- Boost-users - lists.boost.org

boost::asio handler based on received data
by Abdo Haji-Ali 16 Feb '09

16 Feb '09

Greetings, A server that I'm communicating with sends two types of messages identified by an ID. The first message type is sent automatically on specific events. The second type is sent in response to commands. I want to provide an interface that notifies the user when events occur (asynchronously) and at the same time enable performing commands. My first implementation did this (pseudocode): // In initialization socket.async_receive(buffer, event_handler); // In send_command socket.send(command_buffer); socket.receive_async(response_buffer, user_response_handler); event_handler() { user_event_handler(); socket.async_receive(buffer, event_handler); } The problem here is that 'event_handler' might (or will?) receive the response message. An additional problem is that these messages have no specific order. In other words, the following sequence might occur: - Send command (A) - Send command (B) - Receive event notification - Receive response (A) - Receive response (B) A possible solution might be the following: // In initialization socket.async_receive(buffer, msg_handler); // In send_command socket.send(command_buffer); handlers_queue.enqueue(user_command_handler); socket.receive_async(response_buffer, msg_handler); // Line(3) msg_handler() { if (event) user_event_handler(); else handlers_queue.dequeue() (); // Call first user handler socket.async_receive(buffer, msg_handler); } I want to know if boost::asio provides a better way to solve this problem. I'm particularly interested if writing a custom io service can help me in this situation. Thanks you, -- Abdo Haji-Ali Programmer In|Framez

1 0

[Lambda] Misplaced "const" causes runtime error?
by Paul Baker 16 Feb '09

16 Feb '09

Hi, The following program, as expected, outputs "1 1 1": #include <algorithm> #include <iostream> #include <vector> #include <boost/mem_fn.hpp> #include <boost/bind.hpp> #include <boost/lambda/bind.hpp> int main() { std::vector<std::pair<int, int> > pairs; pairs.push_back(std::make_pair(1, 1)); std::vector<int> result; std::transform(pairs.begin(), pairs.end(), std::back_inserter(result), boost::mem_fn(&std::pair<int, int>::first)); std::transform(pairs.begin(), pairs.end(), std::back_inserter(result), boost::bind(&std::pair<int, int>::first, _1)); std::transform(pairs.begin(), pairs.end(), std::back_inserter(result), boost::lambda::bind(&std::pair<int, int>::first, boost::lambda::_1)); std::cout << result[0] << ' ' << result[1] << ' ' << result[2]; } Changing "std::pair<int, int>" to "std::pair<const int, int>" on the line using boost::mem_fn or the line using boost::bind causes a compile-time error. However, performing the same change on the line using boost::lambda::bind leads to a successful compile - yet at run-time result[2] is set to a seemingly random value. This has been tested on Visual C++ 7.1 SP1, 8.0 SP1 and 9.0 SP1. Is this expected behaviour? I was expecting to see either a compile-time error or run-time success. Thanks

2 1

[boost-users][signals]
by Germán Diago 16 Feb '09

16 Feb '09

Hello. I'm trying to implement a similar system to signals/slots found in qt, with connections at runtime. Now I need to know the argument types for a given signal, like boost::function<>::first_arg_type and so on. Is this possible, and how? thank you very much.

3 2

[asio] read with transfer_at_least(300) return after 256 characters with no error
by tomasz jankowski 16 Feb '09

16 Feb '09

hello, how it is possible that size = boost::asio::read(socket_, boost::asio::buffer(serverReply_,300),boost::asio::transfer_at_least(300)); return with no error after reading 256? thanks, tom ---------------------------------------------------- Inżynierskie Targi Pracy 26.02.2009 http://klik.wp.pl/?adr=http%3A%2F%2Fwww.itp.BEST.krakow.pl&sid=643

1 0

Importing namespaced templates
by Esben Mose Hansen 16 Feb '09

16 Feb '09

Hi, from the C++0x standard, 7.3.3 (5) I gather that it is not possible to write a using statement to import a templated function into the global namespace. E.g. #include <boost/units/quantity.hpp> using template<typename X, typename Y> boost::units::quantity_cast<X,Y>; will be malformed. The only alternatives I have been able to come up with are 1) Use boost::units::quantity_cast<X>(Y) which is a bit longwinded 2) Write a new quantity_cast in the global namespace, and forward it to the boost::units one. Is there a better way? (Sorry that this is only borderline a boost question) -- kind regards, Esben

2 1

FindBoost.cmake
by Ruediger Berlich 16 Feb '09

16 Feb '09

Hi there, I'm using CMake as my build system -- not to build boost, but build my own project, which heavily depends on Boost. I habe in the past used the FindBoost.cmake macro to locate the most suitable Boost version. However, there does not seem to be a centrally maintained version of this file (I believe it originated from the KDE project ?), and I was wondering whether "anyone" in Boost was working on a replacement. After all, if CMake is considered as an alternative build system for Boost itself, allowing users to find the installed libraries would be the most logical step. Thanks and Best Regards, Ruediger

