- Boost-users - lists.boost.org

custom allocators, pool_alloc release_memory() seems to be broken
by B Hart 18 Mar '10

18 Mar '10

I tried contacting Stephen the pool_allocator author, but his e-mail is invalid (shouldn't the docs be updated?) Here is what I wrote in case someone else here can offer some help: I'm trying unsuccessfully to use your Boost::pool_alloc. In the documentation at the boost website: http://www.boost.org/doc/libs/1_42_0/libs/pool/doc/interfaces.html it says: // Exiting the function does NOT free the system memory allocated by the pool allocator // You must call // boost::singleton_pool<boost::pool_allocator_tag, sizeof(int)>::release_memory() // in order to force that release_memory() does not seem to work. Not sure why. I called a function in my code to verify my processes memory usage, and see that indeed memory was not being released. I also google searched and found that others experienced the same: http://groups.google.ms/group/boost-list/browse_thread/thread/7d350f4c30d76… http://article.gmane.org/gmane.comp.lib.boost.devel/179196 http://lists.boost.org/Archives/boost/2008/08/141468.php http://archives.free.net.ph/message/20080827.184624.21654313.en.html Your help on this is requested. Thanks.

2 6

shared_ptr with custom allocator but default deleter?
by Rhys Ulerich 18 Mar '10

18 Mar '10

I'm missing something w.r.t. shared_ptr and custom allocators... I'd like to use the following shared_ptr constructor template<class Y, class D, class A> shared_ptr(Y * p, D d, A a); with a custom allocator but "the usual deleter behavior". I'm confused about what to supply for 'd'. Restating my confusion, first pretend the signature was this template<class Y, class A> shared_ptr(Y * p, A a); and ignore the normal shared_ptr two argument constructor for the moment. Now pretend I could implement this hypothetical two argument version by delegating to the three argument version. What would I supply as the 'd' parameter? - Rhys

2 1

[SoC] Summer of Code
by Andrew Sutton 18 Mar '10

18 Mar '10

All, Boost is once again participating in the Google Summer of Code program. If you are a student who is interested in contributing to Boost, now would be a good time to discuss your plans on the *developer's* mailing list ( http://lists.boost.org/mailman/listinfo.cgi/boost) Student applications will be accepted here starting March 29, and can be submitted here: http://socghop.appspot.com/. If you have any questions, please feel free to ask me directly or respond to this message. Cheers, Andrew Sutton andrew.n.sutton(a)gmail.com

1 0

[asio] Problem Assignin a File Descriptor to posix:basic_stream_descriptor
by Nader Salehi 18 Mar '10

18 Mar '10

Hi All, I want to use Boost ASIO as an event manager in a server which, among other things, asynchronously writes into a file. I use open() to create my file and then try to assign it to a posix::stream_descriptor object. However, the following simple code throws an "Operation not permitted" exception. int main () { boost::asio::io_service ioService; int fd = open("./asio_sample.txt", O_WRONLY | O_CREAT | O_ASYNC | O_NONBLOCK, S_IRUSR | S_IWUSR | S_IRGRP); assert(fd != -1); boost::asio::posix::stream_descriptor descriptor(ioService); descriptor.assign(fd); return 0; } I traced down the epoll_reactor::register_descriptor() where the function fails to add the descriptor to the poll list. Any idea what I am doing wrong? I would appreciate any comment. Cheers, Nader

2 3

[graph] Problem compiling simple graph code with recent g++
by Stephen Kell 18 Mar '10

