July 2008 - Boost-users - lists.preview.boost.org

Avoiding a circular reference of boost::functions
by Simon Pickles 13 Jul '08

13 Jul '08

Hello, I am using boost::functions to provide a simple event system, but have come across a problem. I have two classes that need to communicate with each other, CGUI and CEventHandler. Their instances are both shred_ptrs owned by a third class, CCore. To establish comms, I am passing boost::functions which have been bound to members of the two classes: //--------------------- typedef std::vector<boost::any> many // CCore.cpp snippet boost::function<bool(many&)> cb = m_eventHandler->RegisterGUICallback( boost::bind( &IGUIInterface::HandleEventMessage, m_gui, _1 ) ); m_gui->RegisterEventHandler( cb ); // CEventHandler.cpp snippet // boost::function<bool(many&) m_guiTx declared in .h boost::function<bool(many&)> CEventHandler::RegisterGUICallback(boost::function<bool(many&)> f) { // MEM LEAK IS HERE - A circular reference? m_guiTx = f; return boost::bind(&CEventHandler::HandleGUIRx, this, _1); } // CGUI.cpp snippet // boost::function<bool(many&) eventHandlerCallback declared in .h void CGUI::RegisterEventHandler( boost::function<bool(many&)> cb ) { m_eventHandlerCallback = cb; m_log->GUI("EventHandler callback registered. Testing...\n");//, s ); std::string message = "This was sent from CGUI::RegisterEventHandler\n"; int id = 43; many msg; boost::any _message = message; boost::any _id = id; msg.push_back(_id); msg.push_back(_message); m_eventHandlerCallback(msg); } //----------------------------------------------------- As you can see, CCore binds a public member function of m_gui and passes it to the eventhandler, where it is stored and used to send events. The event handler returns its own member function, bound to this, which CCore duly passes to m_gui. Now both classes hold objects with refs to each other, and the program does a big memory dump when it is ended, on win32 with MSVC2005. I envisage this is because the shared_ptrs holding the instances of CGUI and CEvenHandler are not destructing at program close. Commenting out the line: m_guiTx = f; eliminates the issue. What is a better way to achieve this? Thanks Simon

4 6

[serialization, boost 1.36] Changes in BOOST_IS_ABSTRACT
by Ruediger Berlich 13 Jul '08

13 Jul '08

Hi there, I'm trying to prepare my application for the upcoming 1.36 Boost release and understand that there have been a number of substantial changes in the serialization code. I understand the serialization chapter of the 1.36 documentation in such a way that it is recommended to use BOOST_SERIALIZATION_ASSUME_ABSTRACT(T) for compilers that cannot cope with the 1.36 BOOST_IS_ABSTRACT(T) macro. gcc 4.3.1 seems to be one of the compilers that need this - it emits an error for BOOST_IS_ABSTRACT in 1.36 . Now I'm wondering whether BOOST_SERIALIZATION_ASSUME_ABSTRACT will work for other compilers, even if they could also deal directly with BOOST_IS_ABSTRACT. /********************************************************/ #if BOOST_VERSION <= 103500 BOOST_IS_ABSTRACT(GenEvA::GObject) #else BOOST_SERIALIZATION_ASSUME_ABSTRACT(GenEvA::GObject) #endif /* Serialization traits */ /********************************************************/ Thanks and Best Regards, Ruediger

2 1

Re: [Boost-users] [boost-users][multi-index] hashed identity: lookup by hash value
by JOAQUIN M. LOPEZ MUÑOZ 13 Jul '08

13 Jul '08

________________________________________ De: boost-users-bounces(a)lists.boost.org [boost-users-bounces(a)lists.boost.org] En nombre de Igor R [boost.lists(a)gmail.com] Enviado el: domingo, 13 de julio de 2008 21:29 Para: boost-users(a)lists.boost.org Asunto: Re: [Boost-users] [boost-users][multi-index] hashed identity: lookup by hash value [...] > What's not working is the second approach - with the hash_value > template function as a the extractor of an additional key (it's not > compiling). It's weird, cause it ought to work. For instance, the following does it nicely both in GCC 3.4.4 and MSVC++ 8.0: #include <boost/multi_index/global_fun.hpp> #include <cstddef> #include <iostream> struct Element{}; template<typename T> std::size_t my_hash_value(const T&){return 0;} namespace mi=boost::multi_index; typedef mi::global_fun<const Element&,std::size_t,&my_hash_value > key_extractor; int main() { key_extractor k; Element e; std::cout<<k(e)<<std::endl; } What error messages are you getting? You can make your compiler's life easier by being a little more explicit when passing my_hash_value (though it's not obligatory to do so): typedef mi::global_fun<const Element&,std::size_t,&my_hash_value<Element> > key_extractor; If you provide me with a little more info about the errors you're getting maybe I can be more helpful. Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

1 0

Re: [Boost-users] [boost-users][multi-index] hashed identity: lookup by hash value
by JOAQUIN M. LOPEZ MUÑOZ 13 Jul '08

