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

[propertytree] more beginners questions
by Dietmar Hummel 21 Aug '10

21 Aug '10

Hi! I have some more questions regarding property tree (for parsing xml) and maybe someone could help me. Lets assume we have the following xml file to read: <a> <b> <c><d>1</d></c> <c><d>2</d></c> <c><d>3</d></c> <x></x> </b> </a> Now, my idea was to parse all the c-nodes by calling a specific function for each of them. Something like this... void parseDNode( const boost::property_tree::ptree& v ) { //... } void parseCNode( const boost::property_tree::ptree::value_type& v ) { parseDNode( v.second.get_child( "d" ) ); } void load( const std::string& filename ) { boost::property_tree::ptree pt; boost::property_tree::read_xml(filename, pt); BOOST_FOREACH( boost::property_tree::ptree::value_type &v, pt.get_child( "a.b.c" ) ) { parseCNode( v ); } } ---- Now I have several questions for this: 1. The code doesn't work as expected, it just handles the first one of the c nodes 2. The function parseCNode() also doesn't work. I assumed that v is a reference to the c node where I have to claim the d-node by get_child. But it seems v is already pointing to d? 3. I don't understand why BOOST_FOREACH is working with value_type here. As far as I see, get_child returns a self_type which should be a ptree? See also the different parameter lists for parseCNode and parseDNode... Thanks in advance.... Regards Dietmar

2 1

MPICH2 + Boost.MPI Collective Problems
by Stephan Hackstedt 21 Aug '10

21 Aug '10

Hi there, i have a big problem by running MPI programs which use the Boost.MPI library. When i'm trying to run programs on *more *than one node, collective operations like communicator::barrier<http://boost.org/doc/libs/1_44_0/doc/html/boost/mpi/communicator.html#id918…>or broadcast,<http://boost.org/doc/libs/1_44_0/doc/html/boost/mpi/broadcast.html>or even the environment<http://boost.org/doc/libs/1_44_0/doc/html/boost/mpi/environment.html>destructor (cause of FINALIZE, which is colletive) causing the programm to crash. I got errors like this : *[1]terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<boost::mpi::exception> >' [1] what(): MPI_Barrier: Other MPI error, error stack: [1]PMPI_Barrier(362).................: MPI_Barrier(MPI_COMM_WORLD) failed [1]MPIR_Barrier_impl(255)............: [1]MPIR_Barrier_intra(79)............: [1]MPIC_Sendrecv(186)................: [1]MPIC_Wait(534)....................: [1]MPIDI_CH3I_Progress(184)..........: [1]MPID_nem_mpich2_blocking_recv(895): [1]MPID_nem_tcp_connpoll(1746).......: Communication error with rank 0: * I also tested this with the simple broadcast example from the Boost.MPI tutorial - same errors.. But when using the original MPI equivalent without the Boost.MPI library, such as MPI_Barrier<http://www.mpi-forum.org/docs/mpi-11-html/node66.html#Node66>, the programm runs well. I am using MPICH2 on Ubuntu 10.04 platforms. Someone had problems like this or know a fix for that? Regards, stephan

1 2

Using unordered_set or unordered_map to hold small array(s) of ints
by B Hart 21 Aug '10

21 Aug '10

Can someone please provide an example of how to best use unordered_map/set to store/lookup many small arrays of integers. e.g. unsigned int I[5] = {1000,3344455,12455222,8832232}; I'm guessing that there is no facility for using raw C++ arrays, and that one must wrap them in a struct. Thanks in advance.

5 9

Beginner question: Why boost::signal creates another instance of the slot object ?
by Mauricio Gomes 20 Aug '10

20 Aug '10

Hi, I am running the program below (Vista with boost 1.43, mingw/gcc 4.5.1). I would expect slot.x to be 5 but it is zero. Why boost::signal creates another instance of the slot object ? What am I missing here ? Thank you in advance for your attention, Mau. ---------------------------------------------------------------- #include <iostream> #include <boost/signal.hpp> struct S { int x; S (): x(0) {} void operator () (int i) { std::cout << "&slot1 = " << this << std::endl; x = i; std::cout << "slot1.x = " << this->x << std::endl; } }; int main () { S slot; boost::signal<void (int)> sig; std::cout << "&slot2 = " << &slot << std::endl; sig.connect (slot); sig (5); std::cout << "slot2.x = " << slot.x << std::endl; return 0; } -------------------------------------------------------------------- Output: &slot2 = 0x22fee4 &slot1 = 0x341448 slot1.x = 5 slot2.x = 0

