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

How to reuse system_timer of boost::asio with async_wait?
by Bruce 03 Jan '17

03 Jan '17

Hi, I am doing a program that use asio to construct a timer thread. The thread use asio to async_wait many timers. These timers are constructed by boost::asio::system_timer. These timers are mostly designed to fire only once. When they are fired or canceled, I put them to a free queue to reuse. But when I reuse these times, I called expire_from_now() and async_wait() again on the timer, I got the timer fired immediately with error_code set to "ec == boost::asio::error::operation_aborted". And the timer are not fired after like 5 seconds late as I set to expire_from_now(seconds(5)). How to reuse the timer correctly? My codes basically are as follows: m_timer.cancel(); // I tried to cancel the timer first, but no use, I still get the timer fired at once m_timer.expires_from_now(std::chrono::seconds(nInterval)); m_timer.async_wait(boost::bind(&qmInnerTimer::Fire, this, _1)); Thanks. -Bruce

1 0

Boost.Context: SIGSEGV when switch a thread within a catch block
by Александр Пронченков 02 Jan '17

02 Jan '17

Hi, I get SIGSEGV on macOS when I switch an execution_context from a catch block with one thread and than switch back with another thread. Segmentation fault happens in libc++abi.dylib`__cxa_end_catch Does anyone have any ideas, what it could be? Is it legal to switch thread under an excecution_context while it's within a catch block? A minimal example is here: #include <boost/context/all.hpp> #include <future> using Context = boost::context::execution_context<void>; Context Main(Context context) { try { throw std::runtime_error("123"); } catch (const std::exception& ex) { context = context(); } return context(); } int main() { Context context(&Main); context = context(); std::thread([&] { context = context(); }).join(); return 0; } I see the problem on macOS with clang++ but can't reproduce on a linux with g++. $ uname -ap Darwin air.local 16.3.0 Darwin Kernel Version 16.3.0: Thu Nov 17 20:23:58 PST 2016; root:xnu-3789.31.2~1/RELEASE_X86_64 x86_64 i386 $ clang++ --version Apple LLVM version 8.0.0 (clang-800.0.42.1) Target: x86_64-apple-darwin16.3.0 Thread model: posix InstalledDir: /Library/Developer/CommandLineTools/usr/bin -- Thanks, Alexander Pronchenkov

