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

[graph] How to find the number of adjacent nodes
by Suresh Kumar 22 Dec '09

22 Dec '09

Hi Is there any easy method to find out the number of adjacent nodes in a graph? I am using adjacency_list with listS as parameters. Hence I i am unable to subtract the vertex iterator values returned by adjacent_vertices. Thanks suresh -- R. Suresh Kumar PhD Student, Vislab #216, Engineering II Bldg University of California, Riverside,

2 1

How to use boost::asio to implement a synchronous operation with timeouts?
by watsonsong 22 Dec '09

22 Dec '09

In the boost official documents, the example of timeouts in asio is implements by asynchronous operations and a deadline_timer. But what I want is a synchronous operation with a timeout, but I search nearly all of the reference manual and find nothing about it. In an other words, how can I used the asynchronous operations to simulate synchronous operations with timeouts. Maybe I could launch an asynchronous operation and call io_service::run_one immediately, but I don't know whether it is safe when I call io_service::run_one in a complete callback function which is call by io_service::run? Or there is any classic solution of this question. Thank you. -- View this message in context: http://old.nabble.com/How-to-use-boost%3A%3Aasio-to-implement-a-synchronous… Sent from the Boost - Users mailing list archive at Nabble.com.

1 0

storing data
by Sean Farrow 22 Dec '09

22 Dec '09

Hi: I am in need to store data. Basically I have an application, when I have 8 collumns that are named, however thee are 38 possible rows with differing nameson. I also need to store the screen co-ordinates for each column name on each row. So for example: Column 1, name Item, row 1, x 66, y 85, row 2 x 88 y 121. What is the best way of being able to store this data with boost and get access to the item name, and co-ordinates given a row and column number? Any help appreciated. Cheers Sean./

1 0

Echo server demo
by Michael Fowler 22 Dec '09

22 Dec '09

Hello. I tried using the code provided here: http://www.boost.org/doc/libs/1_41_0/doc/html/boost_asio/example/echo/block… to echo text back to a client but i haven't been successful. It seems to hang at size_t length = sock->read_some(boost::asio::buffer(data), error); what is wrong? Thanks

2 2

[Serialization] shared_ptr to const object
by elizabeta petreska 21 Dec '09

21 Dec '09

Hello I am using boost 1.38.0 How to serialize shared_ptr member which points to const object. The following compiles, but is not working i.e the restored boost_shared_ptr does not point to anything. I tried to use const_pointer_cast to cast the const away. I suspect that this is the problematic part. I think the question here is , how should I cast away the const correctly for boost::shared_ptr<const T>? #include <boost/archive/xml_iarchive.hpp> #include <boost/archive/xml_oarchive.hpp> #include <boost/serialization/nvp.hpp> #include <boost/serialization/utility.hpp> #include <fstream> #include <boost/serialization/shared_ptr.hpp> class B { public: B() { } B(int a):ma(a) { } int ma; template<class Archive> void serialize(Archive & ar, const unsigned int version) { using boost::serialization::make_nvp; ar & make_nvp("ma",ma); } }; class BC { public: BC() { } BC(boost::shared_ptr<B>& t):mac(t) { } boost::shared_ptr<const B> mac; template<class Archive> void serialize(Archive & ar, const unsigned int version) { using boost::serialization::make_nvp; ar & make_nvp("mac", boost::const_pointer_cast<B> (mac)); } }; int main() { std::ofstream ofs(L"C:\\MyTest.xml"); std::ifstream ifs(L"C:\\MyTest.xml"); boost::shared_ptr<B> bb(new B(4)); BC aa(bb); { boost::archive::text_oarchive oa(ofs); oa & BOOST_SERIALIZATION_NVP(aa); } BC raa; { boost::archive::text_iarchive ia(ifs); ia & BOOST_SERIALIZATION_NVP(raa); } }