4 8

How to best use unordered_set or unordered_map to hold small array(s) of ints
by B Hart 20 Aug '10

20 Aug '10

I posted this last night, but not sure it went through to the list as I don't see it in my sent items and I never got any responses. Sorry if it is duplicate. I'm hoping someone could offer an example or advice on how to best use unordered_set or unordered_map to hold many small arrays of unsigned integers. I would like to use the built in hash function rather than have to create one. Also, since I have many of these small arrays, keeping the memory footprint small and the hash time at a minimum are important concerns. A typical int array would hold 5 unsigned ints. Maybe someone could offer example code or suggestions. The samples I see are always the trivial case of just entering a single number or string as key or value. Thanks.

2 1

[property tree] silly beginners question
by Dietmar Hummel 20 Aug '10

20 Aug '10

Hi all! I am trying to read an xml file with property tree (first try). Because I have dots int the tag names I must use the extended version of the get functions. So here is the slightly modified example code, which sadly does'nt compile on vc2008. Probably its a totally silly error on my side... // Loads debug_settings structure from the specified XML file void load(const std::string &filename) { // Create an empty property tree object using boost::property_tree::ptree; ptree pt; // Load the XML file into the property tree. If reading fails // (cannot open file, parse error), an exception is thrown. read_xml(filename, pt); // Get the filename and store it in the m_file variable. // Note that we construct the path to the value by separating // the individual keys with dots. If dots appear in the keys, // a path type with a different separator can be used. // If the debug.filename key is not found, an exception is thrown. std::string m_file = pt.get< std::string >( '/', "debug/filename" ); // Get the debug level and store it in the m_level variable. // This is another version of the get method: if the value is // not found, the default value (specified by the second // parameter) is returned instead. The type of the value // extracted is determined by the type of the second parameter, // so we can simply write get(...) instead of get<int>(...). int m_level = pt.get( '/', "debug/level", 0); // Iterate over the debug.modules section and store all found // modules in the m_modules set. The get_child() function // returns a reference to the child at the specified path; if // there is no such child, it throws. Property tree iterators // are models of BidirectionalIterator. std::vector< std::string > m_modules; BOOST_FOREACH( ptree::value_type &v, pt.get_child( '/', "debug/modules" ) ) { m_modules.push_back( v.second.data() ); } } vc says: 1>c:\projekte\proptreetest\proptreetest\proptreetest.cpp(23) : error C2664: 'std::basic_string<_Elem,_Traits,_Ax> boost::property_tree::basic_ptree<Key,Data>::get<std::string>(const boost::property_tree::string_path<String,Translator> &,const Type &) const': Konvertierung des Parameters 1 von 'char' in 'const boost::property_tree::string_path<String,Translator> &' nicht möglich 1> with 1> [ 1> _Elem=char, 1> _Traits=std::char_traits<char>, 1> _Ax=std::allocator<char>, 1> Key=std::string, 1> Data=std::string, 1> String=std::string, 1> Translator=boost::property_tree::id_translator<std::string>, 1> Type=std::string 1> ] 1> and 1> [ 1> String=std::string, 1> Translator=boost::property_tree::id_translator<std::string> 1> ] So, it looks to me like the compiler does'nt know about the extended get functions... Regards Dietmar