3 4

[Stacktrace] Review Manager report
by Niall Douglas 02 Jan '17

02 Jan '17

Firstly happy new year to everyone, and many thanks to all those who participated in the peer review of proposed Boost.Stacktrace before Christmas. I received the following formal reviews: 1. Vladimir Batov (YES) 2. Klemens Morgenstern (YES) 3. Andrey Semashev (NO) 4. Nat Goodspeed (YES) 5. Artyom Beilis (NO) 6. Barrett Adair (YES) Extensive comments without a formal acceptance recommendation were also made by: 1. Peter Dimov 2. Edward Diener There was a very noticeable lack of consensus on the strengths and weaknesses of the proposed library which has made summarising all these reviews into a coherent report with concrete conditions for acceptance very difficult. Antony also responded to many of the concerns throughout the review, agreeing with some points, disagreeing with others. I think this lack of consensus stemmed from multiple opinions on what a stacktrace library ought to provide: Minimum viable product: A portable backtrace() and backtrace_symbols() implementation with a C++ STL idiomatic API. The stack backtrace capture implementation must not allocate memory and be "fast". Note that backtrace_symbols() does NOT symbolise into strings containing source file name and line number. It also allocates memory (backtrace_symbols_fd() does not). Opinion 1: Portable serialisation of the stacktrace with instruction pointers converted to offsets into SO/DLL binary paths so a later external process can use those in combination with their debug symbols to retrieve source file names and line numbers. Ideally the serialisation formats supported would include those consumable by well known tooling for this purpose e.g. addr2line or llvm-symbolizer. Opinion 2: That Stacktrace can in process use debug symbols to retrieve source file names and line numbers. This could be implemented by invoking addr2line or equivalent as a child process which has the enormous advantage of being async safe, but has disadvantages in almost every other area. Or it could be implemented by one's own DWARF parser but with the huge caveat that memory could be badly corrupted by now. Opinion 3: That Stacktrace be "async safe" in that backtraces can be safely captured and parsed from a signal handler or SEH/TEH handler callback or point of C++ exception throw. *This implies no memory allocation during the execution of the handler*. >From the reviews, I think there is consensus that the minimum viable product is achieved in Stacktrace though many mentioned that improvements are needed to the documentation and most of the negative feedback was on features not in the minimum viable product described above. I therefore recommend the ACCEPTANCE of proposed Boost.Stacktrace into Boost with the following conditions: 1. That the documentation more clearly set out what Stacktrace does and especially does not do, and that the default facilities provided are those of the Minimum Viable Product described above i.e. no source file nor line number discovery, no external dependencies on libraries not always provided on the target platform, no invocation of child processes, no async safety during stacktrace parsing. Stacktrace ought to be pared back to the minimum viable product featureset with no external dependencies at all. 2. That suitable predefined macros opt in to additional functionality such as serialisation of offset translated backtraces, retrieval of source code path and line number etc. Enabling these may introduce dependencies on third party libraries not always present on a system such as libunwind, or on third party helper utilities such as addr2line. I would recommend that the APIs the use of which are dependent on such third party items are entirely compiled out of Stacktrace when the macro enabling them is not defined. 3. That the documentation describe ONLY the default minimum viable product functionality in its tutorial and quick start. All extra functionality enabled by macros is to be relocated into a separate documentation section away from the main documentation e.g. "Advanced/optional features". 4. That APIs made available only when a given macro has turned them on are documented as such in the API reference docs. I'd like to congratulate Antony on getting yet another library of his into Boost. I hope some of his magic rubs off on me when I bring Outcome to review later this year! Regards, Niall -- ned Productions Limited Consulting http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/

2 1

[log] disabling logging at compile-time
by Igor R 02 Jan '17

02 Jan '17

Is it possible to completely eliminate the logging code (i.e. the run-time overhead) using some macro? Thanks.

2 1

Problems with nesting boost::container::small_vector and C++98
by Daniel Trstenjak 02 Jan '17

