml_archive with polymrphic classes
by przemyslaw.sliwa＠uk.bnpparibas.com 03 Sep '10

03 Sep '10

Hello, thanks for help with this. This has fixed the problem. However, I have got a much more complicated example which fails. I have a simple inheritance structure with 4 classes, they derive a->b->c->d nothing complex. say a is the mose base and d is the most derived class, Each class has got a virtual destructor. I can easily serialize object of type d but I cannot do this with a pointer to a this class. Basicaly invoking in the code std::vector<b*> tmpCcys; ar & boost::serialization::make_nvp("currencyEvents", tmpCcys); does not work at all! I can get to the point // retrieve the true type of the object pointed to // if this assertion fails its an error in this library assert(NULL != this_type); const boost::serialization::extended_type_info * true_type = i.get_derived_extended_type_info(t); // note:if this exception is thrown, be sure that derived pointer // is either registered or exported. if(NULL == true_type){ boost::serialization::throw_exception( archive_exception( archive_exception::unregistered_class, true_type->get_debug_info() ) ); } in the oserializer.hpp where each element of the vector is dumped to the archive but call to i.get_derived_extended_type_info(t) throws an exception of an unknown type. It is not archive or event std::exception onlt catch(...) {...} can handle it and therefore it is not possible to see what is going on. The data under t looks fine but const extended_type_info * get_derived_extended_type_info(const T & t) const { // note: this implementation - based on usage of typeid (rtti) // only does something if the class has at least one virtual function. BOOST_STATIC_WARNING(boost::is_polymorphic<T>::value); return typeid_system::extended_type_info_typeid_0::get_extended_type_info( typeid(t) ); } simply throws something which is not possible to handle correctly. I cannot step into typeid(t) on windows and VC++2005. As mentioned i have no problems at all with serializing object. Serializing pointers deos not work att all! I am exporting the class using BOOST_CLASS_EXPORT_GUID(d, "CurrencyEvent") I have to mention that the most base class is a template and I export it as well using BOOST_CLASS_EXPORT_GUID(a<Currency>, "EventCurrency") BOOST_CLASS_EXPORT_GUID(a<CurrencyCouple>, "EventCurrencyCouple") similar BOOST_CLASS_EXPORT_GUID(b, "BaseCurrencyEvent") BOOST_CLASS_EXPORT_GUID(c, "RealCurrencyEvent") the classes a,b,c and d are within a namespace Events and defined/declared within a dll but this should not be a problem at all since I can serialize objects of the most derived class. Could someone help me with this issue? Thanks a lot in advance ___________________________________________________________ This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and delete this e-mail. Any unauthorised copying, disclosure or distribution of the material in this e-mail is prohibited. Please refer to http://www.bnpparibas.co.uk/en/information/legal_information.asp?Code=ECAS-… for additional disclosures.

2 1

[Test] BOOST_*_EQUAL_COLLECTIONS ostream problem
by Thomas Suckow 02 Sep '10

02 Sep '10

I am attempting to compare two std::map<string,string> objects in Boost.Test but am finding that it is trying to stream the std::pair< const string, string > that results from dereferencing the iterator. /* Simplified Example */ AUTO_TEST_CASE( Map_Same ) { std::map< std::string, std::string > x; std::map< std::string, std::string > y; BOOST_CHECK_EQUAL_COLLECTIONS( x.begin(), x.end(), y.begin(), y.end() ); } Results in: /usr/include/boost/test/utils/wrap_stringstream.hpp: In function 'boost::basic_wrap_stringstream<CharT>& boost::operator<<(boost::basic_wrap_stringstream<CharT>&, const T&) [with CharT = char, T = std::pair<const std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::basic_string<char, std::char_traits<char>, std::allocator<char> > >]': /usr/include/boost/test/test_tools.hpp:641: instantiated from 'boost::test_tools::predicate_result boost::test_tools::tt_detail::equal_coll_impl(Left, Left, Right, Right) [with Left = std::_Rb_tree_iterator<std::pair<const std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::basic_string<char, std::char_traits<char>, std::allocator<char> > > >, Right = std::_Rb_tree_iterator<std::pair<const std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::basic_string<char, std::char_traits<char>, std::allocator<char> > > >]' In my application the values a particular entry in the map will not reveal the problem, so I chose to suppress printing of the pair: typedef std::pair< const std::basic_string<char>,std::basic_string<char> > pair_type; BOOST_TEST_DONT_PRINT_LOG_VALUE( pair_type ); But alas, this did absolutely nothing. Confused, I tried: BOOST_CHECK_EQUAL( *x.begin(), *y.begin() ); And it compiled. Remove the suppression and the error returns. So the suppression works, just not in conjunction with BOOST_*_EQUAL_COLLECTIONS. Searching the mailing list revealed a post from 2005. http://article.gmane.org/gmane.comp.lib.boost.user/13146 He mentions as I also found that equal_coll_impl uses the stream operator directly on the dereferenced iterators. This *appears* to be bypassing the code that allows suppression. Though I may be wrong (I can rarely tell with boost). Can someone check on this and see if I am crazy? In the mean time I guess I just have to overload the stream operator. Regards, Thomas Suckow

