June 2009 - Boost-users - lists.preview.boost.org

Mixing shared_ptr and regular pointers in containers
by Remko Tronçon 12 Jun '09

12 Jun '09

Hi, This feels like a stupid question, but is there an easy/clean way of mixing shared_ptr<A> and regular A* pointers in one STL container? Boost.Variant comes to mind, but I somehow have the idea that there's a more obvious way and that I'm missing it. thanks, Remko

3 3

C++ Framework Libraries for Programming Popular databases
by biolaser 12 Jun '09

12 Jun '09

I am interested in your recommendation of any open source, cross-platform C++ libraries designed for programming the major databases (Oracle, MySql, Postgres, MS Access, Sybase, etc). Boost does not provide any such database/sql library framework for C++. If you can point me toward a source or sources that do offer such a framework, and possibly one(s) the Boost people favor, it would be very much appreciated. Primarily, I am developing code for Windows and Unix/Linux.

9 8

Spirit: Chset base uint32_t
by Etienne Philip Pretorius 11 Jun '09

11 Jun '09

Hello List, I am trying to define the Char definition in the XML specification and I receive this compiler error: hex escape sequence out of range Here is the relvent code: const static boost::spirit::chset<uint32_t> Char( "\x9\xA\xD\x20-\xD7FF\xE000-\xFFFD\x10000-\x10FFFF" ); Can someone please point me in the right direction to get this working. Thank you, Etienne Pretorius

2 1

shared_ptr binary compatibility
by pcayouette 11 Jun '09

11 Jun '09

Are shared_ptr binary compatible through boost versions? Example: Imagine a "Producer" class that sends shared_ptr to a "Consumer" class. Now imagine an executable that instanciate a Producer and a shared library (.dll or .so) that instanciate a Consumer. The executable is built with boost v1.xx. The shared library is built with boost v1.yy. If xx == yy, obviously there will be no problem since the version of the shared_ptr is the same. However if xx != yy, there might be a problem since the version of the shared_ptr might be binary imcompatible. Thanks, Philippe Cayouette -- View this message in context: http://www.nabble.com/shared_ptr-binary-compatibility-tp23985635p23985635.h… Sent from the Boost - Users mailing list archive at Nabble.com.

2 1

Am I doing something wrong here? Boost threads & intrusive_ptr crash
by Space Ship Traveller 11 Jun '09

11 Jun '09

Hi, I'm using an event loop type system which I've implemented myself. This is due to specific needs. Event loop is fairly simple, it has several types of events: - Timer based events (relative or exact, repeating, etc) - IO based events - Notification based events Notifications are events which can come from any thread, and get processed on in the main loop. Here is the function which is used to post a notification. REF(t) is simply a macro for boost::intrusive_ptr<t>. void Loop::postNotification (REF(INotificationSource) note, bool urgent) { using namespace boost; typedef boost::mutex::scoped_lock scoped_lock; // Lock the event loop notification queue // Add note to the end of the queue // Interrupt event loop thread if urgent // We should probably lock around m_currentThread, it might not be set before we read it..? if (this_thread::get_id() == m_currentThread) { note->processEvents(this, NOTIFICATION); } else { { // Enqueue the notification to be processed scoped_lock lock(m_notifications.lock); m_notifications.sources->push(note); (*) } if (urgent) { // Interrupt event loop thread so that it processes notifications more quickly m_urgentNotificationPipe->notifyEventLoop(); } } } I'm experiencing a very infrequent crash at the line marked with (*). The note is a reference counted pointer (boost::intrusive_ptr), but by the time we get here, this pointer is invalid, even though there are still references to it in the system. The odd thing is, further up the chain (and on the same thread), I'm checking to ensure that note is valid, i.e. //... display update system, separate thread from event-loop. ensure(ctx->notificationSource); loop->postNotification(ctx->notificationSource); //... After writing this far in the email, I'm starting to see something where there could be an error. Is it possible that the object reference count is being messed up due to multi-threading issues? I.e. if the object is being passed around on two separate threads, having the reference count incremented and decremented at the same time? Actually, it seems obvious that this could be a problem... Are there any "patterns" for using reference counted objects across different threads? Any feedback is appreciated. Kind regards, Samuel

