- Boost-users - lists.boost.org

BOOST_STATIC_ASSERT with boost::array.static_size
by Anton Daneyko 12 Jul '13

12 Jul '13

Dear boost users, could someone tell me why this short program does not compile? #include <boost/array.hpp> #include <boost/static_assert.hpp> int main(int argc, const char *argv[]) { boost::array<int, 3> a1 = {{1, 1, 1}}; BOOST_STATIC_ASSERT(a1.static_size == 3); return 0; } With g++ (4.3.4) I get the following error: tmp.cpp: In function ‘int main(int, const char**)’: tmp.cpp:7: error: ‘a1’ cannot appear in a constant-expression tmp.cpp:7: error: `.' cannot appear in a constant-expression tmp.cpp:7: error: template argument 1 is invalid With icc (12.1.6) I get the following error: tmp.cpp(7): error: this operator is not allowed in a template argument expression BOOST_STATIC_ASSERT(a1.static_size == 3); ^ tmp.cpp(7): error: incomplete type is not allowed BOOST_STATIC_ASSERT(a1.static_size == 3); ^ compilation aborted for tmp.cpp (code 2) However I can compile using IBM xlC compiler. I thought that static_size is a compile time constant and can be used in this context...

2 1

Query on Boost.Log
by Leo Carreon 11 Jul '13

11 Jul '13

Hi everyone, I’m trying to learn how to use Boost.Log from Boost 1.54.0 on Fedora 18. When I attempt to call boost::log::init_from_stream() even with BOOST_LOG_DYN_LINK defined, I get the error message”undefined reference to `void boost::log::v2_mt_posix::init_from_stream<char>(std::basic_istream<char, std::char_traits<char> >&)'” for the line where I called the function. Note that I have included the header file “boost/log/utility/setup/from_stream.hpp” and added –lboost_log to my link command line. Am I missing something. I have noticed that boost::log::init_from_stream() is a template. Is an instance of this template added into the shared library? Kind regards, Leo

1 0

Procedure entry point _ZNSt16invalid_argumentD2Ev could not be located in the dynamic link library libstdc++-6.dll
by Joe Hays 11 Jul '13

11 Jul '13

Hello all, I have a newbie question for you. I'm trying Boost for the very first time. I'm also not terribly experienced with C++. I'm using, Eclipse Kepler (latest) Mingw 3.20 (latest) Boost_1_54_0 on a 64-bit Windows 7 machine. I followed this tutorial to configure Boost to be used with Mingw and Eclipse: http://theseekersquill.wordpress.com/2010/08/24/howto-boost-mingw/ Sections 1-6 performed as expected. The 7th and final step is when a "hello world" like application finally compiles and links Boost into a .exe. I cut-n-paste the following code from Boost's Getting Started tutorial ( http://www.boost.org/doc/libs/1_54_0/more/getting_started/windows.html) <snip> #include <boost/regex.hpp> #include <iostream> #include <string> int main() { std::string line; boost::regex pat( "^Subject: (Re: |Aw: )*(.*)" ); while (std::cin) { std::getline(std::cin, line); boost::smatch matches; if (boost::regex_match(line, matches, pat)) std::cout << matches[2] << std::endl; } } </snip> The output of the build is, <snip> 12:20:46 **** Rebuild of configuration Debug for project BoostRegexExample1 **** Info: Internal Builder is used for build g++ "-Ic:\\boost_1_54_0" -O0 -g3 -Wall -c -fmessage-length=0 -o BoostRegexExample1.o "..\\BoostRegexExample1.cpp" g++ "-LC:\\boost_1_54_0\\stage\\lib" -o BoostRegexExample1.exe BoostRegexExample1.o -lboost_regex-mgw46-mt-1_54 12:20:52 Build Finished (took 5s.992ms) </snip> If I try to run that application within Eclipse the terminal simply states the application terminated. If I go to a terminal and run the .exe (using a text file as in input as directed in the Getting Started document) I get, *"Procedure entry point _ZNSt16invalid_argumentD2Ev could not be located in the dynamic link library libstdc++-6.dll"* Any ideas why this is happening? And, what I can do to get past this? Thanks, Joe

