June 2013 - Boost-users - lists.preview.boost.org

Asio and Spirit
by LeMay.Steve 30 Jun '13

30 Jun '13

Can anyone point me to a modern example or description of using Asio with Spirit to parse and asynchronous stream. I keep finding articles that say this can't be done, or at least not easily (but the comments date back to 2007). SGL

3 3

Herb Sutter on C++ Standardization and Visual Studio
by LeMay.Steve 30 Jun '13

30 Jun '13

Herb Sutter on C++ Standardization at Microsoft. http://channel9.msdn.com/Events/Build/2013/2-306 SGL

3 2

future<>.then and promises
by Tobias Furuholm 30 Jun '13

30 Jun '13

I am trying out future<>.then and I am seeing some results that confuses me. I have tried to illustrate this with a few examples. The first example is using a promise together with future<>.then. It generates no output: boost::promise<int> promise; boost::future<int> fut = promise.get_future(); promise.set_value(123); fut.then([](boost::future<int>& f) { std::cout << "The value is: " << f.get() << std::endl; return f.get(); }); boost::chrono::milliseconds duration(1000); boost::this_thread::sleep_for(duration); Why is fut.then not triggered? (assigning fut.then to another future makes no difference) By adding a launch policy I get the output "The value is: 123" which is the output I would expect from all my examples: boost::promise<int> promise; boost::future<int> fut = promise.get_future(); promise.set_value(123); fut.then(boost::launch::async, [](boost::future<int>& f) { std::cout << "The value is: " << f.get() << std::endl; return f.get(); }); boost::chrono::milliseconds duration(1000); boost::this_thread::sleep_for(duration); To make this more real, let's move promise.set_value to after .then boost::promise<int> promise; boost::future<int> fut = promise.get_future(); fut.then(boost::launch::async, [](boost::future<int>& f) { std::cout << "The value is: " << f.get() << std::endl; return f.get(); }); promise.set_value(123); boost::chrono::milliseconds duration(1000); boost::this_thread::sleep_for(duration); This program outputs nothing and never terminates. But when I assign the result of fut.then to another future I get the expected output once again: boost::promise<int> promise; boost::future<int> fut = promise.get_future(); boost::future<int> fut2 = fut.then(boost::launch::async, [](boost::future<int>& f) { std::cout << "The value is: " << f.get() << std::endl; return f.get(); }); promise.set_value(123); boost::chrono::milliseconds duration(1000); boost::this_thread::sleep_for(duration); Finally, as I bonus I add an example where I use async to generate a future instead. boost::future<int> fut = boost::async([]() { return 123; }); fut.then([](boost::future<int>& f) { std::cout << "The value is: " << f.get() << std::endl; return f.get(); }); boost::chrono::milliseconds duration(1000); boost::this_thread::sleep_for(duration); This generates the output "The value is: " (without 123). Once again, by assigning the result of fut.then to another future I get the expected output. I use boost 1.54 beta 1 with clang 3.2 on Linux and. /Tobias

3 7

Re: [Boost-users] Accessing memory in shared segments
by Andy Falanga (afalanga) 28 Jun '13

28 Jun '13

>From: Andy Falanga (afalanga) >Sent: Thursday, June 27, 2013 7:42 AM >To: 'boost-users(a)lists.boost.org' >Subject: Accessing memory in shared segments > >Hi, > >I have some code which shares data with shared memory. To that end, I'm using >boost::interprocess:managed_shared_memory. Without respect to scope and functions, what I have is >roughly: > >boost::interprocess::managed_shared_memory sharedMem(boost::interprocess::create_only, > "SharedMemory", 4096); > >//later on >u_int8_t* dataSegment = sharedMemory.construct<u_int8_t>("data")[5](0); >char* string1 = sharedMemory.construct<char>("String 1")[50]('\0'); >// other stuff > >// placing stuff in the data segment >for(int i(0); i < 5; i++) dataSegment[i] = anEarlierArray[i]; > > >// then, in other process read the data Segment >char ary[5]; >for(int i(0); i< 5; i++) ary[i] = dataSegment.first[i]; > > >For whatever reason I couldn't simply access through dataSegment[n] notation. The weirdest thing is, >the data is being read back to me in the reverse order that it should be. For example, if >[0x2,0x4,0x7,0x1] is written I read it back as [0x1,0x7,0x4,x02] even though I'm writing and reading >in the same order. Why is this? What is it I'm missing about using shared memory? Problem solved: it was between the keyboard and the chair. Further debugging showed that the data wasn't being written in reverse (it made no sense that it would have been) but rather was a result of some HID libraries I'm using under the hood. Posting this in case someone hits on this message while looking through the archives. Andy

1 0

[beta] 1.54 for ARM
by Michael Powell 28 Jun '13

28 Jun '13

Hello, I haven't actually built 1.54 for ARM yet, although I have successfully built and am using 1.53 for ARM today. Is it that much different from 1.53? And/or has someone already built it for ARM to give a little beta feedback? If I can manufacture some time this weekend I may give that a go regardless. Thank you... Regards, Michael Powell

1 0

Re: [Boost-users] [asio] How to deal with no-connect and/or read-timeout
by Marat Abrarov 28 Jun '13

28 Jun '13

> If the timer handler is called first, then you're done: initiate > disconnection, etc. > If the read handler is called first, you simply reset the timer (maybe > cancel and then reset). > I honestly don't know if you have to worry about a race condition (timer > going off while handler for async_read is executing), or if a cancel / > reschedule out of the async_read handler would preempt any other handlers. > You might end up having to stick both the async_read and deadline_timer in > the same strand; that would prohibit concurrent execution. There is a good example: http://www.boost.org/doc/libs/1_53_0/doc/html/boost_asio/example/timeouts/se rver.cpp and a good article: http://blog.think-async.com/2010/04/timeouts-by-analogy.html and a rather heavy example: http://sourceforge.net/projects/asio-samples/ (see project named echo_server). In general, you will need a state machine (or even few of them). Regards, Marat Abrarov.