18 Mar '10

I'm having some problems with the Boost Graph library in 1.42. Specifically, if I include the graph concepts header before declaring my overloads of vertices, out_edges et al, the compiler can't see my overloads. I've produced a small-ish test case (below) which doesn't compile with g++ 4.4.3. (It's not supposed to link...) In file included from test.cpp:3: /home/srk31/opt/include/boost/graph/graph_concepts.hpp: In destructor ?boost::concepts::VertexListGraph<G>::~VertexListGraph() [with G = Gr]?: /home/srk31/opt/include/boost/graph/graph_concepts.hpp:166: instantiated from ?static void boost::concept::requirement<Model>::failed() [with Model = boost::concepts::VertexListGraphConcept<Gr>]? /home/srk31/opt/include/boost/concept_check.hpp:43: instantiated from ?void boost::function_requires(Model*) [with Model = boost::concepts::VertexListGraphConcept<Gr>]? test.cpp:49: instantiated from here /home/srk31/opt/include/boost/graph/graph_concepts.hpp:187: error: no matching function for call to ?vertices(Gr&)? /home/srk31/opt/include/boost/graph/graph_concepts.hpp: In member function ?void boost::concepts::VertexListGraph<G>::const_constraints(const G&) [with G = Gr]?: /home/srk31/opt/include/boost/graph/graph_concepts.hpp:189: instantiated from ?boost::concepts::VertexListGraph<G>::~VertexListGraph() [with G = Gr]? /home/srk31/opt/include/boost/graph/graph_concepts.hpp:166: instantiated from ?static void boost::concept::requirement<Model>::failed() [with Model = boost::concepts::VertexListGraphConcept<Gr>]? /home/srk31/opt/include/boost/concept_check.hpp:43: instantiated from ?void boost::function_requires(Model*) [with Model = boost::concepts::VertexListGraphConcept<Gr>]? test.cpp:49: instantiated from here /home/srk31/opt/include/boost/graph/graph_concepts.hpp:201: error: no matching function for call to ?vertices(const Gr&)? What's odd is that if I move the #include of boost/graph/graph_concepts.hpp to *below* my overloads, it does work. However, that code reordering is not an option in the actual context where I'm having this problem. Can anyone see anything wrong with my code, or verify that they do or don't see the same problem? I'm thinking it's a compiler bug but am not sure... will file it in gcc bugzilla if I'm not proved wrong here. It does build successfully with g++ 4.1.2. If anyone knows a workaround, or has other suggestions, I'd be very grateful. Anyway, thanks for reading! Stephen #include <utility> #include <boost/graph/graph_traits.hpp> #include <boost/graph/graph_concepts.hpp> struct Gr { }; namespace boost { template <> struct graph_traits <Gr> { typedef void *vertex_descriptor; typedef std::pair<void*, void*> edge_descriptor; typedef void **vertex_iterator; typedef directed_tag directed_category; typedef allow_parallel_edge_tag edge_parallel_category; struct traversal_tag : public virtual vertex_list_graph_tag, public virtual incidence_graph_tag {}; typedef traversal_tag traversal_category; typedef size_t vertices_size_type; typedef size_t edges_size_type; typedef size_t degree_size_type; typedef std::pair <void*, void*> *out_edge_iterator; }; typedef std::pair<graph_traits<Gr>::out_edge_iterator, graph_traits<Gr>::out_edge_iterator> edge_iter_pair; typedef std::pair<graph_traits<Gr>::vertex_iterator, graph_traits<Gr>::vertex_iterator> vertex_iter_pair; edge_iter_pair out_edges (graph_traits<Gr>::vertex_descriptor u, const Gr& g); graph_traits<Gr>::degree_size_type out_degree( graph_traits<Gr>::vertex_descriptor u, const Gr & g); graph_traits<Gr>::vertex_descriptor source( graph_traits<Gr>::edge_descriptor e, const Gr & g); graph_traits<Gr>::vertex_descriptor target( graph_traits <Gr>::edge_descriptor e, const Gr & g); vertex_iter_pair vertices (const Gr & g); graph_traits<Gr>::vertices_size_type num_vertices(const Gr & g); } // end namespace boost int main () { Gr g; boost::out_edges (*boost::vertices (g).first, g); boost::function_requires < boost::IncidenceGraphConcept < Gr > >(); boost::function_requires < boost::VertexListGraphConcept < Gr > >(); }

2 2

[Interprocess] create a managed_mapped_file without allocator mutex?
by Andy Wiese 18 Mar '10

18 Mar '10

I'm using managed_mapped_file to hold an index for a small property database. In various operations of boost ::interprocess::rbtree_best_fit<boost::interprocess::mutex_family, boost::interprocess::offset_ptr<void>, 0ul> the process frequently blocks on boost::interprocess::interprocess_mutex::lock I synchronize all access to this file using interprocess mutex utilities, so write access is restricted to one single-threaded process at a time. I have looked for a deadlock between my outer mutex and the allocator's mutex, but I haven't been able to detect any process blocking on my mutex. It also might be that a process crashes while holding the allocator mutex, but in that case I don't know how to flush the allocators mutex (like I can with my own). I'm wondering if I need the allocator mutex at all, since I provide synchronization at a higher level. Is it possible to create a: typedef basic_managed_mapped_file< char,rbtree_best_fit< mutex_family >,iset_index > managed_mapped_file; That uses some kind of no-op mutex family? Can anyone give me a hint how to declare such a thing? Thanks much for any advice, Andy

2 2

Query Regarding mixing two different version of boost with same c++ application at same time
by akhilesh singh 18 Mar '10

18 Mar '10

Hi Boost users, Can we mix two different version of boost library(boost 1.34.1 and boost 1.39.0) with same appliaction at the same time? and if not, whether it will cause issue? Any help will be apprciated. Thanks Akhil