02 Jan '17

Hi, first the used versions: boost 1.63.0 gcc 5.4.0 20160609 Ubuntu 16.04.1 the example I've problems with is: #include <map> #include <boost/container/small_vector.hpp> using boost::container::small_vector; typedef std::pair<int, small_vector<float, 8> > mypair; typedef std::map<int, mypair> mymap; typedef std::map<int, small_vector<mypair, 8> > mymap2; typedef std::map<int, small_vector<small_vector<float, 8>, 8> > mymap3; int main() { mymap m; mypair& sv = m[0]; mymap2 m2; small_vector<mypair, 8>& sv2 = m2[0]; mymap3 m3; small_vector<small_vector<float, 8>, 8>& sv3 = m3[0]; } I can compile it without errors with 'g++ -std=c++11', but without '-std=c++11' I'm getting an error for the 'operator[]' on mymap2: dan@octa ~> g++ -Iboost_1_63_0/ small_vector_test.cpp In file included from boost_1_63_0/boost/container/small_vector.hpp:27:0, from small_vector_test.cpp:2: boost_1_63_0/boost/container/vector.hpp: In instantiation of ‘void boost::container::vector<T, Allocator>::assign(FwdIt, FwdIt, typename boost::move_detail::disable_if_or<void, boost::move_detail::is_same<typename boost::container::container_detail::version<Allocator>::type, boost::move_detail::integral_constant<unsigned int, 0u> >, boost::move_detail::is_convertible<InIt, typename boost::container::allocator_traits<Allocator>::size_type>, boost::container::container_detail::is_input_iterator<FwdIt> >::type*) [with FwdIt = boost::container::container_detail::vec_iterator<std::pair<int, boost::container::small_vector<float, 8ul> >*, true>; T = std::pair<int, boost::container::small_vector<float, 8ul> >; Allocator = boost::container::small_vector_allocator<boost::container::new_allocator<std::pair<int, boost::container::small_vector<float, 8ul> > > >; typename boost::move_detail::disable_if_or<void, boost::move_detail::is_same<typename boost::container::container_detail::version<Allocator>::type, boost::move_detail::integral_constant<unsigned int, 0u> >, boost::move_detail::is_convertible<InIt, typename boost::container::allocator_traits<Allocator>::size_type>, boost::container::container_detail::is_input_iterator<FwdIt> >::type = void]’: boost_1_63_0/boost/container/small_vector.hpp:563:7: required from ‘boost::container::small_vector<T, N, Allocator>::small_vector(const boost::container::small_vector<T, N, Allocator>&) [with T = std::pair<int, boost::container::small_vector<float, 8ul> >; long unsigned int N = 8ul; Allocator = boost::container::new_allocator<std::pair<int, boost::container::small_vector<float, 8ul> > >]’ /usr/include/c++/5/bits/stl_pair.h:113:31: required from ‘std::pair<_T1, _T2>::pair(const _T1&, const _T2&) [with _T1 = const int; _T2 = boost::container::small_vector<std::pair<int, boost::container::small_vector<float, 8ul> >, 8ul>]’ /usr/include/c++/5/bits/stl_map.h:487:23: required from ‘std::map<_Key, _Tp, _Compare, _Alloc>::mapped_type& std::map<_Key, _Tp, _Compare, _Alloc>::operator[](const key_type&) [with _Key = int; _Tp = boost::container::small_vector<std::pair<int, boost::container::small_vector<float, 8ul> >, 8ul>; _Compare = std::less<int>; _Alloc = std::allocator<std::pair<const int, boost::container::small_vector<std::pair<int, boost::container::small_vector<float, 8ul> >, 8ul> > >; std::map<_Key, _Tp, _Compare, _Alloc>::mapped_type = boost::container::small_vector<std::pair<int, boost::container::small_vector<float, 8ul> >, 8ul>; std::map<_Key, _Tp, _Compare, _Alloc>::key_type = int]’ small_vector_test.cpp:18:39: required from here boost_1_63_0/boost/container/vector.hpp:1261:15: error: binding ‘const std::pair<int, boost::container::small_vector<float, 8ul> >’ to reference of type ‘std::pair<int, boost::container::small_vector<float, 8ul> >&’ discards qualifiers *cur = *first; ^ In file included from /usr/include/c++/5/bits/stl_algobase.h:64:0, from /usr/include/c++/5/bits/stl_tree.h:63, from /usr/include/c++/5/map:60, from small_vector_test.cpp:1: /usr/include/c++/5/bits/stl_pair.h:96:12: note: initializing argument 1 of ‘std::pair<int, boost::container::small_vector<float, 8ul> >& std::pair<int, boost::container::small_vector<float, 8ul> >::operator=(std::pair<int, boost::container::small_vector<float, 8ul> >&)’ struct pair ^ Any ideas? Is 'boost:container::small_vector' supposed to work without C++11 support? Thanks! Greetings, Daniel

2 4

How to use Boost libs from projects?
by Andy 01 Jan '17

01 Jan '17

I have project compiled as Debug,x64. And boost 1.63 compiled by: - bootstrap.bat - b2 variant=debug,release link=static runtime-link=static address-model=64 In project I have cpp files with #include "boost/filesystem.hpp" and when compile project I have error: cannot open file 'libboost_filesystem-vc140-mt-gd-1_61.lib' whereas in c:\boost\stage\lib\ I have libboost_filesystem-vc140-mt-sgd-1_63.lib but not libboost_filesystem-vc140-mt-gd-1_61.lib and in VS2015 project I have no Boost lib. (sorry for my English)

4 13

How to compile boost-thread for Universal Windows?
by Petr Hudeček 31 Dec '16

31 Dec '16

Hello, I'm trying to use the boost-thread library in a Windows Runtime (UWP) project. The tutorial at https://blogs.msdn.microsoft.com/vcblog/2014/07/18/using-boost-libraries-in… doesn't seem to work anymore. If I compile boost sources, either from the primary distribution, or from https://github.com/wpbest/Boost, I can compile it as a static library but when I link to it from the project, I receive linker errors that four identifiers could not be found. These identifiers are API functions such as CreateWaitableTimer. Is there a way to use boost-thread from a UWP project despite this? Perhaps setting some preprocessor symbols? Thank you, Petr Hudecek

