April 2012 - Boost-users - lists.preview.boost.org

map of ranges?
by Sven Bauhan 24 Apr '12

24 Apr '12

Hi, i was thinking of something like a map of ranges, but i did not get the range concept in boost yet. So what i want is something like: typedef map< range<time_duration>, Color> AgeColorMap; Is it possible to use something like that with the boost library? Thanks, Sven

2 2

Please tell me how to add PKG-CONFIG Support in bjam ?
by Baba Prasad 24 Apr '12

24 Apr '12

Dear friends, Please tell me how to set the pkg-config in bjam. please see like this I am using while compiling the using gcc and make command 1) gcc -g - o `pkg-config --cflags glib-2.0 dbus-glib-1 --libs glib-2.0 dbus-glib-1` main main.c but I don't know how to add this option in bajm and I don't know where to add this option in either jamfile or boost-build.jam. because you might get doubt why do you want use glib and dbus means I using dbus api's in my program. If it should be add in boost-build.jam file means it is already have the following boost-build /home/kaka/Documents/seamo/boost_1_49_0/tools/build/v2 ; so how to and where to add is my big confusion , please if any one knows could help me get out of this problem. Best Regards BaBu -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.

1 0

passing compiler directives
by Matt Priest 23 Apr '12

23 Apr '12

Hello, In advance, please excuse this newbie question. I am trying to build boost_1_48_0 and it is going well for all but 4 targets, two of which are the disclaimed <iostreams> failures relating to the bzip targets. However, I'm getting errors when trying to build : boost_1_48_0/libs/filesystem/v3/src/unique_path.cpp Here's the error (there's a nearly duplicate error for the statically linked target): gcc.compile.c++ /users/mpriest/Docs/Boost/build/boost/bin.v2/libs/filesystem/build/gcc-4.7.0/release/threading-multi/v3/src/unique_path.o libs/filesystem/v3/src/unique_path.cpp: In function 'void {anonymous}::system_crypt_random(void*, std::size_t, boost::system::error_code*)': libs/filesystem/v3/src/unique_path.cpp:68:49: error: 'read' was not declared in this scope libs/filesystem/v3/src/unique_path.cpp:71:17: error: 'close' was not declared in this scope libs/filesystem/v3/src/unique_path.cpp:79:13: error: 'close' was not declared in this scope "gcc" -ftemplate-depth-128 -O3 -finline-functions -Wno-inline -Wall -m64 -pthread -fPIC -DBOOST_ALL_NO_LIB=1 -DBOOST_FILESYSTEM_DYN_LINK=1 -DBOOST_SYSTEM_DYN_LINK=1 -DNDEBUG -I"." -c -o "/users/mpriest/Docs/Boost/build/boost/bin.v2/libs/filesystem/build/gcc-4.7.0/release/threading-multi/v3/src/unique_path.o" "libs/filesystem/v3/src/unique_path.cpp" ...failed gcc.compile.c++ /users/mpriest/Docs/Boost/build/boost/bin.v2/libs/filesystem/build/gcc-4.7.0/release/threading-multi/v3/src/unique_path.o... It seems like a simple include file availability issue or a pre-processor define issue, but I can't seem to affect a change in that compiler command through the user-config.jam file edits. When I run b2 with --debug-configuration option, I clearly see that change the -L and -I options in the user-config.jam file is reflected in the output, but still only see the -I"." option as shown above. It seems to me that those functions, read and close, should be made available through including the usual <fcntl.h>, and I think it's a bit curious that no other targets fail from the lack of "read" and "close". Any insight would be appreciated... thank you. Matt

1 0

[serialization] Serializing a byte*
by Sean Farrow 23 Apr '12

23 Apr '12

Hi: I have a function in another library that outputs a byte*. Can I safely just serialize this with boost.serialization, or should I be doing something else to make this work. Any help appreciated. Cheers Sean.

2 1

PKG-CONFIG Support in bjam is how
by Baba Prasad 23 Apr '12

23 Apr '12

Dear friends, Please tell me how to set the pkg-config in bjam. please see like this I am using while compiling the using gcc and make command 1) gcc -g - o `pkg-config --cflags glib-2.0 dbus-glib-1 --libs glib-2.0 dbus-glib-1` main main.c but I don't know how to add this option in bajm and I don't know where to add this option in either jamfile or boost-build.jam. because you might get doubt why do you want use glib and dbus means I using dbus api's in my program. If it should be add in boost-build.jam file means it is already have the following boost-build /home/kaka/Documents/seamo/boost_1_49_0/tools/build/v2 ; so how to and where to add is my big confusion , please if any one knows could help me get out of this problem. Best Regards BaBu -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.