6 18

Re: [Boost-users] Mutex synchronised wrapper class for boost::interprocess vector of strings
by Riskybiz 28 Jun '13

28 Jun '13

"It's not an easy task. You should catch the exception and then notify all connected processes that should disconnected from the segmenet (as only offline growing is supported). After that a process can call "grow" to try extend the segment. Then, all processes can attach to the newly grown segment. Think of that as a complex binary file that must be extended and internal structures changed. All users should close the file, the file should be changed by a process and then all users can start using the file again." OK Ion, I see. 1. Can I ask what mechanism you would suggest for notifying the other processes to disconnect before the grow() operation. 2. My wrapper class implementation is still in the experimental stages, I'm trying to build it up to operate in the way I need. I've hacked it around a bit to account for the fact that I cannot control the order in which the client & server instances will ultimately be launched by the 3rd party application I'm building a DLL for; so I've got to make the wrapper class as flexible as possible. What I'm trying to do is get my toy client process to wait, if it should happen to launch first, before the server, and then have the clients notified by the server process when data is ready to be accessed. So....I can get the client to wait using a named_condition, but then I can't get it to wake up. The notifyall() is not taking effect. P.S. In my tests I first launch the client and then the server process. Have I missed something fundamental here, is my notifyall() properly constructed and directed to the right destination? 3. Also, another thought; if the shared memory segment should happen to grow whilst a client is waiting then the client will need to remap the segment on waking, correct? But will it even get the notifyall() signal, because the client is still 'looking' at a different memory location? Could this be my problem? Not sure. Though I did test the code without the grow() commands and the client process still did not wake! Many, many thanks for your help so far, just a few more hurdles to clear and this wrapper class will be operational. Riskybiz. #ifndef SHARED_MEMORY_WRAPPER_H//if not defined already #define SHARED_MEMORY_WRAPPER_H//then define it #include <boost/interprocess/managed_shared_memory.hpp> #include <boost/interprocess/containers/vector.hpp> #include <boost/interprocess/containers/string.hpp> #include <boost/interprocess/allocators/allocator.hpp> #include <boost/interprocess/sync/named_mutex.hpp> #include <boost/interprocess/sync/named_condition.hpp> #include <iostream> #include <stdexcept> using namespace boost::interprocess; class SharedMemoryWrapper { public: SharedMemoryWrapper(const std::string &name, bool server) : m_name(name), m_server(server) { try { //Open the shared segment. Will throw an interprocess exception if fails to open i.e. segment does not exist. Control jumps to the catch block. m_segment = new managed_shared_memory(open_only, m_name.c_str()); //Find the vector using the c-string name and open it sharedSegmentVector = m_segment->find<MyShmStringVector>("sharedSegmentVector").first; //Open the underlying mutex m_mutex = new named_mutex(open_only, "smw_mutex"); //Open the underlying condition variable waitForData = new named_condition(open_only, "smw_named_condition"); if(m_server)//Flag to put client processes to sleep until ready for them { sleep = false; } else { sleep = true; } } catch(interprocess_exception &ipe)//Catch the interprocess exception and instantiate the shared memory resources here. { try { std::cout << "Error: interprocess_exception: " << ipe.what() << "Going to instantiate shared memory resources now...." << std::endl; //Clean up resources from any prior instances named_mutex::remove("smw_mutex"); named_condition::remove("smw_named_condition"); shared_memory_object::remove(m_name.c_str()); //Create shared memory m_segment = new managed_shared_memory(create_only, m_name.c_str(), 100000); //Create allocators CharAllocator charallocator(m_segment->get_segment_manager()); StringAllocator stringallocator(m_segment->get_segment_manager()); //This vector is fully constructed in shared memory. All pointers //buffers are constructed in the same shared memory segment //This vector can be safely accessed from other processes. sharedSegmentVector = m_segment->construct<MyShmStringVector>("sharedSegmentVector")(stringallocat or); if(m_server)//Flag to put client processes to sleep until ready for them { sleep = false; } else { sleep = true; } //Instantiate the underlying mutex m_mutex = new named_mutex(create_only, "smw_mutex"); //Instantiate the underlying condition variable waitForData = new named_condition(create_only, "smw_named_condition"); //......despite the exception code continues executing to the end of the constructor. } catch(interprocess_exception &ipe1) { std::cout << "Error: Whilst instantiating shared memory resources: interprocess_exception: " << ipe1.what() << std::endl; } } catch(...) { std::cout << "Error: Unknown Exception: " << std::endl; } } ~SharedMemoryWrapper() { /* //Destructor: Will remove the shared memory resources if (m_server) { named_mutex::remove("named_mutex"); //Destroy the vector from the managed_shared_memory //m_segment->destroy_ptr(sharedSegmentVector);//This will free all strings that the vector contains m_segment->destroy<MyShmStringVector>("sharedSegmentVector");//Destroy the vector //Is it necessary to destroy the allocators here? //Remove the managed_shared_memory; this may fail if the memory does not exist or is mapped or opened by another process shared_memory_object::remove(m_name.c_str()); } */ //Delete instance members delete m_mutex; delete m_segment; delete waitForData; } //Writer void write(const std::string &inStr) { try { CharAllocator charallocator(m_segment->get_segment_manager()); scoped_lock<named_mutex> lock(*m_mutex); try { if(sleep)//Put client process to sleep when it first tries to access. { std::cout << "Going to wait...." << std::endl; sleep = false; waitForData->wait(lock);//Set condition variable to wait. When ready to allow access then call notify(). } } catch(interprocess_exception &ex) { std::cout << "Error in Writer Process Wait: interprocess_exception: "<< ex.what() << std::endl; } MyShmString shmStr(inStr.c_str(), charallocator); sharedSegmentVector->push_back(shmStr); //Try to grow if free memory is running low if(m_segment->get_free_memory() < 50000) { std::cout << "Free Memory Before: " << m_segment->get_free_memory() << " Elements: " << getSize() << std::endl; delete m_mutex;//delete delete waitForData;//delete delete m_segment;//unmap managed_shared_memory::grow(m_name.c_str(), 50000);//grow m_segment = new managed_shared_memory(open_only, m_name.c_str());//remap sharedSegmentVector = m_segment->find<MyShmStringVector>("sharedSegmentVector").first;//find //Reopen the underlying mutex m_mutex = new named_mutex(open_only, "smw_mutex"); //Reopen the underlying condition variable waitForData = new named_condition(open_only, "smw_named_condition"); std::cout << "Free Memory After Grow: " << m_segment->get_free_memory() << " Elements: " << getSize() << std::endl; } } catch(interprocess_exception ipe) { std::cout << "Error in SharedMemoryWrapper::write(): interprocess_exception: "<< ipe.what() << std::endl; } catch(...) { std::cout << "Error in SharedMemoryWrapper::write(): unknown error" << std::endl; } } //Reader std::string read(size_t &index) { try { scoped_lock<named_mutex> lock(*m_mutex); try { if(sleep)//Put client process to sleep when it first tries to access. { std::cout << "Going to wait...." << std::endl; sleep = false; waitForData->wait(lock);//Set condition variable to wait. When ready to allow access then call notify(). } } catch(interprocess_exception &ex) { std::cout << "Error in Reader Process Wait: interprocess_exception: "<< ex.what() << std::endl; } const MyShmString &shmStr = (*sharedSegmentVector)[index]; std::string stdStr(shmStr.begin(), shmStr.end()); return stdStr; } catch(interprocess_exception &ipe) { std::cout << "Error in SharedMemoryWrapper::read(): interprocess_exception: "<< ipe.what() << std::endl; } catch(std::range_error &re) { std::cout << "Error in SharedMemoryWrapper::read(): range_error: " << re.what() << std::endl; } catch(...) { std::cout << "Error in SharedMemoryWrapper::read(): unknown error "<< std::endl; } return "Error in SharedMemoryWrapper::read()"; } //Notifier void notify() { try { scoped_lock<named_mutex> lock(*m_mutex); std::cout << "Going to Notify...." << std::endl; lock.unlock(); waitForData->notify_all();//notify to release the wait condition; when data is available and ready to be accessed by client processes std::cout << "Notified" << std::endl; } catch(interprocess_exception &ipe) { std::cout << "Error: boost_interprocess exception: in notifier: " << ipe.what() << std::endl; } catch(...) { std::cout << "Unknown Error In Notifier: " << std::endl; } } //Vector Size size_t getSize() const { return this->sharedSegmentVector->size(); } //Remover void release() { try { std::cout << "Going to release shared memory resources...." << std::endl; named_mutex::remove("named_mutex"); named_condition::remove("smw_named_condition"); //Destroy the vector from the managed_shared_memory //m_segment->destroy_ptr(sharedSegmentVector);//This will free all strings that the vector contains m_segment->destroy<MyShmStringVector>("sharedSegmentVector");//Destroy the vector //Is it necessary to destroy the allocators here? //Remove the managed_shared_memory; this may fail if the memory does not exist or is mapped or opened by another process shared_memory_object::remove(m_name.c_str()); std::cout << "Resources Released" << std::endl; } catch(interprocess_exception ipe2) { std::cout << "Error: remove() interprocess_exception: " << ipe2.what() << std::endl; } } private: //Typedefs typedef boost::interprocess::allocator<char, managed_shared_memory::segment_manager> CharAllocator; typedef boost::interprocess::basic_string<char, std::char_traits<char>, CharAllocator> MyShmString; typedef boost::interprocess::allocator<MyShmString, managed_shared_memory::segment_manager> StringAllocator; typedef boost::interprocess::vector<MyShmString, StringAllocator> MyShmStringVector; bool m_server; std::string m_name; bool sleep; managed_shared_memory *m_segment; MyShmStringVector *sharedSegmentVector; named_mutex *m_mutex; named_condition *waitForData; };//class #endif//header guard Server program to prepare the shared vector of strings: // PrepareSharedMemoryVectorOfStrings.cpp : Defines the entry point for the console application. // #include "SharedMemoryWrapper.h" #include <iostream> #include <string> #include <sstream> int main(int argc, char *argv[]) { SharedMemoryWrapper *smw = new SharedMemoryWrapper("SMT", true); size_t cnt(0); while(cnt < 100) { //Convert number to string representation std::string cntStr = static_cast<std::ostringstream*>( &(std::ostringstream() << cnt) )->str(); //Write to shared memory vector smw->write(cntStr); cnt++; } size_t vecSize = smw->getSize(); std::cout << "Elements Present: "<< vecSize << std::endl; smw->notify(); //smw->release(); delete smw; return 0; } Client program to access the shared memory vector of strings: // AccessSharedMemoryVectorOfStrings.cpp : Defines the entry point for the console application. // #include "SharedMemoryWrapper.h" #include <iostream> #include <string> int main(int argc, char *argv[]) { SharedMemoryWrapper *smw = new SharedMemoryWrapper("SMT", false); try { std::string outStr("Blank"); //if(smw->getSize() == 100) //{ outStr = ""; for(size_t index = 0; index < 100; index++) { outStr += (smw->read(index) + " "); } //} std::cout << outStr << std::endl; } catch(...) { std::cout << "Error: Unknown Exception." << std::endl; } smw->release();//removes shared memory resources from the system delete smw; return 0; }