2 3

Seeking writers to develop quarterly digests of Boost newsgroups
by Brian Wood 16 Feb '09

16 Feb '09

I'm planning to start an on line magazine called C++ Quarterly. *http://tinyurl.com/df6hn2 or **http://preview.tinyurl.com/df6hn2 * If you are interested in writing a quarterly digest of Boost Users and/or Boost Developers, please let me know. Alternatively, if you're interested in writing a non-digest article, I'd be happy to hear from you. Just to clarify something from the above link. I want to set things up so that advertisers send money directly to authors. Advertisers would send ads to us but money to authors. Brian Wood Ebenezer Enterprises www.webEbenezer.net <http://www.webebenezer.net/>

1 0

[Regex]static compilation problems
by Moritz Beber 15 Feb '09

15 Feb '09

Dear users, to start using regular expressions I tried a very simple program only declaring one regular expression. #include <boost/regex.hpp> int main(int , char* []) { boost::regex re("\\d+"); return 0; } Linking and compiling this with the command: g++ test.cpp -o test -L ~/libs/boost/lib/ -l boost_regex-gcc43-mt works without problems, but running it then leads to the following: error while loading shared libraries: libboost_system-gcc43-mt-1_37.so.1.37.0: cannot open shared object file: No such file or directory But the directory clearly shows its existence: -rw-rw-r-- 1 engineer engineer 1587006 2008-11-21 07:53 /home/engineer/libs/boost/lib/libboost_regex-gcc43-mt-1_37.a lrwxrwxrwx 1 engineer engineer 38 2008-11-21 07:50 /home/engineer/libs/boost/lib/libboost_regex-gcc43-mt-1_37.so -> libboost_regex-gcc43-mt-1_37.so.1.37.0 -rwxrwxr-x 1 engineer engineer 722184 2008-11-21 07:50 /home/engineer/libs/boost/lib/libboost_regex-gcc43-mt-1_37.so.1.37.0 lrwxrwxrwx 1 engineer engineer 30 2008-11-21 07:53 /home/engineer/libs/boost/lib/libboost_regex-gcc43-mt.a -> libboost_regex-gcc43-mt-1_37.a lrwxrwxrwx 1 engineer engineer 38 2008-11-21 07:52 /home/engineer/libs/boost/lib/libboost_regex-gcc43-mt.so -> libboost_regex-gcc43-mt-1_37.so.1.37.0 Then I tried doing the same using static libraries: g++ -static test.cpp -o test -L ~/libs/boost/lib/ -l boost_regex-gcc43-mt but got an extremely long list of errors: /home/engineer/libs/boost/lib//libboost_regex-gcc43-mt.a(static_mutex.o): In function `boost::scoped_static_mutex_lock::unlock()': static_mutex.cpp:(.text+0x16): undefined reference to `pthread_mutex_unlock' /home/engineer/libs/boost/lib//libboost_regex-gcc43-mt.a(static_mutex.o): In function `boost::scoped_static_mutex_lock::lock()': static_mutex.cpp:(.text+0x46): undefined reference to `pthread_mutex_lock' /home/engineer/libs/boost/lib//libboost_regex-gcc43-mt.a(static_mutex.o): In function `boost::scoped_static_mutex_lock::~scoped_static_mutex_lock()': static_mutex.cpp:(.text+0x76): undefined reference to `pthread_mutex_unlock' /home/engineer/libs/boost/lib//libboost_regex-gcc43-mt.a(static_mutex.o): In function `boost::scoped_static_mutex_lock::~scoped_static_mutex_lock()': static_mutex.cpp:(.text+0xa6): undefined reference to `pthread_mutex_unlock' /home/engineer/libs/boost/lib//libboost_regex-gcc43-mt.a(static_mutex.o): In function `boost::scoped_static_mutex_lock::scoped_static_mutex_lock(boost::static_mutex&, bool)': static_mutex.cpp:(.text+0xdd): undefined reference to `pthread_mutex_lock' /home/engineer/libs/boost/lib//libboost_regex-gcc43-mt.a(static_mutex.o): In function `boost::scoped_static_mutex_lock::scoped_static_mutex_lock(boost::static_mutex&, bool)': static_mutex.cpp:(.text+0x10d): undefined reference to `pthread_mutex_lock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libstdc++.a(locale.o): In function `__gnu_cxx::__scoped_lock::~__scoped_lock()': (.text._ZN9__gnu_cxx13__scoped_lockD1Ev[__gnu_cxx::__scoped_lock::~__scoped_lock()]+0x18): undefined reference to `pthread_mutex_unlock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libstdc++.a(locale.o): In function `std::locale::_Impl::_M_install_cache(std::locale::facet const*, unsigned int)': (.text._ZNSt6locale5_Impl16_M_install_cacheEPKNS_5facetEj+0x2d): undefined reference to `pthread_mutex_lock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libstdc++.a(locale.o): In function `std::locale::_Impl::_M_install_cache(std::locale::facet const*, unsigned int)': (.text._ZNSt6locale5_Impl16_M_install_cacheEPKNS_5facetEj+0x59): undefined reference to `pthread_mutex_unlock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libstdc++.a(locale_init.o): In function `std::locale::locale()': (.text._ZNSt6localeC1Ev+0x2a): undefined reference to `pthread_mutex_lock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libstdc++.a(locale_init.o): In function `std::locale::locale()': (.text._ZNSt6localeC1Ev+0x52): undefined reference to `pthread_mutex_unlock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libstdc++.a(locale_init.o): In function `std::locale::global(std::locale const&)': (.text._ZNSt6locale6globalERKS_+0x2a): undefined reference to `pthread_mutex_lock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libstdc++.a(locale_init.o): In function `std::locale::global(std::locale const&)': (.text._ZNSt6locale6globalERKS_+0x9a): undefined reference to `pthread_mutex_unlock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libstdc++.a(locale_init.o): In function `std::locale::locale()': (.text._ZNSt6localeC2Ev+0x2a): undefined reference to `pthread_mutex_lock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libstdc++.a(locale_init.o): In function `std::locale::locale()': (.text._ZNSt6localeC2Ev+0x52): undefined reference to `pthread_mutex_unlock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libstdc++.a(eh_alloc.o): In function `__cxa_free_exception': (.text.__cxa_free_exception+0x6b): undefined reference to `pthread_mutex_lock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libstdc++.a(eh_alloc.o): In function `__cxa_free_exception': (.text.__cxa_free_exception+0x86): undefined reference to `pthread_mutex_unlock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libstdc++.a(eh_alloc.o): In function `__cxa_allocate_exception': (.text.__cxa_allocate_exception+0x6f): undefined reference to `pthread_mutex_lock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libstdc++.a(eh_alloc.o): In function `__cxa_allocate_exception': (.text.__cxa_allocate_exception+0xd7): undefined reference to `pthread_mutex_unlock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libgcc_eh.a(unwind-dw2-fde-glibc.o): In function `__register_frame_info_bases': (.text+0x60): undefined reference to `pthread_mutex_lock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libgcc_eh.a(unwind-dw2-fde-glibc.o): In function `__register_frame_info_bases': (.text+0x77): undefined reference to `pthread_mutex_unlock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libgcc_eh.a(unwind-dw2-fde-glibc.o): In function `__register_frame_info_table_bases': (.text+0x12a): undefined reference to `pthread_mutex_lock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libgcc_eh.a(unwind-dw2-fde-glibc.o): In function `__register_frame_info_table_bases': (.text+0x141): undefined reference to `pthread_mutex_unlock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libgcc_eh.a(unwind-dw2-fde-glibc.o): In function `_Unwind_Find_FDE': (.text+0x18a8): undefined reference to `pthread_mutex_lock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libgcc_eh.a(unwind-dw2-fde-glibc.o): In function `_Unwind_Find_FDE': (.text+0x18f8): undefined reference to `pthread_mutex_unlock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libgcc_eh.a(unwind-dw2-fde-glibc.o): In function `__deregister_frame_info_bases': (.text+0x1aaa): undefined reference to `pthread_mutex_lock' /usr/lib/gcc/i386-redhat-linux/4.3.0/libgcc_eh.a(unwind-dw2-fde-glibc.o): In function `__deregister_frame_info_bases': (.text+0x1b2f): undefined reference to `pthread_mutex_unlock' collect2: ld returned 1 exit status If anyone could point me to a solution of this problem, or give one here I'd very much appreciate it. Thank you. Best, Moritz