1 0

Subject: [Property Tree] Insert child node at the top of ptree
by johnnyturbo3 23 Apr '12

23 Apr '12

Hello all, I'm trying to find a way to write a child node to the top of a ptree, instead of the bottom as default. I've tried a number of things but for something that should be so simple I've had no luck. The reason I want to write a node to the top of the structure is because when I use ptree::get_child(), it fetches the first node match in the tree i.e. the oldest node written, not the latest written. I've posted the question on stackoverflow.com, it may explain some more detail: http://stackoverflow.com/questions/10272371/add-child-node-to-top-of-xml-st… I hope I've explained my problem clearly. Thanks in advance! Nathan

2 1

Boost.Atomic documentation, Reference Counting
by Václav Zeman 23 Apr '12

23 Apr '12

Hello. Your documentation to Boost.Atomic ([1]) has been given me as an answer to my StackOverflow ([2]) question. Here is a snippet from the documentation: friend void intrusive_ptr_release(const X * x) { if (x->refcount_.fetch_sub(1, boost::memory_order_release) == 1) { boost::atomic_thread_fence(boost::memory_order_acquire); delete x; } } And your rationale: "It is important to enforce any possible access to the object in one thread (through an existing reference) to happen before deleting the object in a different thread. This is achieved by a "release" operation after dropping a reference (any access to the object through this reference must obviously happened before), and an "acquire" operation before deleting the object." I wonder why this would not work well if it looked like this instead: friend void intrusive_ptr_release(const X * x) { if (x->refcount_.fetch_sub(1, boost::memory_order_relaxed) == 1) { boost::atomic_thread_fence(boost::memory_order_acq_rel); delete x; } } [1] <http://www.chaoticmind.net/~hcb/projects/boost.atomic/doc/atomic/usage_exam…> [2] <http://stackoverflow.com/questions/10268737/c11-atomics-and-intrusive-share…> -- VZ

2 2

[multi_index] A multi_index where one index is a subset of the whole collection
by Philipp Moeller 22 Apr '12

22 Apr '12

Hi, I've asked this question [1] two days ago on stackoverflow.com but haven't gotten any answer on it, yet. So I'll try it here. [1] http://stackoverflow.com/questions/10217601/a-multi-index-where-one-index-i… If this is not possible out of the box, can I write my own indices for multi_index that would allow such functionality? Cheers, Philipp Moeller GeometryFactory

2 1

[thread] safe portable static mutex initialization
by Serge Skorokhodov 22 Apr '12

22 Apr '12

Hello, I've read Anthony Williams' articles and documentation but still have no definite answer. Thus two questions: 1. Is it safe to use a block level static boost::mutex on Unix and Windows? 2. Is it safe to use a file/class level static boost::mutex on Unix and Windows? TIA -- The fish that is singing in Ucayali river...

7 12

Re: [Boost-users] boost::filesystem::path operator/= is broken for char[1] in 1.48 (Nat Linden)
by ZNV 21 Apr '12

21 Apr '12

Hi Nat, >> I believe the current operator =/ behavior is outright dangerous and > >> should be fixed. > > Nope, sorry. If you pass a fixed-length array, a proper C++ library > *should* be sensitive to its max. Anything else is unsafe and should > be discouraged. > What if I pass fixed-length array without actually knowing that? There are many C libraries out there assuming char * and char [] are interchangeable. These libraries are well documented. You won't check the source code beforehand if you have the docs, will you? The fact that we're still dealing with legacy C-style APIs is not a > good reason for a modern C++ library to blithely run past the declared > end of a char array. C++ tries very hard to bring type-safety to the > party. It is far more important to try to keep new code from crashing > than it is to accommodate such misleading APIs. > I totally agree with you – type safety is an important C++ feature. It is supposed to help a programmer by eliminating more errors early. But you've achieved just the opposite result with your creative abuse of C++ type system. I would've agreed with your resolution if appending char[1] to boost::filesystem::path resulted in a static_assert. This would allow to apply the proposed workaround immediately (cast to char *). Unfortunately boost::filesystem have chosen to break my code silently instead. I had to hunt the bug in debugger – the same very thing the type safety was supposed to prevent, is not it? So yes technically my program did not crash. But I totally believe that it is not only preventing new code from crashing which is important. Eliminating obscure pitfalls the new code is suffering from is equally important. Every engineering problem is a quest for a better tradeoff. Concerning appending fixed-length array of length one or less to a boost::filesystem::path we are facing the following choices 1) silently treat the operation as a no-op (current) 2) treat it the same way as char * 3) static_assert In my opinion (1) is the worst possible choice since it breaks the code silently in a non-obvious way. When you know you're dealing with code that willfully lies to the > compiler -- even if it's been frozen into an OS API -- use Ovanes's > workaround to pass char* rather than char[size]. > See above.