1 0

global instances for managed_shared_memory & message_queue
by bb.andersen 27 Jun '13

27 Jun '13

Dear Boost Grurus, First of all let me thank you for the huge effort that people like Mr. Gaztañaga has made to keep boost interprocess improved. It is a really time-saver API! I have a very basic question that, despite the large amount of documentation I've looked, I haven't clarified yet. I just want to create a global static variables (i.e. instances) for managed_shared_memory, message_queue, mutex and condition var. By now I have my app prototype (CGI) is running fine with auto vars, as appeared in the several examples: /main() { ... boost::interprocess::permissions perm; perm.set_unrestricted(); managed_shared_memory segment(create_only,"MonitorOEE_SharedMemory", 65536, 0, perm); ... message_queue mq(create_only, "MonitorOEE_MessageQueue", 256, sizeof(t_QRequest), perm); ... named_mutex mutex_waiting(create_only, namedMutex, perm); ... named_condition cdv_wakeup(create_only, namedConditionVar, perm); ... }/ All segment, mq, cdv_wakeup live in the context of the main function... Now I want them to be globally accessible from different functions. But I am not able to. I just partially succeeded creating global pointers to managed_shared_memory and message_queue. But when constructing object to the pointer segmentation faults appear... ///Global Declarations ... static boost::interprocess::message_queue *mq; static boost::interprocess::managed_shared_memory *segment; static boost::interprocess::interprocess_mutex mutex_waiting; static boost::interprocess::interprocess_condition cdv_wakeup; ... main() { .. std::auto_ptr<boost::interprocess::managed_shared_memory>ptr01(new boost::interprocess::managed_shared_memory(open_only,"MonitorOEE_SharedMemory")); segment = ptr01.get(); //this seems to work but fails when constructing objects: IPC_SOAPRequest *pSOAPRequest =segment->construct<IPC_SOAPRequest>(namedSOAPRequest)(); (the autovar counterpart works seamlessly) ... std::auto_ptr<boost::interprocess::message_queue>ptr02(new boost::interprocess::message_queue(open_only,MonitorOEE_MessageQueue")); mq = ptr02.get(); //Seems to work...(send and receive) .. //dunno how to instantiate to mutex and cond var global references } / I am sure it should be an easier and tidier way. Can you give me just the first steps to work this out? I've been sleepless a couple of nights. Eskerrik Asko! (Thanks in the basque language!) -- View this message in context: http://boost.2283326.n4.nabble.com/global-instances-for-managed-shared-memo… Sent from the Boost - Users mailing list archive at Nabble.com.