13 Jul '08

Hi Igor, ________________________________________ De: boost-users-bounces(a)lists.boost.org [boost-users-bounces(a)lists.boost.org] En nombre de Igor R [boost.lists(a)gmail.com] Enviado el: domingo, 13 de julio de 2008 14:12 Para: boost-users(a)lists.boost.org Asunto: [Boost-users] [boost-users][multi-index] hashed identity: lookup by hash value > > Hello, > > Given a multi_index_container with 1 hashed unique identity key (kind > of unordered set) - is there a "trivial" way to find a value by its > hash-code? There is a trivial method but it relies on undocumented implementation features of hashed indices: std::size_t h=boost::hash<Element>()(...); std::size_t n=h%elementSet.bucket_count(); // the element lies in the range [first,last). In most cases (i.e. if the hash // function is doing a good job), this range only contains one element. ElementSet::iterator first=elementSet.begin(n), last=elementSet.end(n); As I said, this relies on the knowledge that an element with hash value h goes to the bucket in the position h%bucket_count. So, my strong advise is *not to use this*. The correct approach is the one you are already exploring below. > Currently, I'm using a trick with supplying custom hashing/equality > functors, but probably there's some more straightforward way? [...] > struct ElementEq > { > bool operator()(size_t x, const Element &y) const > { > return x == boost::hash<Element>()(y); > } > }; Strictly speaking, your ElementEq struct has to provide also the operator()(const Element&,size_t): struct ElementEq { bool operator()(const Element &x, size_t y) const { return boost::hash<Element>()(x) == y; } bool operator()(size_t x, const Element &y) const { return x == boost::hash<Element>()(y); } }; (In practice it works with the second oeprator() only, but this is again undocumented). [...] > int main() > { > ElementSet elementSet; > //.... > // I've got here a hash-value of an element, and I wish to find it in the set. > ElementSet::iterator pos = elementSet.find(hashValue, > ElementHashHasher(), ElementEq()); > } You have to take into account that there is a possibility that find() gets you a distinct element with the same hash value, so you'd better use equal_range() and postprocess the obtained range to find the exact element you're after. Other than this, the technique you describe is absolutely OK and the one I'd recommend to do what you want. Is it not working for you? I've written a small code snippet to try it myself and everything seems to work OK (attached). Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

2 1

Problem with recursion using Boost intrusive list
by Tomáš Šalamon 13 Jul '08

13 Jul '08

Hi, I am a newbie to Boost libraries and I have the following problem with Boost intrusive list: I have a Page class and I need to put its instances into an intrusive set. There is no problem and it is working fine. Beside that, I need every instance of this class had a member intrusive list containing other Pages. The source is as follows: struct pageTag; typedef link_mode<normal_link> NormalLinkMode; typedef set_base_hook<tag<pageTag>, NormalLinkMode> SetPageHook; typedef list_base_hook<tag<pageTag>, NormalLinkMode> ListPageHook; class Page : public ListPageHook, public SetPageHook { public: typedef set<Page, base_hook<SetPageHook>> PageSet; typedef list<Page, base_hook<ListPageHook>> PageList; Page(); ... friend bool operator< (const Page &a, const Page &b) { return a.fullPath < b.fullPath; } friend bool operator> (const Page &a, const Page &b) { return a.fullPath > b.fullPath; } friend bool operator== (const Page &a, const Page &b) { return a.fullPath < b.fullPath; } private: // some stuff here PageList links; /****************** THIS IS THE PROBLEM ************************* }; As I add the last line with the list, I get: "Error 1 error C2248: 'boost::intrusive::list_impl<Config>::list_impl' : cannot access private member declared in class 'boost::intrusive::list_impl<Config>' c:\program files\boost\boost_1_35_0\boost\intrusive\list.hpp 1434". Is there any problem with recursion in Boost intrusive containers? Thank you. Tomas

2 2

[boost-users][multi-index] hashed identity: lookup by hash value
by Igor R 13 Jul '08

13 Jul '08

Hello, Given a multi_index_container with 1 hashed unique identity key (kind of unordered set) - is there a "trivial" way to find a value by its hash-code? Currently, I'm using a trick with supplying custom hashing/equality functors, but probably there's some more straightforward way? struct Element {//...}; size_t hash_value(const Element &e) { //... } typedef mi::multi_index_container<Element, mi::indexed_by<mi::hashed_unique<mi::identity<Element> > > ElementSet; struct ElementEq { bool operator()(size_t x, const Element &y) const { return x == boost::hash<Element>()(y); } }; struct ElementHashHasher { std::size_t operator()(size_t x) const { return x; } }; int main() { ElementSet elementSet; //.... // I've got here a hash-value of an element, and I wish to find it in the set. ElementSet::iterator pos = elementSet.find(hashValue, ElementHashHasher(), ElementEq()); } Another way I tried was defining the hash-code as an explicit key: //... mi::hashed_unique< mi::global_fun<const Element&, std::size_t, &hash_value> > //... The problem however is that my hash_value functions are templates, and the above definition is not compiling... Thanks!

