January 2013 - Boost - lists.preview.boost.org

Static initialization of boost shared objects question
by Juan Carlos Franzoy 02 Jul '13

02 Jul '13

Hello. Thanks for reading my question. I fell into a deadlock in static initialization of boost shared objects. The deadlock occurs in a reentrant call to a function which contains a definition of a static variable. Apparenlty gcc protects the initialization of such variables with a condition variable. Looking at the c++ code nothing at all make anybody suspect about reentrancy. My "interpretation of the fact" may be titled 'something in gcc, ld, ld.so, c++language made the call to be reentrant' - There is a function boost::system::generic_category() that has a static variable V - V type is boost::system::annonymous-namespace::generic_error_category - boost::system::annonymous-namespace::generic_error_category is subclass of boost::system::error_category. - boost::system::annonymous-namespace::generic_error_category has no declared/defined constructor as well as all its superclasses - The compilation process generate the constructor boost::system::error_category::error_category() as a weak symbol inside every shared object that uses it. - There is a static variable boost::system::posix_category in one of the shared objects that is initialized calling boost::system::generic_category(). - when the static initialization of a shared object boost::system::posix_category get initialized - which calls boost::system::generic_category() - Inside boost::system::generic_category() starts the initialization of V which calls boost::system::error_category::error_category() - It may happend that the runtime calls an instance of boost::system::error_category::error_category() that resides in a shared object that is still not loaded - So that .so get loaded and it is called its own static initialization - which can call boost::system::generic_category() - BOOM I would love if you can confirm of refute my humble "interpretation". Details follows ------------------- boost version: 1-52-0 Operating system: uname -a SunOS bcpp-solaris10 5.10 Generic_147441-25 i86pc i386 i86pc $ gcc -v Using built-in specs. COLLECT_GCC=gcc COLLECT_LTO_WRAPPER=/export/home/local/gcc-4.6.2/libexec/gcc/i386-pc-solaris2.10/4.6.2/lto-wrapper Target: i386-pc-solaris2.10 Configured with: ../src/gcc-4.6.2/configure --prefix=/export/home/local/gcc-4.6.2 --with-gmp=/export/home/local/gcc-4.6.2/gmp-4.3.2 --with-mpfr=/export/home/local/gcc-4.6.2/mpfr-3.0.0 --with-mpc=/export/home/local/gcc-4.6.2/mpc-0.8.2 --with-libelf=/export/home/local/gcc-4.6.2/libelf-0.8.12 --enable-shared --enable-threads=posix --enable-checking=release --with-system-zlib --enable-__cxa_atexit --enable-languages=c,c++ --with-as=/usr/local/bin/as --with-gnu-as --with-ld=/usr/ccs/bin/ld --without-gnu-ld --enable-plugin --enable-multilib --enable-version-specific-runtime-libs --enable-bootstrap --enable-ssp --enable-libssp --with-included-gettext --enable-lto --host=i386-pc-solaris2.10 --build=i386-pc-solaris2.10 Thread model: posix gcc version 4.6.2 (GCC) Callstack backtrace (lines begining with = are my comments): = = deadlock here: the threads is waiting itself = #0 0xfd459019 in __lwp_park () from /lib/libc.so.1 #1 0xfd45365e in cond_sleep_queue () from /lib/libc.so.1 #2 0xfd453817 in cond_wait_queue () from /lib/libc.so.1 #3 0xfd453b94 in cond_wait_common () from /lib/libc.so.1 #4 0xfd453d02 in _cond_wait () from /lib/libc.so.1 #5 0xfd453d2d in cond_wait () from /lib/libc.so.1 #6 0xfd453d66 in pthread_cond_wait () from /lib/libc.so.1 #7 0xfd632898 in __gthread_cond_wait (g=0xfd822ea8) at ...build-gcc/out/i386-pc-solaris2.10/libstdc++-v3/include/i386-pc-solaris2.10/bits/gthr-default.h:846 #8 __gthread_cond_wait_recursive (g=0xfd822ea8) at ...work/build-gcc/out/i386-pc-solaris2.10/libstdc++-v3/include/i386-pc-solaris2.10/bits/gthr-default.h:860 #9 __gnu_cxx::__cond::wait_recursive (g=0xfd822ea8) at ...build-gcc/out/i386-pc-solaris2.10/libstdc++-v3/include/ext/concurrence.h:377 #10 __cxa_guard_acquire (g=0xfd822ea8) at .../src/gcc-4.6.2/libstdc++-v3/libsupc++/guard.cc:289 = = Reentrant call to boost::system::generic_category() = #11 0xfd812313 in boost::system::generic_category () at libs/system/src/error_code.cpp:425 #12 0xfe85cd38 in __static_initialization_and_destruction_0(__initialize_p=1, __priority=65535) at ./boost/system/error_code.hpp:214 #13 0xfe85d05f in _GLOBAL__sub_I_instantiate_re2c_lexer_str.cpp () at libs/wave/src/instantiate_re2c_lexer_str.cpp:56 = = static initization of boost.wave library = #14 0xfe8868fd in __do_global_ctors_aux () from ...boost/lib/libboost_wave-gcc46-mt-d-32-SOLARIS-X86-1_52.so.1.52.0 #15 0xfe88692a in _init () from ...boost/lib/libboost_wave-gcc46-mt-d-32-SOLARIS-X86-1_52.so.1.52.0 #16 0xfefce37e in call_init () from /usr/lib/ld.so.1 #17 0xfefce1cf in is_dep_init () from /usr/lib/ld.so.1 #18 0xfefdaa4b in elf_bndr () from /usr/lib/ld.so.1 #19 0xfefc2ab4 in elf_rtbndr () from /usr/lib/ld.so.1 #20 0xfee60688 in ?? () #21 0xfee0c8f6 in boost::exception_detail::get_static_exception_object<boost::exception_detail::bad_alloc_>() at ./boost/exception/detail/exception_ptr.hpp:119 #22 0xfee0c0d2 in __static_initialization_and_destruction_0(__initialize_p=1, __priority=65535) at ./boost/exception/detail/exception_ptr.hpp:138 #23 0xfee0c271 in _GLOBAL__sub_I_future.cpp () at libs/thread/src/future.cpp:60 = = static initization of boost.thread library = #24 0xfee0d98d in __do_global_ctors_aux () from ...boost/lib/libboost_thread-gcc46-mt-d-32-SOLARIS-X86-1_52.so.1.52.0 #25 0xfee0d9ba in _init () from ...boost/lib/libboost_thread-gcc46-mt-d-32-SOLARIS-X86-1_52.so.1.52.0 #26 0xfefce37e in call_init () from /usr/lib/ld.so.1 #27 0xfefce1cf in is_dep_init () from /usr/lib/ld.so.1 #28 0xfefdaa4b in elf_bndr () from /usr/lib/ld.so.1 #29 0xfefc2ab4 in elf_rtbndr () from /usr/lib/ld.so.1 #30 0xfdce0688 in ?? () = = First calll to boost::system::generic_category() = #31 0xfd81232a in boost::system::generic_category () at libs/system/src/error_code.cpp:425 #32 0xfd812386 in __static_initialization_and_destruction_0(__initialize_p=1, __priority=65535) at ./boost/system/error_code.hpp:214 #33 0xfd8123ed in _GLOBAL__sub_I_error_code.cpp () at libs/system/src/error_code.cpp:430 = = static initialization of boost.system = #34 0xfd8126ed in __do_global_ctors_aux () from ...boost/lib/libboost_system-gcc46-mt-d-32-SOLARIS-X86-1_52.so.1.52.0 #35 0xfd81271a in _init () from ...boost/lib/libboost_system-gcc46-mt-d-32-SOLARIS-X86-1_52.so.1.52.0 #36 0xfefce37e in call_init () from /usr/lib/ld.so.1 #37 0xfefcd7d5 in setup () from /usr/lib/ld.so.1 #38 0xfefdba3e in _setup () from /usr/lib/ld.so.1 #39 0xfefc299c in _rt_boot () from /usr/lib/ld.so.1 #40 0x08047540 in ?? () #41 0x08047670 in ?? () More thanks for reading all.