2 1

Re: [Boost-users] boost::filesystem::path operator/= is broken for char[1] in 1.48 (Ovanes Markarian)
by ZNV 21 Apr '12

21 Apr '12

Hello Ovanes, > This over-allocation is pretty much the same as reserving some space at the > end of the data structure to place data together. Am I right? Why isn't it > possible to just write char[128] or char[16] instead of char[1]? I can't change that code because it lives in a system header. > And if > you know that this buffer represents some special case, why don't you pass > it as such to filesystem? &char[0], than boost::filesystem will not assume > it got char[1] and will be looking to '\0' marker... > How would I know that this buffer represents some special case? FTS(3) library documentation says something like that: struct _ftsent { ... char *fts_name; }; But the structure is really defined another way (Linux, Mac OS X): struct _ftsent { ... char fts_name[1]; }; I believe this discrepancy is pretty much negligible anyway. There are huge number of programs utilizing fts(3) APIs and they don't break. Unfortunately boost::filesystem is another story. It does break and it does break silently.

1 0

boost::filesystem::path operator/= is broken for char[1] in 1.48
by Nick Zavaritsky 21 Apr '12

21 Apr '12

Hi! Please consider the code below: === BUG === boost::filesystem::path some_path("first"); char path_component[1]; /* = "second" */ some_path /= path_component; Some_path value is expected to be "first/second" but it is "first". I.e operator /= had no effect. === END BUG === Concerning a longer string in a char[1] buffer. Actually path_component is the *last* field in a heap-allocated structure. The structure is over-allocated to enable storing a longer string in a char[1] buffer. This technique is often used in C code. Here is operator /= callstack. #0 boost::filesystem3::path_traits::empty<char const, 1ul> () at path_traits.hpp:89 #1 0x0000000100001e7e in boost::filesystem3::path::append<char [1]> (this=0x7fff5fbff910, source=@0x100100270, cvt=@0x100100280) at path.hpp:655 #2 0x0000000100001f77 in boost::filesystem3::path::operator/=<char [1]> (this=0x7fff5fbff910, source=@0x100100270) at path.hpp:235 The relevant excerpt from boost::filesystem3::path_traits::empty follows: namespace path_traits { ... template <typename T, size_t N> inline bool empty(T (&)[N]) { return N <= 1; } ... } It appears that boost::filesystem::path treats char[1] as an empty path regardless of the actual content. This is definitely wrong due to so many C libraries storing long strings in char[1] buffers. In particular on Mac OS X fts(3) FTSENT::fts_name is char[1]. I believe the current operator =/ behavior is outright dangerous and should be fixed. By the way the only code currently benifiting from the boost::filesystem::path being overly smart is the code appending "" (empty string literal) to pathes.

4 3

MSM: Separating templatized on_entry/on_exit into cpp/h
by ev_mp 20 Apr '12

20 Apr '12

Hello, I'm trying to modify the Player example and to separate the on_exit/on_entry template function definition from declaration so I can hide the actual implementation details. "My.h" ... namespace myNamespace { ... // Concrete FSM implementation struct player_state_machine : public msm::front::state_machine_def { ... struct Playing : msm::front::state<> { template void on_entry(event const &, player_state_machine & ); // { "My Secret Code } } ...... } } "My.cpp" namespace myNamespace { ... template void myNamespace::player_state_machine::Playing::on_entry(Event const &, player_state_machine & ); { Here I want to put the actual implementation... } Is this feasible ? If not - what are the alternatives ? As I previously said - putting the implementation into the header is not an option. Any help will be appreciated. Evgeni -- View this message in context: http://boost.2283326.n4.nabble.com/MSM-Separating-templatized-on-entry-on-e… Sent from the Boost - Users mailing list archive at Nabble.com.

