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

Re: [Boost-users] Annoying coding
by Alan M. Carroll 20 Jan '09

20 Jan '09

Why not use Boost.Bind? It will bind to data members, returning a functor that accesses that member. E.g. boost::bind(&A::a, _1); will create a functor that takes an instance of A and returns a reference to the a member. You can then pass that to whatever other functor is being used. Something like (not tested!) std::accumulate(m.begin(), m.end(), bind(std::add, _1, bind(&std::map<std::string, int>::value_type::first, _1)); Boost.Lambda can do this as well, but I am not familiar enough with it to provide an example. At 07:22 AM 1/20/2009, you wrote: >Hi again, > >I have a problem. Not really big one but annoying. >Often we need to write code like this: > >struct A{ > int a; > std::string b; >}; > >int a_sum(int i, const A &v){ return v.a + i; }; > >std::vector<A> v; >//fill this vector somewhere > >std::accumulate( v.begin(), v.end(), 0, a_sum ); > >or something like > >int m_sum(int i, const std::pair<int, std::string> p) { return p.first + i; }; > >std::map <std::string, int> m; > >std::accumlate( m.begin(), m.end(), 0, m_sum); > >Or in common words we need to write some kind functor which just get >member from element and apply it to another functor. > >Of course I can write something like: > >template<typename T_, typename V_, V_ T_::*Ptr_> >struct get_mem{ > const V_ &operator()(const T_ &elem){ > return elem->*Ptr_; > }; >}; > >but may be there is library way to do this? >_______________________________________________ >Boost-users mailing list >Boost-users(a)lists.boost.org >http://lists.boost.org/mailman/listinfo.cgi/boost-users

4 15

Annoying coding
by Maxim Koshelev 20 Jan '09

20 Jan '09

Hi again, I have a problem. Not really big one but annoying. Often we need to write code like this: struct A{ int a; std::string b; }; int a_sum(int i, const A &v){ return v.a + i; }; std::vector<A> v; //fill this vector somewhere std::accumulate( v.begin(), v.end(), 0, a_sum ); or something like int m_sum(int i, const std::pair<int, std::string> p) { return p.first + i; }; std::map <std::string, int> m; std::accumlate( m.begin(), m.end(), 0, m_sum); Or in common words we need to write some kind functor which just get member from element and apply it to another functor. Of course I can write something like: template<typename T_, typename V_, V_ T_::*Ptr_> struct get_mem{ const V_ &operator()(const T_ &elem){ return elem->*Ptr_; }; }; but may be there is library way to do this?

1 0

Re: [Boost-users] Generating a large number of class member variables: PP vs mpl
by Hicham Mouline 20 Jan '09

20 Jan '09

But once I have my mpl::vector or fusion::vector, how do i apply a meta function like for_each to generate the class members automatically. for e.g., I define the vector<> with 15 elements, I should be able to have 15 member types in the class, without manually writing at_c<0> .... at_c<1>? I'm now looking again at mpl::inherit_linearly rds, > ----- Original Message ----- > From: "Juraj IvanÄiÄ" <juraj.ivancic(a)gmail.com> > To: boost-users(a)lists.boost.org > Subject: Re: [Boost-users] Generating a large number of class member variables: PP vs mpl > Date: Tue, 20 Jan 2009 11:03:56 +0100 > > > Hicham Mouline wrote: > > > Is it possible to generate the members from the mpl::vector<> ? > > It is certainly possible and is already done in Boost.Fusion. > > See > http://www.boost.org/doc/libs/1_37_0/libs/fusion/doc/html/index.html > and especially > http://www.boost.org/doc/libs/1_37_0/libs/fusion/doc/html/fusion/container.… > > HTH, > Juraj

2 1

Generating a large number of class member variables: PP vs mpl
by Hicham Mouline 20 Jan '09

