- Boost-users - lists.boost.org

Getting multi_index_container in 1.46.1 to work with VC++ 6.0
by tangleman＠secules.net 05 May '11

05 May '11

I've got a legacy MFC project that was built in VC++ 6.0. It needs to stay compatible with VC6 runtime so that I can just drop it into Windows NT4 where the original program was and have it just work. I've finally gotten MFC and VC6 runtime to play nice in VS2003. Now I'm trying to work in boost libraries so that I can use the ethernet and serial aspects of ASIO. Of course, I downloaded and am working with the latest version of boost (1.46.1), and I have it to the point where 311 errors are reported at compile time, all having to do with multi_index_container. I've looked through the compiler specifics pages of the 1.46.1 and 1.45.0 releases with respect to the multi_index and have made the workaround changes that I could (i.e. member->member_offset, boost::multi_index::multi_index_container), but I am still getting syntax errors related to the typedef of my multi_index_container. If it will help, here is my definition: struct ID{}; struct Name{}; typedef boost::multi_index::multi_index_container< PDCUTestMessage, boost::multi_index::indexed_by< boost::multi_index::ordered_unique< boost::multi_index::tag<ID>, boost::multi_index::member_offset<PDCUTestMessage, unsigned char, offsetof(PDCUTestMessage, id)>>, boost::multi_index::ordered_unique< boost::multi_index::tag<Name>, boost::multi_index::member_offset<PDCUTestMessage, string, offsetof(PDCUTestMessage, name)>> > > PDCUMessageList; The first error I'm getting is a syntax error calling out the comma on the line with member_offset PDCUTestMessage id. Then, since that doesn't compile, subsequent error messages snowball for every time I instantiate PDCUTestMessage and every time I add to the multi_index_container. Please help as soon as possible. I'm on a tight schedule, and this is the only major hurdle I see at this point. Thanks! Chris

2 1

[variant] Custom swap() Not Called
by Andreas Grosam 05 May '11

05 May '11

Hi, I'm using a boost::variant whose types have a custom swap() defined. When calling swap() on the variant variable, the custom swap() of the current type will not be called. The "swapable" class has a member function swap(), and there is also a free function swap() defined in the same namespace (see below). To illustrate and test the behavior I defined a special "swapable" class: namespace test { // This class is for testing purpose only! class swapable { public: swapable() : swapped_(0) {} swapable(const swapable& other) : swapped_(0) { } ~swapable() { } int swapped() const { return swapped_; } void swap(swapable& other) { ++swapped_; ++other.swapped_; } swapable& operator= (const swapable& other) { return *this; } private: int swapped_; }; void swap(swapable& lhv, swapable& rhv) { lhv.swap(rhv); } } The calling context looks as follows: void test() { typedef boost::variant< test::swapable > variant_t; variant_t v0( (test::swapable()) ); variant_t v1( (test::swapable()) ); swap(v0, v1); bool swapCalled = boost::get<test::swapable>(v0).swapped() == 1 and boost::get<test::swapable>(v1).swapped() == 1; std::cout << "boost::variant swap test: " << (swapCalled ? "PASSED":"FAILED" ) << std::endl; } So, assuming Koenig lookup works (which in fact does work for the compilers I'm using) calling swap() for the varaint variable should finally call the custom swap(), but it doesn't. Looking deeper into the sources, the culprit seem to be a workaround declaration in file move.hpp (please see below). Note: I'm using gcc, clang and apple's versions of clang. For these compilers __GNUC__ is defined as 4, and in the current config setup BOOST_WORKAROUND(__GNUC__, BOOST_TESTED_AT(2)) evaluates to true. So, when the workaround has been selected, it seems this effectively prevents Koenig lookup which then uses a std::swap like mechanism. Well, now since the compilers I'm using do not have issues with argument-dependent name lookup, how can I fix this - respectively, what is the preferred way to make this work (config setting, maybe)? Note: for testing purposes, I defined BOOST_STRICT_CONFIG 1 (which disables all workarounds) - and this test ran fine. As a side note: I only figured that there is an issue with custom swap() through making performance test - as they ran as slow as a copy. As far as I can see, the boost::variant test suite does not cover such cases. The types which the variant shall hold are copy-expensive - but they are very fast when swap is used (or rvalue reference is enabled some times for boost::variant). Thanks in advance for tips! Regards Andreas ////////////////////////////////////////////////////////////////////////// // function template move_swap // // Swaps using Koenig lookup but falls back to move-swap for primitive // types and on non-conforming compilers. // #if defined(BOOST_NO_ARGUMENT_DEPENDENT_LOOKUP) \ || BOOST_WORKAROUND(__GNUC__, BOOST_TESTED_AT(2)) // [Indicate that move_swap by overload is disabled...] #define BOOST_NO_MOVE_SWAP_BY_OVERLOAD // [...and provide straight swap-by-move implementation:] template <typename T> inline void move_swap(T& lhs, T& rhs) { T tmp( boost::detail::variant::move(lhs) ); lhs = boost::detail::variant::move(rhs); rhs = boost::detail::variant::move(tmp); } #else// !workaround namespace detail { namespace move_swap { template <typename T> inline void swap(T& lhs, T& rhs) { T tmp( boost::detail::variant::move(lhs) ); lhs = boost::detail::variant::move(rhs); rhs = boost::detail::variant::move(tmp); } }} // namespace detail::move_swap