2 1

[Preprocessor] Questiona bout reentrant call and file iteration performance
by Joel Falcou 15 Feb '09

15 Feb '09

I've been playing with Boost PreProcessor for some times now and it really helped. Two corners still looks obscure for me as the doc doesn't provide clear-cut enough use-cases for my taste. 1/ Concerning the _D version of some macro, what's their use case ? I can see the example in the doc but I can't see why useing the normal version of the _D function should fail or behave badly 2/ Concerning file iteration vs BOOST_PP_REPEAT, what's the trade-off of one vs the other ? Is iterating faster at compile time than REPEATing ? Is there features of iteration that can't be done using REPEAT (or the reverse) or is the only bonus of ITERATE the fact that error actually send you in code and not on macro call site ? Thanks in advance -- ___________________________________________ Joel Falcou - Assistant Professor PARALL Team - LRI - Universite Paris Sud XI Tel : (+33)1 69 15 66 35

2 1

boost interprocess / shared memory question
by Naveen Santhanam 14 Feb '09

14 Feb '09

Hello, We have a windows application (a process) that requires a large data file (approx. 100 MB) to perform a task. We have a master that launches several instances of this application. We would like to share the large data file across the processes such that, there is just one instance of the data file in physical memory without any disc access. Hence each application will have a small footprint in memory during execution. We have tried using a managed shared memory object from the boost::interprocess library to achieve this. However, this doesn't help us as the memory footprint of the application goes up upon access of the data file. Is there a way to achieve the desired single footprint memory sharing? Does this work as expected with other OSes? Naveen

2 1