20 Jan '09

Hello, I have a base class with a number of derived classes. Derived classes might have large number of private member variables, each with a different type. Some of these derived classes may have identical lists of member names and types. I seem to have 2 possibilities to generate the list of member variables: class Derived1 { ... typedef mpl::vector< Type1, Type2, Type3, Type4, Type5, Type6 > memberTypes; // then try to generate automatically from memberTypes, as: // const Type1& mType1; // const Type2& mType2;... // const Type6& mType6; // with some meta function }; or // use only PP class Derived1 { ... #define MYNAMESPACE_MEMBER_TYPES (Type1)(Type2)(Type3)(Type4)(Type5)(Type6) // use PP macros to generate the members }; Is it possible to generate the members from the mpl::vector<> ? in particular, willing to use the mpl::size meta function will forbid me from using PP... The advantage I think is a cleaner definition of types, and other derived classes may use the same type memberTypes. The code generating the member types could isolated in some user meta function. regards,

2 1

create type inside boost::variant without calling copy constructor
by Maxim Koshelev 20 Jan '09

20 Jan '09

Hi, all! At first small example which introduce problem and solutions. Let's write class that hold some object and give us access to it (it's not for real usage, only for demonstration): class Holder{ struct Impl{}; template<typename T_> struct Impl{ T_ val; Impl(const T_ &v):val(v){} }; Impl *v_; public template<typename T_> Holder(const T_&val):v_(new Impl<T_>(val)){}; template<T_> T_ get(){ dynamic_cast<T_>(*v_); }; }; now we have a constructor which create Holder with some type inside by copying it. If we want to create Holder with some type without copying, just created by default constructor, we can use "traits" technics: template<typename T_> class Trait{ typedef T_ type; } template<typename T_> Holder(const Trait<T_>):v_(new Impl<T_>){}; and when we want to create Holder with some type inside we use: Holder v( Trait<SomeType>() ); And now problem: is there any way to create boost::variant with a type inside without calling copy constructor? Currently I see only one way: boost::variant< std::string, std::pair<std::string, std::string> > p; p=std::string(); in this code we create "empty" variant variable "p", then copy empty string inside it. But I need just create p with std::string inside, without copying! Any solution? Thanks in advance, and additional thanks if you have read till this sentence :-) P.S. Sorry for my ugly English :-)

1 0

[filesystem] running afoul of _HAS_ITERATOR_DEBUGGING - VC 9.0
by Raman 19 Jan '09

19 Jan '09

Hello, I'm a boost newbie, so please bear with me if this sounds trivial. I'm rebuilding our code that uses boost heavily with VC 9.0 on Windows 2008, 32-bit. All of our code is compiled with _SECURE_SCL=0 & _HAS_ITERATOR_DEBUGGING=0. And I rebuild boost with the same flags. Our code links with boost as static libraries. However, at runtime - I'm seeing access violations occuring in the Microsoft secure STL code that should have been disabled with _HAS_ITERATOR_DEBUGGING=0. A specific example, when we call filesystem::basic_path::normalize(), this calls: // path.hpp template<class String, class Traits> void basic_path<String, Traits>::m_append_separator_if_needed() { if ( # ifdef BOOST_WINDOWS_PATH *(m_path.end()-1) != colon<path_type>::value && # endif // >>>>>> following is the offending line *(m_path.end()-1) != slash<path_type>::value ) { m_path += slash<path_type>::value; } } which leads me to: //xtring reference __CLR_OR_THIS_CALL operator*() const { // return designated object #if _HAS_ITERATOR_DEBUGGING if (this->_Mycont == 0 || _Myptr == 0 // >>>>>> following is the offending line - _MyCont contains garbage || _Myptr < ((_Mystring *)this->_Mycont)->_Myptr() || ((_Mystring *)this->_Mycont)->_Myptr() + ((_Mystring *)this->_Mycont)->_Mysize <= _Myptr) { _DEBUG_ERROR("string iterator not dereferencable"); SCL_SECURE_OUT_OF_RANGE; } __analysis_assume(_Myptr != 0); #else if (this->_Mycont != _IGNORE_MYCONT) { _SCL_SECURE_VALIDATE(this->_Has_container()); _SCL_SECURE_VALIDATE_RANGE(_Myptr < (((_Mystring *)this->_Mycont)->_Myptr() + ((_Mystring *)(this->_Mycont))->_Mysize)); } #endif /* _HAS_ITERATOR_DEBUGGING */ return (*_Myptr); } I'm stumped as to why its even in that piece of code when I've disabled HAS_ITERATOR_DEBUGGING. Any help/guidance is appreciated. Thanks Raman