3 4

De Bruijn Bind (alternate bind syntax) Interest?
by David Sankel 24 Jun '13

24 Jun '13

Hello all, I've been working on an alternate bind syntax based on De Bruijn indices[1]. The syntax is very simple, yet the terms are very powerful. Here is an example of a function const that takes in an argument c and returns another function that, for all input, returns c: //λx.λy.x = λλ1 with De Bruijn indices. auto const_ = abs<1>( abs<1>( var<1,0>() ) ); More examples, further explanation, and an implementation are available here[2]. I'm thinking that this library could also be useful as a core for more syntax heavy bind variations. David [1] http://en.wikipedia.org/wiki/De_Bruijn_index [2] http://bitbucket.org/camior/de-bruijn-bind/ -- David Sankel Sankel Software www.sankelsoftware.com 585 617 4748 (Office)

11 73

[TypeErasure] references
by Chapman, Alec 10 May '13

10 May '13

Hi Steven, I have a somewhat-generic type erasure class that I would like to switch to yours. The main difficulty I can see is in the way capturing by reference is handled. The approach I take in my class is to pass a boost::reference_wrapper to the constructor: typedef mpl::vector<copy_constructible<>, type_id_<>, incrementable<> > requirements; int i = 0; any<requirements> x(boost::ref(i)); // capture i by reference ++x; // i == 1 any<requirements> y(x); // makes a copy of i ++y; // i == 1, any_cast<int>(y) == 2 My motivation is that I have a bunch of template functions already written that I often call directly, but in other cases I need to compile them separately by instantiating them with any<...>: ---------------------- module 1 ------------------------- // some long algorithm that I don't want to modify to be aware of any<...> template<class T> void f(T& t) { ++t; T t2 = t; ++t2; // do something with t2 } // make a type-erased version for separate compilation void any_f(any<requirements>& x) { return f(x); } -------------------- module 2 ---------------------------- // provide a thin wrapper that applies type erasure before calling any_f template<class T> void f_fwd(T& t) { any<requirements> x(boost::ref(t)); any_f(x); } // some function that needs to call f void g(int& i) { // do something with i... // then pass it to f f_fwd(i); } Is something like this possible with your method? Are there advantages that I'm not seeing to having the type signature contain information about whether the object is held by reference? Thanks, Alec

3 6

What Should we do About Boost.Test?
by Dave Abrahams 16 Apr '13

16 Apr '13

Hi All, I was just going through Boost.Test to try to figure out how to teach it, and while it looks to have substantial value, it is also in quite a mess. It contains loads of features that are exercised in the examples/ directory but neither included in any of the tests nor documented. There are facilities for command-line argument parsing! There are "decorators" that turn on/off features for test cases. There is support for mock objects! These are cool and sometimes necessary features, but who knew? The third tutorial page (http://www.boost.org/doc/libs/1_51_0/libs/test/doc/html/tutorials/new-year-…) has a glaring typo in the code examples: "BOOST_AUTO_EST_CASE". There's no reference manual at all. There are nearly-identical files in the examples/ directory called "est_example1.cpp" and "test_example1.cpp" (Did the "t" key on someone's keyboard break?) I could go on, but where would I stop? I don't know what to do about this. Because of the lack of redundancy (i.e. tests and documentation), it's hard to tell whether this library is correct or even to define what "correct" should mean. It seems like, as long as the code is incompletely / incorrectly documented and tested, it's just someone's personal coding project that we happen to keep shipping with Boost, and not really a library for general use. This situation reflects poorly on Boost as a whole and the fact that it centers around a _testing_ library, which is concerned with robustness... well, let's just say that the irony isn't lost on me. I don't mean this posting as an attack on Gennadiy in any way, but I think the situation is unacceptable and therefore am opening a discussion about what should happen. As a straw man, I'll make this suggestion: - Boost.Test is officially deprecated in the next release - Its documentation, such as it is, is removed from the release after that - Meanwhile, other tests in Boost that use this library are rewritten to use a different mechanism - The code is removed from Boost thereafter I am not at all attached to removing Boost.Test from Boost, but IMO rescuing it would require a significant new investment of time and energy from people who are committed to bringing the library up to par with the rest of what we do. (I seriously thought about volunteering for this myself, but realistically speaking, I don't have the time, and volunteering for something you can't actually do is worse than not volunteering at all.) Even if volunteers show up, I'd suggest proceeding with the plan above, subject to reversal at any time the work actually gets done. Thoughts? -- Dave Abrahams BoostPro Computing Software Development Training http://www.boostpro.com Clang/LLVM/EDG Compilers C++ Boost