1 0

Re: [Boost-users] Problem using serialization::xml_archive withpolymrphic classes
by przemyslaw.sliwa＠uk.bnpparibas.com 02 Sep '10

02 Sep '10

Thanks for reply, All classes a, b, c and d are actually defined within one dll (one module). I forgot to mention it. I can do this following thing: d* pD = mew d(); and then ar & boost::serialization::make_nvp("item", pD); but c* pD = new D(); ar & boost::serialization::make_nvp("item", pD); does not work at all! It throws this strange type of exceptions which I cannot handle at all. Another problem is that if I do the first version the member m_key variable of true_type in the call const boost::serialization::extended_type_info * true_type = i.get_derived_extended_type_info(t); is a null pointer which is not right. Do you have any idea what could be wrong? Thanks a lot for help ___________________________________________________________ This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and delete this e-mail. Any unauthorised copying, disclosure or distribution of the material in this e-mail is prohibited. Please refer to http://www.bnpparibas.co.uk/en/information/legal_information.asp?Code=ECAS-… for additional disclosures.

2 1

Boost.Functional/Hash
by Andrej van der Zee 02 Sep '10

02 Sep '10

Hi, I have a question that I hope can be answered by using Boost. I have the following datastructure: struct ip_info { const u_int8_t _ifile_index; const unsigned long _src_ip; const unsigned long _dst_ip; const u_int16_t _id; const u_int16_t _length; const u_int16_t _src_port; const u_int16_t _dst_port; const u_int32_t _seq; const u_int32_t _ack_seq; }; The object is identified by all its members. Now I want to count duplicates. But I have MANY of these objects, I mean billions. I have to reduce the memory footprint of my program, because it soon grows over 4GB and boom! So I rather not keep all the objects in memory and why would I, because I do not need them. I just need to count duplicates. My idea was to hash such objects to a smaller values and use that hash value as key_type in a map. The value_type of the map would be an integer. This something like std::map<int32_t, u_int16_t> or something. Then I though maybe use Boost.Functional/Hash, but are there any guarantees that the hash value is unique? How unique are they? Are there better alternatives? Thank you, Andrej

4 3

[BGL] Using pointers as indexes in graphs
by Bruno Barberi Gnecco 02 Sep '10

02 Sep '10

Hi, I want to store pointers in a graph. Can I somehow use their pointers as indexes? Here's the code I'm trying to execute. It does not work correctly, however; apparently graph add other nodes in the vector storage (to fill the holes?). I suppose I could use listS instead of vecS, but then I can't get the topological_sort code to work (I know I have to provide some form of indices, but I don't understand how to do that). Any suggestions to solve this? Thanks! #include <iostream> #include <boost/graph/graph_traits.hpp> #include <boost/graph/adjacency_list.hpp> #include <boost/graph/topological_sort.hpp> class Plugin { public: // some data here int x; Plugin(int a) : x(a) {} }; typedef std::list< std::pair<Plugin *, Plugin *> > EdgeList; typedef boost::adjacency_list<boost::setS, boost::vecS, boost::directedS, Plugin *> PipelineGraph; typedef boost::graph_traits<PipelineGraph>::vertex_descriptor PipelineVertex; void graph(EdgeList edges) { PipelineGraph graph; for (EdgeList::iterator it(edges.begin()); it != edges.end(); it++) { PipelineVertex a(boost::vertex(reinterpret_cast<unsigned long> (it->first), graph)); PipelineVertex b(boost::vertex(reinterpret_cast<unsigned long> (it->second), graph)); boost::add_edge(a, b, graph); } // sort by topology, so we run the plugins in the correct order std::vector<PipelineVertex> sorted; boost::topological_sort(graph, std::back_inserter(sorted)); std::cout << sorted.size() << std::endl; // REVERSE topological order std::vector<PipelineVertex> vl(sorted.rbegin(), sorted.rend()); // loop for (std::vector<PipelineVertex>::iterator it(vl.begin()); it != vl.end(); it++) { Plugin *p = reinterpret_cast<Plugin *>(*it); std::cout << p->x << std::endl; } } int main() { EdgeList edges; Plugin *a = new Plugin(1); Plugin *b = new Plugin(2); Plugin *c = new Plugin(3); Plugin *d = new Plugin(4); edges.push_back(std::pair<Plugin*, Plugin*>(a, b)); edges.push_back(std::pair<Plugin*, Plugin*>(b, c)); edges.push_back(std::pair<Plugin*, Plugin*>(b, d)); graph(edges); return 0; }