3 5

[range, iterator?] compiler error going from 1.34 to 1.37
by Martin Livesey 19 Jan '09

19 Jan '09

We are upgrading our legacy code to use boost 1.37 from version 1.34 and have hit a slight hitch. The following (massively simplified) code no longer compiles. I tried it on more than one compiler and the error messages aren't too useful in helping to determine the issue. (The target compiler will be VS 2k8). Any assistance would be appreciated. Thanks in advance. #include "tchar.h" #include <boost/iterator/filter_iterator.hpp> #include <boost/range/iterator_range.hpp> #include <list> class A{}; class B { public: typedef std::list<A>::iterator iterator; iterator begin() { return bs.begin(); } iterator end() { return bs.end(); } private: typedef std::list<A> List; List bs; }; template <class Predicate> struct FilteredRange { typedef boost::filter_iterator<Predicate, B::iterator> iterator_type; typedef boost::iterator_range<iterator_type> range_type; }; template <typename Predicate> typename FilteredRange<Predicate>::range_type getFilteredRange(B *list) { return boost::make_iterator_range(boost::make_filter_iterator<Predicate>(list->begin(), list->end()), boost::make_filter_iterator<Predicate>(list->end(), list->end())); } class C { public: bool operator()(const A &a) const {return true;} }; int _tmain(int argc, _TCHAR* argv[]) { getFilteredRange<C>(new B).size(); return 0; }

2 1

[boost-users][function] how to
by Igor R 19 Jan '09

19 Jan '09

Hello, function<void(Base *)> funcB; function <void(Derived *)> funcD; Is it possible to convert somehow funcD --> funcB ? Thanks.

2 3

Boost::interprocess sharing string
by Federico Golfrè Andreasi 19 Jan '09

19 Jan '09

Hi ! I have a problem sharing a string within a shared memory segment using named objects. (Boost.interporcess 1.36.0) I attach here the code of the two programs: one writes the string and the other read it. In the reader a segmentation fault is throw. Where am I wrong? Thank you very much Federico

2 1

[proto] applying proto to real problems
by Michael Marcin 19 Jan '09

19 Jan '09

I haven't used proto before and as I read the documentation everything makes sense but I can't figure out how to apply the information there to solve my problems. I think if I can figure out how to solve a simplified example from my actual problem I can figure out the rest. // typical 3d vector template< typename T > struct vec3 { T x,y,z; }; // typical 3d dot product template< typename T > T dot_product( const vec3<T>& a, const vec3<T>& b ) { return a.x*b.x + a.y*b.y + a.z*b.z; } // typical 16.16 fixed point number struct fixed { boost::int32_t data; }; // typical multiplication for this fixed point number fixed operator*( fixed lhs, fixed rhs ) { fixed result = { static_cast<boost::int32_t> ( ( boost::int64_t(lhs.data)*rhs.data ) >> 16 ) }; return result; } // however using the typical dot product with // fixed numbers is too lossy and inefficient // so we have to specialize dot_product for // these fixed point numbers to do this template<> fixed dot_product( const vec3<fixed>& a, const vec3<fixed>& b ) { fixed result = { static_cast<boost::int32_t> ( ( boost::int64_t(a.x.data) * b.x.data + boost::int64_t(a.y.data) * b.y.data + boost::int64_t(a.z.data) * b.z.data ) >> 16 ) }; return result; } What I want to be able to do is to make the typical dot product generate code that is equivalent to the specialized dot product when instantiated with T = fixed. This seems like something proto should handle easily from the description of the library but I just don't get proto yet. Thanks, -- Michael Marcin

3 2