May 2010 - Boost-users - lists.boost.org

Problem with signals2 and shared_ptr
by Glenn Macgregor 26 May '10

26 May '10

Hi All, I am using boost 1.41.0 on Windows with Visual Studio 2008 and am seeing what seems to be an issue with signals2 and shared_ptr. The problem is that the shared_ptr does not decrement is reference count after I disconnect the signals2::connection. This is probably best illustrated with some code. ===== Test.h===== #ifndef _CLASS1_H #define _CLASS1_H #include <boost/signals2/signal.hpp> #include <boost/smart_ptr.hpp> class Callee : public boost::enable_shared_from_this<Callee> { public: Callee() : mInt(0) {}; virtual ~Callee() {}; void func1() { mInt++; }; private: int mInt; }; class Caller { public: typedef boost::signals2::signal<void (boost::shared_ptr<Callee>)> OnRun; public: Caller() : mOnRun() {}; ~Caller() {}; void run(boost::shared_ptr<Callee> sp) { boost::signals2::connection conn = mOnRun.connect(boost::bind(&Callee::func1, sp)); _sleep(1000); conn.disconnect(); }; private: OnRun mOnRun; }; #endif ===== End Test.h===== ======main.cpp====== int main(int argc, char **argv) { Caller caller; { boost::shared_ptr<Callee> sp(new Callee); caller.run(sp); } } ======End main.cpp====== Running this in the debugger I would expect to hit a breakpoint in the Callee class before the Caller class, this is not the case I destruct Caller before Callee. Having dug in a bit I see the use_count of the shared pointer go to 2 while in the caller.run which I would expect. The strange thing is the use_count does not decrement after the disconnect or exiting the caller.run function. Is there a way to force the connection/signal to release that shared_ptr? Thanks Glenn

2 1

lambda::bind with A::method throw (char*)
by Archie14 26 May '10

26 May '10

I have a problem compiling this: class A { public: void method(); void methodwiththrow() throw (char*); }; std::vector<A*> lst; // compiles OK. std::for_each (lst.begin(), lst.end(), boost::lambda::bind(&A::method, boost::lambda::_1)); // compile fails. std::for_each (lst.begin(), lst.end(), boost::lambda::bind(&A::methodwiththrow, boost::lambda::_1)); What might be the reason?

1 0

[ASIO|Thread] Advice for removing MFC/ afxwin.h dependancies
by smbrown＠teledyne.com 26 May '10

26 May '10

Hi all, I've inherited a windows C++ application which I've extended and that I'm now preparing for porting to Linux. I'm going through the app removing any MFC code replacing it with standard C++. I use Boost asio and thread and they seem to be requiring me to include afxwin.h before any boost includes or it won't build complaining that Winsock.h or Windows.h has alreday been included. I can bodge around those errors, but then the code doesn't link as new and delete have already been referenced from an afx lib. Is there any general advice for Boost/MFC coexistance or MFC removal? Thanks, Simon ++++ IMPORTANT NOTICE ++++ This email may contain confidential information and will be protected by copyright. It is intended only for the use of the person(s) to whom it is addressed and any unauthorised use (including forwarding or dissemination the contents) may be unlawful. If you receive this email by mistake, please advise the sender immediately by using the reply facility on your email software and delete the email from your computer. Opinions, conclusions and other information in this email that do not relate to the official business of this organisation shall be understood as neither given nor endorsed by it. If this message forms part of a quotation for the supply of goods, the quotation is an invitation to treat only. No contract will come into effect until the Company issuing such quotation issues an acknowledgment of order for the supply of such goods. The contract will be on the terms of such acknowledgment of order. ------------------------------------------------------------------------ Teledyne Defence Limited phone: +44 (0) 1274 531602 web: www.teledynedefence.co.uk ------------------------------------------------------------------------

3 9

boost::asio::deadline_timer::expires_at() seems to have a bug
by Hite, Christopher 26 May '10