2 1

Re: [Boost-users] Mutex synchronised wrapper class for boost::interprocess
by Riskybiz 27 Jun '13

27 Jun '13

Ion; thank you!! I found the errors in my prior code thanks to your tip and used the improvements you suggested. I had unwittingly declared variables with very similar names and different types, leading to confused behavior. Anyway that is solved and I now have a working implementation of a mutex synchronised interprocess vector of strings. Have included it below in case it's of use to anyone else trying the same... A further question if I may? What is the best approach to extend a managed_shared_memory_segment? I can dynamically add (via write() function) shared vector elements until such time as an interprocess_exception boost_interprocess::bad_alloc is thrown; when the memory allocated to the shared segment is filled. Is there a way to handle this exception so that the segment can be extended with more memory and then resume adding data elements to the vector? Or perhaps some other way to detect the segment memory is running short and take action to extend it? Thanks again, Riskybiz. Code: Class Definition: #ifndef SHARED_MEMORY_WRAPPER_H//if not defined already #define SHARED_MEMORY_WRAPPER_H//then define it #include <boost/interprocess/managed_shared_memory.hpp> #include <boost/interprocess/containers/vector.hpp> #include <boost/interprocess/containers/string.hpp> #include <boost/interprocess/allocators/allocator.hpp> #include <boost/interprocess/sync/named_mutex.hpp> #include <iostream> #include <stdexcept> using namespace boost::interprocess; class SharedMemoryWrapper { public: SharedMemoryWrapper(const std::string &name, bool server) : m_name(name), m_server(server) { if (server) { //Clean up resources from any prior instances named_mutex::remove("named_mutex"); shared_memory_object::remove(m_name.c_str()); //Create shared memory m_segment = new managed_shared_memory(create_only, m_name.c_str(), 1000000); //Create allocators CharAllocator charallocator(m_segment->get_segment_manager()); StringAllocator stringallocator(m_segment->get_segment_manager()); //This vector is fully constructed in shared memory. All pointers //buffers are constructed in the same shared memory segment //This vector can be safely accessed from other processes. sharedSegmentVector = m_segment->construct<MyShmStringVector>("sharedSegmentVector")(stringallocat or); } else { try { //Open the shared segment m_segment = new managed_shared_memory(open_only, name.c_str()); //Find the vector using the c-string name and open it sharedSegmentVector = m_segment->find<MyShmStringVector>("sharedSegmentVector").first; } catch(interprocess_exception ipe) { std::cout << "Error: Interprocess Exception: " << ipe.what() << std::endl; } catch(...) { std::cout << "Error: Unknown Exception: " << std::endl; } } m_mutex = new named_mutex(open_or_create, "named_mutex"); } ~SharedMemoryWrapper() { /* //Destructor: Will remove the shared memory resources if (m_server) { named_mutex::remove("named_mutex"); //Destroy the vector from the managed_shared_memory //m_segment->destroy_ptr(sharedSegmentVector);//This will free all strings that the vector contains m_segment->destroy<MyShmStringVector>("sharedSegmentVector");//Destroy the vector //Is it necessary to destroy the allocators here? //Remove the managed_shared_memory; this may fail if the memory does not exist or is mapped or opened by another process shared_memory_object::remove(m_name.c_str()); } */ //Delete instance members delete m_mutex; delete m_segment; } //Writer void write(const std::string &inStr) { try { CharAllocator charallocator(m_segment->get_segment_manager()); scoped_lock<named_mutex> lock(*m_mutex); MyShmString shmStr(inStr.c_str(), charallocator); sharedSegmentVector->push_back(shmStr); } catch(interprocess_exception ipe) { std::cout << "Error in SharedMemoryWrapper::write(): interprocess_exception: "<< ipe.what() << std::endl; } catch(...) { std::cout << "Error in SharedMemoryWrapper::write(): unknown error" << std::endl; } } //Reader std::string read(size_t &index) { try { scoped_lock<named_mutex> lock(*m_mutex); const MyShmString &shmStr = (*sharedSegmentVector)[index]; std::string stdStr(shmStr.begin(), shmStr.end()); return stdStr; } catch(interprocess_exception ipe) { std::cout << "Error in SharedMemoryWrapper::read(): interprocess_exception: "<< ipe.what() << std::endl; } catch(std::range_error re) { std::cout << "Error in SharedMemoryWrapper::read(): range_error: " << re.what() << std::endl; } catch(...) { std::cout << "Error in SharedMemoryWrapper::read(): unknown error "<< std::endl; } return "Error in SharedMemoryWrapper::read()"; } //Vector Size size_t getSize() const { return this->sharedSegmentVector->size(); } private: //Typedefs typedef boost::interprocess::allocator<char, managed_shared_memory::segment_manager> CharAllocator; typedef boost::interprocess::basic_string<char, std::char_traits<char>, CharAllocator> MyShmString; typedef boost::interprocess::allocator<MyShmString, managed_shared_memory::segment_manager> StringAllocator; typedef boost::interprocess::vector<MyShmString, StringAllocator> MyShmStringVector; bool m_server; std::string m_name; managed_shared_memory *m_segment; MyShmStringVector *sharedSegmentVector; named_mutex *m_mutex; };//class #endif//header guard Test Programmes: // PrepareSharedMemoryVectorOfStrings.cpp : Defines the entry point for the console application. // #include "SharedMemoryWrapper.h" #include <iostream> #include <string> #include <sstream> int main(int argc, char *argv[]) { SharedMemoryWrapper *smw = new SharedMemoryWrapper("SharedMemoryTest", true); size_t cnt(0); while(cnt < 100) { //Convert number to string representation std::string cntStr = static_cast<std::ostringstream*>( &(std::ostringstream() << cnt) )->str(); //Write to shared memory vector smw->write("Iteration: " + cntStr); cnt++; } size_t vecSize = smw->getSize(); std::cout << "Elements Present: "<< vecSize << std::endl; delete smw; return 0; } // AccessSharedMemoryVectorOfStrings.cpp : Defines the entry point for the console application. // #include "SharedMemoryWrapper.h" #include <iostream> #include <string> int main(int argc, char *argv[]) { SharedMemoryWrapper *smw = new SharedMemoryWrapper("SharedMemoryTest", false); std::string outStr; for(size_t index = 0; index < 100; index++) { outStr = smw->read(index); std::cout << outStr << std::endl; } delete smw; return 0; }

2 1

[1.54.0] Release candidates available
by Marshall Clow 27 Jun '13

27 Jun '13

Release candidate files for 1.54.0 are available at http://boost.cowic.de/rc/ As always, the release managers would appreciate it if you download the candidate of your choice and give building it a try. Please report both success and failure, and anything else that is noteworthy. This helps ensure the candidates build OK before we push them out to SourceForge. The files (and associated md5s) are: efbfbff5a85a9330951f243d0a46e4b9 *boost_1_54_0_rc1.tar.gz 15cb8c0803064faef0c4ddf5bc5ca279 *boost_1_54_0_rc1.tar.bz2 78a35834c45220a6164310e280abe675 *boost_1_54_0_rc1.zip 89e734ffd2afe0982b2d9fc21d033dba *boost_1_54_0_rc1.7z Thanks! -- The release managers

2 1

ASIO: Problem with building boost 1.53 ASIO examples under cygwin
by mschulze＠ivs.cs.ovgu.de 27 Jun '13