2 1

[Log] How to use?
by Olivier Tournaire 20 Aug '10

20 Aug '10

Hi all, i am currently trying to use Boost.Log in my application. After fighting the two last days, I still do not understand how to correctly use the lib. I would like the result to be the same as the one provided by the example advanced_use.cpp. So, I wrote a simple class to make some tests. As I want to use the logging library in a shared library (an application plugin), I made a simple test: compile the class which uses Log in a shared library and link an executable against this library. I also have created an executable which directly compiles the class (so does not link against the shared library). Basically, I have a test_boost_log class. The constructor is a copy/paste of the code from the example advanced_use.cpp: #include "test_boost_log.hpp" #include <cassert> #include <iostream> #include <fstream> #include <boost/shared_ptr.hpp> #include <boost/log/common.hpp> #include <boost/log/filters.hpp> #include <boost/log/formatters.hpp> #include <boost/log/attributes.hpp> #include <boost/log/sinks.hpp> #include <boost/log/utility/empty_deleter.hpp> #include <boost/log/utility/scoped_attribute.hpp> namespace logging = boost::log; namespace fmt = boost::log::formatters; namespace flt = boost::log::filters; namespace sinks = boost::log::sinks; namespace attrs = boost::log::attributes; namespace src = boost::log::sources; namespace keywords = boost::log::keywords; using boost::shared_ptr; // Here we define our application severity levels. enum severity_level { normal, notification, warning, error, critical }; // The formatting logic for the severity level template< typename CharT, typename TraitsT > inline std::basic_ostream< CharT, TraitsT >& operator<< ( std::basic_ostream< CharT, TraitsT >& strm, severity_level lvl) { static const char* const str[] = { "normal", "notification", "warning", "error", "critical" }; if (static_cast< std::size_t >(lvl) < (sizeof(str) / sizeof(*str))) strm << str[lvl]; else strm << static_cast< int >(lvl); return strm; } int foo(src::logger& lg) { BOOST_LOG_FUNCTION(); BOOST_LOG(lg) << "foo is being called"; return 10; } test_boost_log::test_boost_log() { // For now we only create a text output sink: typedef sinks::synchronous_sink< sinks::text_ostream_backend > text_sink; shared_ptr< text_sink > pSink(new text_sink); { text_sink::locked_backend_ptr pBackend = pSink->locked_backend(); shared_ptr< std::ostream > pStream(&std::clog, logging::empty_deleter()); pBackend->add_stream(pStream); shared_ptr< std::ofstream > pStream2(new std::ofstream("sample.log")); assert(pStream2->is_open()); pBackend->add_stream(pStream2); } logging::core::get()->add_sink(pSink); src::logger lg; BOOST_LOG(lg) << "Hello, World!"; pSink->locked_backend()->set_formatter(fmt::stream << fmt::attr("LineID") // First an attribute "LineID" is written to the log << " [" << fmt::date_time< boost::posix_time::ptime >("TimeStamp", "%d.%m.%Y %H:%M:%S.%f") << "] [" << fmt::attr< severity_level >("Severity", std::nothrow) << "] [" << fmt::time_duration< boost::posix_time::time_duration >("Uptime") << "] [" // then this delimiter separates it from the rest of the line << fmt::if_(flt::has_attr("Tag")) [ fmt::stream << fmt::attr< std::string >("Tag") << "] [" // yet another delimiter ] << fmt::named_scope("Scope", keywords::iteration = fmt::reverse) << "] " << fmt::message()); // here goes the log record text shared_ptr< logging::attribute > pCounter(new attrs::counter< unsigned int >(1)); // Since we intend to count all logging records ever made by the application, // this attribute should clearly be global. logging::core::get()->add_global_attribute("LineID", pCounter); // And similarly add a time stamp shared_ptr< logging::attribute > pTimeStamp(new attrs::local_clock()); logging::core::get()->add_global_attribute("TimeStamp", pTimeStamp); // And an up time stopwatch BOOST_LOG_SCOPED_THREAD_ATTR("Uptime", attrs::timer); // Let's also track the execution scope from which the records are made boost::shared_ptr< logging::attribute > pNamedScope(new attrs::named_scope()); logging::core::get()->add_thread_attribute("Scope", pNamedScope); // We can mark the current execution scope now - it's the 'main' function BOOST_LOG_FUNCTION(); // Let's try out the counter attribute and formatting BOOST_LOG(lg) << "Some log line with a counter"; BOOST_LOG(lg) << "Another log line with the counter"; { BOOST_LOG_NAMED_SCOPE("Tagging scope"); // Now these lines will be highlighted with the tag BOOST_LOG(lg) << "Some tagged log line"; BOOST_LOG(lg) << "Another tagged log line"; } // And this line is not highlighted anymore BOOST_LOG(lg) << "Now the tag is removed"; pSink->set_filter( flt::attr< severity_level >("Severity", std::nothrow) >= warning // Write all records with "warning" severity or higher || flt::attr< std::string >("Tag", std::nothrow).begins_with("IMPORTANT")); // ...or specifically tagged src::severity_logger< severity_level > slg; // These two lines test filtering based on severity BOOST_LOG_SEV(slg, normal) << "A normal severity message, will not pass to the output"; BOOST_LOG_SEV(slg, error) << "An error severity message, will pass to the output"; { // Next we try if the second condition of the filter works // We mark following lines with a tag BOOST_LOG_SCOPED_THREAD_TAG("Tag", std::string, "IMPORTANT MESSAGES"); BOOST_LOG(slg) << "Some really urgent line"; } pSink->reset_filter(); // And moreover, it is possible to nest logging records. For example, this will // be processed in the order of evaluation: BOOST_LOG(lg) << "The result of foo is " << foo(lg); } I have added another method to test the logger state: void test_boost_log::a_test_method_to_see_if_the_logger_is_still_working() { src::severity_logger< severity_level > slg; BOOST_LOG_SEV(slg, error) << "An error severity message, will pass to the output"; } My main: #include "test_boost_log.hpp" int main(int argc, char** argv) { test_boost_log tbl; tbl.a_test_method_to_see_if_the_logger_is_still_working(); return 0; } Here is the output of the program: Hello, World! 1 [17.08.2010 21:22:43.083302] [] [00:00:00.000076] [test_boost_log::test_boost_log()] Some log line with a counter 2 [17.08.2010 21:22:43.083445] [] [00:00:00.000217] [test_boost_log::test_boost_log()] Another log line with the counter 3 [17.08.2010 21:22:43.083484] [] [00:00:00.000256] [Tagged line] [Tagging scope<-test_boost_log::test_boost_log()] Some tagged log line 4 [17.08.2010 21:22:43.083517] [] [00:00:00.000289] [Tagged line] [Tagging scope<-test_boost_log::test_boost_log()] Another tagged log line 5 [17.08.2010 21:22:43.083556] [] [00:00:00.000328] [test_boost_log::test_boost_log()] Now the tag is removed 6 [17.08.2010 21:22:43.083598] [error] [00:00:00.000370] [test_boost_log::test_boost_log()] An error severity message, will pass to the output 7 [17.08.2010 21:22:43.083633] [normal] [00:00:00.000405] [IMPORTANT MESSAGES] [test_boost_log::test_boost_log()] Some really urgent line 9 [17.08.2010 21:22:43.083673] [] [00:00:00.000444] [int foo(boost::log_mt_posix::sources::logger&)<-test_boost_log::test_boost_log()] foo is being called 8 [17.08.2010 21:22:43.083668] [] [00:00:00.000440] [test_boost_log::test_boost_log()] The result of foo is 10 *terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<boost::log_mt_posix::missing_value> >'* * what(): Requested attribute value not found* So, the logger throws, sample.log is created, but empty. The library and the executable are compiled with -DBOOST_LOG_USE_CHAR and -DBOOST_LOG_DYN_LINK If someone could help... Best regards, Olivier PS: boost_1.42, log_1.0 (from sourceforge), g++4.3.4, 64bits ubuntu