26 May '10

See the #if below. I'm pretty sure these should do the same thing. The expires_from_now() version does what I'd expect and calls onTimer() once a second. The expires_at() version never calls onTimer(). Even though 'expires_at() const' returns 1s from now: doit now =2010-May-21 15:07:24.882854 expires_at =2010-May-21 15:07:25.882854 <and nothing more> //********************************************************************** *********** #include <boost/asio.hpp> #include <boost/date_time/posix_time/posix_time.hpp> #include <iostream> using std::cout; using namespace boost::asio; using namespace boost::posix_time; io_service ioService; deadline_timer timer(ioService); seconds s1(1); void doit(); void onTimer(const boost::system::error_code&){ doit(); //repeat forever } void doit(){ ptime now=microsec_clock::local_time(); cout<< "doit now ="<< now <<std::endl; #if 0 timer.expires_from_now(s1); //works #else timer.expires_at( now+s1); //doesn't work #endif timer.async_wait( onTimer ); cout<< "expires_at ="<< timer.expires_at()<< std::endl ; } void timerTest(){ doit(); io_service::work w(ioService); ioService.run(); } int main(){ timerTest(); return 0; }

2 2

[preprocessor] one comma too many
by Manfred Doudar 26 May '10

26 May '10

Hello, Near sure this is trivial, but I don't see it, and hope someone might point out the obvious: #include <boost/preprocessor/iteration/local.hpp> #include <boost/preprocessor/repetition.hpp> #include <boost/preprocessor/repetition/enum_params.hpp> #define OPERATOR(z, n, _) \ \ BOOST_PP_ENUM_PARAMS(n, T r) /**/ #define BOOST_PP_LOCAL_MACRO(n) OPERATOR(_, n, _) #define BOOST_PP_LOCAL_LIMITS(1, 3) #include BOOST_PP_LOCAL_ITERATE() #undef BOOST_PP_LOCAL_MACRO #undef BOOST_PP_LOCAL_LIMITS #undef OPERATOR ..The above works as expected: T r0 T r0, T r1 T r0, T r1, T r2 However, I would like to replace "T" with the following instead: std::pair<T, T> ie: I want to write something to the effect of: BOOST_PP_ENUM_PARAMS(n, std::pair<T, T> r) Of course, BOOST_PP_ENUM_PARAMS now sees one too many arguments, and errors. Any ideas on how to achieve this would be most appreciated. With thanks in advance, -- Manfred

3 3

Exception on windows::stream_handle!
by Lloyd 26 May '10

26 May '10

Hello, We are trying to create a "windows::stream_handle" from a native Windows HANDLE. But it throws an exception with "The parameter is incorrect". The code is given below. What could be the reason? (Another IOservice is already running in the system) HANDLE FileHandle=CreateFile("MyFile",FILE_WRITE_DATA,0,NULL,CREATE_NEW,FILE_ATTRIBUTE_NORMAL,NULL); if(FileHandle==INVALID_HANDLE_VALUE) { return; } try { io_service ios; windows::stream_handle StreamHandle(ios,FileHandle); .... .... } catch(std::exception& e) { Log(e.what()); //The parameter is incorrect } Thanks, Lloyd ______________________________________ Scanned and protected by Email scanner

1 0

PropertyTree compatibility with Visual C++
by Warren Seine 26 May '10

26 May '10

Hi, I'm trying to port a C++ application to Microsoft Windows using the Visual C++ environment (Express edition, if it matters) and I'm having some trouble with Boost.PropertyTree. The following code works with GCC but fails to compile on Visual C++ 10.0: #include <boost/property_tree/ptree.hpp> int main() { using boost::property_tree::ptree; ptree pt; return 0; } Here is a sample of the output: [..]\boost\property_tree\detail\ptree_implementation.hpp(30): error C2440: 'specialization': cannot convert from 'const std::string std::_Pair_base<_Ty1,_Ty2>::* ' to 'const std::basic_string<_Elem,_Traits,_Ax> std::pair<_Ty1,_Ty2>::* ' with [ _Ty1=const std::string, _Ty2=boost::property_tree::basic_ptree<std::string,std::string> ] and [ _Elem=char, _Traits=std::char_traits<char>, _Ax=std::allocator<char>, _Ty1=const std::string, _Ty2=boost::property_tree::basic_ptree<std::string,std::string> ] Can someone reproduce the bug? If so, is there some kind of workaround, even if it requires hacking the library? Thank you, -- Warren Seine

