August 2005 - Boost-users - lists.preview.boost.org

Re: [Boost-users] Maybe a problem with boost::object_pool?
by Labour, Matthieu (MLPRO SF) 09 Aug '05

09 Aug '05

Caleb Thank you for taking the time to answer my email. I actually did use a mutex to protect my queue. I did another test very simple to reproduce the problem. A program creates N threads that construct & destroy objects in an object pool With N = 9, it crashes on linux Following is the source code of the program & the gdb trace. Thank you for your help #include <iostream> #include <boost/thread/thread.hpp> #include <boost/pool/poolfwd.hpp> #include <boost/pool/object_pool.hpp> class Foo; static boost::object_pool<Foo>* _pool; class Foo{ public: void handle(){ _pool->destroy(this); } }; class Bar{ public: Bar(){} void operator()(){ for(;;){ Foo* foo = _pool->construct(); foo->handle(); } } }; int main(int argc, char* argv[]) { if(argc != 2){ std::cout << "Usage: nbThrd" << std::endl; return 0; } int nbThrd = atoi(argv[1]); _pool = new boost::object_pool<Foo>(1000); boost::thread_group _threads; Bar bar; for (int i=0; i<nbThrd; ++i){ _threads.create_thread(bar); } _threads.join_all(); return 0; } Reading symbols from /lib/libdl.so.2...done. Loaded symbols for /lib/libdl.so.2 Reading symbols from /lib/tls/libpthread.so.0...done. Loaded symbols for /lib/tls/libpthread.so.0 Reading symbols from /lib/tls/librt.so.1...done. Loaded symbols for /lib/tls/librt.so.1 Reading symbols from /usr/lib/libstdc++.so.5...done. Loaded symbols for /usr/lib/libstdc++.so.5 Reading symbols from /lib/tls/libm.so.6...done. Loaded symbols for /lib/tls/libm.so.6 Reading symbols from /lib/libgcc_s.so.1...done. Loaded symbols for /lib/libgcc_s.so.1 Reading symbols from /lib/tls/libc.so.6...done. Loaded symbols for /lib/tls/libc.so.6 Reading symbols from /lib/ld-linux.so.2...done. Loaded symbols for /lib/ld-linux.so.2 #0 boost::pool<boost::default_user_allocator_new_delete>::ordered_malloc_ne ed_resize (this=0x8fc94c0) at simple_segregated_storage.hpp:46 46 { return *(static_cast<void **>(ptr)); } (gdb) where #0 boost::pool<boost::default_user_allocator_new_delete>::ordered_malloc_ne ed_resize (this=0x8fc94c0) at simple_segregated_storage.hpp:46 #1 0x08049559 in boost::object_pool<Foo, boost::default_user_allocator_new_delete>::construct (this=0x8fca760) at object_pool.hpp:76 #2 0x08049492 in boost::detail::function::void_function_obj_invoker0<Bar, void>::invoke (function_obj_ptr= {obj_ptr = 0x8fc9628, const_obj_ptr = 0x8fc9628, func_ptr = 0x8fc9628, data = "("}) at custom.cpp:26 #3 0x00138ef5 in boost::function0<void, std::allocator<boost::function_base> >::operator() (this=0x40417a70) at function_template.hpp:581 #4 0x00138145 in thread_proxy (param=0xbfffc950) at /home/tjamet/dev/boost_1_32_0/libs/thread/src/thread.cpp:113 #5 0x00454dec in start_thread () from /lib/tls/libpthread.so.0 #6 0x00224a2a in clone () from /lib/tls/libc.so.6 -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Caleb Epstein Sent: Saturday, August 06, 2005 11:31 AM To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] Maybe a problem with boost::object_pool? On 8/5/05, Labour, Matthieu (MLPRO SF) <matthieu_labour(a)ml.com> wrote: > There are 2 threads. > One thread creates Objects using boost::object_pool and enqueues them into a > stl queue. Are you using a mutex to protect the push/pop operations on the queue? If not, then this is your problem. -- Caleb Epstein caleb dot epstein at gmail dot com _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users -------------------------------------------------------- If you are not an intended recipient of this e-mail, please notify the sender, delete it and do not read, act upon, print, disclose, copy, retain or redistribute it. Click here for important additional terms relating to this e-mail. http://www.ml.com/email_terms/ --------------------------------------------------------