34 85

future<>::then() Error returning future<void> (on trunk)
by Fernando Pelliccioni 07 Apr '13

07 Apr '13

Hi Vicente, hi all, I noticed that in the trunk there is a implementation of future<>::then(). Thank you very much for your efforts to implement it! The following example, compiles OK boost::future<int> f1 = boost::async([]() { return 123; }); f1.then([](boost::future<int> const& f) { return std::string(""); }); But this one fails to compile: boost::future<int> f1 = boost::async([]() { return 123; }); f1.then([](boost::future<int> const& f) { }); Here future<>::then() have to returns future<void> like ... boost::async([]() { }); Am I right? On the other hand, have you taken into account the following paper? http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2012/n3451.pdf Regards, Fernando Pelliccioni.

3 5

[fusion] What do we do with nil ?
by Joel de Guzman 31 Mar '13

31 Mar '13

Hello, This is a constantly nagging issue. There are a couple of places where nil is a PP define especially on the Mac: - objc.h #ifndef nil #define nil __DARWIN_NULL #endif and - MacTypes.h #ifndef nil #define nil NULL #endif So people complain because fusion uses nil as: namespace boost { namespace fusion { struct nil; }} So, nil is a reserved word in Objective-C. But this is C++! And that define is SUCH A BAD PRACTICE in C++! Some folks, like Paul Mensonides maintains a strong stance against this: I would just use #define nil nil as a sanity check and call it done. If the symbol is already defined, compilation error (or at least warning on overly permissive compilers). The symbol will be need to be undefined in between the inclusion of the two libraries. Workarounds like nil_ and (min)(...) just perpetuate the problem. Fix the root problem rather than treating the symptom indefinitely. Otherwise, where is the line drawn? Even if you can't provide a transparent fix (such as defining nil as a constant and min/max as functions), the backwards compatibility argument fails in the long run. The amount of effort required to make the necessary changes to existing code will eventually be outweighed by the effort to implement workarounds put into new code (as well as educate everyone that they are necessary)--especially in cases like these which can mostly be emulated in a way that causes most existing code to work. As I said over and over again, I'm listening and I would want to "fix" this issue if someone comes up with a reasonable (tested) patch that: 1) Detects the platform 2) Renames fusion::nil to something else (only for the affected platform) Yet, no-one wants to do this. Instead we get nasty remarks such as those found here: https://svn.boost.org/trac/boost/ticket/5010#comment:16 What do we do? I'm seeking for a resolution. Regards, -- Joel de Guzman http://www.boostpro.com http://boost-spirit.com

4 5

[config] consexpr workaround
by vicente.botet 21 Mar '13

21 Mar '13

Hi, in order to make easier the introduction of constexpr in a portable way I would like we add a Boost Config workarround. We can define some macros depending on whether BOOST_NO_CONSTEXPR is defined or not. #if defined(BOOST_NO_CONSTEXPR) #define BOOST_CONSTEXPR #define BOOST_CONSTEXPR_OR_CONST const #else #define BOOST_CONSTEXPR constexpr #define BOOST_CONSTEXPR_OR_CONST constexpr #endif For most of the constexpr functions or constructors BOOST_CONSTEXPR will work. constexpr tuple(); =versus= BOOST_CONSTEXPR tuple(); When used for static constants BOOST_CONSTEXPR_OR_CONST is needed as we need to preserv the const keyword when consexpr is not available. static constexpr UIntType xor_mask = a; =versus= static BOOST_CONSTEXPR_OR_CONST UIntType xor_mask = a; We could also add a STATIC_CONSTEXPR macro #define BOOST_STATIC_CONSTEXPR static BOOST_CONSTEXPR_OR_CONST static constexpr UIntType xor_mask = a; =versus= BOOST_STATIC_CONSTEXPR UIntType xor_mask = a; For this case the existing macro BOOST_STATIC_CONSTANT could be modified to use BOOST_CONSTEXPR_OR_CONST when static const was used. static constexpr UIntType xor_mask = a; =versus= BOOST_STATIC_CONSTANT(UIntType, xor_mask = a); Can we introduce this workaround? Best regards, Vicente _____________________ Vicente Juan Botet Escribá http://viboes.blogspot.com/