3 6

Bug using mpl with <ciso646> and MSVC?
by ebryan 21 Dec '09

21 Dec '09

Hello, I am using Spirit with some code that uses the keywords 'and', 'or', etc. MSVC (at least my version, 9.0) does not implement these keywords, but support can be added by including <ciso646>, which just has a bunch of macros, ie. "#define and &&". The problem I am running into is that when I try to compile something that uses Spirit and ciso646, I get this compiler error: boost\mpl\aux_\include_preprocessed.hpp(37) : fatal error C1083: Cannot open include file: 'boost/mpl/aux_/preprocessed/plain/&.hpp': No such file or directory The offending line in include_preprocessed.hpp is: # include BOOST_PP_STRINGIZE(boost/mpl/aux_/preprocessed/AUX778076_PREPROCESSED_HEADER) It looks like the macro 'bitand' is being expanded in here, so it tries to include '&.hpp'. Now, I think this problem has already been solved (mostly). See this archived discussion, where this problem was worked around for 'and' and 'or': http://lists.boost.org/Archives/boost/2004/05/65563.php I have checked the header files in boost/mpl and only see three more conflicts between header names and ciso646 macros: 'bitand', 'bitor', and 'not'. If someone with more expertise than myself could implement a similar workaround for these headers, I would be very grateful. Thanks, Erik -- View this message in context: http://old.nabble.com/Bug-using-mpl-with-%3Cciso646%3E-and-MSVC--tp26851153… Sent from the Boost - Users mailing list archive at Nabble.com.

1 1

Using Fusion adapted struct with mpl::is_sequence
by John Smith 21 Dec '09

21 Dec '09

Hi, I have adapted a struct to be a fusion sequence like so: struct s { uint8_t i; uint8_t j; uint8_t k; }; BOOST_FUSION_ADAPT_STRUCT( s, (uint8_t, i) (uint8_t, j) (uint8_t, k) ) which works fine. However, I pass the adapted struct to my MPL algorithm and it seems to always return false on mpl::is_sequence. If I typedef boost::fusion::fusion_sequence_tag tag in the struct, then it all works, also if I pass in a proper fusion sequence, then that works as well. I was wondering if there was a different way of doing things, I don't really want to require the tag to be typedef'ed in the structure. Thanks, Dave

4 5

Serialization of derived class in a plugin
by Jean-Charles Quillet 21 Dec '09

21 Dec '09

Hi, I've been using the boost serialization on many projects. But I can't manage to make it work on my current project. Basically, I've a main application and an abstract class located in a separate dll. The application loads dynamically plugins which defines classes derived from the abstract class. Then I'm trying to serialize a class created by a plugin using a pointer to the abstract class, I always get a "Unregistered class" exception. I've used the BOOST_CLASS_EXPORT macro in the plugins trying to make the boost serialization library aware of my classes without any success so far. Looking in the source directory and specially in the "test" directory, I've seen that what I'm trying to do is the same as the test described by the file test_dll_plugin.cpp, unfortunately, it seems the test is not compiled and I can't manage to make it link without error. So my main concern is: is this feature supposed to be working ? If not I'll have to review my design. If it does what is the magic to make it work (like how to compile test_dll_plugin.cpp or another simple working example) ? Thanks for any help, Jean-Charles

2 8

How can I create a ip::tcp::iostream from a SOCKET descriptor?
by Eric J. Holtman 21 Dec '09

21 Dec '09

I'm trying to convert my hand-rolled marshalling code into boost::serialization and boost::ip classes. My code gets handed a SOCKET handle (win XP). I have no control over this, the connection/server/client stuff is handled elsewhere, I just get the SOCKET after everything's already open. How can I create an iostream from that, so I can then use the Archive classes to send and receive objects over the socket? I understand the "harder" stuff (writing the overloads for my classes for input/output to a generic stream). I'm just stuck on finding that construction (or function) that will take an already opened SOCKET handle and wrap an iostream around it. any clues?