2 3

how to get std::min_element() working with boost::multi_array?
by Valmor de Almeida 02 Sep '10

02 Sep '10

Hello, Could someone help getting std::min_element() (or a variation of) to work with multi_array? boost::multi_array<int, 2> A; // code for dimensioning A and filling in; deleted cout << *min_element(A.begin(),A.end()) << endl; Thanks, -- Valmor

1 1

Looking for Halfedge or other topologic representation library
by Lang Stefan 02 Sep '10

02 Sep '10

Hello everyone, While looking for an existing implementation of halfedge data structures or similar topological representations of 2-dimensional manifolds I found one article pointing to boost:graph. I've checked the documentation, but it seem that this library is way to generic/abstract for my purpose, and yet seems to miss a number of important functions that probably only make sense when working with a topological representation of a two-dimensional oriented manifold. Conceptually, Half-edge data structures would fit my needs very well, and I've found several sites with half-done or half-documented implemetations. But those that actually had code were either too specific to be useful for my purposes, or half of the site couldn't be retrieved, probably because the relevant pages are no longer hosted. So I was wondering, if there is, maybe, another library within boost that is dealing more with the topological aspect of graphs, whether or not based on boost::graph? What I am looking for is a topological framework for oriented surfaces that lets me retrieve adjacency information regarding faces, edges and vertices, disregardful of whether or not the set of surfaces represent the full surface of a 3D object or just a part of it. I also need to directly navigate between the objects defining the actual 3D points, curves, and surfaces, and the corresponding topological objects. But that, I hope, is just a minor detail. I suspect it won't be so hard to implement myself, but building on a well-designed library is so much less error prone... Cheers, Stefan

3 2

Re: [Boost-users] does pool::release_memory() really require an ordered
by Lang Stefan 02 Sep '10

02 Sep '10

-----Ursprüngliche Nachricht----- Message: 5 Date: Wed, 1 Sep 2010 17:32:46 -0400 From: Mark Rajcok <mrajcok(a)gmail.com> To: boost-users(a)lists.boost.org Subject: [Boost-users] does pool::release_memory() really require an ordered pool? Message-ID: <AANLkTim0t9KTDJowTbn2viBu+RSTqnwfKA1c7qRbv_X2(a)mail.gmail.com> Content-Type: text/plain; charset=ISO-8859-1 On page http://www.boost.org/doc/libs/1_44_0/libs/pool/doc/interfaces/pool.html there is this precondition for release_memory(): "t must be ordered". Does that mean that ordered_malloc() must always be used, if I want to use release_memory()? Looking at the implementation for release_memory(), it appears to be iterating through chunks in each block, which I wouldn't expect if ordered_malloc() were always used. What I'm really after: I want to use object_pool, but periodically I'd like to reclaim/release any memory blocks that are no longer in use. In my application, I may have traffic spikes, where I allocate a lot of memory, but then later (after spiking) I'd like to release it. (The object pools will persist for the life of the application, and there will probably always be some "active objects" -- i.e., allocated from the pool and not free()d yet.) (I'm hoping the documentation may be out of date, and release_memory() can be used on pools that don't use ordered_malloc() and on object_pools.) -- Mark ------------------------------ Mark, I cannot check the repository right now, but last time I did there were no changes to the version of boost::pool since I last tried to work with it, about three years ago. Back then I found that memory wasn't released at all due to a relevant part of the code being commented out. But even uncommenting it didn't help all that much: just like you, I found that the effort of releasing a single object was, essentially, proportional to the size of the pool, at best at an order of O(log(N)), maybe even worse. So just using standard new and delete operators would at some point outperform this pool implementation! I've done my own implementation in the meantime, with all operations at constant time, and using shared pointers to objects from the pool, which would make sure unused memory blocks can be released ASAP. Never got around to actually use it though, so I cannot say for sure if it has errors still. The tests and benchmarks I ran did look rather good, but I found that using malloc/free instead of new/delete got me about the same performance boost, which is why I ultimately discarded it. (malloc and free were already being used anyway, not that I was happy with it. But unfortunately I ran out of arguments to replace them...) Cheers, Stefan

