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

[Interprocess] "You cannot bind an lvalue to an rvalue reference" in 1.44 using VC10
by Phil Ratzloff 20 Aug '10

20 Aug '10

When I attempt to compile a project using boost::interprocess I get compile errors when I include managed_mapped_file.hpp when using VC10. I can create a simple project that is just the following code: #include <boost\interprocess\managed_mapped_file.hpp> int main(int argc, char* argv[]) { return 0; } When compiling, I get the following errors: Empty.cpp c:\lib\boost_1_44_0\boost\interprocess\detail\move.hpp(342): error C2440: 'return' : cannot convert from 'boost::interprocess::mapped_region' to 'boost::interprocess::mapped_region &&' You cannot bind an lvalue to an rvalue reference c:\lib\boost_1_44_0\boost\interprocess\mapped_region.hpp(159) : see reference to function template instantiation 'boost::interprocess::mapped_region &&boost::interprocess::move<boost::interprocess::mapped_region&>(T)' being compiled with [ T=boost::interprocess::mapped_region & ] c:\lib\boost_1_44_0\boost\interprocess\detail\move.hpp(342): error C2440: 'return' : cannot convert from 'boost::interprocess::detail::file_wrapper' to 'boost::interprocess::detail::file_wrapper &&' You cannot bind an lvalue to an rvalue reference c:\lib\boost_1_44_0\boost\interprocess\detail\file_wrapper.hpp(63) : see reference to function template instantiation 'boost::interprocess::detail::file_wrapper &&boost::interprocess::move<boost::interprocess::detail::file_wrapper&>(T)' being compiled with [ T=boost::interprocess::detail::file_wrapper & ] c:\lib\boost_1_44_0\boost\interprocess\detail\move.hpp(342): error C2440: 'return' : cannot convert from 'boost::interprocess::file_mapping' to 'boost::interprocess::file_mapping &&' You cannot bind an lvalue to an rvalue reference c:\lib\boost_1_44_0\boost\interprocess\file_mapping.hpp(62) : see reference to function template instantiation 'boost::interprocess::file_mapping &&boost::interprocess::move<boost::interprocess::file_mapping&>(T)' being compiled with [ T=boost::interprocess::file_mapping & ] When I look at move.hpp(342) I find the following code: #if defined(BOOST_MOVE_DOXYGEN_INVOKED) ...unused code... #else template <class T> inline typename remove_reference<T>::type&& move(T&& t) { return t; } #endif It is complaining about the "return t;" statement. I have just moved up from boost 1.43, where I have been working without problems with VC10. Is there something I should be doing to prevent this, or is it a bug?

1 0

[log] calling init_from_stream() or init_from_settings() more than once (reinitialize at runtime)
by Oskar Nordquist 19 Aug '10

19 Aug '10

Hello, I've just started experimenting with the unofficial Boost.Log by Andrey Semashev. I would like to know the defined behaviour of calling init_from_stream() and init_from_settings() more than once, i.e. is it possible to re-initialize with new settings at run time? Basically I would like to do this: boost::log::settings setts; setts["Sink:File"]["Filter"] = "%Severity% > 4"; setts["Sink:File"]["Destination"] = "TextFile"; setts["Sink:File"]["FileName"] = "MyApp_%3N.log"; setts["Sink:File"]["AutoFlush"] = true; setts["Sink:File"]["RotationSize"] = 10 * 1024 * 1024; // 10 MiB boost::log::init_from_settings(setts); // ... later on change severity filter to >= 0 setts["Sink:File"]["Filter"] = "%Severity% >= 0"; boost::log::init_from_settings(setts); This does not work. It seems I am able to add new sinks without problem, but modifying existing sinks is not possible. I think that being able to reinitialize at runtime is an absolutely vital feature. Being able to change a settings file and then just reinit to turn on logging! Please tell me this is already supported and I am doing something wrong, or at least that this will be supported in the future :) Regards, Oskar

1 0

[mpl] Reconstructing an mpl::list from mpl::l_item's.
by AJG 19 Aug '10

19 Aug '10

Hi there, Using mpl::remove, et al. I can eliminate types from an mpl::list. However, it appears that the result is not an actual mpl::list, but some sort of indexed, cons-like structure composed of mpl::l_item types. Is there a way to recreate an mpl::list proper using these l_item's? E.g. typedef mpl::list<foo, bar, qux> L; typedef mpl::remove<L, bar>::type R; typedef mpl::???<R>::type => mpl::list<foo, qux>; Thanks!

2 1

filesystem lambda and for_each
by Jean-Christophe Roux 19 Aug '10

19 Aug '10

hello, Despite some research, I am struggling to combine boost filesystem, boost lambda and STL for_each. 32 std::string str = ""; 33 std::string path = "/tmp"; 34 boost::filesystem::directory_iterator it(path); 35 while(it != boost::filesystem::directory_iterator()) 36 { 37 str += it->path().filename() + "<br>"; 38 ++it; 39 } works fine but 40 std::for_each(boost::filesystem::directory_iterator(path), boost::filesystem::directory_iterator(), 41 str += *boost::lambda::_1.path().parent_path() ); does not compile with the error message that : error: ‘const struct boost::lambda::lambda_functor<boost::lambda::placeholder<1> >’ has no member named ‘path’ I don't understand why and I don't really see how I could have this work. Also, I am wondering why with _1 instead of boost::lambda::_1 I am getting: error: reference to ‘_1’ is ambiguous /usr/local/include/boost/lambda/core.hpp:69: thanks!