2 2

[serialization] "unregistered class" with 1.41, o.k. with 1.38
by Ruediger Berlich 21 Dec '09

21 Dec '09

Hi there, Consider a hierarchy of classes that is accessed through an abstract base class. The most derived class is a template which is "filled" with the basic types bool, char, boost::int32_t and double. I use typedefs of the kind "typedef GParameterT<char> GChar;" for ease of use. This is held in its own header file, in this case GChar.hpp . A corresponding GChar.cpp holds the export statement needed to register the class (or typedef in this case): #include <boost/serialization/export.hpp> BOOST_CLASS_EXPORT(Gem::GenEvA::GChar) There are also identical (except for base types) files for GDouble, GBoolean and GInt32 (holding the tell-tale typedefs and exports for the corresponding types). I'm using unit tests for my library, and there you will find code like: // Test (de-)serialization in different modes { // plain text format T gpt4(gpt0); BOOST_CHECK(gpt4.isEqualTo(gpt0)); BOOST_CHECK_NO_THROW(gpt4.fromString(gpt1.toString(TEXTSERIALIZATION), TEXTSERIALIZATION)); BOOST_CHECK(!gpt4.isEqualTo(gpt0)); BOOST_CHECK(gpt4.isSimilarTo(gpt1, exp(-10))); } // [ ... further serialization modes removed ... ] T can again be GChar, GBoolean, GDouble or GInt32 . to/fromString is a simple wrapper to the serialization calls (all happening through the base pointer). Now I have the problem that the toString/fromString throws an "unregistered class" with Boost 1.4.1, but **only** if GParameterT is fed a double or a boost::int32_t . (De-)Serialization works fine with char and bool as template parameter. With Boost 1.38, no errors occur at all, and also my production code works nicely with 1.41 (and it uses Boost.Serialization extensively). Any idea ? In any case: Thanks for a great library and have a nice christmas time and a happy new year! Best Regards, Ruediger

1 0

bimap find
by Bosko Markovic 21 Dec '09

21 Dec '09

Hi, I have one simple question for which I could not find answer in Bimap documentatio nor on the list. When searching bimap using find method what is the value of returned iterator if nothing matches search query? Example code: BimapDWORDVariant::right_map::const_iterator right_iter = bm.right.find(sSyncMsg); Thank you. BoskoM.

2 2

[boost::interprocess] Do I need to use lock before allocate shared memory?
by blp330 21 Dec '09

21 Dec '09

Hi, Do I need to use lock before allocate shared memory? I have two processes that will allocate string to the same shared memory, do I need to use lock? Thanks. -- View this message in context: http://old.nabble.com/-boost%3A%3Ainterprocess--Do-I-need-to-use-lock-befor… Sent from the Boost - Users mailing list archive at Nabble.com.

2 1

boost::iostreams
by Anand Patel 20 Dec '09

20 Dec '09

I have implemented iostream::filter and would like to set the stream(source/sink) error flags in the event of unexpected conditions. I did not find anyway to set the flags. Can someone please help. -Anand

1 0

asio: sys/epoll.h
by Pascal Kesseli 20 Dec '09

20 Dec '09

Hi everyone Trying to cross-compile my embedded code to a new platform, I hit the following error message: ... /home/beo/Software/C++/UC-8410/boost/include/boost-1_39/boost/asio/detail/epoll_reactor.hpp:27:23: warning: sys/epoll.h: No such file or directory ... My cross toolchain obviously does not feature such a file. Any way to circumvent or solve this issue? These are my compiler specs: g++ (GCC) 4.2.1 Copyright (C) 2007 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. It's the UC-8410 MOXA toolchain. Thanks in advance for any help and best regards Pascal _________________________________________________________________ Ski-Weltcup: Alle Rennen, alle Resultate und News auf MSN Sport http://sport.ch.msn.com/skialpin/

2 1

undefined reference to `boost::thread_resource_error::thread_resource_error()'
by Stefano 20 Dec '09

