July 2007 - Boost-users - lists.boost.org

Re: [Boost-users] [Serialization] is_serializable trait?
by Roman Perepelitsa 24 Jul '07

24 Jul '07

> Re: [Serialization] is_serializable trait? (Chris Fairles) > > I also took a stab at making a boostified version: > > #include <iostream> > #include <boost/type_traits/config.hpp> > #include <boost/utility/enable_if.hpp> > #include <boost/type_traits/is_fundamental.hpp> > #include <boost/type_traits/detail/yes_no_type.hpp> > #include <boost/type_traits/detail/bool_trait_def.hpp> > > namespace boost { > namespace detail { > > template <class T> > struct has_serialize_mem_fun { > typedef void (T::*SerializationFun)(int &, unsigned); > > template <SerializationFun> > struct A {}; > > template <class Q> > static ::boost::type_traits::yes_type > has_serialize_mem_fun_tester(A<&Q::serialize>*); > > template <class Q> > static ::boost::type_traits::no_type has_serialize_mem_fun_tester(...); > > BOOST_STATIC_CONSTANT(bool, value = > (sizeof(has_serialize_mem_fun_tester<T>(0)) == > sizeof(::boost::type_traits::yes_type))); > > }; > template <typename T, typename IsFundamental = void> > struct is_serializable_impl > { > BOOST_STATIC_CONSTANT(bool, value = has_serialize_mem_fun<T>::value); > }; > > template <typename T> > struct is_serializable_impl<T, typename ::boost::enable_if< > ::boost::is_fundamental<T> >::type> > { > BOOST_STATIC_CONSTANT(bool, value = true); > }; > } > > BOOST_TT_AUX_BOOL_TRAIT_DEF1(is_serializable,T,::boost::detail::is_serializable_impl<T>::value) > > } > > struct A { > template <class Stream> > void serialize(Stream & strm, unsigned version) {} > }; > > struct B {}; > > struct C { void serialize() {} }; > > int main() > { > using namespace std; > cout > << boolalpha > << boost::is_serializable<A>::value << '\n' // true > << boost::is_serializable<int>::value << '\n' // true > << boost::is_serializable<float>::value << '\n' // true > << boost::is_serializable<double>::value << '\n' // true > << boost::is_serializable<char>::value << '\n' // true > << boost::is_serializable<B>::value << '\n' // false > << boost::is_serializable<C>::value << '\n' // false > << boost::is_serializable<string>::value << endl; //false but > should be true > } > > Needs specializations for standard containers as well but this suits > my needs currently. Thoughts? I tried to implement is_non_intrusively_serializable metafunction, but found no general solution. Although I've impelented it for all types which are instances of some class template. For example, it works correctly for vector<int>, some_class<double, string>, etc. Note that it is possible to create metafunction which determines whether type is instance of class template or not. So, we have the following: - is_intrusivelly_serializable metafunction. - is_class_template_instance metafunction. - is_non_intrusively_serializable metafunction, which currently works only for class template instances. Probably someone can implement the last missing case? That is we need is_non_intrusively_serializable implementation for types, which are not class template instances (int, some_nontemplate_class, etc). Here is is_non_intrusively_serializable restricted to class templates. #include <boost/serialization/serialization.hpp> #include <boost/type_traits/remove_cv.hpp> #include <boost/type_traits/detail/yes_no_type.hpp> #include <boost/preprocessor.hpp> #include <boost/static_assert.hpp> #include <boost/serialization/vector.hpp> namespace detail { struct tag {}; using boost::serialization::serialize; #define BOOST_PP_LOCAL_LIMITS (1, 10) #define BOOST_PP_LOCAL_MACRO(N) \ template < \ class Archive, \ template <BOOST_PP_ENUM_PARAMS(N, class I)> class T, \ BOOST_PP_ENUM_PARAMS(N, class T) \ > \ tag serialize( \ Archive & ar, \ T<BOOST_PP_ENUM_PARAMS(N, T)> &, \ const unsigned \ ); #include BOOST_PP_LOCAL_ITERATE() tag operator,(tag, int); boost::type_traits::no_type check(tag); template <class T> boost::type_traits::yes_type check(T const&); template <class T> struct impl { static typename boost::remove_cv<T>::type & x; static const int & archive; static const unsigned version; static const bool value = sizeof(check((serialize(archive, x, version), 0))) == sizeof(boost::type_traits::yes_type); }; } template <class T> struct is_non_intrusively_serializable : detail::impl<T> {}; template <class T> struct A {}; int main() { BOOST_STATIC_ASSERT( is_non_intrusively_serializable<std::vector<int> >::value); BOOST_STATIC_ASSERT( !is_non_intrusively_serializable<A<int> >::value); } Regards, Roman Perepelitsa Deutsche Bank Moscow +7 (495) 660-74-08 --- This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and delete this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden. Please refer to http://www.db.com/en/content/eu_disclosures.htm for additional EU corporate and regulatory disclosures.