3 3

[thread] race condition detected with valgrind drd
by Jorge Moraleda 26 May '10

26 May '10

Dear all, I must be doing something wrong, but I can't figure out what. I would appreciate some help. Thank you in advance. The following program: #include <boost/thread.hpp> void threadEntry() { } int main (int argc, char *argv[]) { for(size_t t=0; t<10; t++) { boost::thread(threadEntry); } } runs cleanly with valgrind drd, but this other program produces race conditions when run with valgrind drd (see below). The only difference is "new". #include <boost/thread.hpp> void threadEntry() { } int main (int argc, char *argv[]) { for(size_t t=0; t<10; t++) { new boost::thread(threadEntry); } } This problem is not due to the leak, because this other program: #include <boost/thread.hpp> #include <boost/shared_ptr.hpp> void threadEntry() { } int main (int argc, char *argv[]) { std::vector<boost::shared_ptr<boost::thread> > threads; for(size_t t=0; t<10; t++) { boost::shared_ptr<boost::thread> thread(new boost::thread( threadEntry )); threads.push_back(thread); } for (size_t i = 0; i < threads.size(); ++i) { threads[i]->join(); } } which properly manages the heap also produces the same race conditions. I compile using: g++ -g -L /opt/boost/boost/stage/lib -lboost_thread threads.cpp and I test the program using valgrind using: valgrind --tool=drd ./a.out The output of valgrind is: ==3822== drd, a thread error detector ==3822== Copyright (C) 2006-2009, and GNU GPL'd, by Bart Van Assche. ==3822== Using Valgrind-3.5.0-Debian and LibVEX; rerun with -h for copyright info ==3822== Command: ./a.out ==3822== ==3822== Thread 3: ==3822== Conflicting store by thread 3 at 0x0504b350 size 8 ==3822== at 0x4E417B3: ??? (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E41B28: boost::detail::set_current_thread_data(boost::detail::thread_data_base*) (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E42124: thread_proxy (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4C32870: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==3822== by 0x5D74A03: start_thread (pthread_create.c:300) ==3822== by 0x58D680C: clone (clone.S:112) ==3822== Allocation context: BSS section of /usr/local/lib/libboost_thread.so.1.42.0 ==3822== Other segment start (thread 2) ==3822== at 0x58D67D1: clone (clone.S:84) ==3822== by 0x5D7493F: ??? (allocatestack.c:743) ==3822== by 0x6B8A90F: ??? ==3822== Other segment end (thread 2) ==3822== at 0x5D7A7C4: pthread_once (pthread_once.S:119) ==3822== by 0x4C28ADD: pthread_once (drd_pthread_intercepts.c:518) ==3822== by 0x4E473B3: boost::detail::get_once_per_thread_epoch() (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E4172F: ??? (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E41B28: boost::detail::set_current_thread_data(boost::detail::thread_data_base*) (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E42124: thread_proxy (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4C32870: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==3822== by 0x5D74A03: start_thread (pthread_create.c:300) ==3822== by 0x58D680C: clone (clone.S:112) ==3822== ==3822== Conflicting store by thread 3 at 0x0504b350 size 8 ==3822== at 0x4E417F9: ??? (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E41B28: boost::detail::set_current_thread_data(boost::detail::thread_data_base*) (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E42124: thread_proxy (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4C32870: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==3822== by 0x5D74A03: start_thread (pthread_create.c:300) ==3822== by 0x58D680C: clone (clone.S:112) ==3822== Allocation context: BSS section of /usr/local/lib/libboost_thread.so.1.42.0 ==3822== Other segment start (thread 2) ==3822== at 0x58D67D1: clone (clone.S:84) ==3822== by 0x5D7493F: ??? (allocatestack.c:743) ==3822== by 0x6B8A90F: ??? ==3822== Other segment end (thread 2) ==3822== at 0x5D7A7C4: pthread_once (pthread_once.S:119) ==3822== by 0x4C28ADD: pthread_once (drd_pthread_intercepts.c:518) ==3822== by 0x4E473B3: boost::detail::get_once_per_thread_epoch() (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E4172F: ??? (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E41B28: boost::detail::set_current_thread_data(boost::detail::thread_data_base*) (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E42124: thread_proxy (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4C32870: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==3822== by 0x5D74A03: start_thread (pthread_create.c:300) ==3822== by 0x58D680C: clone (clone.S:112) ==3822== Other segment start (thread 2) ==3822== at 0x58D67D1: clone (clone.S:84) ==3822== by 0x5D7493F: ??? (allocatestack.c:743) ==3822== by 0x6B8A90F: ??? ==3822== Other segment end (thread 2) ==3822== at 0x5D7A7C4: pthread_once (pthread_once.S:119) ==3822== by 0x4C28ADD: pthread_once (drd_pthread_intercepts.c:518) ==3822== by 0x4E473B3: boost::detail::get_once_per_thread_epoch() (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E4172F: ??? (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E41B28: boost::detail::set_current_thread_data(boost::detail::thread_data_base*) (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E42124: thread_proxy (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4C32870: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==3822== by 0x5D74A03: start_thread (pthread_create.c:300) ==3822== by 0x58D680C: clone (clone.S:112) ==3822== ==3822== Conflicting load by thread 3 at 0x05f85288 size 8 ==3822== at 0x5D73AEE: __nptl_deallocate_tsd (pthread_create.c:153) ==3822== by 0x5D74A11: start_thread (pthread_create.c:307) ==3822== by 0x58D680C: clone (clone.S:112) ==3822== Allocation context: BSS section of /lib/libpthread-2.10.1.so ==3822== Other segment start (thread 2) ==3822== at 0x58D67D1: clone (clone.S:84) ==3822== by 0x5D7493F: ??? (allocatestack.c:743) ==3822== by 0x6B8A90F: ??? ==3822== Other segment end (thread 2) ==3822== at 0x5D7A7C4: pthread_once (pthread_once.S:119) ==3822== by 0x4C28ADD: pthread_once (drd_pthread_intercepts.c:518) ==3822== by 0x4E473B3: boost::detail::get_once_per_thread_epoch() (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E4172F: ??? (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E41B28: boost::detail::set_current_thread_data(boost::detail::thread_data_base*) (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E42124: thread_proxy (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4C32870: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==3822== by 0x5D74A03: start_thread (pthread_create.c:300) ==3822== by 0x58D680C: clone (clone.S:112) ==3822== ==3822== Thread 4: ==3822== Conflicting load by thread 4 at 0x0504b350 size 8 ==3822== at 0x4E41724: ??? (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E41B28: boost::detail::set_current_thread_data(boost::detail::thread_data_base*) (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4E42124: thread_proxy (in /usr/local/lib/libboost_thread.so.1.42.0) ==3822== by 0x4C32870: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==3822== by 0x5D74A03: start_thread (pthread_create.c:300) ==3822== by 0x58D680C: clone (clone.S:112) ==3822== Allocation context: BSS section of /usr/local/lib/libboost_thread.so.1.42.0 ==3822== Other segment start (thread 3) ==3822== (thread finished, call stack no longer available) ==3822== Other segment end (thread 3) ==3822== (thread finished, call stack no longer available) ==3822== Other segment start (thread 3) ==3822== (thread finished, call stack no longer available) ==3822== Other segment end (thread 3) ==3822== (thread finished, call stack no longer available) ==3822== ==3822== ==3822== For counts of detected and suppressed errors, rerun with: -v ==3822== ERROR SUMMARY: 11 errors from 4 contexts (suppressed: 126 from 76)