20 Dec '09

Hello, On ubuntu, I have a piece of software linking against the boost thread library. Linking with boost version 1.40 I don't have any problem, whereas switching to 1.41 I got the undefined reference to `boost::thread_resource_error::thread_resource_error()' In the sources I see that this constructor is now inlined and if I do nm -C -D libboost_thread.so | grep thread_resource_error from this (1.40) 0000ee80 T boost::thread_resource_error::thread_resource_error(int) 0000ee00 T boost::thread_resource_error::thread_resource_error() 0000ee40 T boost::thread_resource_error::thread_resource_error(int) 0000edc0 T boost::thread_resource_error::thread_resource_error() 0000f4c0 T boost::thread_resource_error::~thread_resource_error() 0000f500 T boost::thread_resource_error::~thread_resource_error() 0000f530 T boost::thread_resource_error::~thread_resource_error() 0000eec0 T boost::thread_resource_error::what() const 00013e00 V typeinfo for boost::thread_resource_error 0000ffc0 V typeinfo name for boost::thread_resource_error 00013d98 V vtable for boost::thread_resource_error I got this (1.41) 0000b6e0 W boost::thread_resource_error::~thread_resource_error() 0000b720 W boost::thread_resource_error::~thread_resource_error() 0000b480 W boost::thread_resource_error::what() const 0001282c V typeinfo for boost::thread_resource_error 0000ed00 V typeinfo name for boost::thread_resource_error 00012848 V vtable for boost::thread_resource_error The symbol is weak now. I should be not a problem, right? But I got the unresolved ref What is the best option to have my stuff compiled without the unresolved refs? Thanks Stefano

4 4

[iostreams] writing to filtering_ostream
by Igor R 20 Dec '09

20 Dec '09

Hello, I've got an application that originally used IOStreams of Boost 1.34. Now I recompiled it with 1.41, and everything stopped functioning. Meanwhile, I figured out that one of the failures occurs due to the following line: iostreams::copy(istream(&someStreambuf), someFilteringStream); The intention was just to copy a chunk of binary data from an input stream to a filtering stream, but it turns out that iostreams::copy() now *closes* the sink. I tried to re-write it like this: someFilteringStream->write(sourceData, sourceSize); or this: iostreams::write(someFilteringStream, sourceData, sourceSize); But neither of the above executes the filter. So what's the right way to do this in 1.41? Thanks.