27 Jun '13

Hi all, I tried building boost 1.53 ASIO examples under cygwin and it didn't worked with bjam or b2 respectively. Then I used "bjam -n" to get the invoked commands in order to further track down the issue. I end up with the following command to compile an example "g++" -ftemplate-depth-128 -O0 -fno-inline -Wall -g -mthreads -DBOOST_ALL_NO_LIB=1 -DBOOST_CHRONO_DYN_LINK=1 -DBOOST_SYSTEM_DYN_LINK=1 -DBOOST_SYSTEM_NO_DEPRECATED -DBOOST_THREAD_BUILD_DLL=1 -DBOOST_THREAD_POSIX -DBOOST_THREAD_USE_DLL=1 -I"../../../.." -o "blocking_udp_echo_server" "blocking_udp_echo_server.cpp" -D__USE_W32_SOCKETS -D_WIN32_WINNT=0x0501 ../../../../bin.v2/libs/thread/build/gcc-4.5.3/release/link-static/threading-multi/libboost_thread.a ../../../../bin.v2/libs/system/build/gcc-4.5.3/release/link-static/threading-multi/libboost_system.a -lws2_32 -lmswsock and I got the following errors: In file included from ../../../../boost/asio/detail/impl/posix_tss_ptr.ipp:24:0, from ../../../../boost/asio/detail/posix_tss_ptr.hpp:77, from ../../../../boost/asio/detail/tss_ptr.hpp:27, from ../../../../boost/asio/detail/call_stack.hpp:20, from ../../../../boost/asio/impl/handler_alloc_hook.ipp:19, from ../../../../boost/asio/handler_alloc_hook.hpp:80, from ../../../../boost/asio/detail/handler_alloc_helpers.hpp:22, from ../../../../boost/asio/detail/bind_handler.hpp:19, from ../../../../boost/asio/detail/wrapped_handler.hpp:18, from ../../../../boost/asio/io_service.hpp:24, from ../../../../boost/asio/basic_io_object.hpp:19, from ../../../../boost/asio/basic_socket.hpp:19, from ../../../../boost/asio/basic_datagram_socket.hpp:20, from ../../../../boost/asio.hpp:20, from blocking_udp_echo_server.cpp:13: ../../../../boost/asio/error.hpp:77:17: error: a function call cannot appear in a constant-expression ../../../../boost/asio/error.hpp:130:15: error: a function call cannot appear in a constant-expression ../../../../boost/asio/error.hpp:135:19: error: a function call cannot appear in a constant-expression ../../../../boost/asio/error.hpp:149:23: error: a function call cannot appear in a constant-expression ../../../../boost/asio/error.hpp:163:15: error: a function call cannot appear in a constant-expression ../../../../boost/asio/error.hpp:244:45: warning: boost::asio::error::system_category defined but not used ../../../../boost/asio/error.hpp:246:45: warning: boost::asio::error::netdb_category defined but not used ../../../../boost/asio/error.hpp:248:45: warning: boost::asio::error::addrinfo_category defined but not used ../../../../boost/asio/error.hpp:250:45: warning: boost::asio::error::misc_category defined but not used ../../../../boost/asio/detail/winsock_init.hpp:78:30: warning: boost::asio::detail::winsock_init_instance defined but not used The error messages boils down to the use of "__MSABI_LONG(x)" within an enumeration definition. __MSABI_LONG is a macro that has to be resolved during the cpp run but isn't but it seems not to be known at this point. I also tested the boost version 1.52 and there are none of these issues and it worked like it is intended. Best regards, Michael

2 4

Accessing memory in shared segments
by Andy Falanga (afalanga) 27 Jun '13

27 Jun '13

Hi, I have some code which shares data with shared memory. To that end, I'm using boost::interprocess:managed_shared_memory. Without respect to scope and functions, what I have is roughly: boost::interprocess::managed_shared_memory sharedMem(boost::interprocess::create_only, "SharedMemory", 4096); //later on u_int8_t* dataSegment = sharedMemory.construct<u_int8_t>("data")[5](0); char* string1 = sharedMemory.construct<char>("String 1")[50]('\0'); // other stuff // placing stuff in the data segment for(int i(0); i < 5; i++) dataSegment[i] = anEarlierArray[i]; // then, in other process read the data Segment char ary[5]; for(int i(0); i< 5; i++) ary[i] = dataSegment.first[i]; For whatever reason I couldn't simply access through dataSegment[n] notation. The weirdest thing is, the data is being read back to me in the reverse order that it should be. For example, if [0x2,0x4,0x7,0x1] is written I read it back as [0x1,0x7,0x4,x02] even though I'm writing and reading in the same order. Why is this? What is it I'm missing about using shared memory? Thanks, Andy

1 0

[System] Getting error messages in English
by Igor R 27 Jun '13

27 Jun '13

I'd like to always get English error messages on Win7 with localized UI. As far as I can see, system_error_category::message() calls FormatMessageA passing it MAKELANGID(LANG_NEUTRAL, SUBLANG_DEFAULT). I tried to set locale, but it doesn't affect FormatMessageA for some reason. I'd appreciate any idea.

1 0

[1.54.0] Beta 1 Available
by Marshall Clow 27 Jun '13

27 Jun '13

Boost release 1.54.0 beta 1 is now available from SourceForge See http://sourceforge.net/projects/boost/files/boost/1.54.0.beta.1/ This release contains 3 new libraries: • Log: Logging library, from Andrey Semashev. • TTI: Type Traits Introspection library, from Edward Diener. • Type Erasure: Runtime polymorphism based on concepts. and numerous enhancements and bug fixes for existing libraries. For details of what's in the release, see http://www.boost.org/users/history/version_1_54_0.html. Note that the links to files on this web page are for the final release - use the SourceForge link above to get the beta files. Please download the beta, give it a try, and report any problems you encounter. Thanks, -- The Boost Release Team

5 4

Re: [Boost-users] [boost] [thread] synchronized_value: value and move semantics
by Klaim - Joël Lamotte 26 Jun '13

26 Jun '13

On Wed, Jun 26, 2013 at 12:37 PM, Ben Pope <benpope81(a)gmail.com> wrote: > > I think this should just not exist in C++11 and instead be replaced by > something like monitor<T> described by Herb Sutter [1]: I'm using several variants of Monitor in my projects. It don't work well in cases where what you really want is syncrhonization, which happen sometime when the design suggests it. In this case you need a mutex, so synchronized_value helps with that. In my opinion, If the multiple access is a bottleneck in real application, it is very easy to spot with a profiler and modify the code as you pointed. Joel Lamotte

1 0

Problem Iterating a directory
by Sarma Tangirala 26 Jun '13

26 Jun '13

Hello, I am using boost filesystem to iterate through my directory and have encountered a problem that I am unsure of how to debug. try { //for ( boostfs::directory_iterator dir(boostfs::path("/home/vtangira/Desktop/person_finder_v20130513/data/dummy1")), end_itr; for ( boostfs::directory_iterator dir(boostfs::path(paths_to_files.at(0).c_str())), end_itr; dir != end_itr; dir++ ) { std::cout << "File: " << dir->path() << std::endl; } } catch (boostfs::filesystem_error e) { cout << "Error Code: " << e.code() << endl << "Error Msg: " << e.code().message() << endl; } In this snippet, the line with the for loop that has been commented out works but the the one below it with the vector does not. The vector version represents the same string. Any advice is appreciated. Thanks -- 0 1 0 0 0 1 1 1 1