2 1

Re: [Boost-users] [multi_index] - iterator validity?
by JOAQUIN LOPEZ MU?Z 09 Aug '05

09 Aug '05

----- Mensaje original ----- De: Stuart Dootson <stuart.dootson(a)gmail.com> Fecha: Martes, Agosto 9, 2005 11:01 am Asunto: [Boost-users] [multi_index] - iterator validity? > Hi - I've just started using multi_index. I'm wondering - does > multi_index satisfy the same iterator validity constraints as the > standard associative containers, i.e: > > The insert members shall not affect the validity of iterators and > references to the container, and the erase members shall invalidate > only iterators and references to the erased elements. > > I *have* looked in the documentation, but haven't been able to find > anything which tells me. > This guarantee (stability) is held for all types of indices --ordered, sequenced and hashed (the last case is not so obvious, as TR1 does not require hashed containers to be stable.) There is no statement on the docs ensuring that multi_index_containers are stable because this is not a property of the container, but rather of each particular type of index: conceivably, a future index type may not guarantee stability for whatever reason, though currently this is not the case. So, you can see stability is referred to in the reference section of each type of index. For instance, for ordered indices stability is (succintly) stated in http://boost.org/libs/multi_index/doc/reference/ord_indices.html#ord_in dices Does this clear things up for you? Best, Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

2 1

gcc 3.4.4 ded_typename test failure
by Steven Solie 09 Aug '05

09 Aug '05

I'm stumped as to why the following is happening with GCC 3.4.4 on the AmigaOS PPC platform with Boost 1.32.0: no_ded_typename_pass (passed) no_ded_typename_fail (failed) All the other config tests pass except for non_ded_typename_fail which always compiles. According to the Jamfile it should not compile. I'm using v1 of bjam if that makes any difference. Here is the contents of config_test.output for reference: GNU C++ version 3.4.4 (AmigaOS build 20050716) __cplusplus =1 __STDC__ =1 __GNUC__ =3 __GNUC_MINOR__ =4 __GNUG__ =3 __VERSION__ ="3.4.4 (AmigaOS build 20050716)" __CHAR_UNSIGNED__ =1 __REGISTER_PREFIX__ [no value] __USER_LABEL_PREFIX__ [no value] __EXCEPTIONS =1 __ELF__ =1 __powerpc__ =1 __STDC_HOSTED__ =1 ********************************************************************* GNU libstdc++ version 20050519 __GLIBCXX__ =20050519 _GLIBCXX_USE_LONG_LONG =1 _GLIBCXX_USE_C99_MATH =1 _GLIBCXX_HOSTED =1 _GLIBCXX_SJLJ_EXCEPTIONS =1 ********************************************************************* Detected Platform: AmigaOS Type char is unsigned Type wchar_t is signed byte order for type short =8 0 byte order for type int =24 16 8 0 byte order for type long =24 16 8 0 sizeof(wchar_t) =4 sizeof(short) =2 sizeof(int) =4 sizeof(long) =4 sizeof(size_t) =4 sizeof(ptrdiff_t) =4 sizeof(void*) =4 sizeof(void(*)(void)) =4 sizeof(float) =4 sizeof(double) =8 sizeof(long double) =8 CHAR_BIT =8 CHAR_MAX =255 WCHAR_MAX =65535 SHRT_MAX =32767 INT_MAX =2147483647L LONG_MAX =2147483647L LLONG_MAX =0x7fffffffffffffffLL ULLONG_MAX =0xffffffffffffffffULL ********************************************************************* Boost version 103200 BOOST_USER_CONFIG =<boost/config/user.hpp> BOOST_COMPILER_CONFIG ="boost/config/compiler/gcc.hpp" BOOST_STDLIB_CONFIG ="boost/config/stdlib/libstdcpp3.hpp" BOOST_PLATFORM_CONFIG ="boost/config/platform/amigaos.hpp" BOOST_DEDUCED_TYPENAME =typename BOOST_DISABLE_THREADS [no value] BOOST_HAS_DIRENT_H [no value] BOOST_HAS_GETTIMEOFDAY [no value] BOOST_HAS_LONG_LONG [no value] BOOST_HAS_NRVO [no value] BOOST_HAS_PARTIAL_STD_ALLOCATOR [no value] BOOST_HAS_STDINT_H [no value] BOOST_HAS_UNISTD_H [no value] BOOST_MSVC6_MEMBER_TEMPLATES [no value] BOOST_NESTED_TEMPLATE =template BOOST_NO_CWCHAR [no value] BOOST_NO_CWCTYPE [no value] BOOST_NO_HASH [no value] BOOST_NO_INTRINSIC_WCHAR_T [no value] BOOST_NO_MS_INT64_NUMERIC_LIMITS [no value] BOOST_NO_SLIST [no value] BOOST_NO_STD_WSTREAMBUF [no value] BOOST_NO_STD_WSTRING [no value] BOOST_NO_SWPRINTF [no value] BOOST_STD_EXTENSION_NAMESPACE =std BOOST_UNREACHABLE_RETURN(0) [no value] Any ideas? --Steven Solie