1 0

Detecting Valid Expressions for a Type at Compile-Time
by Martin Ecker 24 Jul '07

24 Jul '07

Hi, Is there a way, using SFINAE tricks and potentially the techniques used in the Boost Concept Check Library, to detect at compile-time if a type T supports a certain expression? In my case, I'm interested if T supports operator ==. So I'm basically looking for a compile-time predicate so that this compiles: struct my_struct { int i; }; struct my_struct2 { int i; }; bool operator == (my_struct2 const&, my_struct2 const&) const; BOOST_STATIC_ASSERT((supports_comparison_operator<int>::value)); BOOST_STATIC_ASSERT((!supports_comparison_operator<my_struct>::value)); BOOST_STATIC_ASSERT((supports_comparison_operator<my_struct2>::value)); Thanks, Martin

3 4

Serialization
by Alejandro Rubinos Carbajal 23 Jul '07

23 Jul '07

How can i serialize an object of the class tm that belongs to time.h, and also i will like to serialize a char * object. Thanks

1 0

[intrusive][interprocess] Online documentation site changed
by Ion Gaztañaga 23 Jul '07

23 Jul '07

Hi to all, Due to recent problems with the automatically generated Boost documentation in Boost CVS and some other problems with my old website, I've just uploaded up-to-date (CVS HEAD documentation) Boost.Intrusive and Boost.Interprocess documentation to these sites: http://igaztanaga.drivehq.com/intrusive http://igaztanaga.drivehq.com/interprocess These sites have no ads, so I expect better user experience. I've requested the cancellation of the old site (www.prohosting.com) so I hope this helps Intrusive and Interprocess user experience until Boost 1.35 is released. Regards, Ion

1 0

Re: [Boost-users] [Serialization] is_serializable trait?
by Roman Perepelitsa 23 Jul '07

23 Jul '07

> Chris Fairles" <chris.fairles(a)gmail.com> on Fri, 20 Jul 2007 18:03:18 -0400 > How can you determine (compile-time) that a class has a serialize function? > > struct foo { > template <typename Archive> > void serialize (Archive &, const unsigned int); > }; > > "&foo::serialize" isn't a type and I can't figure out a way to make it > a type (besides a typedef in the class) and also a constant > expression. > > Tried to play around with this as well but theres no substitution > failure here and static functions can't be constant expressions > (commonnnn c++0x!!) > > struct serialize { > template <typename T> > static void value(T const&) { void(T::*f)(archive::xml_oarchive &, > const unsigned int) = &T::template serialize<archive::xml_oarchive>; > (void)f; } > }; > > Basically, in the code below I want to see "GoodGood" printed out and > the only thing I can change is Q. > > struct A{ > void f(int,int){} > }; > struct B{}; > > template <class T,class F=void> > struct Q{enum{value=false};}; > > template<class T> struct Q<T,void(T::serialize)(int,int)>{enum{value=true};}; > > int main(){ > if(Q<A>::value) { cout << "Good";} > if(!Q<B>::value){ cout << "Good";} > } > > Help/hints/remotely possible? You can try something like this. #include <iostream> #include <iomanip> typedef char yes_type; struct no_type { char padding[8]; }; template <class T> struct HasSerializeMember { typedef void (T::*SerializationFun)(int &, unsigned); template <SerializationFun> struct A {}; template <class Q> static yes_type Test(A<&Q::serialize>*); template <class Q> static no_type Test(...); static const bool value = sizeof(Test<T>(0)) == sizeof(yes_type); }; struct A { template <class Stream> void serialize(Stream & strm, unsigned version) {} }; struct B { }; struct C { void serialize() {} }; int main() { using namespace std; cout << boolalpha << HasSerializeMember<A>::value << '\n' // true << HasSerializeMember<B>::value << '\n' // false << HasSerializeMember<C>::value << endl; // false } Regards, Roman Perepelitsa Deutsche Bank Moscow +7 (495) 660-74-08 --- This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and delete this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden. Please refer to http://www.db.com/en/content/eu_disclosures.htm for additional EU corporate and regulatory disclosures.