2 3

boost mpi and mvapich
by Cornelius Steinhardt 20 Apr '12

20 Apr '12

Hi, have somebody compiled the boost_1.48 with a mvapich compiler? The test from boost mpi get started, which tests the mpi installation works fine. Than i installed the whole library, with adding using mpi : path_to/mpiCC ; into the user-config.jam After this i run ./bootstrap --prefix=where_to_install and ./bjam --prefix=where_to_install install My next step was to test the installation with the first program code from the boost mpi tutorial. ("I am process " << world.rank() << " of " << world.size()) But i get a deadlock in the line mpi::environment env(argc, argv); Can somebody explain me why? or where my failure is? Regards Cornelius

1 0

How to send list from Python to boost python
by Guruswamy B M 20 Apr '12

20 Apr '12

Hi, I have been using boost python for one of my application. In the same I have to send list from python to boost wrapper. Ex:. >From python. a = [1, 2, 3, 4] test(a); In wrapper my C++ fucntion should be like below: void test(list a) { list processing. } I tried similar way, but getting an signature error. I need a help on how to send list from the python, so that can access each element of list in C++ wrapper. Thanks and Regards, Guru

2 2

How to send python list to c++ function
by Guruswamy B M 20 Apr '12

20 Apr '12

Hi, I have been using boost python for one of my application. In the same I have to send list from python to boost wrapper. Ex:. >From python. a = [1, 2, 3, 4] test(a); In wrapper my C++ fucntion should be like below: void test(list a) { list processing. } I tried similar way, but getting an signature error. I need a help on how to send list from the python, so that can access each element of list in C++ wrapper. Thanks and Regards, Guru

2 2

Re: [Boost-users] Timeout on read/write using Boost.Asio
by boost 20 Apr '12

20 Apr '12

So, it implies that it is a bad idea to implement timeout read/write using timer and "cancel". Is there any better solution? Best Regards Zhanwei Wang ----- Original Message -----From: Igor R <boost.lists(a)gmail.com> To: boost-users(a)lists.boost.org > What if I cancel all the async operation, is there any partial data which has been read from or write to socket after the cancelation? > If partial data has been read from or write to socket, how can I issue both async read/write again? Right, some data might have been sent/received, and you have to re-synchronise somehow your client and server. Even if it's possible, it would be pretty complicated. Besides, cancellation is not quite portable: http://www.boost.org/doc/libs/1_49_0/doc/html/boost_asio/reference/basic_st… That's why it's usually a bad idea to cancel operations on a socket. Instead, you can just close the socket and then re-establish connection from scratch. > I cancel the read/write on a tcp connection because the current node I connected to may be too busy to response, so I can try another node. I do not want cancel all async request because the other pending operations may finish successfully, the workload on peer node is very heavy instead of failure. Well, I don't know what you application does, but I'm afraid it's not the right way to go. First of all, why can't you connect all the "nodes" you need? Anyway, you should decide for every particular connection whether it's ok or "faulty" (returned error, takes too much time, no ping, no heartbeat, etc...): if it's ok - wait for response, if it's faulty - just close it. _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

Re: [Boost-users] Timeout on read/write using Boost.Asio
by boost 19 Apr '12

19 Apr '12

hi, Thanks for your reply. I cancel the read/write on a tcp connection because the current node I connected to may be too busy to response, so I can try another node. I do not want cancel all async request because the other pending operations may finish successfully, the workload on peer node is very heavy instead of failure. Anyway, I can cancel all request, but I still have the question about the cancel. What if I cancel all the async operation, is there any partial data which has been read from or write to socket after the cancelation? If partial data has been read from or write to socket, how can I issue both async read/write again? Best Regards Zhanwei Wang ----- Original Message -----From: Igor R <boost.lists(a)gmail.com> To: boost-users(a)lists.boost.org > My question is: > the "cancel" call acutely cancel all the pending async operation, but only one of them is timeout. > For example, there is a pending async read and a pending async write operation, only read is timeout, I do not want cancel write any way. Yes, you're right. You'll have to issue both async read and write again. But usually "timeout" means that something goes wrong with the connection (i.e. you expected something, but it failed to happen). _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

2 1

Questions about using "mcregor_common_subgraphs" algorithm in Boost Graph Library
by Bin Ren 19 Apr '12