6 17

Help in setupping regression tests on ARM
by Antony Polukhin 16 Mar '13

16 Mar '13

I've successfully run nests on x86, but on ARM I've got the following message: # Running Boost.Build tests Traceback (most recent call last): File "run.py", line 60, in <module> runner(root) File "/home/pi/boost_maintain/run_trunk/tools_regression_src/regression.py", line 205, in __init__ self.main() File "/home/pi/boost_maintain/run_trunk/tools_regression_src/regression.py", line 566, in main getattr(self,action_m)() File "/home/pi/boost_maintain/run_trunk/tools_regression_src/regression.py", line 531, in command_regression self.command_test() File "/home/pi/boost_maintain/run_trunk/tools_regression_src/regression.py", line 329, in command_test self.command_test_boost_build() File "/home/pi/boost_maintain/run_trunk/tools_regression_src/regression.py", line 369, in command_test_boost_build "boost/bin.v2/libs/any/test/any_test.test")); OSError: [Errno 2] No such file or directory: '/home/pi/boost_maintain/run_trunk/results/boost/bin.v2/libs/any/test/any_test.test' I try to run tests on RaspberryPi, it is armhf. Compilller is GCC-4.6.3, 25Gb of free space on SD card. It looks strange, that I see only a few folders in /home/pi/boost_maintain/run_trunk/results/boost/bin.v2/libs/ (there are only 6 folders) -- Best regards, Antony Polukhin

2 7

[thread] thread_specific_ptr leaks on VS2012 when using std::async()
by Klaim - Joël Lamotte 05 Mar '13

