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

[phoenix] lazy function to reference or const reference
by alfC 04 Apr '11

04 Apr '11

Hi, Is it possible to use references as the type of the function wrapped by lazy function. The reason I need this is because I need to instatiate a lazy function inside a class and to a class while the class does not have yet a complete definition. The application may sound exotic but I see no intrinsic limitation why the following doesn't work. If this is not possible because an intrinsic limitation of phoenix, is there a workaround? -- Thanks -- Alfredo ... is_odd_impl is defined as in http://svn.boost.org/svn/boost/sandbox/SOC/2010/phoenix3/libs/phoenix/doc/h… #include<iostream> #include<boost/phoenix/core.hpp> #include<boost/phoenix/function.hpp> using namespace std; using boost::phoenix::arg_names::arg1; int main(){ int cuatro = 4.; is_odd_impl ioi; clog << ioi(cuatro) << endl; boost::phoenix::function<is_odd_impl> fioi; clog << fioi(arg1)(cuatro) << endl; boost::phoenix::function<is_odd_impl&> fioi_ref(ioi); clog << fioi_ref(arg1)(cuatro) << endl; //doesn't compile !!! boost::phoenix::function<is_odd_impl const&> fioi_cref(ioi); clog << fioi_cref(arg1)(cuatro) << endl; //doesn't compile !!! return 0; } example error message: usr/include/boost/phoenix/core/preprocessed/expression_10.hpp:101:17: error: invalid initialization of reference of type ‘const is_odd_impl&’ from expression of type ‘const boost::proto::result_of::make_expr<boost::proto::tag::function, boost::phoenix::default_domain_with_basic_expr, const is_odd_impl, const boost::phoenix::actor<boost::proto::exprns_::expr<boost::proto::tag::terminal, boost::proto::argsns_::term<boost::phoenix::detail::argument<1> >, 0l> >, void, void, void, void, void, void, void, void, void, void>::type’

2 2

[Thread] Strange problem.
by Germán Diago 04 Apr '11

04 Apr '11

Hello. I'm designing a class with a locking policy. I have some code that does exaclty this: if (cond) { locking_policy_.lock(); out << message.message(); locking_policy_.unlock(); } return *this; And the lockingpolicy is a class like this one: struct ThreadSafe { boost::mutex io_mutex_; boost::unique_lock<boost::mutex> lock_; ThreadSafe() : io_mutex_(), lock_(io_mutex_, boost::defer_lock) {} void lock() { std::cout << "Locking from " << boost::this_thread::get_id() << std::endl; lock_.lock(); } void unlock() { std::cout << "Unlocking from " << boost::this_thread::get_id() << std::endl; lock_.unlock(); } }; Then, I create 3 threads and join them in the same order I created them: boost::thread t1(&doSomething, "Hello1", boost::ref(logsink)); boost::thread t2(&doSomething, "Hello2", boost::ref(logsink)); boost::thread t3(&doSomehting, "Hello3", boost::ref(logsink)); t1.join(); t2.join(); t3.join(); and I get a lock_error. Any help here, please? I can't figure out the problem.

3 3

[asio] port question
by Pace, Antony [Harman Pro Group UK] 04 Apr '11

04 Apr '11

I would like to use boost::asio on my unusual platform which supports Berkeley sockets and has an RTOS and I have full control of low level interrupts etc. I have seen socket_types.hpp and feel that it is likely that I could find a set of defines that match my stack. I also have a timer interrupt that I imagine could be fed into the asio timer somehow but wonder if anyone could tell me how much work I am letting myself in for with that approach or if I would be better off using the Berkeley sockets directly? Many thanks Ant [cid:image001.png@01CBF081.1EA6DAD0] Cranborne House, Cranborne Road, Potters Bar, Herts, EN6 3JN Phone: +44 1707 668025 Email: antony.pace(a)harman.com Skype: antony.pace.harman Web: www.harman.com ANTONY PACE Senior Software Engineer Soundcraft Studer R&D Confidentiality Notice: This e-mail message, including any attachments, is for the sole use of the intended recipient(s) and may contain confidential and privileged information. Any unauthorised review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message.