2 1

How to: get low level file descriptor for tcp::iostream
by Prince Singh 25 May '10

25 May '10

How to: get the low level file descriptor for a tcp::iostream ? -- Regards, Prince Singh Drishti-Soft Solutions Pvt Ltd 62-A, First Floor, Maruti Industrial Area, Sector - 18, Gurgaon - 122016 Haryana, India. W: http://www.drishti-soft.com B: http://blog.drishti-soft.com

2 2

[lambda] extract a constant subexpression by introspection
by alfC 25 May '10

25 May '10

Hello, I have a lambda expression that is suppoalways constructed as [some lambda expression] / ( _1 - c); where c is a double constant, for example c = 3.14. What I need is a function that given the expression returns the internal value of c, if that is possible: template<class LambdaExp> double c(LambdaExp expr){ ... return c in expr assuming that expr has the form [...]/(_1 - c) } The function may throw an exception or whatever if the expr is not of the form [...]/(_1 - c) Some additional infomation that maybe useful to simplify the solution: the lambda expression returns a double, _1 is meant to be a placeholder for double. optional feature: Also it would be ideal if the function could also recognize *simple* alternative expressions like [..]/(_1 + k) (and return -k) or [...]/( k + _1) (and return -k). Solutions using other lambda-like libraries are welcomed too. Thank you, Alfredo

1 1