October 2011 - Boost-users - lists.preview.boost.org

Does Boost suit me?
by Nouf M. 12 Oct '11

12 Oct '11

Hello, I'm a phd student. I'm working in the graph mining field. I have a huge text file that I would like to convert to a graph. Then, I will write some code to implement my algorithm (something similar to shortest paths algorithms). I'm planning to write my code in C++. Do you think Boost will suit me? Best regards, Nouf

2 2

Re: [Boost-users] Compiling version 1.43 of boost/graph with SunCC 5.8
by Rob Riggs 12 Oct '11

12 Oct '11

You won't get far with that compiler. Your only hope is SS12 (CC 5.10), but even that's a stretch. The Solaris compiler was not exactly standard compliant. Rob Renaud Kauffmann <renaud(a)ekauffmann.net> wrote: >#include <boost/graph/adjacency_list.hpp> > >using namespace boost; > >typedef adjacency_list <vecS, vecS, undirectedS> BoostGraph; >typedef graph_traits < BoostGraph >::vertex_descriptor vertex_t; > >int main() { > > BoostGraph g; > vertex_t v1 = add_vertex(g); > vertex_t v2 = add_vertex(g); > vertex_t v3 = add_vertex(g); > > return 0; >} > >The above code doesn't seem to compile with Sun Studio 11, that is Sun C++ >compiler 5.8 and boost 1.43 >Here is the error: > >"/home/renaudk/boost/boost_1_43_0/boost/unordered/detail/table.hpp", line >38: Error: Could not find a match for >boost::unordered_detail::hash_table<boost::unordered_detail::T>::find_iterator<boost::unordered_detail::hash_table<boost::unordered_detail::T>::Key, >boost::unordered_detail::hash_table<boost::unordered_detail::T>::Pred>(bucket_ptr, >const boost::unordered_detail::Key&, const boost::unordered_detail::Pred&) >const. >"/home/renaudk/boost/boost_1_43_0/boost/unordered/detail/table.hpp", line >38: Error: Too many args in template, from boost::unordered_detail::Pred on >ignored. >"/home/renaudk/boost/boost_1_43_0/boost/unordered/detail/table.hpp", line >595: Error: Could not find a match for >boost::unordered_detail::hash_table<boost::unordered_detail::T>::find(const >boost::unordered_detail::Key&, const boost::unordered_detail::Hash&, const >boost::unordered_detail::Pred&) const. >"/home/renaudk/boost/boost_1_43_0/boost/unordered/detail/table.hpp", line >595: Error: Too many args in template, from boost::unordered_detail::Hash on >ignored. >4 Error(s) detected. > >I have tried some of the switch I found in the forums such as >-library=stlport4 >-features=tmplife -features=tmplrefstatic > >But no luck. > >Does anybody know the trick ? > >Thanks. > >_______________________________________________ >Boost-users mailing list >Boost-users(a)lists.boost.org >http://lists.boost.org/mailman/listinfo.cgi/boost-users

2 1

Anyone using boost.MPI with latest MPICH2 stuff
by Jonathan Bishop 12 Oct '11

12 Oct '11

Linked with MPICH2 1.4.1p1 but got some run time errors. Is anyone actively doing this? Thanks, Jon

1 0

Re: [Boost-users] [serialization] serialization of a class as pointer and non-pointer
by Daniela Lambertz 11 Oct '11

11 Oct '11

1 0

[serialization] serialization of a class as pointer and non-pointer
by Daniela Lambertz 11 Oct '11

11 Oct '11

3 2

Using boost with explicit template instance
by ranadheer p 11 Oct '11

11 Oct '11

We have built a C++ library using Boost and when we try to integrate the library in a binary we are having problem in linking because of “--instances=explicit” option used for building the binary. The problem is, when we use --instances=explicit option (of Sun CC compiler) it expects us to instantiate the templates explicitly. As Boost internally uses templates when we use Boost with this option the linker is throwing “Undefined symbol” errors - which is expected as we are not instantiating the templates used in Boost explicitly. Though I understand why the linker is throwing the error here, it looks tricky to instantiate all boost templates explicitly in the code. So, want to check if there any solution/workaround to use Boost with “-instances=explicit” option of Sun CC compiler. Any help is much appreciated. We can reproduce the problem with simple sample code as - #include<iostream> #include<boost/lexical_cast.hpp> intmain() { std::cout<<boost::lexical_cast<int>(“8”)<<std::endl; } With “--instances=explicit” option Undefined first referenced symbol in file __type_0 boost::detail::lexical_cast<unsigned,std::string,false,char>(__type_0,__type_3*,unsigned) lexical.o ld:fatal:Symbolreferencing errors.Nooutput written to a.outWithout “--instances=explicit” option this works fine

4 3

ptr_vector<> and vector<shared_ptr<>> compatibility
by Jan Vogelgesang 11 Oct '11

