September 2011 - Boost-users

[boost.thread?][1.46.1] "memory clobbered before allocated block"
by Igor Gorbunov 07 Sep '11

07 Sep '11

Hello, I've ran across a strange bug. GCCs mcheck reports "memory clobbered before allocated block". Reproducible (not every run) with the following test program when executed from gdb: ----- #include <boost/thread.hpp> void proc_func(){ std::vector<int> v1(1024); } int main() { std::vector<int> v2(1024); boost::thread t1(proc_func); t1.join(); } ----- The same test program runs fine (or I'm not been able to trigger this bug) when recompiled with std.thread. I'm in doubt that there is a bug in boost.thread, so maybe build system misconfiguration or gdb bug? stack trace: /home/gie/work/tests/memd1/debug/test [18157] [cores: 0] Thread [2] 18160 [core: 0] (Suspended : Signal : SIGABRT:Aborted) *__GI_raise() at raise.c:64 0x7ffff70d0405 *__GI_abort() at abort.c:92 0x7ffff70d3680 __libc_message() at libc_fatal.c:189 0x7ffff7106dbb *__GI___libc_fatal() at libc_fatal.c:200 0x7ffff7106e9e mabort() at mcheck.c:366 0x7ffff7117aea checkhdr() at mcheck.c:117 0x7ffff71172eb freehook() at mcheck.c:190 0x7ffff71175dc __nptl_deallocate_tsd() at pthread_create.c:155 0x7ffff6e88129 start_thread() at pthread_create.c:311 0x7ffff6e88b4e clone() at clone.S:112 0x7ffff717336d 0x0 Thread [1] 18157 [core: 0] (Suspended : Container) pthread_join() at pthread_join.c:89 0x7ffff6e89e65 boost::thread::join() at thread.cpp:234 0x7ffff7bd2bba main() at main.cpp:12 0x409212 build log: /usr/bin/make all /usr/bin/cmake -H/home/gie/work/tests/memd1/src -B/home/gie/work/tests/memd1/debug --check-build-system CMakeFiles/Makefile.cmake 0 /usr/bin/cmake -E cmake_progress_start /home/gie/work/tests/memd1/debug/CMakeFiles /home/gie/work/tests/memd1/debug/CMakeFiles/progress.marks /usr/bin/make -f CMakeFiles/Makefile2 all make[1]: Entering directory `/home/gie/work/tests/memd1/debug' /usr/bin/make -f CMakeFiles/test.dir/build.make CMakeFiles/test.dir/depend make[2]: Entering directory `/home/gie/work/tests/memd1/debug' cd /home/gie/work/tests/memd1/debug && /usr/bin/cmake -E cmake_depends "Unix Makefiles" /home/gie/work/tests/memd1/src /home/gie/work/tests/memd1/src /home/gie/work/tests/memd1/debug /home/gie/work/tests/memd1/debug /home/gie/work/tests/memd1/debug/CMakeFiles/test.dir/DependInfo.cmake --color= make[2]: Leaving directory `/home/gie/work/tests/memd1/debug' /usr/bin/make -f CMakeFiles/test.dir/build.make CMakeFiles/test.dir/build make[2]: Entering directory `/home/gie/work/tests/memd1/debug' /usr/bin/cmake -E cmake_progress_report /home/gie/work/tests/memd1/debug/CMakeFiles 1 [100%] Building CXX object CMakeFiles/test.dir/main.cpp.o /usr/bin/c++ -g -I/home/gie/work/tests/memd1/debug -std=c++0x -Wall -Wextra -pedantic -o CMakeFiles/test.dir/main.cpp.o -c /home/gie/work/tests/memd1/src/main.cpp Linking CXX executable test /usr/bin/cmake -E cmake_link_script CMakeFiles/test.dir/link.txt --verbose=1 /usr/bin/c++ -g CMakeFiles/test.dir/main.cpp.o -o test -rdynamic -lmcheck -lboost_thread-mt make[2]: Leaving directory `/home/gie/work/tests/memd1/debug' /usr/bin/cmake -E cmake_progress_report /home/gie/work/tests/memd1/debug/CMakeFiles 1 [100%] Built target test make[1]: Leaving directory `/home/gie/work/tests/memd1/debug' /usr/bin/cmake -E cmake_progress_start /home/gie/work/tests/memd1/debug/CMakeFiles 0 -- With best regards, Igor.