19 Apr '12

Hi, I am a beginner of C++ Boost Library. I am doing a project about "Maximum Common Subgraph" Algorithm, and I found there is an implementation in Boost Graph Library: "boost/graph/mcgregor_common_subgraphs.hpp" I use it like this: (and my question is marked as red) ---------------------------------------------------------------------------------------------------- //Compare the input graphs against the Graph Forest for (int i = 0; i < graphInputs.size(); ++i){ graph_t * pInputGraphi = graphInputs[i]; for (int j = 0; j < graphForest.size(); ++j) { std::cout << "Common Sub graph of input graph" << i << " and graph" << j << " :" << std::endl; //Compare one to one among graphi and graphj graph_t * pGraphj = graphForest[j]; //Here I am using the callback function from the example code: "mcgregor_subgraphs_example.cpp" example_callback<graph_t> my_callback(*pInputGraphi); property_map<graph_t, vertex_name_t>::type vname_mapi = get(vertex_name, *pInputGraphi); property_map<graph_t, vertex_name_t>::type vname_mapj = get(vertex_name, *pGraphj); property_map<graph_t, edge_name_t>::type ename_mapi = get(edge_name, *pInputGraphi); property_map<graph_t, edge_name_t>::type ename_mapj = get(edge_name, *pGraphj); //Print out Maximum connected Common Subgraphs between graphi and graphj mcgregor_common_subgraphs_maximum_unique(*pInputGraphi, *pGraphj, true, my_callback, //edges_equivalent(make_property_map_equivalent(ename_mapi, ename_mapj)), //Here is my problem: if I use both edges_equivalent and vertices_equivalent, the code can not be compiled //However, if I only use vertices_equivalent, I doubt whether the maximum common subgraph result is correct. //If I use edges_equivalent only, the code can be compiled, however, it runs into an infinite loop. vertices_equivalent(make_property_map_equivalent(vname_mapi, vname_mapj))); } } ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ Is there any way to use both the edges_equivalent and the vertices_equivalent (Or it is not necessary at all)? BTW, it seems the performance of this algorithm is not perfect; for a small test case below, it needs around 5 seconds by vertices_equivalent, while it loops for a very long time for edges_equivalent (it might be a infinite loop...). Looking forward to your feedback. Thanks and have a good night, Bin ----------------------------------------------------------------------------------------------- Small Test Case: For test, I only use one graph for "graphInputs", and one graph for "graphForest": graphInputs: (v: vertex, 1st number: the vertex id, 2nd number: the vertex name; e: edge, 1st: source vertex, 2nd: target vertex, 3rd: edge name) t # 0 v 0 0 v 1 0 v 2 0 v 3 0 v 4 0 v 5 0 v 6 1 v 7 1 v 8 1 v 9 1 v 10 2 v 11 0 v 12 0 v 13 2 v 14 0 v 15 0 v 16 0 v 17 0 v 18 1 v 19 1 v 20 1 v 21 3 v 22 3 v 23 4 v 24 5 v 25 5 e 0 1 3 e 1 2 3 e 2 3 3 e 3 4 3 e 4 5 3 e 5 0 3 e 0 6 0 e 1 7 0 e 4 8 0 e 5 9 0 e 2 10 0 e 10 11 0 e 11 12 3 e 12 13 0 e 13 3 0 e 11 14 3 e 14 15 3 e 15 16 3 e 16 17 3 e 17 12 3 e 14 18 0 e 15 19 0 e 17 20 0 e 13 21 1 e 10 22 1 e 16 23 0 e 23 24 0 e 23 25 0 graphForest: t # 1 v 0 0 v 1 0 v 2 0 v 3 0 v 4 0 v 5 0 v 6 1 v 7 1 v 8 1 v 9 1 v 10 1 v 11 4 v 12 4 v 13 6 v 14 5 v 15 3 e 0 1 3 e 1 2 3 e 2 3 3 e 3 4 3 e 4 5 3 e 5 0 3 e 0 6 0 e 1 7 0 e 2 8 0 e 3 9 0 e 5 10 0 e 4 11 0 e 11 12 0 e 11 13 0 e 13 14 0 e 12 15 1 If we only use vertex_equivalent, we can get some result like this: This is input Graph 0 <--> 1 5 6 1 <--> 0 2 7 2 <--> 1 3 10 3 <--> 2 4 13 4 <--> 3 5 8 5 <--> 4 0 9 6 <--> 0 7 <--> 1 8 <--> 4 9 <--> 5 10 <--> 2 11 22 11 <--> 10 12 14 12 <--> 11 13 17 13 <--> 12 3 21 14 <--> 11 15 18 15 <--> 14 16 19 16 <--> 15 17 23 17 <--> 16 12 20 18 <--> 14 19 <--> 15 20 <--> 17 21 <--> 13 22 <--> 10 23 <--> 16 24 25 24 <--> 23 25 <--> 23 This is the Graph in Forest: 0 <--> 1 5 6 1 <--> 0 2 7 2 <--> 1 3 8 3 <--> 2 4 9 4 <--> 3 5 11 5 <--> 4 0 10 6 <--> 0 7 <--> 1 8 <--> 2 9 <--> 3 10 <--> 5 11 <--> 4 12 13 12 <--> 11 15 13 <--> 11 14 14 <--> 13 15 <--> 12 Start to do Maximum Common Subgraph Search: Found common subgraph (size 10) 0 <--> 1 5 6 1 <--> 0 2 7 2 <--> 1 3 3 <--> 2 4 4 <--> 3 5 8 5 <--> 4 0 9 6 <--> 0 7 <--> 1 8 <--> 4 9 <--> 5 ....