05 Mar '13

Hi, The title says it all. I think I found a memory leak with the help of boost.log author Andrei Semashev. I was using Boost.Log in Visual Studio 2012 (released one) and tried to use Visual Leak Detector because Visual Studio basic leak detector was reporting leaks. I isolated the leaks to usage of log and switched to VLD to check if the basic detector wasn't giving me false positive. I contacted Andrei to see if he could check it but he don't have VS2012 available so I setup some tests to isolate the problem. In the end of our exchange we concluded it was a false positive (it only leak on end of program). However it is still an annoying noise in the leak report logs, so people not understanding the source of this problem might be interested in this email. I attach the whole email exchange to this email, tell me if leak logs are necessary too. The last version of my test is at the end of the email (first program). Basically, the leak appear only when using std::async() in combination of boost::thread_specific_ptr : - using exclusively boost threading libraries (both with dynamic and static linking) result in no leak - using exclusively std threading libraries but not std::async() result in no leak - using std::async() result in a leak The second test program shows that using Intel Threading Building Blocks, I get leaks only if : - I use std::async() in combination with boost::thread_specific_ptr, as before (I mean I use tbb without a leak, then I add a std::async() call to the same task function and there is a leak) - I don't use std::async() but I don't manage the lifetime of tbb task scheduler We think the threads spawned by both tbb (that is supposed to be a superset of ppl) and std::async() (that is supposed to use ppl/concrt task scheduler) are not ended yet when the program reach the end of main (which seems logical as the task scheduler is then managed as a static global), making boost::thread_specific_ptr not release some data on thread destruction, as it is notified too late and any leak detection tool willl generate noise due to this. A way to fix this is to explicitely end the task scheduler, being tbb or ppl/concrt. I will use tbb task scheduler explicitely initialized/terminated in my project so it will solve the problem on my side. Joel Lamotte ------------------------------------ // this program was used to see the memory leak #define TEST_ASYNC // comment this to make it linear and not leak //#define WITH_BOOST_THREADS #define WITH_STD_ASYNC // uncomment this to leak //#define WITH_BOOST_LOG // boost log also leak if std::async is used #include <boost/thread.hpp> #ifdef WITH_BOOST_THREADS # include <boost/thread/future.hpp> # include <boost/chrono.hpp> # define THIS_THREAD boost::this_thread # define THREAD_LIB boost::thread # define CHRONO_LIB boost::chrono # define MY_FUTURE boost::future #else # include <future> # include <thread> # include <chrono> # define THIS_THREAD std::this_thread # define THREAD_LIB std::thread # define CHRONO_LIB std::chrono # define MY_FUTURE std::future #endif #ifdef WITH_BOOST_LOG # include <boost/log/trivial.hpp> # define MY_LOG BOOST_LOG_TRIVIAL(info) #else # include <iostream> # define MY_LOG std::cout << '\n' #endif #define BASIC_VS_LEAK_DETECTOR 0 #define VISUAL_LEAK_DETECTOR 1 #define LEAK_DETECTOR VISUAL_LEAK_DETECTOR #if LEAK_DETECTOR == BASIC_VS_LEAK_DETECTOR # define _CRTDBG_MAP_ALLOC # include <stdlib.h> # include <crtdbg.h> #else # include "vld.h" #endif struct my_data { char data[1024]; my_data() { MY_LOG << "CONSTRUCTOR my_data - " << THIS_THREAD::get_id(); } ~my_data() { MY_LOG << "DESTRUCTOR my_data - " << THIS_THREAD::get_id(); } }; boost::thread_specific_ptr< my_data > ts_ptr; void my_thread_foo() { ts_ptr.reset(new my_data()); } int work() { for( int i = 0; i < 42; ++i ) { MY_LOG << "THIS IS A MESSAGE - I = " << i; MY_LOG << "ts_ptr = " << ts_ptr.get(); my_thread_foo(); THIS_THREAD::sleep_for( CHRONO_LIB::milliseconds(10) ); } return 0; } int main(int argc, char *argv[]) { #if LEAK_DETECTOR == BASIC_VS_LEAK_DETECTOR // First simple memory leak detection for Visual Studio // use _crtBreakAlloc to put a breakpoint on the provided memory leak id allocation //_crtBreakAlloc = 1149; _CrtSetDbgFlag ( _CRTDBG_ALLOC_MEM_DF | _CRTDBG_LEAK_CHECK_DF ); #endif //int* k = new int[42]; // to check that memory leak detection does work MY_LOG << "Hello, World!"; #ifdef TEST_ASYNC # ifndef WITH_BOOST_THREADS std::vector< std::future<int> > work_in_progress(2); # ifdef WITH_STD_ASYNC for( auto& ftr : work_in_progress ) { ftr = std::async( work ); } # else for( auto& ftr : work_in_progress ) { std::packaged_task< int() > task(work); ftr = task.get_future(); std::thread thread( std::move(task) ); thread.detach(); } # endif for( auto& ftr : work_in_progress ) { ftr.get(); } # else std::vector< boost::unique_future<int> > work_in_progress(42); for( auto& ftr : work_in_progress ) { boost::packaged_task<int> task(work); // boost's