1 0

Re: [Boost-users] How to use boost's shared_ptr.hpp?
by Richard L. Aurbach 12 Jul '08

12 Jul '08

On Jul 12, 2008, at 2:28 AM, boost-users-request(a)lists.boost.org wrote: > On Fri, Jul 11, 2008 at 6:06 PM, Igor R <boost.lists(a)gmail.com> wrote: >> Does it compile if you define BOOST_NO_TYPEID ? >> >>> My question is why I need to enable 'rtti' in order to use boost >>> shared_ptr ? >> >> IIRC, get_deleter() function uses typeid(), which is unavailable in >> your case. > > I have added '#define BOOST_NO_TYPEID ' Before I include > 'shared_ptr.hpp', but i still get the following error: > > /usr/include/boost/detail/sp_counted_impl.hpp: In member function > 'virtual void* boost::detail::sp_counted_impl_pd<P, > D>::get_deleter(const std::type_info&)': > /usr/include/boost/detail/sp_counted_impl.hpp:150: error: cannot use > typeid with -fno-rtti > > #define BOOST_NO_TYPEID > #include <boost/shared_ptr.hpp> > > using namespace std; > using namespace boost; > > Can you please tell me if there is a way to use boost smart pointer > without rtti? > Thank you. Actually, it isn't clear to me that you CAN use shared_ptr without RTTI. The shared_ptr package (both boost and TR1) includes dynamic_pointer_cast and static_pointer_cast, which are RTTI-related functions. Perhaps if you could share with us why it is important to use the -fno- rtti switch in your work, someone could provide more substantive suggestions. Cheers, Rick Aurbach

2 1

[GIL] Point concept
by Felipe Magno de Almeida 12 Jul '08

12 Jul '08

Hi, PointNDConcept is defined as having axis_value member functions. But the GIL code in the boost trunk doesn't comply with this requirement. A free axis_value function is just useless, because it can't be found by ADL when template parameters are passed explicitly. Also, member template axis doesn't model Metafunction. Am I missing something? Regards, -- Felipe Magno de Almeida

2 3

binding arguments for boost:function
by James Sutherland 12 Jul '08

12 Jul '08

Any thoughts on how to bind some arguments for a function? Consider the following: struct A { double operator()( const double a, const double b, const double c ) const { return a+b+c; } }; int main() { typedef function<double(double,double,double)> F3arg; typedef function<double(double,double) > F2arg; typedef function<double(double) > F1arg; typedef function<double() > F0arg; A a; const double x1=0.0, x2=1.0, x3=4.0; F3arg f3 = bind<double>( a, _1, _2, _3 ); cout << f3(x1,x2,x3) << endl; } How would I take "f3" and bind one argument? Something like: F2arg f2 = bind<double>( f3, _2, x2 ); // bind x2 to second argument of f3...

3 10

Re: [Boost-users] [general] boost 1.35.0 library naming and bjam arguments priorities
by damny＠web.de 12 Jul '08

12 Jul '08

> > hi, > > > > for building boost 1.35.0 with bjam 3.1.16 and mingw gcc 3.4.5 on winxp sp2 i created a *.BAT-file > > like follows: > > > > @echo off > > SETLOCAL > > SET MINGW=C:\Development_Compiler\MinGW > > SET PYTHN=C:\Development_Compiler\Python25 > > SET BUILD=C:\Development_Libs\Boost_v1.35.0 > > SET path=C:\WINDOWS;C:\WINDOWS\system32;%MINGW%\Bin > > cls > > bjam.exe -a "-sTOOLS=mingw" "-sGCC_ROOT_DIRECTORY=%MINGW%" "-sPYTHON_ROOT=%PYTHN%" > > "-sPYTHON_VERSION=2.5" > > All the -s option above have no effect. > > > --link=static --runtime-link=static --threading=single --optimization=on > > --toolset=gcc --build-dir=%BUILD% --build-type=complete install ENDLOCAL > > Remove "--" in front of all properties. "--build-dir" and "--build-type" are options > and need "--". Everything else is properties, which tell what flavour of binaries you > want to get, and don't need "--". > > - Volodya hi volodya and all others, thanks for your responses. a) but there is one thing i don't understand. why i got libs like boost_xyz.a boost_xyz.dll libboost_xyz.a without doing any rename while builing boost 1.34.1 with bjam 3.1.15? b) why the -s options have no effect? version "-sPYTHON_VERSION=2.5" of desired python and libpython25.a in "-sPYTHON_ROOT=%PYTHN%" (...\lib) is found by bjam. finally it compiles and works well (boost 1.34.1 with bjam 3.1.15 and nearly identical BAT-file). aside from this, http://www.boost.org/doc/tools/build/doc/html/jam/usage.html says -s var=value set the variable var to value, overriding both internal variables and variables imported from the environment. Command-line and Environment Variable Quoting ... Boost Jam remedies that by treating all such settings as a single string if they are surrounded by double-quotes. ... (all surrounding by me, and not only the value part) daniel ____________________________________________________________________ Ihre Messenger, Communities und E-Mails jetzt in einem Programm! WEB.DE MultiMessenger http://www.produkte.web.de/messenger/?did=3071

2 1