1 0

[BGL] Segmentation fault in clear_vertex
by Daniel Trebbien 02 Sep '10

02 Sep '10

I am attempting to modify my implementation of the Stoer–Wagner min-cut algorithm to modify the input graph. In the code sample below, t and s are vertex_descriptors that were calculated by a call to boost::detail::stoer_wagner_phase: boost::tie(s, t, w) = boost::detail::stoer_wagner_phase(g, assignments, weights, pq); assert(s != t); // ... std::multimap<vertex_descriptor, edge_descriptor> sOutEdges; BGL_FORALL_OUTEDGES_T(s, e, g, UndirectedGraph) { const vertex_descriptor u = target(e, g); if (t != u) { sOutEdges.insert(std::make_pair(u, e)); } } BGL_FORALL_OUTEDGES_T(t, e, g, UndirectedGraph) { const vertex_descriptor u = target(e, g); const weight_type eWeight = get(weights, e); typename std::multimap<vertex_descriptor, edge_descriptor>::iterator sOutEdgesIt, sOutEdgesEnd; boost::tie(sOutEdgesIt, sOutEdgesEnd) = sOutEdges.equal_range(u); if (sOutEdgesIt != sOutEdgesEnd) { for(; sOutEdgesIt != sOutEdgesEnd; ++sOutEdgesIt) { put(weights, sOutEdgesIt->second, get(weights, sOutEdgesIt->second) + eWeight); } } else if (t != u) { // need to update e to set the source to s edge_descriptor ePrime; boost::tie(ePrime, boost::tuples::ignore) = add_edge(s, u, g); put(weights, ePrime, eWeight); } } clear_vertex(t, g); The segmentation fault occurs in the call to clear_vertex, but does not occur if I comment out the line that adds an edge to the graph g. The type of g, by the way, is: boost::adjacency_list<boost::listS, boost::vecS, boost::undirectedS, boost::no_property, boost::property<boost::edge_weight_t, int> > Why does adding an edge between vertices s and target(e, g), both of which are not equal to t, cause a segmentation fault? For reference, gdb says that the fault occurs on line 999 of boost/graph/detail/adjacency_list.hpp: // O(E/V * E/V) template <class Config> inline void clear_vertex(typename Config::vertex_descriptor u, undirected_graph_helper<Config>& g_) { typedef typename Config::global_edgelist_selector EdgeListS; BOOST_STATIC_ASSERT((!is_same<EdgeListS, vecS>::value)); typedef typename Config::graph_type graph_type; typedef typename Config::edge_parallel_category Cat; graph_type& g = static_cast<graph_type&>(g_); typename Config::OutEdgeList& el = g.out_edge_list(u); typename Config::OutEdgeList::iterator ei = el.begin(), ei_end = el.end(); for (; ei != ei_end; ++ei) { detail::erase_from_incidence_list (g.out_edge_list((*ei).get_target()), u, Cat()); g.m_edges.erase((*ei).get_iter()); *// line 999* } g.out_edge_list(u).clear(); } Program received signal SIGSEGV, Segmentation fault. std::_List_node_base::unhook (this=0x3d7a38) at ../../../../gcc-4.4.0/libstdc++-v3/src/list.cc:134 134 ../../../../gcc-4.4.0/libstdc++-v3/src/list.cc: No such file or director y. in ../../../../gcc-4.4.0/libstdc++-v3/src/list.cc (gdb) bt #0 std::_List_node_base::unhook (this=0x3d7a38) at ../../../../gcc-4.4.0/libstdc++-v3/src/list.cc:134 #1 0x0045fa26 in std::list<boost::list_edge<unsigned int, boost::property<boost ::edge_weight_t, int, boost::no_property> >, std::allocator<boost::list_edge<uns igned int, boost::property<boost::edge_weight_t, int, boost::no_property> > > >: :_M_erase (this=0x22fa9c, __position=...) at c:/mingw/bin/../lib/gcc/mingw32/4.4.0/include/c++/bits/stl_list.h:1424 #2 0x0045fa0c in std::list<boost::list_edge<unsigned int, boost::property<boost ::edge_weight_t, int, boost::no_property> >, std::allocator<boost::list_edge<uns igned int, boost::property<boost::edge_weight_t, int, boost::no_property> > > >: :erase (this=0x22fa9c, __position=...) at c:/mingw/bin/../lib/gcc/mingw32/4.4.0/include/c++/bits/list.tcc:111 #3 0x00413ec9 in boost::clear_vertex<boost::detail::adj_list_gen<boost::adjacen cy_list<boost::listS, boost::vecS, boost::undirectedS, boost::no_property, boost ::property<boost::edge_weight_t, int, boost::no_property>, boost::no_property, b oost::listS>, boost::vecS, boost::listS, boost::undirectedS, boost::no_property, boost::property<boost::edge_weight_t, int, boost::no_property>, boost::no_prope rty, boost::listS>::config> (u=5, g_=...) at c:/mingw/bin/../lib/gcc/mingw32/4.4.0/../../../../include/boost/graph/det ail/adjacency_list.hpp:999 #4 0x0041b0cc in boost::detail::stoer_wagner_min_cut<boost::adjacency_list<boos t::listS, boost::vecS, boost::undirectedS, boost::no_property, boost::property<b oost::edge_weight_t, int, boost::no_property>, boost::no_property, boost::listS> , boost::adj_list_edge_property_map<boost::undirected_tag, int, int&, unsigned i nt, boost::property<boost::edge_weight_t, int, boost::no_property>, boost::edge_ weight_t>, boost::associative_property_map<std::map<int, bool, std::less<int>, s td::allocator<std::pair<int const, bool> > > >, boost::shared_array_property_map <unsigned int, boost::vec_adj_list_vertex_id_map<boost::no_property, unsigned in t> >, boost::d_ary_heap_indirect<unsigned int, 4u, boost::shared_array_property_ map<unsigned int, boost::vec_adj_list_vertex_id_map<boost::no_property, unsigned int> >, boost::shared_array_property_map<int, boost::vec_adj_list_vertex_id_map <boost::no_property, unsigned int> >, std::greater<int>, std::vector<unsigned in t, std::allocator<unsigned int> > > > (g=..., weights=..., parities=..., assignments=..., pq=...) ...