2 2

boost::uuid
by John Emmas 11 Jul '13

11 Jul '13

Hello there, I'm developing an app (originally written for Linux) which uses some UUID functionality that originally came from a set of e2fs utilities for Linux. I'm porting the app to Windows so naturally, the e2fs stuff doesn't build. I need to find a UUID implemntation that's more cross-platform - and of course, the obvious one to try is boost::uuid. I first installed boost when it was at version 1.40 (1-2 years ago?) Anyway, my installed version doesn't include UUID so I decided to update to a more recent version. The most recent Windows installer I could find was for version 1.51, from BoostPro:- http://www.boostpro.com/download/ Unfortunately, it probably isn't the absolute latest because BoostPro is apparently now discontinued - but it should be recent enough to include the UUID modules. From what I can gather, boost::uuid has been around since at least version 1.42. However, I hit a couple of problems:- 1) Firstly, I seem to get errors, no matter which mirror I use. I'll try again later, in case the errors go away. 2) Despite the errors, something does seem to get downloaded. However, rather than installing all the source files / DLLs etc, I just get a subdirectory called "lib" which contains a bunch of zip files (over 100). My original (1.40) installation seemed to unpack everything for me. 3) AFAICT even ver 1.51 doesn't include any UUID functionality. There was nothing for UUID that was selectable in the installer - and none of the modules look like UUID. I seem to have boost::regex, boost::filesystem, boost:date_time and a whole load of other stuff - but no boost::uuid :-( I then came across this web page which seems to suggest that all I need to do is add one header file to my existing installation (boost/uuid.hpp). Am I over-thinking this?? Is that all there is to it?? Thanks. John

3 3

[graph] in-depth parsing and modifications of the graph
by Orianne Siret 11 Jul '13

11 Jul '13

Hello, I would like to know if it is possible to parse a boost graph in depth while being able to modify the properties of the verices you parse ? I saw there is a in-depth parser but it seems to not allow the modifications of properties ? Thanks in advance for your help ! Orianne Siret

2 1

[asio] Incorrect reporting of bytes transferred in write handler in 1.54
by Darren 11 Jul '13

11 Jul '13

Is anybody else seeing problems with the bytes_transferred reported to the write_handler after a successful async_write? I am seeing a value of zero despite the fact that the peer receives the bytes and no error is reported to the write_handler. My example code when built against 1.54 exhibits the problem whereas building against 1.53 works correctly. I am running on Windows 7 with MSVC 2010. A couple of notes: * I only see the problem when the number of threads is > 1, even though the handlers are invoked on a strand. * Defining BOOST_ASIO_DISABLE_IOCP makes everything work properly. When running my code a failure shows itself like this: Client: Writing value 85 Server: Received 4 bytes, value 85 Client: bytesWritten = 0 In other words, the client used async_write to send 4 bytes with long value 85 the server. The server received these bytes. The client write_handler was called with no error but a bytes_transferred value of 0. Is there something I don't understand or is this a problem introduced in 1.54? #include <boost/asio.hpp> #include <boost/thread.hpp> #include <boost/thread/mutex.hpp> #include <iostream> using namespace boost::asio; const std::size_t NumThreads = 2; const unsigned short Port = 12345; // Prevent interleaved output boost::mutex logMutex; struct Server { Server(io_service& svc) : m_socket(svc), m_acceptor(svc, ip::tcp::endpoint(ip::tcp::v4(), Port)), m_buffer(0) { } void Run() { m_acceptor.accept(m_socket); // Simple blocking reads while(std::size_t bytesRead = read(m_socket, buffer(&m_buffer, sizeof(m_buffer)))) { boost::mutex::scoped_lock lock(logMutex); std::cout << "Server: Received " << bytesRead << " bytes, value " << m_buffer << std::endl; } } ip::tcp::socket m_socket; ip::tcp::acceptor m_acceptor; unsigned long m_buffer; }; struct Client { Client(io_service& svc) : m_strand(svc), m_socket(svc), m_buffer(0) { m_socket.connect( ip::tcp::endpoint(ip::address_v4::from_string("127.0.0.1"), Port)); ScheduleWrite(); } void WriteHandler(const boost::system::error_code& err, std::size_t bytesWritten) { assert(!err); if(bytesWritten != sizeof(m_buffer)) { boost::mutex::scoped_lock lock(logMutex); std::cout << "Client: bytesWritten = " << bytesWritten << std::endl; return; } ScheduleWrite(); } void ScheduleWrite() { { boost::mutex::scoped_lock lock(logMutex); std::cout << "Client: Writing value " << ++m_buffer << std::endl; } async_write(m_socket, buffer(&m_buffer, sizeof(m_buffer)), m_strand.wrap( std::bind(&Client::WriteHandler, this, std::placeholders::_1, std::placeholders::_2))); } io_service::strand m_strand; ip::tcp::socket m_socket; unsigned long m_buffer; }; int main(int argc, char* argv[]) { boost::asio::io_service svc; Server server(svc); Client client(svc); boost::thread_group threads; for(int n = 0; n < NumThreads; ++n) threads.create_thread(std::bind(&io_service::run, &svc)); server.Run(); return 0; }