1 1

[boost::variant] problem with swap
by Ondrej Kuncar 20 Dec '09

20 Dec '09

Hello! I have problems to compile this code with Visual Studio 9.0 SP1 and Boost 1.41: ------------------------------------------------------------------- #include <boost/variant.hpp> #include <deque> typedef boost::variant<std::deque<int>::iterator> SemanticValue; int main() { SemanticValue v1; SemanticValue v2; v1.swap(v2); } ------------------------------------------------------------------- I got this error: c:\program files\boost\boost_1_41_0\boost\variant\detail\move.hpp(155) : error C2668: 'std::swap' : ambiguous call to overloaded function c:\program files\boost\boost_1_41_0\boost\variant\detail\move.hpp(141): could be 'void boost::detail::variant::detail::move_swap::swap<T>(T &,T &)' with [ T=T0 ] c:\program files\microsoft visual studio 9.0\vc\include\utility(16): or 'void std::swap<T>(_Ty &,_Ty &)' [found using argument-dependent lookup] with [ T=T0, _Ty=T0 ] while trying to match the argument list '(T0, T0)' c:\program files\boost\boost_1_41_0\boost\variant\variant.hpp(698) : see reference to function template instantiation 'void boost::detail::variant::move_swap<T>(T &,T &)' being compiled with [ T=T0 ] c:\program files\boost\boost_1_41_0\boost\variant\variant.hpp(825) : see reference to function template instantiation 'void boost::detail::variant::swap_with<Variant>::operator ()<T>(T &) const' being compiled with [ Variant=boost::variant<std::_Deque_iterator<int,std::allocator<int>,true>>, T=T0 ] c:\program files\boost\boost_1_41_0\boost\variant\detail\visitation_impl.hpp(141) : see reference to function template instantiation 'void boost::detail::variant::invoke_visitor<Visitor>::internal_visit<T>(T &,int)' being compiled with [ Visitor=boost::detail::variant::swap_with<boost::variant<std::_Deque_iterator<int,std::allocator<int>,true>>>, T=T0 ] c:\program files\boost\boost_1_41_0\boost\variant\detail\visitation_impl.hpp(169) : see reference to function template instantiation 'void boost::detail::variant::visitation_impl_invoke_impl<Visitor,VoidPtrCV,T>(int,Visitor &,VoidPtrCV,T *,boost::mpl::false_)' being compiled with [ Visitor=boost::detail::variant::invoke_visitor<boost::detail::variant::swap_with<boost::variant<std::_Deque_iterator<int,std::allocator<int>,true>>>>, VoidPtrCV=void *, T=T0 ] c:\program files\boost\boost_1_41_0\boost\variant\detail\visitation_impl.hpp(252) : see reference to function template instantiation 'void boost::detail::variant::visitation_impl_invoke<Visitor,VoidPtrCV,T0,NoBackupFlag>(int,Visitor &,VoidPtrCV,T *,NoBackupFlag,int)' being compiled with [ Visitor=boost::detail::variant::invoke_visitor<boost::detail::variant::swap_with<boost::variant<std::_Deque_iterator<int,std::allocator<int>,true>>>>, VoidPtrCV=void *, NoBackupFlag=boost::variant<std::_Deque_iterator<int,std::allocator<int>,true>>::has_fallback_type_, T=T0 ] c:\program files\boost\boost_1_41_0\boost\variant\variant.hpp(1769) : see reference to function template instantiation 'void boost::detail::variant::visitation_impl<first_which,first_step,Visitor,VoidPtrCV,boost::variant<T0_>::has_fallback_type_>(const int,const int,Visitor &,VoidPtrCV,boost::mpl::false_,NoBackupFlag,Which *,step0 *)' being compiled with [ Visitor=boost::detail::variant::invoke_visitor<boost::detail::variant::swap_with<boost::variant<std::_Deque_iterator<int,std::allocator<int>,true>>>>, VoidPtrCV=void *, T0_=std::_Deque_iterator<int,std::allocator<int>,true>, NoBackupFlag=boost::variant<std::_Deque_iterator<int,std::allocator<int>,true>>::has_fallback_type_, Which=first_which, step0=first_step ] c:\program files\boost\boost_1_41_0\boost\variant\variant.hpp(1780) : see reference to function template instantiation 'void boost::variant<T0_>::internal_apply_visitor_impl<Visitor,void*>(int,int,Visitor &,VoidPtrCV)' being compiled with [ T0_=std::_Deque_iterator<int,std::allocator<int>,true>, Visitor=boost::detail::variant::invoke_visitor<boost::detail::variant::swap_with<boost::variant<std::_Deque_iterator<int,std::allocator<int>,true>>>>, VoidPtrCV=void * ] c:\program files\boost\boost_1_41_0\boost\variant\variant.hpp(1803) : see reference to function template instantiation 'void boost::variant<T0_>::internal_apply_visitor<boost::detail::variant::invoke_visitor<Visitor>>(boost::detail::variant::invoke_visitor<Visitor> &)' being compiled with [ T0_=std::_Deque_iterator<int,std::allocator<int>,true>, Visitor=boost::detail::variant::swap_with<boost::variant<std::_Deque_iterator<int,std::allocator<int>,true>>> ] c:\program files\boost\boost_1_41_0\boost\variant\variant.hpp(1637) : see reference to function template instantiation 'void boost::variant<T0_>::apply_visitor<boost::detail::variant::swap_with<Variant>>(Visitor &)' being compiled with [ T0_=std::_Deque_iterator<int,std::allocator<int>,true>, Variant=boost::variant<std::_Deque_iterator<int,std::allocator<int>,true>>, Visitor=boost::detail::variant::swap_with<boost::variant<std::_Deque_iterator<int,std::allocator<int>,true>>> ] c:\program files\boost\boost_1_41_0\boost\variant\variant.hpp(1631) : while compiling class template member function 'void boost::variant<T0_>::swap(boost::variant<T0_> &)' with [ T0_=std::_Deque_iterator<int,std::allocator<int>,true> ] d:\dokumenty\visual studio 2008\projects\pokus4\pokus4\pokus4.cpp(20) : see reference to class template instantiation 'boost::variant<T0_>' being compiled with [ T0_=std::_Deque_iterator<int,std::allocator<int>,true> ] Do you have any ideas? Thanks. OK