1 0

Need help with compiling boost::function and ublas::matrix
by archie14 05 May '11

05 May '11

I am getting compiler error using boost::function and boost::numeric::ublas as function return value. I am sure I am not declaring something right, but I can't quite catch where (probably the "const" member function specifier). Any help will be greatly appreciated. Code to reproduce the problem is shown below. If #include <vector> #include <boost/function.hpp> #include <boost/numeric/ublas/matrix.hpp> using namespace boost::numeric::ublas; class Foo { public: int bar() const { return 1; } const matrix<double>& stdev() const { return _stdev; } private: matrix<double> _stdev; }; typedef std::vector<Foo> container; class invoker { public: invoker (container* cont, boost::function<int (Foo*) > t) : f(t), _cont(cont) { for (container::iterator it = _cont->begin(); it != _cont->end(); ++it) { Foo& foo = *it; int i = f(&foo); } } // compiler generates error if following code // is uncomennted together with the call in _tmain // invoker (container* cont, // boost::function<const matrix<double>& (Foo* const) > fmatrix) // : _fmatrix(fmatrix), _cont(cont) // { // for (container::iterator it = _cont->begin(); // it != _cont->end(); ++it) // { // Foo& foo = *it; // const matrix<double>& stddev = _fmatrix(&foo); // } // } private: boost::function<int (Foo*) > f; boost::function<const matrix<double>& (Foo* const) > _fmatrix; container* _cont; }; int _tmain(int argc, _TCHAR* argv[]) { container lst (5); std::fill(lst.begin(), lst.begin() + 5, Foo()); invoker invbar(&lst, &Foo::bar); // uncomment this line together with method in "invoker" to generate error // invoker invstdev(lst, &Foo::stdev); return 0; }

1 0

Adjacency List vertex access
by Carlo jardinier 05 May '11

05 May '11

I try to use the adjacency list from boost. I can fill the list with vertices and edges, but somehow I cannot access a specific vertex or edge. For example this is not working (see test.cpp): ** struct EdgeProperties { bool bConnectedBond; EdgeProperties(){bConnectedBond=false;} }; struct VertexProperties{ int index; int spin; bool bOnCluster; VertexProperties(): spin(){ bOnCluster = false;} ~VertexProperties(){ } }; typedef adjacency_list< vecS, vecS, bidirectionalS, VertexProperties, EdgeProperties > Graph; typedef graph_traits< Graph >::vertex_iterator Vit; Graph g(9); unsigned int i = 3; Vit vi; vi = vertex(i,g); ** I want vi to bi an iterator that points on the 3. vertex of the graph g. I attached the error message from g++ when compiling. Can anyone help?

4 7

[Signals] Is connection::disconnect() "const-correct"?
by Greg Hickman 05 May '11

05 May '11

For a connection 'c', if c.connected() == true prior to calling c.disconnect(); after which c.connected() == false, how can connection::disconnect() be considered a "const" member function since it clearly changes the observable state of a connection?

3 3

dereferencing type-punned pointer will break strict-aliasing rules
by Hossein Haeri 05 May '11

05 May '11

Dear all, [Boost 1.46, GCC 4.5.1, MinGW32, WinXP SP3] I use a number of Boost libraries that I see emit the following warning in GCC under release build: "dereferencing type-punned pointer will break strict-aliasing rules" I have traced the code (of SmartPtr and Function specifically) and it's always with an operations like reinterpret_cast<T*>(...)->~T(). Now, I'm wondering how safe this reinterpret_cast is? How serious should I take this warning here? (I do already know how to turn it off or the strict aliasing check.) And, more than anything else: Is this warning suggesting an improvement in my own code? TIA, --Hossein