2 1

[multi_index] - iterator validity?
by Stuart Dootson 09 Aug '05

09 Aug '05

Hi - I've just started using multi_index. I'm wondering - does multi_index satisfy the same iterator validity constraints as the standard associative containers, i.e: The insert members shall not affect the validity of iterators and references to the container, and the erase members shall invalidate only iterators and references to the erased elements. I *have* looked in the documentation, but haven't been able to find anything which tells me. TIA Stuart Dootson

1 0

Stack based "pimpl"
by Michal Kandulski 09 Aug '05

09 Aug '05

Hi! I'd like to implement a pimpl, pointing to a stack "allocated" buffer by means of palcement new/destroy. There would be no problem but there are alignment issues. What really I need is: - a way to check by te means of Boost wheater that an object of class X can be placed (with placement new) in a buffer of given adress and size: - a way of getting a pointer to X (within the buffer) where the object could be paced // file.hpp: struct X { X(); ~X(); private: struct Impl; const static int max_size = 100; Impl* m_pimpl; char m_impl_buffer[max_size]; bool m_on_stack; }; // file.cpp: #include "file.hpp" struct X::Impl { // members }; X::X() { m_on_stack = value_of(max_size is big enough for an object of X); if (m_on_stack) m_pimpl = new (m_impl_buffer)Impl; // any alignment stuff here??? else m_pimpl = new Impl; } X::~X() { if (m_on_stack) m_pimpl->~Impl(); else delete m_pimpl; } I'll be gratefull for any help, Michal

2 2

Simple Boost Lambda problem using argument as object. part III
by Thaddeus Olczyk 08 Aug '05

08 Aug '05

Here I go again. This time the code is: #include <boost/lambda/lambda.hpp> #include <boost/lambda/bind.hpp> #include <boost/lambda/if.hpp> #include <vector> #include <set> #include <string> #include <iostream> using boost::lambda::_1; //using boost::lambda::bind; class A { std::string s; public: A(std::string sIn):s(sIn){} void print_A() { std::cout<<"A="<<s<<std::endl; } bool isLong() { return s.length()>5; } std::string get(){return s;} }; typedef std::vector<A *> AList; typedef std::set<std::string> StrList; typedef StrList::const_iterator (StrList::* StrListFind)(const StrList::key_type&) const; typedef std::pair <StrList::iterator,bool> (StrList::* StrListInsert)(const StrList::value_type& _Val); StrListFind strListFind=&StrList::find; StrListInsert strListInsert=&StrList::insert; int main() { std::string in; AList alist; StrList excluded; while(in!="end") { in=""; std::getline(std::cin,in); alist.push_back(new A(in)); } std::for_each(alist.begin(),alist.end(), boost::lambda::if_then((bind(strListFind,&excluded,bind(&A::get,_1))) !=excluded.end()), bind(strListInsert,&excluded,bind(&A::get,_1))); std::for_each(alist.begin(),alist.end(), bind(&A::print_A,_1)); return 1; } The compiler output is: g++ -g -o test.exe -I/cy/f/Programs/development/boost_1_32_0/nocygwin/include/boost-1_32/ main2.cpp main2.cpp: In function `int main()': main2.cpp:50: error: no matching function for call to `if_then(const boost::lambda::lambda_functor<boost::lambda::lambda_functor_base<boost::lambda::relational_action<boost::lambda::notequal_action>, boost::tuples::tuple<boost::lambda::lambda_functor<boost::lambda::lambda_functor_base<boost::lambda::action<3, boost::lambda::function_action<3, boost::lambda::detail::unspecified> >, boost::tuples::tuple<std::_Rb_tree_const_iterator<std::string> (std::set<std::string, std::less<std::string>, std::allocator<std::string> >::*const)(const std::string&) const, StrList* const, const boost::lambda::lambda_functor<boost::lambda::lambda_functor_base<boost::lambda::action<2, boost::lambda::function_action<2, boost::lambda::detail::unspecified> >, boost::tuples::tuple<std::string (A::*const)(), const boost::lambda::lambda_functor<boost::lambda::placeholder<1> >, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type> > >, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type> > >, const std::_Rb_tree_const_iterator<std::string>, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type> > >)' make: *** [test.exe] Error 1 Any idea what is wrong? Thanks -- Thad

