- Boost-users - lists.boost.org

Waiting for a thread on another thread
by Daniel Lidström 05 Mar '10

05 Mar '10

Hello! I have created a simple file logger using the active object pattern. This is intended to offload logging operations to a separate thread, to not disturb the main thread with time- consuming file operations. It works very nicely with boost::thread, but now I have a new desired functionality. The file logger keeps an internal queue of strings that need to be written to file. I want to know, from the main thread, when this queue is empty. To illustrate I have created a minimal, hopefully compilable, sample that shows what I am trying to achieve. #include <boost/thread/thread.hpp> #include <boost/thread/xtime.hpp> #include <boost/thread/condition.hpp> #include <boost/foreach.hpp> #include <boost/tr1/functional.hpp> #include <vector> #include <fstream> #include <string> class FileLogger { typedef boost::unique_lock<boost::mutex> unique_lock; std::ofstream logfile; bool running; boost::thread running_thread; boost::mutex mutex; boost::condition_variable condition; std::vector<std::string> queue; public: FileLogger(const std::string& filename) : logfile(filename.c_str()) , running(true) , running_thread(std::tr1::bind(&FileLogger::Run, this)) { } ~FileLogger() { running = false; condition.notify_one(); running_thread.join(); } void Log(const std::string& s) { unique_lock lock(mutex); queue.push_back(s); } void WaitForQueueFlush() { // ?? } private: void Run() { while( running ) { std::vector<std::string> copyList; { // wait 0.1 s here, for queue to fill up boost::xtime sleepTime; boost::xtime_get(&sleepTime, boost::TIME_UTC); sleepTime.nsec += 100*1000*1000; unique_lock lock(mutex); condition.timed_wait(lock, sleepTime); // fetch the current message list copyList.swap(queue); } BOOST_FOREACH(const std::string& s, copyList) { logfile << s << '\n'; } logfile.flush(); } } }; int main() { FileLogger fileLogger("logfile.txt"); fileLogger.Log(std::string(1000*1000, '*')); // this call should wait, synchronously, until // current queue has been flushed to disk fileLogger.WaitForQueueFlush(); } So the problem is to implement FileLogger::WaitForQueueFlush. How can I signal the main thread when the queue has been emptied? Thanks in advance! Regards, Daniel Lidström Stockholm, Sweden

2 1

[interprocess] Sharing data in a peer-to-peer fashion
by Brett Gmoser 05 Mar '10

05 Mar '10

Hello, I'm working on what seems to be a fairly interesting problem, and I'm looking for any interprocess experts to lend any advice. After reading all of the documentation for Interprocess, it seems that all of the examples work with a "one parent, many children" type model. The problem that I am having is that I am working with a "many children, no parent" model. My processes are spawned by the web server (and FastCGI), so there is no easy way to modify that code to manage the interprocess data. At least one instance of my program is meant to stay alive forever, so I'm not very concerned about deleting the interprocess data at any time. My problem however is the mutexes, and the stale locks that result if one of the processes crashes or exits abnormally. It seems that once an instance of my application crashes, no other instance may ever get a lock on the shared mutex, because the original process still holds a lock. I've thought of a few different ideas to solve this problem: * Using an interprocess shared_ptr for the mutex and data, with a custom deleter to remove all instances once the last application is exiting and the use count is zero. However this suffers from the same problem - it seems the use count is never decremented when the application exits abnormally (for example, kill -9, crash, or CTRL+C). * Using a "heart beat" type system to keep track of processes. My idea was to do something like this: Keep an interprocess associative array of process ID's mapped to last heartbeat time, with each process updating it's own heartbeat. At the same time, every other process (since there is no one parent process) must keep track of the heartbeats, and remove processes from the array which have not responded with a heartbeat in awhile. The problem is this - what about the mutex for the associative array? And the locks that another application might have on them? I wind up back at the point that I started at. The stale locks are destroyed if the mutex is deleted (via boost::interprocess::named_mutex::delete), and I can detect the stale lock pretty reliably with a timed try lock (wait 30 seconds or so to aquire the lock, if it can never be acquired then there is obviously a problem). But what of thread safety, since I now have to delete the lock and re-create it, where other processes might wind up doing the same thing? And what if those other processes are also trying to obtain a lock on the mutex at the same time I'm deleting and recreating it? Those are pretty much the only ideas I've had. Does anybody have anything better? Maybe some of you have tackled a similar problem? Thanks!

1 0

question about a asio example
by pedro chaparro 05 Mar '10

05 Mar '10

Hi, inthe example of a chat server usin asio library, http://www.boost.org/doc/libs/1_42_0/doc/html/boost_asio/example/chat/chat_… there is a fuction socket() , that socket_ value the function returns i use to each new connection with the server right? so if i have n users , i'll have n opened sockets?? i'll aprecciate to get me clear on this thanks tcp::socket& socket() { return socket_; } -- Ing Pedro Alonso Chaparro Valero Ciudad Politecnica de la Innovación iTEAM - Mobile Communications Group Polytechnic University of Valencia C\ Camino de Vera S/N, Edificio 8G 46022 Valencia, Spain

2 1

lambda::bind and abstract class - how to?
by Archie14 05 Mar '10

05 Mar '10

Here is the sample that fails to compile. Complain is about foo being abstract in IA. Any advise of how to make it work will be greatly appreciated. #include <boost/lambda/lambda.hpp> #include <boost/lambda/bind.hpp> #include <boost/ptr_container/ptr_vector.hpp> class IA { public: virtual int foo() const = 0; }; class B : public IA { public: B (int a) : _a(a) {}; int foo() const {return _a;} private: int _a; }; int _tmain(int argc, _TCHAR* argv[]) { boost::ptr_vector<IA> lst; lst.push_back(new B(1)); lst.push_back(new B(2)); boost::ptr_vector<IA>::iterator it = std::find_if (lst.begin(), lst.end(), boost::lambda::bind(&IA::foo, boost::lambda::_1) == 1); return 0; }