3 5

[boost::mutex] Suspicious EINVAL error
by Panagiotis Foteinos 03 Apr '11

03 Apr '11

Hello users. I am running a single thread which locks and unlocks mutexes, and I get the following error: * /usr/include/boost/thread/pthread/mutex.hpp:55: void boost::mutex::lock(): Assertion `!pthread_mutex_lock(&m)' failed, *after a boost::shared_mutex::lock() or a boost::shared_mutex::try_lock(). In fact, I checked the error code that pthread_mutex_lock(&m) returned and it was the EINVAL error (code number 22). The posix manual gives 2 possibilities for this error: 1) the mutex has not been properly initialized, or 2) The *mutex* was created with the protocol attribute having the value PTHREAD_PRIO_PROTECT and the calling thread's priority is higher than the mutex's current priority ceiling. Sadly, the mutex is correctly initialized. As far as possibility (2) is concerned, well I cannot interpret it... My code is pretty big, so I cannot provide a complete snippet, but roughly (just in case it helps): **************************************************************************************************** I have the following class: class Vertex { boost::shared_mutex mtx; Vertex() {} }; whose objects call Vertex::mtx->lock(), Vertex::mtx->unlock(), Vertex::mtx->lock_shared(), and Vertex::mtx->unlock_shared(). The above error is produced whenever a specific object (residing in memory) locks or even try_locks its mutex. It always happens for a specific object. **************************************************************************************************** I am using the GNU C++ compiler 4.4.3 on Ubuntu x86_64. I do realize that the information I provide in this email is little, but I, being new in boost::mutexes and having spent so much time on it, decided to give it a shot. Best Regards, Panagiotis Foteinos

2 2

Copy-on-write STL containers
by Alexandr Sterligov 03 Apr '11

03 Apr '11

Is there any interest in a library of STL-compilant containers (vector, deque, map, set, list) with copy-on-write? -- Best regards, Alexander Sterligov

2 1

[BGL] Confusion about iterator_property_map and initialization of OffsetMap
by Cedric Laczny 03 Apr '11

03 Apr '11

Hi, I have modified the astar-cities.cpp example to use an external rank_map based on iterator_property_map. The following should give you an idea: vector<cost> v_vec(num_vertices(g), 0.0); typedef property_map< mygraph_t, vertex_index_t>::type VertexIndexMap; VertexIndexMap v_index = get(vertex_index, g); // Initialization of interest // Create the external property map typedef iterator_property_map< std::vector< cost >::iterator, VertexIndexMap > CostMap; CostMap c_map(v_vec.begin(), v_index); and the function-call of astar_search is as follows: astar_search (g, start, distance_heuristic<mygraph_t, cost, location*, const char**> (locations, goal, name), predecessor_map(&p[0]).distance_map(&d[0]). visitor(a_star_vis). rank_map(c_map)); // Here we use the "global" cost map Basically, the code is the same as in the example, with only the addition of the explicit rank_map. Now I am fairly confused that the function call seems to behave the same if I have "get(vertex_index, g)" and if I don't have it. This comes unexpected to me as I would think that it needs OffsetMap (v_index) to correctly find the index in the random access container (v_vec). But when I simply declare an object it is not yet initialized and will not contain any meaningfull offsets (not using "get(vertex_index, g)"). So why does it work, when the map is not initialized? Or what am I missing here please? Best, Cedric

2 7

[MPL] no invoke<T>?
by Noah Roberts 03 Apr '11

03 Apr '11

I had to write an invoke metafunction because I couldn't seem to find one that looked like it was what I needed: template < typename MF > struct invoke { typedef typename MF::type type; }; I've looked through the documentation TOC: http://www.boost.org/doc/libs/1_42_0/libs/mpl/doc/refmanual/refmanual_toc.h… Is there something I'm missing?