1 0

How are template arguments to rbtree_best_fit resolved
by Eric Whitcombe 18 Mar '10

18 Mar '10

I am stepping through interprocess shared memory code and I was a little confused by the offset_ptr code that I was hitting. I cannot figure out how that is specified to the template of rbtree_best_fit. I see in boost/interprocess/interprocess_fwd.hpp declaring managed_shared_memory by typedef'ing basic_managed_shared_memory with rbtree_best_fit<mutex_family> as the AllocationAlgorithm argument but I can't figure out how offset_ptr is is supplied as the VoidPointer argument to the rbtree_best_fit template. That template has 3 arguments but the forward declaration only provides one. Are the other two somehow inferred? If so, how?

2 1

Copy MPL sequence at runtime?
by Rhys Ulerich 18 Mar '10

18 Mar '10

Evening all, I'd like to copy the contents of a boost::mpl::vector_c (or similar) into an output iterator at runtime. I've not found a clean way to get an MPL sequence into a runtime-usable form. So far, I've gotten the included ugliness to work. 0) Is there some simple, idiomatic way to accomplish this goal that I've missed? 1) Any suggestions for how to clean this up? Specifically ways to reduce the assumptions about the underlying MPL sequence or fixes for braindead mistakes I've made? 2) Is there some way to have syntax like 'seqcopy<S>(iter)' where the iterator type is deduced but the sequence type is explicit? Thanks, Rhys #define BOOST_TEST_MODULE $Id: test_mpl.cc 996 2010-03-18 07:04:23Z rhys $ #include <boost/mpl/empty.hpp> #include <boost/mpl/front.hpp> #include <boost/mpl/pop_front.hpp> #include <boost/mpl/vector_c.hpp> #include <boost/utility.hpp> #include <boost/test/included/unit_test.hpp> using boost::disable_if; using boost::mpl::empty; using boost::mpl::front; using boost::mpl::pop_front; template<typename S, typename FI, class Enable = void> struct seqcopy { void operator()(FI fi) {} }; template<typename S, typename FI> struct seqcopy<S, FI, typename disable_if<typename empty<S>::type>::type> { void operator()(FI fi) { *fi++ = front<S>::type::value; seqcopy<typename pop_front<S>::type, FI>()(fi); } }; BOOST_AUTO_TEST_CASE( testing ) { typedef boost::mpl::vector_c<int,2,4,6> values; int a[3]; seqcopy<values, int*>()(a); BOOST_CHECK_EQUAL(a[0], 2); BOOST_CHECK_EQUAL(a[1], 4); BOOST_CHECK_EQUAL(a[2], 6); }

4 4

Problem Using managed_shared_memory
by med ennemri 18 Mar '10

18 Mar '10

Hi all I have a problem using managed_shared_memory , I can only read strings that has size less that 16 characters. This is my server code ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// #include <boost/interprocess/managed_shared_memory.hpp> #include <utility> #include <iostream> int main () { using namespace boost::interprocess; typedef std::pair<std::string, int> FileState; try{ shared_memory_object::remove("MySharedMemory"); managed_shared_memory segment(create_only, "MySharedMemory", 655360); std::string str1 = "aaaaaaaaaaaaaaabbb";//client can only read <16 sized strings std::string str2 = "baaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"; std::string str3 = "caaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"; std::string *instance1 = segment.construct<std::string>("1")(str1); std::string *instance2 = segment.construct<std::string>("2")(str2); std::string *instance3 = segment.construct<std::string>("3")(str3); while(segment.get_num_named_objects() > 0) std::cout<<"Memory is ready"<<std::endl; } catch(...){ shared_memory_object::remove("MySharedMemory"); throw; } shared_memory_object::remove("MySharedMemory"); return 0; } ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// Here is my client code ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// #include <boost/interprocess/managed_shared_memory.hpp> #include <cstddef> #include <utility> #include <cassert> #include <iostream> #include <fstream> int main (int argc, char *argv[]) { using namespace boost::interprocess; try{ managed_shared_memory segment(open_only, "MySharedMemory"); std::pair<std::string*, std::size_t> res; int a = 1; std::ostringstream oss; oss<<a; res = segment.find<std::string> (const_cast<char*>(oss.str().c_str())); if(res.second != 0) { std::cout<<*(res.first)<<std::endl; segment.destroy<std::string>(const_cast<char*>(oss.str().c_str())); } } catch(...){ shared_memory_object::remove("MySharedMemory"); throw; } shared_memory_object::remove("MySharedMemory"); return 0; } ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// Best Regards

2 2