3 4

Re: [Boost-users] [Asio] Does stop() cancel handlers post()ed to io_service?
by Marat Abrarov 26 Jun '13

26 Jun '13

> However, I'm still interested whether stop() should always cancel the > queued handlers, i.e. if this is a bug (on windows). The io_service > documentation says this regarding the use of a work object > (http://alturl.com/98rvq) > > "To effect a shutdown, the application will then need to call the > io_service > object's stop() member function. This will cause the io_service run() > call > to return as soon as possible, abandoning unfinished operations and > without > permitting ready handlers to be dispatched. > Alternatively, if the application requires that all operations and > handlers > be allowed to finish normally, the work object may be explicitly > destroyed." I think (I've just inspected the Asio code) this behavior is designed in such a way for performance reasons - to not check "stopped" flag too often. Even the "stopped" flag is just an atomic value - it's Windows IOCP where OS uses specific technics for scheduling IOCP-tied threads and, may be, the author of Asio thinks (or even knows/has experience) that even atomic-value-check degrades performance or brakes specific scheduling of IOCP-tied threads. Regards, Marat Abrarov.

1 0

[Boost.Process] Status request
by Damien Kick 26 Jun '13

26 Jun '13

Thought that perhaps somebody on this list would know if/where a/the most current version of Boost.Process might be found. Or has this library been abandoned? Begin forwarded message: From: Damien Kick <dkick(a)shoretel.com<mailto:dkick@shoretel.com>> Subject: Boost.Process Date: June 25, 2013 5:16:10 PM CDT To: <boris(a)highscore.de<mailto:boris@highscore.de>> I've found your version of Boost.Process at <http://www.highscore.de/cpp/process/index.html>. I noticed that the date on this page is from 2009; are you still working on this library? If not, thank you for your effort so far. If you are still working on the library, I was hoping you might have some insights and/or help for a compilation error I'm finding when trying to use Boost.Process with Boost 1.53 on Windows 7 with Visual Studio 2010 SP1. First of all, I had to add a header file to .../process/operation.hpp; i.e. I had to add #include <boost/filesystem/operation.hpp> to get the declaration of boost::filesystem::filesystem_error. But then I'm finding the following error 1>------ Build started: Project: ZERO_CHECK, Configuration: Debug Win32 ------ 2>------ Build started: Project: test-cma, Configuration: Debug Win32 ------ 2> cma.cpp 2>Z:\Documents\perforce\dkick-mac\projects\DandM\DMTarget\windows-6.1\msvc-10.0\include\boost/process/detail/pipe.hpp(129): error C2665: 'boost::system::system_error::system_error' : none of the 7 overloads could convert all the argument types 2> Z:\Documents\perforce\dkick-mac\projects\DandM\DMTarget\windows-6.1\msvc-10.0\include\boost/system/system_error.hpp(39): could be 'boost::system::system_error::system_error(int,const boost::system::error_category &,const std::string &)' 2> Z:\Documents\perforce\dkick-mac\projects\DandM\DMTarget\windows-6.1\msvc-10.0\include\boost/system/system_error.hpp(43): or 'boost::system::system_error::system_error(int,const boost::system::error_category &,const char *)' 2> while trying to match the argument list '(DWORD, overloaded-function, const char [54])' ========== Build: 1 succeeded, 1 failed, 3 up-to-date, 0 skipped ========== Which is coming from .../boost/process/detail/pipe.hpp (the bold italic line below). From the quick skim I've given some of the documentation, I think the fact that BOOST_PROCESS_WINDOWS_USE_NAMED_PIPE is not defined means that I would find some features disabled, right? How is this determined; i.e. whether or not to define BOOST_PROCESS_WINDOWS_USE_NAMED_PIPE? # if defined(BOOST_PROCESS_WINDOWS_USE_NAMED_PIPE) static unsigned int nextid = 0; std::string pipe = "\\\\.\\pipe\\boost_process_" + boost::lexical_cast<std::string>(::GetCurrentProcessId()) + "_" + boost::lexical_cast<std::string>(nextid++); hs[0] = ::CreateNamedPipeA(pipe.c_str(), PIPE_ACCESS_INBOUND | FILE_FLAG_OVERLAPPED, 0, 1, 8192, 8192, 0, &sa); if (hs[0] == INVALID_HANDLE_VALUE) boost::throw_exception(boost::system::system_error(::GetLastError(), boost::system::system_category, "boost::process::detail::pipe::pipe: CreateNamedPipe failed")); hs[1] = ::CreateFileA(pipe.c_str(), GENERIC_WRITE, 0, NULL, OPEN_EXISTING, FILE_FLAG_OVERLAPPED, NULL); if (hs[1] == INVALID_HANDLE_VALUE) boost::throw_exception(boost::system::system_error(::GetLastError(), boost::system::system_category, "boost::process::detail::pipe::pipe: CreateFile failed")); OVERLAPPED overlapped; ZeroMemory(&overlapped, sizeof(overlapped)); overlapped.hEvent = ::CreateEvent(NULL, TRUE, FALSE, NULL); if (!overlapped.hEvent) boost::throw_exception(boost::system::system_error(::GetLastError(), boost::system::system_category, "boost::process::detail::pipe::pipe: CreateEvent failed")); BOOL b = ::ConnectNamedPipe(hs[0], &overlapped); if (!b) { if (::GetLastError() == ERROR_IO_PENDING) { if (::WaitForSingleObject(overlapped.hEvent, INFINITE) == WAIT_FAILED) { ::CloseHandle(overlapped.hEvent); boost::throw_exception(boost::system::system_error(::GetLastError(), boost::system::system_category, "boost::process::detail::pipe::pipe: WaitForSingleObject failed")); } } else if (::GetLastError() != ERROR_PIPE_CONNECTED) { ::CloseHandle(overlapped.hEvent); boost::throw_exception(boost::system::system_error(::GetLastError(), boost::system::system_category, "boost::process::detail::pipe::pipe: ConnectNamedPipe failed")); } } ::CloseHandle(overlapped.hEvent); # else if (!::CreatePipe(&hs[0], &hs[1], &sa, 0)) boost::throw_exception(boost::system::system_error(::GetLastError(), boost::system::system_category, "boost::process::detail::pipe::pipe: CreatePipe failed")); # endif ________________________________ This e-mail and any attachments are confidential. If it is not intended for you, please notify the sender, and please erase and ignore the contents.