2 4

[Interprocess] Opposite of find_or_construct()
by Aaron_Wright＠selinc.com 19 Apr '12

19 Apr '12

I feel like I'm missing something simple. I have multiple processes that are going to use an object in shared memory. The find_or_construct() method lets these processes get a pointer to the object without worrying about startup order. What about the opposite? When a process is done with the object in shared memory it can't destroy it because the other processes may be using it. Is there a destroy_or_not() method that would only destroy the object when everyone is done with it? Normally I would solve this with a smart pointer like shared_ptr, but I think I'm miss understanding smart_ptr and I can't get it to work right. It seems like the shared_ptr is created in shared memory and needs to be explicitly destroyed, which would bring me back to the original problem, of who actually gets to destroy it. Should I be looking at intrusive_ptr? --- Aaron Wright

1 0

Calling dijkstra on a filtered graph with bundled properties?
by Berit Dangaard Brouer 18 Apr '12

18 Apr '12

Hi there, First of all thank you for the boost graph library - I really enjoy working with it. However I have become challenged in calling Dijkstra on a filtered graph with bundled properties: My graph structure "mcf_graph" looks like this: #ifndef MCF_GRAPH_HPP #define MCF_GRAPH_HPP #include <boost/config.hpp> #include <boost/graph/adjacency_list.hpp> #include <iostream> using namespace boost; using namespace std; struct Port { Port(int index=-1, int type=0, int port_code=0, string UN="NN", string name="NN", int service_id=-1 ): m_index(index), m_type(type), m_port_code( port_code), UNLOCODE(UN), m_name(name), m_service_id(service_id) {} int m_index; int m_type;//0 port, 1 port call int m_port_code; string UNLOCODE; string m_name; int m_service_id; }; struct Edge { Edge( int id=0, int type=0, double weight=0.0, double dual_cost=0.0, double util=0.0, double cap=0.0, double r_cap=0.0 ) : m_idx(id), m_type(type), m_weight(weight) , m_dual_cost(dual_cost), m_utilization(util), m_capacity(cap), m_res_cap(r_cap) {} unsigned int m_idx; int m_type; double m_weight; double m_dual_cost; double m_utilization; double m_capacity; double m_res_cap; //residual capacity }; template <typename res_cap_map> struct positive_res_cap { positive_res_cap() { } positive_res_cap( res_cap_map cap) : res_cap(cap) { } bool operator()(const Edge& e) const { return 0 < e.m_res_cap; } res_cap_map res_cap; }; //Graph typedef adjacency_list< vecS , vecS , bidirectionalS, Port, Edge> mcf_graph; private: mcf_graph& g; }; #endif I am filtering an mcf_graph like this: //filter positive residual capacity graph positive_res_cap<res_cap_map> filter(get(&Edge::m_res_cap, m_mcf->graph())); filtered_graph<mcf_graph, positive_res_cap<res_cap_map> > fg(m_mcf->graph(), filter); and then calling Dijkstra like this: vertex_descriptor o_vd; std::vector<vertex_descriptor> p ( num_vertices ( fg ) ); std::vector<double> d ( num_vertices ( fg ) ); dijkstra_shortest_paths ( fg, o_vd, predecessor_map ( &p[0] ).distance_map ( &d[0] ).weight_map ( get ( &Edge::m_weight,fg ) ).vertex_index_map ( get ( vertex_index,fg ) ) ); and I get the following compilation error: /usr/include/boost/graph/filtered_graph.hpp:75:63: error: no match for call to '(const positive_res_cap<boost::bundle_property_map<boost::adjacency_list<boost::vecS, boost::vecS, boost::bidirectionalS, Port, Edge>, boost::detail::edge_desc_impl<boost::bidirectional_tag, unsigned int>, Edge, double> >) (const boost::detail::edge_desc_impl<boost::bidirectional_tag, unsigned int>&)' /home/berit/svn/MCF/src/IPH/include/mcf_graph.h:54:8: note: candidate is: /home/berit/svn/MCF/src/IPH/include/mcf_graph.h:57:7: note: bool positive_res_cap<res_cap_map>::operator()(const Edge&) const [with res_cap_map = boost::bundle_property_map<boost::adjacency_list<boost::vecS, boost::vecS, boost::bidirectionalS, Port, Edge>, boost::detail::edge_desc_impl<boost::bidirectional_tag, unsigned int>, Edge, double>] /home/berit/svn/MCF/src/IPH/include/mcf_graph.h:57:7: note: no known conversion for argument 1 from 'const boost::detail::edge_desc_impl<boost::bidirectional_tag, unsigned int>' to 'const Edge&' The above call to Dijkstra works on a non-filtered graph. Can anyone help me out in both understanding why it does not work for the predicate (Is it because the iterators are different???) and how to solve my problem? Thanks in advance ! Berit -- Best Regards, Berit Dangaard Brouer (formerly Løfstedt) PhD. Student DTU Management Engineering, Operations Research Technical University of Denmark Department of Management Engineering Produktionstorvet, Bygning 426 2800 Kgs. Lyngby blof(a)dtu.dk <mailto:blof@man.dtu.dk> www.or.man.dtu.dk <http://www.or.man.dtu.dk>