3 4

troubles building 1.39.0 with intel compiler under linux
by Frédéric Bron 11 Jun '09

11 Jun '09

Hi, I try to upgrade from 1.38 to 1.39 and have to rewrite my building scripts. For gcc I had no particular issue. But for the intel compiler, it does not work: $ ./bootstrap.sh --without-icu --without-libraries=python --prefix=/soft/bal/usr/icc/boost --with-toolset=intel-linux $ ./bjam --layout=system --libdir=/soft/bal/usr/icc/boost/lib link=static threading=multi runtime-link=shared address-model=64 instruction-set=nocona release install However, I obtain the following error from bjam: error: at /soft/bal/usr/builds/icc/boost_1_39_0/tools/build/v2/build/project.jam:874 error: intel-linux initialization: parameter 'version' inconsistent error: an explicit value was specified in an earlier initialization error: no value is specified now Any clue? Frédéric PS: the content of the automatically generated project-config.jam: import option ; import feature ; if ! intel-linux in [ feature.values <toolset> ] { using intel-linux ; } project : default-build <toolset>intel-linux ; libraries = --without-python ; option.set prefix : /soft/bal/usr/icc/boost ; option.set exec-prefix : /soft/bal/usr/icc/boost ; option.set libdir : /lib ; option.set includedir : /soft/bal/usr/icc/boost/include ;

3 3

Problem with throwing exceptions in async handler functions of ASIO library
by J.W.F. Thirion 11 Jun '09