1 0

Re: [Boost-users] [Asio] Does stop() cancel handlers post()ed to io_service?
by Marat Abrarov 26 Jun '13

26 Jun '13

> The program below gives different results on Windows (Visual Studio > 2008) and Linux. While stop() does cancel the handlers on Linux, the same > is not true for Windows. I also tried to destroy the io_service object > instead of calling stop, but this does not cancel the handlers on either > system, even though the documentation seems to state it should. It would be more accurate to say "on Windows asio::io_service::stop() does not cancel already queued (!) handlers". See example below. ~~~~~~~~~~~~~~~~~~~~~~~ #if defined(WIN32) #include <tchar.h> #endif #include <cstdlib> #include <cstddef> #include <iostream> #include <boost/ref.hpp> #include <boost/asio.hpp> #include <boost/bind.hpp> #include <boost/thread.hpp> #include <boost/function.hpp> namespace { typedef boost::function<void (boost::asio::io_service&)> work_factory; void run_test(work_factory wf) { { boost::asio::io_service io_service; { wf(io_service); std::cout << "io_service work created" << std::endl; { std::cout << "work thread starting..." << std::endl; boost::thread work_thread(boost::bind( &boost::asio::io_service::run, boost::ref(io_service))); boost::this_thread::sleep(boost::posix_time::millisec(300)); io_service.stop(); std::cout << "io_service signalled to stop" << std::endl; work_thread.join(); std::cout << "work thread stopped" << std::endl; } } std::cout << "destroying io_service..." << std::endl; } std::cout << "io_service destroyed" << std::endl; } void test_function1() { static int counter = 0; std::cout << "handler #" << counter++ << std::endl; boost::this_thread::sleep(boost::posix_time::millisec(100)); } void post_test_function1(boost::asio::io_service& io_service) { for (std::size_t i = 0; i != 10; ++i) { io_service.post(&test_function1); } } void post_test_function2(boost::asio::io_service&); void test_function2(boost::asio::io_service& io_service) { static int counter = 0; std::cout << "handler #" << counter++ << std::endl; boost::this_thread::sleep(boost::posix_time::millisec(100)); post_test_function2(io_service); } void post_test_function2(boost::asio::io_service& io_service) { io_service.post(boost::bind(test_function2, boost::ref(io_service))); } } // anonymous namespace #if defined(WIN32) int _tmain(int /*argc*/, _TCHAR* /*argv*/[]) #else int main(int /*argc*/, char* /*argv*/[]) #endif { std::cout << "Test 1" << std::endl; run_test(post_test_function1); std::cout << std::endl << "Test 2" << std::endl; run_test(post_test_function2); return EXIT_SUCCESS; } ~~~~~~~~~~~~~~~~~~~~~~~ Regards, Marat Abrarov.

2 1

Declaring tricky templated variables such as disjoint_sets.
by Jeremy Murphy 25 Jun '13

25 Jun '13

Hi, this is a somewhat general C++ question, but I'm specifically concerned about a Boost class in this case. How do I declare a boost::disjoint_sets<Rank, Parent, FindCompress> variable without some kind of automatic template deduction? Usually I see such classes declared taking their template parameters from within a function such as: template <typename Rank, typename Parent> void foo(Rank rank, Parent parent) { disjoint_sets<Rank, Parent> dset; } but what if I want to declare dset at file scope? I've been trying to manually construct the Rank and Parent types, but it's slightly doing my head in and the iterator_property_map example seems to be stale. Is there a general C++ technique or a Boost-specific one that can help me out here? Or is it just a matter of laboriously building up the typedefs? Thanks, cheers. Jeremy

2 1

[Asio] Does stop() cancel handlers post()ed to io_service?
by Martin Gernhard 25 Jun '13

25 Jun '13

Hello, I want to cancel all pending handlers of an io_service. Is io_service::stop() supposed to cancel handlers that have been added using post()? The documentation for stop() is somewhat ambiguous on this matter, only stating that run() "should return as soon as possible". Alternatively, should the destructor of io_service cancel the handlers? I am trying to build a "worker thread" using Boost.Asio (1.53) and Boost.Thread as outlined at http://think-async.com/Asio/Recipes. I have already looked at the documentation and searched the bug tracker, the mailing list archives, Google and Stack Overflow but only found an unanswered message from almost two years ago (http://lists.boost.org/boost-users/2011/09/70596.php) as well as another message that describes different behavior of stop() than what I am experiencing (http://thread.gmane.org/gmane.comp.lib.boost.user/73762) The program below gives different results on Windows (Visual Studio 2008) and Linux. While stop() does cancel the handlers on Linux, the same is not true for Windows. I also tried to destroy the io_service object instead of calling stop, but this does not cancel the handlers on either system, even though the documentation seems to state it should. Result on Windows for stop() and ~io_service and on Linux for ~io_service: Start 0 1 2 Stop 3 4 5 6 7 8 9 Result on Linux for stop(): Start 0 1 2 Stop For now I am using the workaround of querying a boost::atomic<bool> flag from every handler and setting it to false after calling stop(). Is there a better way to achieve this, maybe by implementing a custom IO object / service? It feels like overkill for something seemingly simple. Regards, Martin #include <boost/asio.hpp> #include <boost/bind.hpp> #include <boost/function.hpp> #include <boost/scoped_ptr.hpp> #include <boost/thread.hpp> #include <iostream> using namespace boost; using namespace boost::asio; void testFunction() { static int counter = 0; std::cout << counter++ << std::endl; this_thread::sleep( posix_time::millisec( 100 ) ); } int main() { shared_ptr<io_service> m_ioService(new io_service()); scoped_ptr<io_service::work> m_work(new io_service::work(*m_ioService)); scoped_ptr<thread> m_thread; for ( unsigned int i = 0; i != 10; ++i ) { m_ioService->post( &testFunction ); } std::cout << "Start" << std::endl; m_thread.reset( new thread( bind( &io_service::run, m_ioService ) ) ); this_thread::sleep( posix_time::millisec( 300 ) ); std::cout << "Stop" << std::endl; m_work.reset(); m_ioService->stop(); m_thread->join(); m_thread.reset(); m_ioService->reset(); return 0; }

1 0

[asio] synchronous read from sslstream with unknown length
by "Torsten Müller" 25 Jun '13

25 Jun '13

1 0

[asio] How to deal with no-connect and/or read-timeout
by Michael Powell 25 Jun '13

25 Jun '13

Hello, I posted previously, but this is a bit of a follow-on thought. Did some searching, and SO has this note: http://stackoverflow.com/questions/13126776/asioread-with-timeout Pardon me while I stare in disbelief. Seriously? Okay, I'd like to use Boost.Asio to facilitate things. Is this the way (one proposed way) of handling read-with-timeout? Our scenario goes similar to this I think: * Client may (or may not) connect, depending on the integration scenario. The main service thread should tolerate either, with plausibly the socket connection/session operating in another thread "asynchronously". * When client connects, the session logic is pretty straightforward: read if/when there is something to read, and pass it through a message framing unwrapper. Completed message gets deserialized and published into a thread-safe queue. * Events may be generated from the main service thread and subsequently published to the client: effectively a write operation, after passing a serialized message through a message framing wrapper. Doable? Feasible, possibly with some sort of a ping-ponging-back-and-forth type of session handling callback mechanism? I am cautiously optimistic that this has more to do with connection/session callback/state-machine type things, than it does Asio? Regards, Michael

