- Boost-users - lists.boost.org

[Signals2] Signal disconnect wrapper won't work
by Marc dela Cruz 07 Jul '09

07 Jul '09

Hopefully this is not another simple oversight on my part but I have code like below and the disconnect wrapper does not work. Thanks. typedef boost::signals2::signal<void ()> VoidSignal; typedef VoidSignal::slot_type VoidSlot; class x{ public: void somefuncA() { cerr << "Bada bing" << endl; } void somefuncB() { cerr << "Bada boom " << endl; } }; void disconnect(VoidSlot const& rF, VoidSignal& rSig) { rSig.disconnect(rF); //won't compile //rSig.disconnect(&rF); //compiles but doesn't disconnect } int main() { x myX; VoidSignal sig; sig.connect(VoidSlot(&x::somefuncB, &myX)); disconnect(VoidSlot(&x::somefuncB, &myX, sig); //doesn't work/compile as expected disconnect(boost::bind(&x::somefuncB, &myX, sig); //doesn't work/compile either sig(); // will still invoke somefuncB return 0; }

4 5

[test] How do run a specific test suite?
by Christian Henning 07 Jul '09

07 Jul '09

Hi there, I believe I once read that one can specify test suite to be run as a command line parameter. By default all test suites are running, but, the user should be able to specify one in the case there are several suites. I've tried to look up that information inside the boost::test documentation but couldn't find it. Am I wrong that such functionality exists? Regards, Christian

5 8

[Asio]How to peocess signal with asio?
by sjdf 07 Jul '09

07 Jul '09

//---------------------------------------------------------------------------- // Hi all, // I have a program like below. // Two questions: // 1. The program can process multi connections, each connection is a // session that has some states. The program should catch signal SIGINT // and SIGTERM, then each session should do different actions according to // their state. But how to do this? Does it have to have a connection // manager? // // 2. When the function Server::handle_accept is called, a SessionController // object will be created. The constructor of SessionController will apply for // some resource, if failed, the constructor would throw an exception. // So, if there is an exception thrown, what should we do? I don't want to // exit the process, may be the resource will be available after some seconds, // then I hope the program continue to accept new connections. //---------------------------------------------------------------------------- class SessionController; typedef shared_ptr<SessionController> SessionController_Ptr; Server::Server(boost::asio::io_service& io_service) : io_service_(io_service), acceptor_(io_service, tcp::endpoint( boost::asio::ip::address::from_string("0.0.0.0"), 10001)) { SessionController_Ptr new_connection(new SessionController(io_service_)); acceptor_.async_accept(new_connection->socket(), boost::bind(&Server::handle_accept, this, new_connection, boost::asio::placeholders::error)); } //---------------------------------------------------------------------------- void Server::handle_accept(SessionController_Ptr new_connection, const boost::system::error_code& error) { if (!error) { new_connection->start(); try { new_connection.reset(new SessionController(io_service_)); acceptor_.async_accept(new_connection->socket(), boost::bind(&Server::handle_accept, this, new_connection, boost::asio::placeholders::error)); } catch (...) { // What should we do? } } }

1 0

[Range.Ex / String.Algo] Ambiguous call to overload function "find"
by 梁绍池 07 Jul '09

07 Jul '09

I use vc 9.0 to compile the following code with boost 1.39.0 and range_ex. #include<string> #include<boost/algorithm/string.hpp> #include<boost/range/algorithm.hpp> int main() { std::string s = "hello"; boost::find_first(s, "lo"); return 0; } The compiler will complain ambiguous call to overload function "find". I found there's also a "find" in boost/range/algorithm.hpp. Even worse when I change the type of "s" to const std::string, the compiler will resolute "find" call as the one in boost/range/algorithm.hpp and reports an error "const_iterator is not a member of std::_String_const_iterator". Now I just indicated the "find" call in boost::find_first explicitly. Is there a better way to solve it? Thanks! S.C. Leung

4 11

[Signals2] Compile error saying slot_type is not a type
by Marc DelaCruz 07 Jul '09