2 1

Throwing an exception in a packaged_task: is there a clear example?
by Mark Wilson 19 Aug '10

19 Aug '10

Hello, I am throwing an exception from a packaged_task, and then trying to catch it with f.get() (where f is a unique_future); it doesn't work, the program just terminates with: terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::unknown_exception>' what(): std::exception Is there a simple example somewhere, or could someone provide one, that makes clear how this mechanism is supposed to work? The Threads documentation does not show how to do this... Here is my simple program, based on the example in the docs: #include <boost/thread.hpp> #include <boost/thread/future.hpp> #include <boost/exception/all.hpp> struct MyException { MyException() { } }; int calculate_the_answer_to_life_the_universe_and_everything() { throw MyException(); return 42; } int main() { boost::packaged_task<int> task(calculate_the_answer_to_life_the_universe_and_everything); boost::unique_future<int> f(task.get_future()); boost::thread t(boost::move(task)); try { f.get(); } catch(MyException& ex) { std::cout << "EXCEPTION!" << std::endl; } } Thanks, Mark Wilson

2 2

[serialization] tuples and portable_binary_archive
by Richard Ulrich 19 Aug '10

19 Aug '10

I used boost serialize with some smaller projects, which always worked nicely. But now I want to use it on a bigger one. And I wonder why some things seem to be missing. - tuple: There is code on the Internet in various places for bridging boost::serialize with boost::tuple. But why isn't there an official version in either boost::serialize or boost::tuple? - portable_binary_archive: There are also some implementations around, one even in the examples of boost::serialization. But why is it not in the official library itself? Sure I could use the code I found, but it would feel better if it was part of the library. That's especially true for serialization where I want to be 99.99999999% sure, the data can always be read back in. Rgds Richard

2 1

[serialization] Different behaviors between archive and its polymorphic variant
by Maxime van Noppen 19 Aug '10

19 Aug '10

Hi, I'm trying to make a custom archive class that works either with the template flavor or via the polymorphic_archive. It compiles but produces a different behavior and I can't find why. Basically I did a trivial oarchive class that overloads (via save_override) how are treated NVP. When used with the template version it works, whereas with the polymorphic version it doesn't. The code is attached. There is a preprocessor switch to choose the template or polymorphic version (-DPOLY). $ g++ poly.cc -lboost_serialization && ./a.out <name=u><name=id_>42</name=id_> </name=u> $ g++ -DPOLY poly.cc -lboost_serialization && ./a.out 42 This is with Boost 1.43 and g++ 4.5.0. I looked at various archive implementations but couldn't find any code that I could relate to this problem. Thanks, -- Maxime

3 9

A problem in using "prim_minimum_spanning_tree.hpp"
by 王伟 19 Aug '10

19 Aug '10

Hi, I'm studying MST algorithms in GPU recently.I need BGL implementation as coparison,But I have a problem in using prim_minimum_spanning_tree.hpp.The problem is:when I use large input(e.g. >1M),the result is wrong,all the parent of vertices is "no parent".But when the input(e.g. 1K) is small,the result is noproblem. I haven't found know the reason yet.My complier is gcc 4.1.2,OS is RHEL5. Help would be appreciated. Thanks in advance. Wang wei The complete code is: #include <iostream> #include <boost/graph/adjacency_list.hpp> #include <boost/graph/prim_minimum_spanning_tree.hpp> #include <fstream> #include <string.h> using namespace std; int main() { using namespace boost; typedef adjacency_list < vecS, vecS, undirectedS, property<vertex_distance_t, long>, property < edge_weight_t, int > > Graph; typedef std::pair < long, long >E; //read graph data from file long num_nodes,num_edges,u,v; int w; ifstream infile; infile.open("convertResult.txt"); if(!infile) { cerr<<"error:unable to open file."<<endl; return -1; } infile>>num_nodes>>num_edges; E edges[num_edges]; int weights[num_edges]; for(long i=0;i<num_edges;i++) { infile>>u>>v>>w; edges[i] = make_pair(u,v); weights[i] = w; } //construct input parameters Graph g(edges, edges + sizeof(edges) / sizeof(E), weights, num_nodes); property_map<Graph, edge_weight_t>::type weightmap = get(edge_weight, g); std::vector < graph_traits < Graph >::vertex_descriptor > p(num_vertices(g)); //call the BGL PRIM prim_minimum_spanning_tree(g, &p[0]); //output the result ofstream outfile; outfile.open("primResult.txt"); for (std::size_t i = 0; i != p.size(); ++i) if (p[i] != i) outfile<< "parent[" << i << "] = " << p[i] << std::endl; else outfile<< "parent[" << i << "] = no parent" << std::endl; return EXIT_SUCCESS; }

1 0

[Serialization] serializing derived class through base class pointer
by Dante Stroe 19 Aug '10

19 Aug '10

Hi, I have a problem with serializing a derived classes through a pointer to a base class. I tried registering with the syntax from the documentation: |ar.template register_type<derived_one>(); | , which works well in case the derived class has a default constructor (just like the example). I just don't know what the syntax is for the case where the derived class does not have a default constructor (if I use the same syntax to register the derived class then I get a compilation error about the fact that there is no. I can provide an example in case this is not clear enough. Or, is it an absolute requirement for all derived classes to have a default constructor? Thank you and best regards, Dante

1 0

[Filesystem] is_directory( ) for a non-existent path.
by Robert Jones 19 Aug '10

19 Aug '10

This is probably something of a shot in the dark, but somewhere between 1.33.1 and 1.37 the behaviour of is_directory() for a non-existent directory changed from failing (or throwing) in 1.33.1 to simply returning false (1.37). Can anyone enlighten me when this was. Trac references would be lovely. Thx - Rob.

1 0