1 1

does pool::release_memory() really require an ordered pool?
by Mark Rajcok 02 Sep '10

02 Sep '10

On page http://www.boost.org/doc/libs/1_44_0/libs/pool/doc/interfaces/pool.html there is this precondition for release_memory(): "t must be ordered". Does that mean that ordered_malloc() must always be used, if I want to use release_memory()? Looking at the implementation for release_memory(), it appears to be iterating through chunks in each block, which I wouldn't expect if ordered_malloc() were always used. What I'm really after: I want to use object_pool, but periodically I'd like to reclaim/release any memory blocks that are no longer in use. In my application, I may have traffic spikes, where I allocate a lot of memory, but then later (after spiking) I'd like to release it. (The object pools will persist for the life of the application, and there will probably always be some "active objects" -- i.e., allocated from the pool and not free()d yet.) (I'm hoping the documentation may be out of date, and release_memory() can be used on pools that don't use ordered_malloc() and on object_pools.) -- Mark

2 1

Struggling to implement bzip2 on windows
by Jared James Thompson 02 Sep '10

02 Sep '10

Hi All, I'm a VS2008 user, trying to use the ESBTL library for a project, which implements Boost to do many operations. ESBTL #includes <boost/iostreams/filter/bzip2.hpp>, which has been disabled for the Boostpro installation of windows. Quite simply, I'm struggling to get bzip2 installed and set up so that boost runs with bzip2 enabled. Has anybody done this before? Can anyone offer me more information than "compile from source?" best, Jared -- Jared James Thompson Department of Medicinal Chemistry and Molecular Pharmacology Laboratory for Computational Drug Design and Biology RHPH 504C Heine Pharmacy Building 575 Stadium Mall Drive West Lafayette, IN 47907-2091

3 3

lamba: QStringList <<
by Hicham Mouline 02 Sep '10

02 Sep '10

Hello, the following for_each call with _1 fails to compile #include <QStringList.h> #include <boost/lambda/lambda.hpp> using namespace boost::lambda; QStringList items_list; std::for_each(begin_items(),end_items(), items_list << _1 ); QStringList has these << operators defined: QStringList& operator<< ( const QString& str ); QStringList& operator<< ( const QStringList& other ); the type of item iterator returned by begin_items() is: const char* const* QString has a ctor: QString (const char * str) The compilation error is: error C2678: binary '<<' : no operator found which takes a left-hand operand of type 'const QStringList' (or there is no acceptable conversion) I don't see why it tries to find a << that takes a ***const*** QStringList left-hand operand? Regards,

2 1