packaged_task don't take the full signature as parameter, it takes the return type only ftr = task.get_future(); boost::thread thread( boost::move(task) ); thread.detach(); } for( auto& ftr : work_in_progress ) { ftr.get(); } # endif #else for( int i = 0; i < 42; ++i ) { work(); } #endif MY_LOG << "END"; return EXIT_SUCCESS; } -------------------------------------- // this program tests if the memory leaks appear with Threading Building Blocks #define BASIC_VS_LEAK_DETECTOR 0 #define VISUAL_LEAK_DETECTOR 1 #define LEAK_DETECTOR VISUAL_LEAK_DETECTOR #define WITH_ROOT_TASK //#define WITH_BOOST_LOG //#define WITH_STD_ASYNC // uncomment to leak #ifdef WITH_STD_ASYNC # include <future> #endif #include <thread> #include <atomic> #include <boost/thread.hpp> #include <tbb/tbb.h> #include <tbb/task_scheduler_init.h> #ifdef WITH_BOOST_LOG # include <boost/log/trivial.hpp> # define MY_LOG BOOST_LOG_TRIVIAL(info) #else # include <iostream> # define MY_LOG std::cout << '\n' #endif #if LEAK_DETECTOR == BASIC_VS_LEAK_DETECTOR # define _CRTDBG_MAP_ALLOC # include <stdlib.h> # include <crtdbg.h> #else # include "vld.h" #endif namespace test { struct my_data { char data[1024]; }; boost::thread_specific_ptr< my_data > ts_ptr; void my_thread_foo() { ts_ptr.reset(new my_data()); } class UpdateTask : public tbb::task { public: UpdateTask() : idx( s_idx++ ) { MY_LOG << "[UpdateTask " << idx << " ] " << std::endl; } ~UpdateTask() { MY_LOG << "[/UpdateTask " << idx << " ]" << std::endl; } tbb::task* execute() { MY_LOG << "[" << std::this_thread::get_id() << "] Tick "<< m_count <<std::endl; my_thread_foo(); ++m_count; if( m_count < 100 ) { // make sure this is re-executed this->increment_ref_count(); this->recycle_as_safe_continuation(); } return nullptr; } private: static std::atomic<int> s_idx; std::atomic<int> m_count; const int idx; }; std::atomic<int> UpdateTask::s_idx = 0; class TaskScheduler // using TBB { public: #ifdef WITH_ROOT_TASK TaskScheduler() : m_root_task( *new (tbb::task::allocate_root()) tbb::empty_task ) { m_root_task.increment_ref_count(); } ~TaskScheduler() { m_root_task.destroy(m_root_task); } #endif void register_update_task() { #ifdef WITH_ROOT_TASK m_root_task.increment_ref_count(); UpdateTask& updater = *new(m_root_task.allocate_child()) UpdateTask(); #else UpdateTask& updater = *new(tbb::task::allocate_root()) UpdateTask(); #endif m_task_list.push_back( updater ); } void run_and_wait() { #ifdef WITH_ROOT_TASK m_root_task.spawn( m_task_list ); m_root_task.wait_for_all(); #else tbb::task::spawn_root_and_wait( m_task_list ); #endif } private: tbb::task_list m_task_list; // uncomment the following line to not leak tbb //tbb::task_scheduler_init m_engine_instance; // We explicitly create and destroy the TBB task scheduler to avoid memory leak warning noises. #ifdef WITH_ROOT_TASK tbb::task& m_root_task; #endif }; void tbb_test() { TaskScheduler task_scheduler; for( int i = 0; i < 50; ++i ) { task_scheduler.register_update_task(); } task_scheduler.run_and_wait(); } } int main(int argc, char *argv[]) { #if LEAK_DETECTOR == BASIC_VS_LEAK_DETECTOR // First simple memory leak detection for Visual Studio // use _crtBreakAlloc to put a breakpoint on the provided memory leak id allocation //_crtBreakAlloc = 148; _CrtSetDbgFlag ( _CRTDBG_ALLOC_MEM_DF | _CRTDBG_LEAK_CHECK_DF ); #endif #ifdef WITH_STD_ASYNC std::atomic<bool> finished = false; std::async( [&]{ while( !finished ) { MY_LOG << "POP"; test::my_thread_foo(); std::this_thread::sleep_for( std::chrono::milliseconds(500) ); } } ); #endif test::tbb_test(); #ifdef WITH_STD_ASYNC finished = true; #endif MY_LOG << "\nEND"; std::this_thread::sleep_for( std::chrono::seconds(4) ); return 0; }

3 5

[thread] boost::thread and thread safety annotations
by Wladimir 04 Mar '13

04 Mar '13

Hello, The Clang C++ compiler project is currently working on a thread safety analyzer. By making use of so-called thread safety annotations within the source code, several problems with synchronization primitives can be found statically. For example: usage of a data structure without acquiring its guarding lock. See here for an overview of the available attributes: http://clang.llvm.org/docs/LanguageExtensions.html#threadsafety It is customary to use these attributes by defining macros, for compatibility: http://www.mail-archive.com/linuxkernelnewbies@googlegroups.com/msg01455.ht… At one time, it was also supposed to end up in G++, but this was abandoned in favor of clang (see http://gcc.gnu.org/wiki/ThreadSafetyAnnotation) But at least the annotations are fairly standardized. I think it would be useful to add these annotations to boost::thread, so that projects using that library can make use of the static checking. As I understand it, is only a matter of marking data structures that serve as lock/scoped_lock (such as boost::mutex, boost::recursive_mutex, ...), and methods that act upon the lock with the appropriate attributes. Any interest in this? Regards, Wladimir

2 1