2 2

VxWorks support in boost 1.41
by Geoff Shapiro 19 Dec '09

19 Dec '09

Having recently upgraded to just released 1.41 version of boost, I noticed the new support for the VxWorks platform. I am now in the process of porting over a body of code using boost to a VxWorks/PPC platform and obviously want to take advantage of the builtin support, but I've run into a problem and would like advice on the "boost" way of rectifying the problem. I am using VxWorks 6.7 to be running on an MPC8548 PPC. When I try to compile my project files in the standard manner for incorporating boost with tr1 support (.../boost/boost/tr1/tr1 and .../boost first in the header search paths, #include <memory> as first thing in files), I am getting a compile warning. The complaint is for already defined preprocessor macros -- in boost/config/vxworks.hpp at lines 23 and 24 there are definitions: #define _POSIX_TIMERS 1 #define _POSIX_THREADS 1 A little further down, this header file includes boilerplate code in boost/config/posix_features.hpp and that header file then includes unistd.h, a system file. For VxWorks 6.7, unistd.h also has defines for the above preprocessor macros, but sets their values to 200112L. While the warning may be able to ignored and is probably benign, IF unistd.h is always included after vxworks.hpp (I am not sure this will be the case for all of our files), I would rather see this fixed. What would be the best way to fix this so as to remain conformant with the "boost" philosophy for platform independence?

2 2

metafunction to 'rangify'
by Hicham Mouline 19 Dec '09

19 Dec '09

Hi, struct inputs { double input1; int input2; bool input3; }; Given type 'inputs', how to I write a metafunction that can return to me this type: rangify<inputs>::type identical to struct inputs_rangified { union { double fixed; boost::tuple<double, double, double> range; } input1; // range is min, max, increment union { int fixed; boost::tuple <int, int, int> range; } input2; union { bool fixed; std::pair<bool, bool> range; } input3; std::bitset<3> is_range; }; 1. Is it possible to get the number of members of a struct at compile-time? 2. What would be the easiest way to fill inputs_rangified from an input stream? Typically, my application will have hundreds different 'inputs' types and I would like to automate this process. Regards,

3 3

[mpl][type_traits] Getting the minimum and maximum value for a buit-in type at compile time
by Edward Diener 19 Dec '09

19 Dec '09

I would have thought that there must be a way to get the minimum and maximum value for a C++ built-in type but my search through mpl and type_traits yields nothing and numeric_limits requires a run-time function call. Surely this information is available to the compiler so it must be available to the TMP programmer somewhere. Where is it ?

3 4

[mpl] Strange integral constant overflow warning in mpl::less_equal
by Edward Diener 19 Dec '09

19 Dec '09

This simplified version of some TMP code I am writing using mpl and integer_traits is giving me a strange warning. I'm using VC9 and Boost 1.40: ------------------------------------------------------------------------------------- struct ResultTypeIntegerValues { typedef boost::mpl::long_<0L> lzero; typedef boost::mpl::long_<boost::integer_traits<long>::const_min> lmin; typedef boost::mpl::less_equal<lmin,lzero>::type type; }; ------------------------------------------------------------------------------------- with warning message: ------------------------------------------------------------------------------------- c:\utilities\boost\boost_1_40_0\boost\mpl\aux_\integral_wrapper.hpp(73) : warning C4307: '-' : integral constant overflow 1> c:\utilities\boost\boost_1_40_0\boost\mpl\aux_\preprocessed\plain\less_equal.hpp(60) : see reference to class template instantiation 'boost::mpl::long_<N>' being compiled 1> with 1> [ 1> N=-2147483648 1> ] 1> c:\utilities\boost\boost_1_40_0\boost\mpl\aux_\preprocessed\plain\less_equal.hpp(70) : see reference to class template instantiation 'boost::mpl::less_equal_tag<T>' being compiled 1> with 1> [ 1> T=nmspace::ResultTypeIntegerValues::lmin 1> ] 1> c:\programming\programs\xxx\yyy.h(64) : see reference to class template instantiation 'boost::mpl::less_equal<N1,N2>' being compiled 1> with 1> [ 1> N1=nmspace::ResultTypeIntegerValues::lmin, 1> N2=nmspace::ResultTypeIntegerValues::lzero 1> ] ------------------------------------------------------------------------------------- As far as I can see I am doing a perfectly valid comparison. Is this a minor bug in mpl or am I doing something wrong or should this warning just be expected ? Warning C4307 gives: ------------------------------------------------------------------------------------- Error Message 'operator' : integral constant overflow The operator is used in an expression that results in an integer constant overflowing the space allocated for it. You may need to use a larger type for the constant. A signed int holds a smaller value than an unsigned int because the signed int uses one bit to represent the sign. The following sample generates C4307: // C4307.cpp // compile with: /W2 int i = 2000000000 + 2000000000; // C4307 int j = (unsigned)2000000000 + 2000000000; // OK int main() { } -------------------------------------------------------------------------------------