11 Jun '09

Dear Boost community, We're having a problem using the ASIO class with multiple IO service threads and throwing exceptions inside the handle_xxx functions, as used by the async_xxx functions of the ASIO library. It seems that when any error is thrown in a normal/non-handler function, the try/catch block of the main program is able to catch the error. However, when the handler functions are called (I would assume by the io service, now running in a separate thread), the main function's try/catch blocks never receive the execeptions, and a std::terminate() is called automatically. To demonstrate this, we've taken the HTTP Server 2 example from the ASIO examples page and reduced its complexity. Below are the sources of the example. Basically, to duplicate the problem, run the main executable and then in another terminal, simply "telnet 127.0.0.1 60000" to see the std::terminate() being called, even though there are try/catch blocks in the main function (main.cpp). Here is an output: $ ./main terminate called after throwing an instance of 'int' Aborted $ We've also taken the HTTP Server 2 example "as is" and added a "throw 123;" at the top of handle_accept in server.cpp and a "catch (...)" block in posix_main.cpp. The same issue occurred here, which indicates that the exceptions are not propagated through all the way (maybe because the io service runs in a separate thread). Does anyone have a suggestion on how we can throw exceptions inside the handler functions of the async functions used? We'd like to handle all our errors with custom exception classes (e.g. ExceptionAccept is thrown if there is an error detected in handle_accept, which can then be handled inside the main function of main.cpp). We've compiled the sources below by typing: g++ -I/usr/local/boost/include/boost-1_38 -L/usr/local/boost/lib -lboost_thread-gcc41-mt-1_38 -lboost_system-gcc41-mt-1_38 io_service_pool.cpp server.cpp main.cpp -o main where /usr/local/boost is the local boost installation path. Many thanks in advance! Kind regards, Derik Thirion main.cpp -------- #include <iostream> #include <string> #include <boost/asio.hpp> #include <boost/thread.hpp> #include <boost/bind.hpp> #include <boost/lexical_cast.hpp> #include "server.hpp" #include <pthread.h> #include <signal.h> int main(int argc, char* argv[]) { try { // Block all signals for background thread. sigset_t new_mask; sigfillset(&new_mask); sigset_t old_mask; pthread_sigmask(SIG_BLOCK, &new_mask, &old_mask); server s("127.0.0.1", "60000", 2); // 2 threads boost::thread t(boost::bind(&server::run, &s)); // 2 threads // Restore previous signals. pthread_sigmask(SIG_SETMASK, &old_mask, 0); // Wait for signal indicating time to shut down. sigset_t wait_mask; sigemptyset(&wait_mask); sigaddset(&wait_mask, SIGINT); sigaddset(&wait_mask, SIGQUIT); sigaddset(&wait_mask, SIGTERM); pthread_sigmask(SIG_BLOCK, &wait_mask, 0); int sig = 0; sigwait(&wait_mask, &sig); s.stop(); t.join(); } catch (std::exception& e) { std::cerr << "exception: " << e.what() << "\n"; } catch (...) { std::cerr << "all exception" << "\n"; } return 0; } server.hpp ---------- #ifndef SERVER_HPP #define SERVER_HPP #include <boost/asio.hpp> #include <string> #include <vector> #include <boost/noncopyable.hpp> #include <boost/shared_ptr.hpp> #include "io_service_pool.hpp" /// The top-level class of the HTTP server. class server : private boost::noncopyable { public: /// Construct the server to listen on the specified TCP address and port, and /// serve up files from the given directory. explicit server(const std::string& address, const std::string& port, std::size_t io_service_pool_size); /// Run the server's io_service loop. void run(); /// Stop the server. void stop(); private: /// Handle completion of an asynchronous accept operation. void handle_accept(const boost::system::error_code& e); /// The pool of io_service objects used to perform asynchronous operations. io_service_pool io_service_pool_; /// Acceptor used to listen for incoming connections. boost::asio::ip::tcp::acceptor acceptor_; }; #endif // SERVER_HPP server.cpp ---------- #include "server.hpp" #include <boost/bind.hpp> server::server(const std::string& address, const std::string& port, std::size_t io_service_pool_size) : io_service_pool_(io_service_pool_size), acceptor_(io_service_pool_.get_io_service()) { // Open the acceptor with the option to reuse the address (i.e. SO_REUSEADDR). boost::asio::ip::tcp::resolver resolver(acceptor_.io_service()); boost::asio::ip::tcp::resolver::query query(address, port); boost::asio::ip::tcp::endpoint endpoint = *resolver.resolve(query); acceptor_.open(endpoint.protocol()); acceptor_.set_option(boost::asio::ip::tcp::acceptor::reuse_address(true)); acceptor_.bind(endpoint); acceptor_.listen(); acceptor_.async_accept(*(new boost::asio::ip::tcp::socket(io_service_pool_.get_io_service())), boost::bind(&server::handle_accept, this, boost::asio::placeholders::error)); } void server::run() { io_service_pool_.run(); } void server::stop() { io_service_pool_.stop(); } void server::handle_accept(const boost::system::error_code& e) { throw 123; // this one never gets caught in main // instead std::terminate() gets called } io_service_pool.cpp ------------------- #include <stdexcept> #include <boost/thread.hpp> #include <boost/bind.hpp> #include <boost/shared_ptr.hpp> #include "io_service_pool.hpp" io_service_pool::io_service_pool(std::size_t pool_size) : next_io_service_(0) { if (pool_size == 0) throw std::runtime_error("io_service_pool size is 0"); // Give all the io_services work to do so that their run() functions will not // exit until they are explicitly stopped. for (std::size_t i = 0; i < pool_size; ++i) { io_service_ptr io_service(new boost::asio::io_service); work_ptr work(new boost::asio::io_service::work(*io_service)); io_services_.push_back(io_service); work_.push_back(work); } } void io_service_pool::run() { // Create a pool of threads to run all of the io_services. std::vector<boost::shared_ptr<boost::thread> > threads; for (std::size_t i = 0; i < io_services_.size(); ++i) { boost::shared_ptr<boost::thread> thread(new boost::thread( boost::bind(&boost::asio::io_service::run, io_services_[i]))); threads.push_back(thread); } // Wait for all threads in the pool to exit. for (std::size_t i = 0; i < threads.size(); ++i) threads[i]->join(); } void io_service_pool::stop() { // Explicitly stop all io_services. for (std::size_t i = 0; i < io_services_.size(); ++i) io_services_[i]->stop(); } boost::asio::io_service& io_service_pool::get_io_service() { // Use a round-robin scheme to choose the next io_service to use. boost::asio::io_service& io_service = *io_services_[next_io_service_]; ++next_io_service_; if (next_io_service_ == io_services_.size()) next_io_service_ = 0; return io_service; } io_service_pool.hpp: -------------------- #ifndef IO_SERVICE_POOL_HPP #define IO_SERVICE_POOL_HPP #include <boost/asio.hpp> #include <vector> #include <boost/noncopyable.hpp> #include <boost/shared_ptr.hpp> /// A pool of io_service objects. class io_service_pool : private boost::noncopyable { public: /// Construct the io_service pool. explicit io_service_pool(std::size_t pool_size); /// Run all io_service objects in the pool. void run(); /// Stop all io_service objects in the pool. void stop(); /// Get an io_service to use. boost::asio::io_service& get_io_service(); private: typedef boost::shared_ptr<boost::asio::io_service> io_service_ptr; typedef boost::shared_ptr<boost::asio::io_service::work> work_ptr; /// The pool of io_services. std::vector<io_service_ptr> io_services_; /// The work that keeps the io_services running. std::vector<work_ptr> work_; /// The next io_service to use for a connection. std::size_t next_io_service_; }; #endif // IO_SERVICE_POOL_HPP