1 0

Compiler errors using boost/program_options with Boost v 1.47
by Jakob S. 07 Sep '11

07 Sep '11

Dear all, I use boost/program_options in one of my programs (I compile under Linux with gcc 4.4.3). With Boost 1.46 everything worked fine, but now I installed Boost 1.47 and I get a bunch of error messages: btw, include/Config.hh:28 is simply: Sorry for the plain question, but: Could you help me to fix this? Am I missing something in my installation? Has boost/program_options not be correctly updated to 1.47? Many thanks for your help... -- View this message in context: http://boost.2283326.n4.nabble.com/Compiler-errors-using-boost-program-opti… Sent from the Boost - Users mailing list archive at Nabble.com.

3 2

Example HTTP Server 3, Requests are processed one by one on Win32?
by Sven 07 Sep '11

07 Sep '11

Hi, I'm new to boost and it's mailing list and I really like the library! As my first example project, I want to run the boost HTTP Server 3. Interesting points for me are Socket and Threading stuff. To get into the threading stuff, I did add some output to the std::cout in the request_handler part at the top of the method "handle_request" following by a sleep(5000). After this I compile (without errors/warnings) and run the program providing Ip/Port, docroot and threadpool size 20. When I start a request to it with my browser, as expected I get the output on the console. But when I start some requests concurrent, the requests seems processed one by one ( not concurrent as expected). I don't know, but I expect, that the Requests would be processed concurrent (1 Request = 1 Thread?) I'm using Visual Studio 10 with boost 1.47.0 on windows server 2k8 rc2 and following example code: http://www.boost.org/doc/libs/1_47_0/doc/html/boost_asio/examples.html#boost _asio.examples.http_server_3 Any suggestions? Thanks in advance Sven

3 6

[MultiArrrat] Unbalanced multi_array<>
by Júlio Hoffimann 07 Sep '11

07 Sep '11

Dear all, By reading Boost.MultiArray docs, i haven't figured out any possible way to deal with unbalanced matrices. The following picture explains what i mean by unbalanced: // two dimensional unbalanced matrix: a00 a01 a02 a03 a04 a05 a06 a07 a08 a10 a11 a12 a13 a14 a20 a21 a22 a30 a31 a40 Each row has it's own extent (defined at runtime). Seems multi_array<> classes are not designed to increase it's subcontainers decoupled, and at a given depth, they always share the same extent? Could you confirm or guide me in how to achieve this behavior? I really need to save memory due to large 3D matrices. Regards, Júlio.

2 2

Wave Preprocessor without Using a Lexer
by Alexander Feldman 07 Sep '11

07 Sep '11

Does anybody know how to use the wave preprocessor *without* using the cpp lexer? Best, -- Alex

1 0

boost::lexical_cast: where to declare operator<<
by Nagaev Boris 07 Sep '11

07 Sep '11

Hello! I'm implementing boost::lexical_cast for class A. When operator<< is declared in the same namespace, as class A, it works: * #include <boost/lexical_cast.hpp> * #include <string> * * namespace abc { * * struct A { * int a; * }; * * std::ostream& operator<<(std::ostream& o, const A& a) { * return o << a.a; * } * * } * * int main() { * abc::A a; * std::string a_str = boost::lexical_cast<std::string>(a); * } But if operator<< is declared in root namespace, it will not compile: * #include <boost/lexical_cast.hpp> * #include <string> * * namespace abc { * * struct A { * int a; * }; * * } * * std::ostream& operator<<(std::ostream& o, const abc::A& a) { * return o << a.a; * } * * int main() { * abc::A a; * std::string a_str = boost::lexical_cast<std::string>(a); * } Part of error message: In file included from error.cpp:2:0: /usr/include/boost/lexical_cast.hpp: In member function ‘bool boost::detail::lexical_stream<Target, Source, Traits>::operator<<(const Source&) [with Target = std::basic_string<char>, Source = abc::A, Traits = std::char_traits<char>]’: /usr/include/boost/lexical_cast.hpp:1151:13: instantiated from ‘Target boost::detail::lexical_cast(typename boost::call_traits<Source>::param_type, CharT*, size_t) [with Target = std::basic_string<char>, Source = abc::A, bool Unlimited = true, CharT = char, typename boost::call_traits<Source>::param_type = const abc::A&, size_t = long unsigned int]’ /usr/include/boost/lexical_cast.hpp:1174:77: instantiated from ‘Target boost::lexical_cast(const Source&) [with Target = std::basic_string<char>, Source = abc::A]’ error.cpp:19:59: instantiated from here /usr/include/boost/lexical_cast.hpp:595:48: error: no match for ‘operator<<’ in ‘((boost::detail::lexical_stream<std::basic_string<char>, abc::A, std::char_traits<char> >*)this)->boost::detail::lexical_stream<std::basic_string<char>, abc::A, std::char_traits<char> >::stream << input’ /usr/include/c++/4.5/ostream:108:7: note: candidates are: So, it does not see my declaration. However, <code>std::cout << a << std::endl;</code> works in both cases. What is the reason of this error? Why operator<< can not be declared in root namespace? Sorry for my question. Maybe, it is connected less to boost than to c++. Regards.