2 1

[serialization] serializing dynamic arrays
by Ivan Rachev 08 Aug '05

08 Aug '05

Does anyone have an idea how to serialize dynamic arrays of objects? An example follows at the end. Its weakness shows up when an outside pointer points to an element in the array. The problem is that the type being serialized is T but not T*. Thanks, Ivan #include <fstream> #include <boost/archive/text_iarchive.hpp> #include <boost/serialization/access.hpp> #include <boost/serialization/split_member.hpp> class A { friend class boost::serialization::access; template<class Archive> void serialize(Archive & ar, const unsigned int version) {} }; //IVAN Element always points to the beginning of an array of // 'Size' elements. The array is NOT null-terminated. If // size == 0, Elements may hold an undefined value. User // is responsible for initializing, new-ing and delete-ing. // Assumption: if there is an outside pointer to an object // inside this array, after loading the array // that pointer will hold a different object. template <typename T> struct DynamicArray { int Size; T* Element; friend class boost::serialization::access; template<class Archive> void save(Archive & ar, const unsigned int version) const { ar & Size; for (int i = 0; i < Size; ++i) ar & Element[i]; } template<class Archive> void load(Archive & ar, const unsigned int version) { ar & Size; //TODO: assert(Size >=0); if (Size > 0) { Element = new T[Size]; for (int i = 0; i < Size; ++i) ar & Element[i]; } } BOOST_SERIALIZATION_SPLIT_MEMBER() }; int main() { DynamicArray<A> ArrayOfAs; A* secondElement; std::ifstream ifs("test.txt"); boost::archive::text_iarchive ia(ifs); ia & ArrayOfAs; ia & secondElement; // this guy will point to a copy of the // 2nd element but NOT the 2nd element itself delete [] ArrayOfAs.Element; }

3 20

boost::iostreams - user defined format flags
by Alan M. Carroll 08 Aug '05

08 Aug '05

I'm trying to use the boost::iostreams library but I've run in to a problem that makes that difficult. The basis is that I want to write manipulators to manipulate the behavior of the filters in a filtering iostream. The standard mechanism for such manipulators are user defined format flags for streams, accessed via xalloc() and iword()/pword(). These are not available in a filter. It is also not feasible to store the formatting state in the filters, as there are not accessible from the filtering stream (and hence inaccessible to the manipulators). I exchanged an e-mail with the author in which he states that a forthcoming version of boost::iostreams will provide access to the filters from the stream object. Manipulators could then reach in to the filter objects and modify them to adjust formatting behavior. However, while this is desirable for other reasons, I'd prefer to be able to use iword/pword from the filters, i.e. be able to call these methods on the Sources and Sinks (which, presumably, would forward them to the actual stream). It seems more "standard" to me, although one could argue for storing state directly in the filters from a clean design point of view. Does anyone else have an opinion on this? Or have a suggestion for an alternate method of creating manipulators to adjust the behavior of the filters using stream local state? I searched around for previous discussion of this issue but didn't find one. If there was, a pointer to it would be much appreciated. P.S. For those interested, the details of what I want follow. I'm doing some work with XML which involves user input. Because one never knows what a user might put in his data, I need to encode/decode problematic characters such as angle brackets. The problem with having streams to do the encode/decode is that one doesn't always want this performed on the data. For instance, during output one needs to send literal angle brackets around the element tags and I don't want to get in to trying to second guess the encoder to work around it. The solution I decided to implement is to have manipulators that turn the coding on and off, so one would do something like this: xml::coding::ostream s; s << "<tag>" << xml::coding::on << node.content() << xml::coding::off << "</tag>"; This allows low level code fine grained control over what gets coded while leaving the details up to the stream provided by the application. The hassle of setting up strings for in place conversions is also avoid, along with the problem of making sure every body is using the same coding algorithms. However, it doesn't help all that much on input as a parser has to read from a non-filtered stream to correctly detect the tags. Element content will end up in a string and it's no easier to use a filtering stream on that than to just call a decode function.