4 8

questions regarding boost::mem_fn and boost::bind
by Littlefield, Tyler 02 Apr '11

02 Apr '11

Hello all: I have a quick question. I was trying to learn how to use boost::mem_fn to work as a sort of function object, because I need something like this for my library. So: If I understand this correctly, it creates a function object that has () overloaded, correct? I was curious if someone could then help me figure this out, the documentation is confusing for me. lets say I have this class: class Foo { public: int Add(int a, int b) { return (a+b); } }; and I could do Foo bar; how would I create a mem_fn named f, that would let me call it for add? so: f(3,1). I would have to tie the instance of foo (bar) to this somehow, correct? Next, I see that mem_fn creates a function object which basically does what I want. What then is the point in boost::bind? It looks to do about the same as mem_fn, just makes my head hurt a little bit more. I am trying to learn boost by finding uses for some of the libraries. Some of them sort of makes sense, but I see that mem_fn are the bases for a lot, so I wanted to learn how to use those. After bind and mem_fn, what other libraries would it be best to learn that is used all over? Last, this is the use I will be putting mem_fn to. If this is incorrect, I still want to learn, but it looked like a great example. I am writing a quick socket class that will have connect, close, and send capabilities. I am hooking this to my game engine, and wanted to use a mem_fn to receive the data. So the server polls all user input, then it will poll the sockets that were added to another list. It will pass off any data read by calling the function object with the data as an argument, as well as the socket. Sorry for all the questions, and thanks in advance for the help/clarification. -- Thanks, Ty

2 1

VS2010+boost 1.44 - error in move.hpp
by archie14 02 Apr '11

02 Apr '11

While compiling the project in VS2010 previously built without errors in VS2008 I am getting following error: c:\libraries\boost\boost_1_44_0\boost\interprocess\detail\move.hpp(342): error C2440: 'return' : cannot convert from 'boost::interprocess::file_mapping' to 'boost::interprocess::file_mapping &&' You cannot bind an lvalue to an rvalue reference c:\libraries\boost\boost_1_44_0\boost\interprocess\file_mapping.hpp (62) : see reference to function template instantiation 'boost::interprocess::file_mapping &&boost::interprocess::move<boost::interprocess::file_mapping&>(T)' being compiled with [ T=boost::interprocess::file_mapping & ] I tried to search the forum for the reports on this error and did not find anything. I'll appreciate any help in this matter.

2 3

[boost] [program options] Potential bug in parsing arguments where one is suffix of the other ?
by Sanjit Jhala 02 Apr '11

02 Apr '11

Hi, I have the following code (modified from example/first.cpp): *#include <boost/program_options.hpp>* *#include <string>* *#include <vector>* *#include <iostream>* * * *namespace po = boost::program_options;* * * *using namespace std;* * * *typedef vector<string> Strings;* * * *// A helper function to simplify the main part.* *template<class T>* *ostream& operator<<(ostream& os, const vector<T>& v)* *{* * copy(v.begin(), v.end(), ostream_iterator<T>(cout, " "));* * return os; * *}* * * * * *int main(int argc, char *argv[]) {* * * * po::options_description cmdline_desc("Usage: %s [Options] [args]\nOptions");* * cmdline_desc.add_options()* * ("str", po::value< string >(), "a string arg")* * ("strs", po::value< Strings >(), "a list of strings")* * ("foo", po::value< string >(), "a string arg")* * ("bar", po::value< Strings >(), "a list of strings")* * ; * * * * po::variables_map vm; * * po::store(po::parse_command_line(argc, argv, cmdline_desc), vm);* * po::notify(vm);* * * * if (vm.count("str"))* * cout << "Got str arg=" << vm["str"].as< string >() << "\n";* * if (vm.count("strs"))* * cout << "Got strs arg=" << vm["strs"].as< Strings >() << "\n";* * if (vm.count("foo"))* * cout << "Got foo arg=" << vm["foo"].as< string >() << "\n";* * if (vm.count("bar"))* * cout << "Got bar arg=" << vm["bar"].as< Strings >() << "\n";* *}* * * When I run the program I see: *./po_test --str "a string" --strs "a, vector, of, strings" --foo "a foo" --bar "a, bar"* *Got strs arg=a string a, vector, of, strings * *Got foo arg=a foo* *Got bar arg=a, bar * It appears the parser is getting confused between the "str" and the "strs". It doesn't seem to be related to ambiguity in handling vectors since the "foo" and "bar" arguments are fine. I see this problem on 32-bit CentOS and Ubuntu machines, but not on 64-bit. I am running Boost 1.43 and from the /usr/local/include/boost/program_options/version.hpp: #define BOOST_PROGRAM_OPTIONS_VERSION_HPP_VP_2004_04_05 #define BOOST_PROGRAM_OPTIONS_VERSION 2 Is this a bug or a expected behaviour? -Sanjit

2 2