2 1

[Interprocess] hang locking p_hdr->m_mutex
by David Byron 06 Sep '11

06 Sep '11

I've got my machine (Win 7) into a state where using a message queue results in a hang. More specifically, calls to: scoped_lock<interprocess_mutex> lock(p_hdr->m_mutex); hang forever. As a bit more background, there's a message queue that holds up to 5 x 1024 byte messages, and I believe currently has 2. I'm having a hard time verifying that because calls to get_num_msg() hang trying on calls like the one above, but that's what I see when I look in c:/ProgramData/boost_interprocess/20110816195406.550541/CliIPC. I have feeling what happened is that some process that was using the message queue got terminated while holding the p_hdr->m_mutex. Does anyone have any suggestions for recovering from this? I'm using version 1.47. Thanks for your help. -DB

2 10

Which encryption algorithm is used in Boost::Asio and SSL?
by Andi Clemens 06 Sep '11

06 Sep '11

Hi, I searched the web for some time now, but I was not able to find out which encryption algorithm is used when using SSL with Boost::Asio. Is there a standard algorithm or do the communicating sockets handle out some method? Andi

1 0

[Spirit] compilation problem
by Igor R 06 Sep '11

06 Sep '11

Hello, I've upgraded to 1.47, and any attempt to include a Spirit header produces the following compiler error (MSVC 10): #include <boost/spirit/include/qi.hpp> int main() { } "../spirit/home/support/detail/endian.hpp(22): fatal error C1083: Cannot open include file: 'boost/integer/endian.hpp': No such file or directory" The error is in the following lines: // If Boost has the endian library, use it, otherwise use an adapted version // included with Spirit #if BOOST_VERSION >= 104600 #include <boost/integer/endian.hpp> #else #include <boost/spirit/home/support/detail/integer/endian.hpp> #endif Actually, there's no boost/integer/endian.hpp file in the zip that I've downloaded. What am I doing wrong? Thanks.

1 1

[serialization] boost::serialization adds huge amounts of exports to resultant Windows PE file
by Matheus Peschke de Azevedo 06 Sep '11

06 Sep '11

Chris, Boost serialization defaults lots of classes/functions to be automatically exported. This is done in \boost\serialization\force_include.hpp: #if defined(BOOST_HAS_DECLSPEC) && !defined(__COMO__) # if defined(__BORLANDC__) # define BOOST_DLLEXPORT __export # else # define BOOST_DLLEXPORT __declspec(dllexport) # endif #elif ! defined(_WIN32) && ! defined(_WIN64) # if defined(__MWERKS__) # define BOOST_DLLEXPORT __declspec(dllexport) # elif defined(__GNUC__) && (__GNUC__ >= 3) # define BOOST_USED __attribute__ ((used)) # elif defined(__IBMCPP__) && (__IBMCPP__ >= 1110) # define BOOST_USED __attribute__ ((used)) # elif defined(__INTEL_COMPILER) && (BOOST_INTEL_CXX_VERSION >= 800) # define BOOST_USED __attribute__ ((used)) # endif #endif The easiest workaround for this is: 1. Place your serialization #include lines in a precompiled header. 2. Right after all your boost serialization #includes, write this: #ifdef BOOST_DLLEXPORT #undef BOOST_DLLEXPORT // it was '__declspec(dllexport)' #define BOOST_DLLEXPORT // now it's '' #endif Just rebuild your project and you'll see all exports gone. No need to recompile boost serialization lib as functions affected are templates. Regards, Matheus

1 0