2 1

Timeout on read/write using Boost.Asio
by boost 18 Apr '12

18 Apr '12

Hi everyone I'm using boost asio to implement a tcp client in my project, and it is my first time of using asio. I need timeout when read/write on the socket and I found the solution of using async io plus a timer, if timeout, cancel the async io operation. My question is: the "cancel" call acutely cancel all the pending async operation, but only one of them is timeout. For example, there is a pending async read and a pending async write operation, only read is timeout, I do not want cancel write any way. Any suggestion of the solution? Or I' m wrong? Thanks

2 1

shared_ptr and container iterators
by Johannes Totz 18 Apr '12

18 Apr '12

Hi all, just started playing with boost::shared_ptr. How do I get a const iterator pointing to a shared_ptr pointing to a const object? See code below, I'd like to prevent a MyItem being changed through const iterators. struct MyItem { int value; }; struct MyContainer { typedef std::set<boost::shared_ptr<MyItem> > MySet; MySet set; }; int main(int argc, char* argv[]) { MyContainer mc; boost::shared_ptr<MyItem> p(new MyItem); p->value = 5; mc.set.insert(p); MyContainer::MySet::const_iterator i = mc.set.begin(); // cannot mutate shared_ptr pointed to by iterator: ok //(*i).reset(new MyItem); // cannot change ptr but can change value: ok (*i)->value = 6; MyContainer::MySet::iterator j = mc.set.begin(); (*j).reset(new MyItem); // how to do this? // should not change ptr nor MyItem std::set<boost::shared_ptr<const MyItem> >::const_iterator k = mc.set.begin(); return 0; }