1 0

[bcp]minimal option
by Thomas Matelich 08 Aug '05

08 Aug '05

I was really shocked when a bcp shared_ptr foo/ copied 274 files. I expected a lot, but not that many. Then I read the doc's more closely and it says that if you depend on any header from a lib, you get the whole lib. Could it be possible to add a -minimal option to have only the actual headers used included? Thanks, Tom

3 3

Re: [Boost-users] [Serialization] 1.33 RC1 Problem linking void_cast_register
by Richard Jennings 08 Aug '05

08 Aug '05

Thanks for your response on this. I have tried moving the void_cast_register to the derived class constructor (without success) and I have looked further in to my use case. This was probably brought about by me removing the archive and serialization headers from the precompiled header file that Borland uses because there were other compilation problems in keeping them in. The use case is as follows: Classes A and B are derived from class Z. Class M has a pointer to class Z as member data. This member data is set to a pointer to an instance of A or B as appropriate. Class S serializes M. Class S includes the Archive headers as it creates the archive class instance. S includes the headers for M that includes the headers for Z. However S does not include the headers for A or B, because it does not need to, and in fact does not need to know of the existence of A or B as they are used polymorphically. In our case S does not use A/B/Z but others use them via M. I have a free function called RegisterTypes in a module that has BOOST_CLASS_EXPORT(A) and void_cast_register<A,Z>(0,0) in it. It's run once at application start. Which means that A.hpp is never included in a compilation unit with any archive header and therefore the void_cast_register function is not generated (as I understand it). I realise that serializing classes via a pointer to the base class and void_cast_register<T> is a seldom used corner of serialization. I can fix the problem by including archive headers where RegisterTypes is defined even though they may not be needed, and then the application links ok. I'm not sure if there is a need for a change, you are best to judge. I hope this explains my use-case. Richard -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Robert Ramey Sent: 04 August 2005 17:57 To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] [Serialization] 1.33 RC1 Problemlinkingvoid_cast_register I believe this is a side effect of our efforts to be sure that the mere inclusion of a class that includes serialization headers doesn't trigger (through auto-linking) the requirement for inlcuding a library module. The idea is that one would never need "void_cast_register<T>" to be generated unless there is an archive somewhere involved. So that raises the question of what the context is here. Your test illustrates the link error but it doesn't illustrate the use case. That is, what is the context of explicitly invoking void_cast_register without an archive header present? I'm not saying its wrong. Its just that I don't see how this comes up. When we implemented changes which resulted in this symptom - we enivisioned a pattern of usage like that shown below. In this way, the *.hpp files could be included in other programs without triggering auto-link and requiring linking with the boost library to include code that in fact will never be used. Were we missing something here? What is it? Robert Ramey Z.hpp ==== #include <boost/serialization/is_abstract.hpp> #include <boost/serialization/serialization.hpp> #include <boost/serialization/export.hpp> class Z ... }; BOOST_IS_ABSTRACT(Z) A.hpp ==== #include ... // other serialization headers #include "Z.hpp" class A : public Z { A(){ boost::serialization::void_cast_register<A, Z>(0,0); } }; BOOST_CLASS_EXPORT(A) B.hpp ==== #include ... // other serialization headers #include "Z.hpp" class B : public Z { ... B(){ boost::serialization::void_cast_register<B, Z>(0,0); } }; BOOST_CLASS_EXPORT(B) main.cpp ====== #include <boost/serialization/text_oarchive.hpp> #include <boost/serialization/text_iarchive.hpp> #include "Z.hpp" #include "A.hpp" #include "B.hpp" int main(int argc, char* argv[]){ return 0; } _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

2 1