2 2

[asio] Grasping the proto-typical socket server
by Michael Powell 25 Jun '13

25 Jun '13

Hello, I've dealt with sockets, and socket servers, in the past, so am familiar with them. Abstracting out a prototypical server details, drawing from the plethora of examples, this is what I can determine. I see three primary abstractions, the server itself, with endpoint binding, socket connection acceptor, so far so good. Then there's a connection, which is a sort of io_service to socket adapter. Followed closely by a session concern, where reads and writes are actually performed. Studying the examples, however, it seems to me that by design io_service wants to be *the* first class citizen that runs, with asynchronous callbacks being the mode of choice. How might we go about mixing this code, presumably in the main thread (?), with any worker service-oriented thread that might be going on. For instance, I start a thread and wait for it to finish and rejoin for a main service loop. Socket connection sessions are secondary to this service behavior taking place IMO. If a client connects, great, we'll interact with it. If not, that should be transparent to the application. Possibly session is a second thread worker with a thread-local-scope io_service, that we pub/sub messages to the main service loop. Thoughts on this type of approach? Much appreciated... Regards, Michael

2 1

Mutex synchronised wrapper class for boost::interprocess shared memory vector of strings
by Riskybiz 24 Jun '13

24 Jun '13

I'm trying to create a wrapper class to control a managed shared memory segment using Boost::Interprocess. I'm having trouble reading data from the shared memory vector of strings... When the vector is instantiated in the managed shared segment I place in the first vector element a test string; "testing testing 1 2 3". Immediately after placing it there and in the scope of the same code block I can access the vector element and print out the string using std::cout. However when I try to read the vector element using the read() function the test program crashes. The problem is with the line: std::string stdStr = sharedSegmentVector->at(index).c_str(); Question is how to fix this? I'm confused because at instantiation time this same code will access the shared memory vector; later on in the read() function it will not. I've as yet been unable to test the write() function; because I can't yet read() the vector! Any help to get this working is much appreciated, With thanks, Riskybiz Class Definition: #ifndef SHARED_MEMORY_WRAPPER_H//if not defined already #define SHARED_MEMORY_WRAPPER_H//then define it #include <boost/interprocess/managed_shared_memory.hpp> #include <boost/interprocess/containers/vector.hpp> #include <boost/interprocess/containers/string.hpp> #include <boost/interprocess/allocators/allocator.hpp> #include <boost/interprocess/sync/named_mutex.hpp> #include <iostream> #include <stdexcept> using namespace boost::interprocess; class SharedMemoryWrapper { public: SharedMemoryWrapper(const std::string &name, bool server) : m_name(name), m_server(server) { if (server) { named_mutex::remove("named_mutex"); shared_memory_object::remove(m_name.c_str()); //Create shared memory managed_shared_memory m_segment(create_only, m_name.c_str(), 10000); //Create allocators CharAllocator charallocator(m_segment.get_segment_manager()); StringAllocator stringallocator(m_segment.get_segment_manager()); //This string is in only in this process (the pointer pointing to the //buffer that will hold the text is not in shared memory). //But the buffer that will hold "this is my text" is allocated from //shared memory MyShmString mystring(charallocator); mystring = "testing testing 1 2 3"; //This vector is fully constructed in shared memory. All pointers //buffers are constructed in the same shared memory segment //This vector can be safely accessed from other processes. MyShmStringVector *sharedSegmentVector = m_segment.construct<MyShmStringVector>("sharedSegmentVector")(stringallocato r); sharedSegmentVector->push_back(mystring); std::string outStr = sharedSegmentVector->at(0).c_str(); std::cout << "Shared Memory Vector Instantiated: Index 0 Data is: " << outStr << std::endl; } else { //Open the shared segment managed_shared_memory m_segment(open_only, name.c_str()); //Find the vector using the c-string name and open it MyShmStringVector *sharedSegmentVector = m_segment.find<MyShmStringVector>("sharedSegmentVector").first; } m_mutex = new named_mutex(open_or_create, "named_mutex"); } ~SharedMemoryWrapper() { if (m_server) { named_mutex::remove("named_mutex"); //Destroy the vector from the managed_shared_memory //m_segment->destroy_ptr(sharedSegmentVector);//This will free all strings that the vector contains m_segment->destroy<MyShmStringVector>("sharedSegmentVector");//Destroy the vector //Is it necessary to destroy the allocators here? //Remove the managed_shared_memory; this may fail if the memory does not exist or is mapped or opened by another process shared_memory_object::remove(m_name.c_str()); } delete m_mutex; delete m_segment; } //Writer void write(const std::string &in) { try { CharAllocator charallocator(m_segment->get_segment_manager()); scoped_lock<named_mutex> lock(*m_mutex); MyShmString shmStr(in.c_str(), charallocator); sharedSegmentVector->push_back(shmStr); } catch(...) { std::cout << "Error in SharedMemoryWrapper::write(): unknown error" << std::endl; } } //Reader std::string read(int index) { try { scoped_lock<named_mutex> lock(*m_mutex); std::cout << "Shared Memory Vector Entered Read(), Passed Index: " << index << std::endl; std::string stdStr = sharedSegmentVector->at(index).c_str(); std::cout << "Shared Memory Vector Read At: Index" << index << "Data is: " << stdStr << std::endl; return stdStr; } catch(std::range_error err) { std::cout << "Error in SharedMemoryWrapper::read(): range_error" << std::endl; } catch(...) { std::cout << "Error in SharedMemoryWrapper::read(): unknown error " << std::endl; } return "ERROR IN SharedMemoryWrapper::read()"; } private: //Typedefs typedef boost::interprocess::allocator<char, managed_shared_memory::segment_manager> CharAllocator; typedef boost::interprocess::basic_string<char, std::char_traits<char>, CharAllocator> MyShmString; typedef boost::interprocess::allocator<MyShmString, managed_shared_memory::segment_manager> StringAllocator; typedef boost::interprocess::vector<MyShmString, StringAllocator> MyShmStringVector; bool m_server; std::string m_name; managed_shared_memory *m_segment; MyShmStringVector *sharedSegmentVector; named_mutex *m_mutex; };//class #endif//header guard Test Programme: #include "stdafx.h" #include "SharedMemoryWrapper.h" #include <iostream> #include <string> int main(int argc, char *argv[]) { SharedMemoryWrapper *smw = new SharedMemoryWrapper("SharedMemory1", true); std::string outStr = smw->read(0); std::cout << outStr << std::endl; return 0; }

2 1