1 0

Specifying an outgoing interface for IPv6 multicast
by Jay Katti 04 May '11

04 May '11

Folks, Has anyone figured out how to specify the outgoing interface for IPv6 multicast using the boost asio library socket options on Linux ? See snippets below... What value does one specify for "unsigned int ipv6_interface" in the network_interface constructor ? Tried specifying the interface index (1..3 on my host) but that did not work. Thanks, Jay .. >From multicast.hpp * Implements the IPPROTO_IP/IP_MULTICAST_IF socket option. * * @par Examples * Setting the option: * @code * boost::asio::ip::udp::socket socket(io_service); * ... * boost::asio::ip::address_v4 local_interface = * boost::asio::ip::address_v4::from_string("1.2.3.4"); * boost::asio::ip::multicast::outbound_interface option(local_interface); * socket.set_option(option); * @endcode * * @par Concepts: * SettableSocketOption. */ #if defined(GENERATING_DOCUMENTATION) typedef implementation_defined outbound_interface; #else typedef boost::asio::ip::detail::socket_option::network_interface< IPPROTO_IP, IP_MULTICAST_IF, IPPROTO_IPV6, IPV6_MULTICAST_IF> outbound_interface; #endif ========================================================== From: socket_options.hpp ... // Construct with IPv6 interface. explicit network_interface(unsigned int ipv6_interface) { ipv4_value_.s_addr = boost::asio::detail::socket_ops::host_to_network_long( boost::asio::ip::address_v4::any().to_ulong()); ipv6_value_ = ipv6_interface; }

1 0

Re: [Boost-users] [msm] EUML and State construction
by Christophe Henry 04 May '11

04 May '11

> I'm trying to construct some euml states with data at the time of the state > machine construction. The below boils down what I'm trying to do and > fails to compile. Should I expect this mechanism to work with the euml > frontend ? Yes you should but there is a bug, thanks for reporting it. <snip code> > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > With gcc 4.5 and Boost 1.46.1, I get: > > simpletest.cpp: In function 'int main()': > simpletest.cpp:30:75: error: no match for 'operator<<' in > 'boost::msm::back::states_ << test_state_with_constructor(1)' Ok, I see what it is, I did not solve correctly my conflict between proto and serialization. I just committed a fix into the trunk (revision 71719), but it's not working, I simply run into my next proto problem :( I'll continue working on this tomorrow. Thanks, Christophe

1 0

[unordered_map] Trying to understand some errors
by Craig Longman 04 May '11

04 May '11

I'm using VS v9, just in case that matters in this case. I'm : using namespace std; using namespace boost; unordered_map< string, uint32_t > xrf1; unordered_map< string, uint32_t >::const_iterator& iter = xrf1.find( "hello" ); This is something I've used many times with the other containers (well, std::map at least), but here I'm getting a conversion error. This is resolved by either removing the reference marker, or removing the const qualifier. Can someone explain this please? I'd always thought that applying const was easily done, and if so, then a reference should be sufficient, right? Unless the const cast is marked as explicit somehow? And as a side(ish) question, is there any real need to worry about have an object versus reference to something like an iterator? I've always tried to use references where I can, but I wonder if I should "just say no" and use regular objects from now on? Thanks, -- Craig Longman If it's not *signed* by me, then please don't assume it *is* from me.

3 2

[BGL] Extending Floyd Warshall (was: All-pairs SP algorithms -> Determine the path)
by Nicholas Mario Wardhana 04 May '11

04 May '11

On 16 April 2011 02:42, Jeremiah Willcock <jewillco(a)osl.iu.edu> wrote: > Hi Jeremiah, My apology for digging up an old post. I'd like to ask how Boost's Floyd Warshall can extended to provide further functionality. I cannot find any notion of visitor for Floyd Warshall, unlike Dijkstra and A*, and I am a bit confused about the types of the arguments that have to be passed to the function. > One thing you can do in BGL is to compute the paths using the distances; > look for edges where the source and target distances from a particular > starting node differ by the edge's weight, and that will be in the SSSP tree > for that starting node. You can also change the distances to <distance, > predecessor> pairs (just by changing the types being passed into the > algorithm); Will that be a variable of type std::vector<std::vector<std::pair<float, boost::vertex_descriptor>>> ? I tried this and the code failed to compile. > you would still compare just the distances, and the weight of an > edge would be a pair <actual_weight, source>. Will that be a std::pair<float, vertex_descriptor> variable? Additionally, is it also possible to know the number of nodes in a shortest path between two vertices in the graph by using Boost's Floyd-Warshall implementation? > > -- Jeremiah Willcock Thank you. Best regards, Nicholas

2 3