4 5

Echo AsyncReadStream/AsyncWriteStream example
by Adam Zegelin 30 Dec '16

30 Dec '16

Hi. Is there a boost::asio example that shows an implementation of AsyncReadStream/AsyncWriteStream that copies write->read? i.e. echoes what was sent I’m trying to emulate serial device that takes input commands separated by \r, executes them, and then outputs the result. i.e., I'm trying to create a fake asio::serial_port that can read input, process it and then provide output. The device I'm trying to emulate echoes the sent characters. Commands are executed once '\r' is received. Ideally the fake serial port also echos the received data immediately. I've got the read part working by implementing async_write_some(…), but I'm having trouble implementing the async_read_some(…) method so that it can be called at anytime, especially when io_service.run() is run from multiple threads. Thanks, Adam

1 0

Boost 1.63.0 is released!
by Marshall Clow 28 Dec '16

28 Dec '16

Release 1.63.0 of the Boost C++ Libraries is now available. These open-source libraries work well with the C++ Standard Library, and are usable across a broad spectrum of applications. The Boost license encourages both commercial and non-commercial use. This release contains numerous enhancements and bug fixes for existing libraries. For details, including download links, see http://www.boost.org/users/news/version_1.63.0 You can also download directly from SourceForge: https://sourceforge.net/projects/boost/files/boost/1.63.0/ To install this release on your system, see http://www.boost.org/doc/libs/release/more/getting_started/index.html Thanks to everyone who participated in this release. Thanks, --The Boost release team Vladimir Prus Rene Rivera Marshall Clow Eric Niebler Daniel James Beman Dawes

2 1

[msm] Transitions not atomic with orthogonal regions
by Stian Zeljko Vrba 26 Dec '16

26 Dec '16

I have spent almost an hour debugging the following issue. I have a "main" orthogonal region with a couple of states and an "error" region with RUNNING and PAUSED states, like this: struct PAUSED : public msm::front::interrupt_state<new_job> {}; struct RUNNING : public msm::front::state<> {}; There are transitions PAUSED --new_job--> RUNNING and RUNNING --failed--> PAUSED. Then, in one of the main states, I have struct MAIN_STATE1 : public msm::front::state<> { template<typename Event, typename FSM> void operator()(const Event& e, FSM& fsm) { if (error_condition) fsm.process_event(failed{}); } } and in the FSM I have using initial_state = mpl::vector<MAIN_STATE0, PAUSED>; Now, when new_job is processed and MAIN_STATE1::on_entry called, the machine is in (MAIN_STATE1, PAUSED). In this state, process_event(failed{}) is ignored and not queued because PAUSED is interrupt state. Then, the next orthogonal region is processed and the machine goes into (MAIN_STATE1, RUNNING), which destroys the rest of the logic around which the FSM is designed... Adding using active_state_switch_policy = msm::active_state_switch_after_exit; to the machine did not help in any way. Reordering the states in initial_state fixed the problem, but this is a rather unsatisfactory fix because it doesn't solve the underlying problem of transition atomicity. This is counter-intuitive, being on the border between a bug and lack of clear documentation. Specifically: 1. State switch should be atomic. By using active state switch after exit, I had *expected* that first all exit handlers of the current state are called, the state is atomically switched to the new state, and then all entry handlers are called. What I have *observed* is that on_entry handler of a state in one orthogonal region is called before the other orthogonal region has switched *its* state. Thus, with orthogonal regions, the machine can be in an "impossible" intermediate combination of states. Because not all side-effects have run, this has consequences for all actions/guards acting on state or FSM data, which is illustrated by the following: 2. process_event eagerly evaluates is_event_handling_blocked_helper for each orthogonal region in order. The event should be queued unconditionally, and only after the transition has completed, the machine should dequeue events and check whether processing is blocked etc. This, I think, would make at least process_event to appear atomic wrt state switch. Which leads me to conclude that MSM is seductively declarative on the surface, but that it's rather tricky operationally.

1 0