2 2

Re: [Boost-users] [Boost.Test] valgrind complains about invalid reads
by Scott Banachowski 18 Dec '09

18 Dec '09

--- Steven Watanabe wrote: > joel falcou wrote: > > I'm jumping in the wagon but this very piece of code > leads to a > > "incorrect free" on Snow Leopard using g++ 4.2.1 > Apple. > > Maybe it is related to the virtual matter > > We've already discussed this many times. The code above > is wrong for reasons unrelated to the use of a virtual > destructor, > although adding a virtual destructor masks the problem. > This has been fixed in the trunk. Thanks Steven, with some searching I think I found the patch you are referring to: https://svn.boost.org/trac/boost/changeset/56467 Seeing this patch, I now see how the virtual function is a red herring (although I don't know why it happens to work, in this case). Nevertheless, I still like the idea of making it virtual, just to avoid the necessity of switching on type and casting to delete, but that's my opinion. Scott

1 0

sizeof of boost::int8_t
by eial＠cs.bgu.ac.il 18 Dec '09

18 Dec '09

running sizeof on a boost::int8_t returns 1, but what is the type? int/char? thanks.

8 7

[containers] defining the insertion point of non-unique keys
by Olaf Krzikalla 18 Dec '09

18 Dec '09

Hi @boost, I'm looking for the solution to the following problem: there is a collection of objects and I want to find a particular object by two different keys. The first key is unique, so there is no problem. The second key however is non-unique and if I search for an object using this key I want to find the element that was inserted last in that sequence. Neither boost::multindex nor boost:intrusive seems to have solutions right at hand. If using boost::multindex I could try to insert a new element in the non-unique sequence using the lower_bound iterator of that sequence as a hint. However it is not documented that in that case the element is always inserted before the hint. Same is true for boost:intrusive::multi_set/multi_map. And I'm not able to swap the position of two adjacent elements in a multi_set or multi_map, if they have equal keys. To illustrate the problem, here is some code: struct A { int a; // non-unique key void* b; // unique key // stuff }; using namespace boost::multi_index; typedef multi_index_container<A, indexed_by< ordered_non_unique<member<A,int,&A::a> >, ordered_unique<member<A,void*,&A::b> > > > tCollection; void foo(tCollection& collection, int key) { typedef tCollection::nth_index<0>::type tItByInt; tItByInt i = collection.get<0>().lower_bound(key); // use the current lower bound as a hint: tItByInt insertPos = collection.get<0>.insert(i, getObject(key)); // the big question: does the assertion holds? assert(collection.get<0>().lower_bound(key) == insertPos); } Best regards Olaf Krzikalla

4 9

Re: [Boost-users] Boost Graph and GraphViz - Names and Subgraphs
by Graham Hemingway 18 Dec '09

18 Dec '09

I am building graphs with boost::graph and want to generate GraphViz output. Most everything works as expected, but I am having trouble getting two things to work correctly: 1) My subgraphs are not visualized in the GraphViz output 2) I am setting the graph_name property of both the root graph and all subgraphs, but the names never appear in the GraphViz output. Any help would be appreciated. Cheers, Graham

1 0