4 7

Implementing python's enumerate with iterator adaptors
by Jens Auer 18 Apr '12

18 Apr '12

Hi, I am implementing something similar to python's enumerate using boost iterator adapter, but I am running into two problems I am unable to solve. Both problems use my Enumerate-adaptor defined as: #include <utility> #include <boost/iterator/iterator_adaptor.hpp> namespace Iterator { // some simple wrapper to have better names than std::pair template<typename T, typename X> struct Enumerated { Enumerated(T i, X x): count(i), value(x) { } T count; X value; }; template< typename Iterator, typename T = typename Iterator::difference_type> class Enumerator: public boost::iterator_adaptor<Enumerator<Iterator, T>, Iterator, Enumerated<T, typename Iterator::value_type>, boost::use_default, Enumerated<T, typename Iterator::reference> > { public: Enumerator(): mCount(), mPos(), mStep() { } template<typename It> explicit Enumerator(It i, T start=T(), T step=T(1) ): mCount(start), mPos(i), mStep(step) { } private: friend class boost::iterator_core_access; Enumerated<T, typename Iterator::reference > dereference() const { return Enumerated<T, typename Iterator::reference >(mCount, *mPos); } Enumerated<T, typename Iterator::reference > dereference() { return Enumerated<T, typename Iterator::reference >(mCount, mPos); } bool equal(Enumerator<Iterator, T> const& other) const { return mPos == mPos; } void increment() { mCount += mStep; mPos += 1; } void decrement() { mCount -= mStep; mPos -= 1; } void advance(typename Iterator::difference_type step) { mCount += step * mStep; mPos += step; } typename Iterator::difference_type distance_to(Enumerator<Iterator, T> const& other) const { return std::distance(mPos, other.mPos); } T mCount; Iterator mPos; typename Iterator::difference_type mStep; }; template<typename Iterator, typename T> Enumerator<Iterator, T> enumerate(Iterator i, T start=0, T step=1) { return Enumerator<Iterator, T>(i, start, step); } When I compile the following sample program, msvc (VS2010) prints an error, but g++ does not have any problems. #include <vector> #include <iostream> #include <algorithm> void foo() { std::vector<int> v; auto i = Iterator::enumerate(v.begin(), 0, 1); auto j = Iterator::enumerate(v.end(), 0, 1); std::for_each( i, j, [](decltype(*i) e) { std::cout << e.count << ", " << e.value << std::endl; }); } Error 1 error C2665: 'std::_Debug_range2' : none of the 2 overloads could convert all the argument types d:\programme\microsoft visual studio 10.0\vc\include\xutility 728 My second problem can be reproduced with the following sample program on both msvc and g++: #include <vector> #include <iostream> #include <algorithm> void foo() { std::vector<int> v; auto i = Iterator::enumerate(v.begin(), 0, 1); auto y = i->count; auto z = i->value; } MSVC: Error 3 error C2664: 'boost::implicit_cast' : cannot convert parameter 1 from 'Iterator::Enumerated<T,X> *' to 'boost::detail::operator_arrow_proxy<T>' c:\software development\externals\include\boost\iterator\iterator_facade.hpp 326 G++: In file included from /usr/include/boost/iterator/iterator_adaptor.hpp:15:0, from Enumerate.h:18, from Enumerate.cpp:1: /usr/include/boost/iterator/iterator_facade.hpp: In static member function 'static boost::detail::operator_arrow_result<ValueType, Reference, Pointer>::type boost::detail::operator_arrow_result<ValueType, Reference, Pointer>::make(Reference) [with ValueType = Iterator::Enumerated<int, int>, Reference = Iterator::Enumerated<int, int&>, Pointer = Iterator::Enumerated<int, int>*, boost::detail::operator_arrow_result<ValueType, Reference, Pointer>::type = boost::detail::operator_arrow_proxy<Iterator::Enumerated<int, int> >]': /usr/include/boost/iterator/iterator_facade.hpp:648:49: instantiated from 'boost::iterator_facade<I, V, TC, R, D>::pointer boost::iterator_facade<I, V, TC, R, D>::operator->() const [with Derived = Iterator::Enumerator<__gnu_cxx::__normal_iterator<int*, std::vector<int> >, int>, Value = Iterator::Enumerated<int, int>, CategoryOrTraversal = boost::random_access_traversal_tag, Reference = Iterator::Enumerated<int, int&>, Difference = int, boost::iterator_facade<I, V, TC, R, D>::pointer = boost::detail::operator_arrow_proxy<Iterator::Enumerated<int, int> >]' Enumerate.cpp:27:14: instantiated from here /usr/include/boost/iterator/iterator_facade.hpp:327:49: error: no matching function for call to 'implicit_cast(Iterator::Enumerated<int, int&>*)' Any help would be greatly appreciated, Bets regards, Jens

3 7

Stack objects and polymorphic access
by asif saeed 18 Apr '12

18 Apr '12

Hi, I want to maintain stack semantics in my application and avoid heap. For very few requirements, I have to write polymorphic code but I still want to create objects on stack and want to assign their addresses to some abstract base class for polymorphic access. So, all stack variables will be destroyed at the end of their scope and, yet, I'll get polymorphic access. However, functions creating/returning concrete objects have concrete class types. Can boost help me here? Best regards, Asif

1 0