3 4

boost::test test case levels
by Anatoli Sakhnik 11 Jun '09

11 Jun '09

Hi there! Our automatic tests are of two kinds: short (just like BOOST_CHECK_EQUAL(2*2, 4);), and complicated, which process bulks of random data (for instance, to check if a map with random numbers is always sorted). So we’d like to select which test cases to run: either simple ones or all. I see two ways: 1. Default variants release, debug build and run only simple test cases. We define another feature (or even a build variant) that builds and runs all the tests. This will require additional build efforts, and we’d like to select test cases in run-time. 2. We try to get use of --run_test=??? and test case name mangling. This is a hack, of course. Is there another way to select test cases in run-time? — Anatoli

2 1

[Boost.Thread] Join not waiting?
by James Auger 11 Jun '09

11 Jun '09

This is my first time working with Boost.Thread and I'm having some difficulty in that Join does not seem to work (obviously I'm doing something wrong). In a nutshell what I'm doing is: Instantiating the Example class (which has a boost::thread as a member variable) calling Example::run() (which starts the thread) do some stuff (both in main and in the thread) calling Example::terminate() (which sets the condition for the thread function to stop, and calls Join) end. The problem is that Join returns instantly, before the thread has finished. In the code below I've commented out the stop condition (making the thread run an infinite loop) to demonstrate that it keeps running after join returns. By my understanding of how boost threads and join works, this is not the correct behavior; my understanding is that the call to join should not return until the thread has stopped running. Either my understanding is wrong, or my code to make that happen is wrong. Can you guys help me? The example is stripped down/simplified code that I wrote to demonstrate and test what was going on in my actual program. Example Output: --- thread: main: 00 thread: main: 11 thread: main: 22 thread: main: 33 main: thread: 44 main: After Terminate thread: 5 thread: 6 thread: 7 thread: 8 --- Note the interleaved output 0 - 4, that's expected and fine. The problem is the thread is still running (as evidenced by writing to cout) after terminate() (and thus join()) has been called on that thread. Example Code: --- #include <iostream> #include <boost/thread/thread.hpp> #include <boost/thread/mutex.hpp> #include <boost/bind.hpp> class Example { public: Example() { flag = 0; }; void run() { flag = 1; boost::thread _myThread(boost::bind(&Example::threadLoop, this)); }; void terminate() { //flag = 3; //commented out stop condition _myThread.join(); }; private: void threadLoop() { flag = 2; for(int i = 0; flag == 2; i++) { boost::this_thread::sleep(boost::posix_time::seconds(1)); std::cout<<"thread: "<<i<<std::endl; } flag = 4; }; boost::thread _myThread; int flag; }; int main() { Example anExample; anExample.run(); for(int i = 0; i < 5; i++) { boost::this_thread::sleep(boost::posix_time::seconds(1)); std::cout<<"main: "<<i<<std::endl; } anExample.terminate(); std::cout<<"main: After Terminate"<<std::endl; boost::this_thread::sleep(boost::posix_time::seconds(5)); //sleep so we can see the thread still outputing return 0; } --- Thanks!

3 3

size of serialized ptime object?
by Archie14 11 Jun '09

11 Jun '09

when serializing ptime without using any custom facets, just by using >> or << : how to find the size of the serialized data? Seems to me that sizeof(ptime) should not return the proper value of serialized object size.

1 0