2 2

[CRC] Differences in crc_16_type and crc_32_type
by Ryan McConnehey 05 Mar '10

05 Mar '10

Work is wanting to use standard CRC algorithms for both 16 and 32 bit results. This would simplify documentation by not needing to use examples to define the process. The CRC library has predefined four common CRC algorithms. I've included three of the definitions for ease of discussion. Using Wikipedia and the truncated polynomial values, I identified the following algorithms. typedef crc_optimal<16, 0x8005, 0, 0, true, true> crc_16_type; typedef crc_optimal<16, 0x1021, 0xFFFF, 0, false, false> crc_ccitt_type; typedef crc_optimal<32, 0x04C11DB7, 0xFFFFFFFF, 0xFFFFFFFF, true, true> crc_32_type; crc_16_type = crc-16-IBM crc_ccitt_type = crc-16-ccitt crc_32_type = crc-32-ieee 802.3 Now for the questions. How do I know the CRC parameters for each algorithm? For example, if the reflection input for crc_ccitt_type was changed from false to true, is the algorithm result still crc-16-ccitt? Why are the parameters for crc_16_type and crc_32_type different? I expected only the truncated polynomial to change, not the initial remainder and final XOR value. Is it true that crc-32-ieee 802.3 is also called crc-32-ccitt? If that is correct why is the crc_16_type so different? Any clarification to my understanding would be helpful. Ryan

1 0

Boost ASIO issue
by A.Agrawal＠patrick.com.au 05 Mar '10

05 Mar '10

Hello All, Thanks for resolving my ASIO related issues so far, special thanks to Igor R. Now I am having a very strange problem. My program (based on chat server example in ASIO documentation) work fine on a simulator machine and have been tested thoroughly. But when I put it on another machine, it starts giving strange problems some seems to be random but I observed one of them very consistent. std::vector<boost::asio::const_buffer> buffers; buffers.push_back(boost::asio::buffer(_outbound_header)); buffers.push_back(boost::asio::buffer(_outbound_message)); buffers.push_back(boost::asio::buffer(_outbound_data)); boost::asio::async_write(*_socket, buffers, boost::bind(&DJClient::handle_write, this, boost::asio::placeholders::error)); For the above code, handle_write is never called. I can see on the server side that the server has received the request and has also sent the response. But I do not receive the response in the client side. It seems something is going wrong with io_service but don't know what. The same code has worked on a different machine. Any help will be highly appreciated. Regards, Anil Agrawal Patrick Technology & Systems Phone: +61 2 8333-6340

3 2

Boost.Thread and Boost.Function problems?
by anony 05 Mar '10

05 Mar '10

I have this situation: void my_thread(std::string const& a, std::string const& b) { } void spawner() { boost::thread(boost::function<void (std::string const&, std::string const&>(my_thread), "bla", "lala"); } This works, but is it possible that the string temporaries are destroyed before the thread starts running?

2 4

Problem with boost:thread
by Daniel Veneros 04 Mar '10

04 Mar '10

Hey, I'm trying to start a thread from a Class but the only way I did it was using a pointer. The thing is that when I assign the thread to the pointer it wont start unless I use join(). I need it to star right away is created just like when I create a thread (boost::thread myThread(funcName);) I was trying to use thread_group to do this, but create_thread is not working. *Here's an example of what I was doing:* *(Here I used the pointer)* class Thread { public: Thread(){}; virtual ~Thread(){}; virtual void run() = 0; void start(){m_Thread = new boost::thread( &Thread::run, this ); }; //wont start unless I add m_Thread.join(); private: boost::thread* m_Thread; }; *and now... I am trying to use thread_group:* class Thread { public: Thread(){}; virtual ~Thread(){}; virtual void run() = 0; void start(){g_Thread.create_thread(&Thread::run); }; private: boost::thread_group g_Thread; }; *But I receive an error:* /usr/local/include/boost/thread/detail/thread.hpp: In member function 'void boost::detail::thread_data<F>::run() [with F = void (Thread::*)()]': main.cpp:61: instantiated from here /usr/local/include/boost/thread/detail/thread.hpp:56: error: must use '.*' or '->*' to call pointer-to-member function in '((boost::detail::thread_data<void (Thread::*)()>*)this)->boost::detail::thread_data<void (Thread::*)()>::f (...)' make[2]: *** [build/Debug/GNU-MacOSX/main.o] Error 1 Im using netBeans on Mac OS X 10.6.2 Anyone knows how should I create this? or how to use simple boost::thread and launch it without using join()?? Thanks!!! Dann

4 8

[Boost.Multi-index] empty container allocates 1 unused node
by Jakub Jablonski 04 Mar '10

04 Mar '10

Hi, I was trying out some examples of using multi_index_container with custom allocator. I found out that even an empty container, on creation, allocates memory for a single node (but no element is stored inside). In other words, if I store N elements, the container allocates memory for N+1 elements. Is it a fault? If not, what is this memory used for? Is there any way to avoid this allocation? Regards, Jakub

2 1

Compiling Boost 1.42 in OS X
by Andrea Moglia 03 Mar '10

03 Mar '10

Hi, I tried many times to compile Boost 1.42 under OS X. However I always get relase mode. I inserted these commands: ./Bootstrap.sh --prefix=destination path ./bjam install I am looking for a way to compile Boost 1.42 in both debug and release mode. Please, could you tell me how to do that? Thank you. Andrea

3 2