2 1

Using boost with cray c++
by avib369 . 11 Jul '13

11 Jul '13

Does any one have any experience of using boost with cray C++ compiler ? Are any known problematic areas/issues ? Ok I know its not in the official list of supported compilers, but neither is the HPUX acc, which seems to work fine for most libs. Any help appreciated . Ta, Avi

2 1

[Boost.Log] Some few newbee questions
by Ovanes Markarian 10 Jul '13

10 Jul '13

Hi *, Thanks for the really great logging library. Just used it a bit and was amazed how useful and easy to use it was. As I start using it more and more I have some questions... 1. In my application I would like to start with the trivial logger and extend this logger with more options, i.e. attributes etc. Currently I see the following input when doing a trace: [2013-07-03 22:17:32.462265] [0x76f73180] [trace] thread_pool: creating 1 threads[2013-07-03 22:17:32.462303] [0x76f73180] [trace] thread_pool: created thread number: 1[2013-07-03 22:17:32.462312] [0x76f73180] [trace] thread_pool: all threads started[2013-07-03 22:17:32.462320] [0x76f73180] [trace] thread_pool: constructor finished My question is what is the second attribute? Thread id? As my code runs on a 64bit machine, that does not seem to be an address... 2. Is it possible to use a scoped logging with the trivial logger? I tried to do it as described in the example but see not output. How can I produce the output from the scoped logger? 3. In my example I wrote thread_pool prefix before each log entry. Can I do it automatically? How would I go on if I only have trivial logging? May be I misunderstand some concepts, like how to use multiple loggers and connect them with each other? Any help is very appreciated. Thanks, Ovanes

2 2

Re: [Boost-users] Mutex synchronised wrapper class for boost::interprocess vector of strings
by Riskybiz 09 Jul '13

09 Jul '13

"It's not an easy task. You should catch the exception and then notify all connected processes that should disconnected from the segment (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, thank you for that, I see.... 1. Can I ask what mechanism you would suggest for notifying the other processes to disconnect before the grow() operation? 2. Considering how to deal with the issue of dynamically growing a shared memory vector where there is one producer process and several consumers; I hit upon the notion of having the producer process use a std::vector to conventionally store data and then using a boost shared memory segment to store a pointer to the std::vector. Consumer processes would access the boost shared memory to retrieve the pointer value thus enabling access to the std::vector data store in the producer process. This, I hoped, might remove the complexities associated with dynamically growing a shared memory segment? The big question is: would it work? Would this be a valid means of interprocess communication? Or fraught with complexities? Thanks, Riskybiz.

2 1

[msm] interrupt_state with more then one event that ends interrupt?
by Alexander Mingalev 09 Jul '13

09 Jul '13

Hi, Subj, Is that possible? for example, one event continues processing, the another one finishes the whole state-machine (some clean-up actions etc). I tried struct Interrupt : public msmf::interrupt_state< boost::mpl::vector<Foo, Bar> > {}; and struct Interrupt : public msmf::interrupt_state< msmf::ActionSequence_< boost::mpl::vector<Foo, Bar> > > {}; - has no effect. Is there another way to achieve this?

3 5