3 3

[Boost.Test] warning C4265 when including <boost/test/unit_test.hpp> and friends
by Pierre-Jules Tremblay 23 Jul '07

23 Jul '07

Just trying out Boost 1.34.1RC3, I found this in <boost/test/detail/config.hpp>, lines 65-70: #ifdef __GNUC__ #define BOOST_TEST_PROTECTED_VIRTUAL virtual #else #define BOOST_TEST_PROTECTED_VIRTUAL #endif perhaps the condition should be changed to include MSVC8(SP1), to eliminate multiple copies of this annoying warning: xxxxx.hpp(54) : warning C4265: 'boost::unit_test::test_observer' : class has virtual functions, but destructor is not virtual Cheers, pj

2 2

Re: [Boost-users] Problem with unit test BOOST_AUTO_TEST_MAIN
by Sakharuk, Vlad (GMI Development) 23 Jul '07

23 Jul '07

I have found the problem (and the answer). Someone recompiled TinyXml lib from executable with main function included. And that function was called instead of test one. Gennadiy, Thanks for help. Regards Vlad. -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Gennadiy Rozental Sent: Thursday, July 19, 2007 1:47 PM To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] Problem with unit test BOOST_AUTO_TEST_MAIN "Sakharuk, Vlad (GMI Development)" <Vlad_Sakharuk(a)ml.com> wrote in message news:AAA38056FF35DE4CB4A65657784B08FAA9E0A6@MLNYC730MB.amrs.win.ml.com.. . > Yes, if I remove all lib it does run successfully. > I identified that tests run from TinyXML lib (opensource). > > I have to link with other libraries because the code I have to test > depends on them. I am just creating tast case to narrow down the > problem. > > I will look in 1.34.o but probably I could not. > > I will apriceate any solution how to disable particular testing suit > (from TinyXML). I could not physically remove those test from the lib. > > Thanks Vlad This is impossible naturally in current version of Boost.Test. I will look into allowing you to do this in 1.35.0. The only thing you can try to do (in 1.34) is massage master test suite somehow in test module init function. Gennadiy _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users -------------------------------------------------------- This message w/attachments (message) may be privileged, confidential or proprietary, and if you are not an intended recipient, please notify the sender, do not use or share it and delete it. Unless specifically indicated, this message is not an offer to sell or a solicitation of any investment products or other financial product or service, an official confirmation of any transaction, or an official statement of Merrill Lynch. Subject to applicable law, Merrill Lynch may monitor, review and retain e-communications (EC) traveling through its networks/systems. The laws of the country of each sender/recipient may impact the handling of EC, and EC may be archived, supervised and produced in countries other than the country in which you are located. This message cannot be guaranteed to be secure or error-free. This message is subject to terms available at the following link: http://www.ml.com/e-communications_terms/. By messaging with Merrill Lynch you consent to the foregoing. -------------------------------------------------------- "Please consider the environment before printing this email."

1 0

iterator_facade and custom reference types
by Daniel Oberhoff 23 Jul '07

23 Jul '07

hi I am struggling with iterator_face. I have a container of bits with a smart reference type. This is because the exposed type is bool, but there is never a full boolean variable around for an index, just a single bit. So whenever a reference to a bool is requested I want to pass my smart reference that knows how to get and set this bit. But iterator_facade is too smart for this: it wants to make a true reference to a supplied Value type, which results in references to temporaries. Is there any way to tell iterator_facade to use my type as a ref type? Daniel Oberhoff

2 1

shared_ptr + vector + std::sort
by bringiton bringiton 23 Jul '07

23 Jul '07

hi guys, just wondering if you have any shared_ptr tips for sorting vectors [code] std::vector< boost::shared_ptr<int> > v; //.. insert items into vector std::sort(v.begin(), v.end()); [/code] OK this doesn't sort by integer. I was hoping shared_ptrs < operator actually used the < operator of it's template type. however, i guess it's sorting by pointer value... the following works OK: [code] template <class T> struct X { bool operator()(const boost::shared_ptr<T>& a, const boost::shared_ptr<T>& b) { return *a < *b; } }; int main() { std::vector< boost::shared_ptr<int> > v; //.. insert items into vector std::sort(v.begin(), v.end(), X<int>()); return 0; } [/code] Does anyone have a better way of doing this? Maybe boost already defines a function similar to my X function. Maybe there is a way without using a predicate...

2 1

Non-throwing version of filesystem::create_directories?
by Olaf van der Spek 23 Jul '07

23 Jul '07

Hi, Is there a non-throwing version of filesystem::create_directories? Before I used normal C/POSIX APIs that returned an error value and I don't want to wrap each call in a try/catch statement. Olaf

1 0