11 Oct '11

Hi, today i had to switch from ptr_vector<MyType> to vector<shared_ptr<MyType> > (needed weak_ptr's to these elements somewhere else). Unfortunately the interfaces of these two containers are quite different when it comes to element access, e.g., ptr_vector<MyType> returns a reference to MyType vs. vector<shared_ptr<MyType> >operator[] returns a reference to a shared_ptr<MyType>. In my case i had a hard time replacing countless occurrences of "BOOST_FOREACH(MyType& e, getMyTypeContainerFromSomewhere())". I'am wondering whats your opinion about a shared_ptr_vector<> that will comply to a ptr_sequence_adapter, using shared_ptr<> to store the elements, and including the possibility to get access to the shared_ptr<> ? Is there any reason why such a container is a bad idea (and thus not already part of the library)? greetings Jan

1 0

[MPL] return type
by Istvan Buki 11 Oct '11

11 Oct '11

Dear Boost experts, I'm trying to find a solution to the following problem. I'm iterating over an MPL sequence and once I found the appropriate type in the sequence I call a static function that returns a pointer to a newly created object that I need to return to the caller of the ModelDispatcher::parse member function with its exact type. In other words I need to propagate the exact type of the object created somewhere in a recursive chain of function calls back to the original caller. Unfortunately I can't figure out how to specify the return type of this recursive function. Every type in the sequence has the following form: struct sequence_item { typedef the_real_type type ; static the_real_type *parse( const QDomElement & ) ; } ; The code for iterating over the sequence is the following: template < typename It, typename End > ?? *dispatchM( const QDomElement &, mpl::true_ /* endofvec */ ) { return 0 ; } template < typename It, typename End > ?? *dispatchM( const QDomElement &element, mpl::false_ ) { typedef typename mpl::deref< It >::type MapEntry ; if ( element.tagName() == MapEntry::first::name() ) { // return type is MapEntry::second::type return MapEntry::second::parse( element ) ; } else { typedef typename mpl::next< It >::type Next ; return dispatchM< Next, End >( element, typename boost::is_same< Next, End >::type() ) ; } } template < typename ModelMap > struct ModelDispatcher { static ?? *parse( const QDomElement &element ) { mpl::false_ f ; return dispatchM< typename mpl::begin< ModelMap >::type, typename mpl::end< ModelMap >::type >( element, f ) ; } } ; Thanks a lot for your help. Istvan

3 6

[Graph] troubles with pointer properties
by Paolo Fogliaroni 11 Oct '11

11 Oct '11

Hello everyone, my first post here, so I hope not to miss some mailing list rule. I'm experiencing some problems in using graph internal properties. I have to generate a graph from a database and I need to differentiate nodes in my graph, namely I will have 2 different node categories, say cat1_node and cat2_node. Each category should have its own property attached. Ex: say N1 is a cat1_node and N2 is a cat2_node. I want N1 to have a "name" (string) property while N2 a "length" (int) property and both have a "label" (string) property. To achieve this I though I could use inheritance in the following manner: //FATHER PROPERTY CLASS class generic_vertex_property{ public: generic_vertex_property(){type = 0;} int type; std::string label; }; //SON PROPERTY CLASS class cat1_property: public generic_vertex_property{ public: std::string name; }; //SON PROPERTY CLASS class cat2_property: public generic_vertex_property{ public: int length; }; In this way, I though, I can define my graph something like: typedef adjacency_list< vecS, vecS, directedS, property<vertex_p_t, generic_vertex_property*> > Graph; and add new vertices like: cat1_property *v1 = new cat1_property(); cat2_property *v2 = new cat2_property(); v1->label = "v1_label"; v1->name = "v1_name"; v2->label = "v2_label"; v2->length = 10; add_vertex(VertexProperty(v1),G); add_vertex(VertexProperty(v2),G); This actually compiles and runs without errors. The problem comes when I try to access the verteces through a property_map, like: vertex_p_map_t vertex_p_map = get(vertex_p, G); //where vertex_p_map_t and vertex_p have been define before Indeed from such get function I have objects of type generic_vertex_property, thus, when trying to access a property of the sone classes I got this compile error: error: ‘class generic_vertex_property’ has no member named ‘name’ I hope I was clear enough, if necessary I can also send the full code. In the meanwhile here are my questions: Is it possible to do what I want to do? Is there a way to get a property map with the "original" object types (as inserted)? Do you have some good websites with practical examples, code snippets and so on about BGL? Last question comes as, in my opinion, the reference manual is quite complete, but 1) lot of examples are not compiling ( look like they have not been updated to the last boost version) 2) examples are usually very basic and it is very hard to get the general rules behind it. hope someone can help me out with this. bests, Paolo

2 1

bind with weak_ptr
by Christopher Pisz 11 Oct '11

11 Oct '11

I see many posts when I google this problem. I wonder, since most of them are dated 1.35 and previous, if there was ever a solution to this problem? I think this fellow says it best when he posted, "My problem is, that when i bind member functions, if i specify the instance of the class as a shared_ptr, then i get memory leaks because the binding's shared_ptr to the instance keeps it alive, and it potentially produces cyclic references that in the current model, would be complicated to break. And of course, if i use raw pointers to the binding instances, then i'll get a crash if the instance has been destroyed and i try to execute the binding function. So what i would need, is to use bind with weak pointers, and just do nothing if locking it fails." http://www.gamedev.net/topic/494089-help-boostbind-and-shared-pointers/ I've almost went to just using raw pointers and letting the program crash if the instance to which the member bound no longer exists. I hate to do that.

2 6