07 Jul '09

Hi all, I'm new to boost and I get the following compile error: ~/FileDesc.h:254: error: `class boost::signals2::signal<void ()(ProcessKit::FileDescBase&, M&), boost::signals2::optional_last_value<typename boost::function_traits<void ()(ProcessKit::FileDescBase&, M&)>::result_type>, int, std::less<int>, boost::function<void ()(ProcessKit::FileDescBase&, M&)>, typename boost::signals2::detail::extended_signature< boost::function_traits<void ()(ProcessKit::FileDescBase&, M&)>::arity, void ()(ProcessKit::FileDescBase&, M&)>::function_type, boost::signals2::mutex>::slot_type' is not a type ~/FileDesc.h:255: error: ISO C++ forbids declaration of `rReadSlot' with no type I have code like the following: template<typename M> class FileDesc : public FileDescBase { ... void setReadCallBack( boost::signals2::signal<void ( FileDescBase&, M&) >::slot_type const& rReadSlot) { ... } ... boost::signals2::signal<void ( FileDescBase&, M&) > mReadSignal; }; I have a feeling it has to do with how/when the templates and the metaprogramming stuff gets "unrolled/expanded" during compile time. However I really don't understand the metaprogramming stuff yet. Thanks.

3 2

[Signals2] Wrapper function for connecting slot won't compile
by Marc DelaCruz 07 Jul '09

07 Jul '09

I have code similar to the following: template<typename A> class SomeClass { void memFunc(FileDescBase&) { ... } }; template<typename B> class OtherClass { void setOpenCallBack(boost::signals2::signal<void (FileDescBase&) >::slot_type const& rOpenSlot) { mOpenSignal.connect(rOpenSlot); } void setCallBacks() { SomeClass myClass; setOpenCallBack(boost::bind(&SomeClass<B>::memFunc, &myClass)); //does not compile setOpenCallBack(boost::signals2::signal<void (FileDescBase&) >::slot_type(&SomeClass<B>::memFunc, &myClass)); //does not compile either } boost::signals2::signal<void (FileDescBase&) > mOpenSignal; }; Even a direct call to mOpenSignal.connect(boost::signals2::signal<void (FileDescBase&) >::slot_type(&SomeClass<B>::memFunc, &myClass)) does not comiple. Thanks.

3 2

Boost ASIO - SO_LINGER Socket option is not working
by Nick Aschberger 07 Jul '09

07 Jul '09

Hi All, I have a situation where at termination of my program, I am closing my socket, and then deleting the socket object. Note, we are using non-asynchronous (blocking) accesses at this point, boost is version 1.36. In some cases, the socket has had some data written to it, but that data is never received by the connected client - because I destroy the socket before the data is sent. So, SO_LINGER is the socket option that is supposed to solve all my worries (it should wait on close for any unsent data to be sent), however I am using it and there is no effect. I presume because I'm doing something wrong. Here is my psuedo-code flow: /* Creation of socket */ m_socket = new boost::asio::ip::tcp::socket(m_io_service); /* Connect the socket with something like: */ acceptor.accept(*m_socket, error); /* Set SO_LINGER option on connected socket */ boost::asio::socket_base::linger linger_op(true, 2000); m_socket->set_option(linger_op); /* program executes, does some stuff */ /* At termination */ // close the socket m_socket->close(); delete(m_socket); I experimented with m_socket->shutdown() instead of close, and it made little difference. In fact, from the documentation it seems shutdown() is the one I definitely don't want to call, as close() is supposed to wait for data to be sent. If anyone has some suggestions as to what would be correct, I would much appreciate it. cheers Nick Aschberger

3 5

STLport Support
by biolaser 06 Jul '09

06 Jul '09

Can anyone point me to a place to obtain support/technical answers (tech forum, etc) for STLport? The STLport forum at SourceForge appears to be largely abandoned. Thanks much! William Stroupe

1 0

[pool] singlton_pool::malloc returning bad address
by Jonathan Leonard 06 Jul '09

06 Jul '09

Hi all, In a long running process on Windows, I'm seeing singleton_pool::malloc eventually return addresses such as 0x100f34b0 (which in this case [according to vmmap] is an address into a 'read-only' section (.rdata) of a DLL which was loaded for 'execute/copy on write'). The system [rightly] barfs at the request to write to that address. Anyone have an idea what could be going on here? Many thanks, Jonathan

2 10

Licensing issues for boost users
by Avi Bahra 06 Jul '09

06 Jul '09

I am trying to understand the boost licensing in terms of users of boost. Is there any where I can find out additional information about this ? For the project I am working on, we will only ever* use* boost, are there an licensing issues/restrictions for *users* of boost ? Am I right is assuming that boost software license only applies to organization that contribute to boost, and not users? i.e I am assuming that users of boost do not need to embed the boost license in our own code that uses the boost libraries ? in essence we are free to license/copyright our own code in any manner, even though it uses boost? Any help in this matter would be greatly appreciated. Avi

3 2