2 10

A question about "prim_minimum_spanning_tree.hpp"
by 王伟 20 Aug '10

20 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),there's a "segmnetation error" problem.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 in the attachment.

2 1

Some trouble using mutexes
by Ernest Galbrun 20 Aug '10

20 Aug '10

Dear boost users, I would be really grateful if someone could give me a hint about what I am doing wrong with the mutex and thread class (I am using boost 1.43, on windows XP running visual c++ 2010, the library has been built using the bjam utility). I have a main thread that creates an object of class robot : class Robot { private: boost::mutex tcpLock; boost::thread continuousThread; ... } Robot k3; this thread launches a thread that makes continuous calls to a member function of the robot. Since this function uses network ressources, I want it to be safe so i use a lock The function called is the only function in my program that uses this lock : int Robot::sendMsg(string msg, int n, vector<string>* answer) { tcpLock.lock(); (do stuff with the tcp io service) tcpLock.unlock(); return 0; } Of course, I also call this function from the main thread. And I have disruptive behavior, which I don't understand. When debugging, I see that the 'active_count' of the tcpLock object stays negative or ==0, which seem absolutely wrong to me, but I don't understand how it could ever be anything other than 0 or 1. Any clue, advice, web link, or soft injures are welcome ! -- Ernest Galbrun

2 1

ptree info format - block commenting
by Hite, Christopher 20 Aug '10

20 Aug '10

http://www.boost.org/doc/libs/1_41_0/doc/html/boost_propertytree/parsers .html#boost_propertytree.parsers.info_parser I'm using the ptree info format. How about C-style commenting out of a block of config stuff like? /* Stuff I { don't need } */ Chris

1 0