February 2022 - Boost-users - lists.stage.boost.cppalliance.org

Boost::coroutine2 and compiling for ASAN problems
by tjisana (泰志) 16 Feb '22

16 Feb '22

Team, I'm trying to get a simple coroutine2 program to work with address sanitizer and I'm running into a block. I'm building boost_1_78_0 as follows: ./b2 context-impl=ucontext and compiling with -DBOOST_USE_ASAN and -DBOOST_USE_CONTEXT. I've built and rebuilt the library thinking I've made an error somewhere and it boils down to jump_fcontext and make_fcontext. What am I doing wrong/missing? Cheers, TJ g++ -DBOOST_USE_ASAN -DBOOST_USE_UCONTEXT -I/home/tjisana/Downloads/boost_1_78_0 -O0 -g3 -fsanitize=address -pedantic -Wall -c -fmessage-length=0 -o src/Testbed II.o ../src/Testbed II.cpp g++ -L/home/tjisana/Downloads/boost_1_78_0/stage/lib -fsanitize=address -o Testbed II src/Testbed II.o -lboost_context -lboost_coroutine /usr/lib64/gcc/x86_64-suse-linux/11/../../../../x86_64-suse-linux/bin/ld: /home/tjisana/Downloads/boost_1_78_0/stage/lib/libboost_coroutine.so: undefined reference to `jump_fcontext' /usr/lib64/gcc/x86_64-suse-linux/11/../../../../x86_64-suse-linux/bin/ld: /home/tjisana/Downloads/boost_1_78_0/stage/lib/libboost_coroutine.so: undefined reference to `make_fcontext' collect2: error: ld returned 1 exit status #include <queue> #include <memory> #include "boost/coroutine2/coroutine.hpp" typedef boost::coroutines2::coroutine<int> coro_t; int main() { int l = 0; std::queue<int> y({1,2,3,4}); std::queue<std::unique_ptr<int>> w; coro_t::push_type sink([&](coro_t::pull_type& source) { for (const auto& i : source) { while (!w.empty()) { auto ptr = std::move(w.front()); w.pop(); if (*ptr == 1) { std::cout << "first: " << i << std::endl; auto item = 21; w.push(std::move(std::make_unique<int>(item))); } else if (*ptr == 21) { std::cout << "second: " << i+2*i << std::endl; auto item = 32; w.push(std::move(std::make_unique<int>(item))); } else if (*ptr == 32) { std::cout << "third: " << i*i << std::endl; } } } source(); }); while (!y.empty()) { if (!y.empty()) { l = y.front(); y.pop(); w.push(std::move(std::make_unique<int>(1))); sink(l); } } return 0; }

3 3

std::launder and its usage in Boost
by Mikhail Kremniov 16 Feb '22

16 Feb '22

Hi, I wonder, what is the Boost community's opinion of std::launder? In particular, of its necessity when accessing an object that was created via placement-new in an aligned_storage. As I can see, it's used by Boost.Beast in its implementation of the variant type, but not in other parts of Boost. The reason I'm asking this is that I'm working on a C++14 code-base that uses Boost.Variant and Boost.Optional extensively (as well as other parts of Boost that use Variant and Optional internally). Now we're trying to switch to C++17 at least, and I worry whether it could potentially break things. To put things into context, C++ standard states in the [basic.life] section that if you had an object and then created a new one in the same location, you can use the pointer to the old object to refer to the new one only if (among other things) "the new object is of the same type as the original object (ignoring the top-level cv-qualifiers)". This means, as I understand it, that it's technically illegal to reinterpret_cast the pointer to aligned_storage to the pointer to the actual type of the object created via placement-new, because placement-new has already ended the lifetime of the aligned_storage object. So, implementations of Boost.Optional and Boost.Variant are illegal then (?). But it's been like that since C++03 and Boost worked fine all that time, so it looks like this UB existed only "on paper". But C++17 then added "Note: If these conditions are not met, a pointer to the new object can be obtained from a pointer that represents the address of its storage by calling std::launder". So now the language has the ability to deal with that UB, and a question arises, is it possible that compilers could start to use the UB to perform additional optimizations and make it a real UB? Also, I've seen a couple of times on stackoverflow.com people saying that it's actually fine to reinterpret_cast the storage in C++14, but in C++17 it's not (they didn't explain why though). So, can switching from -std=c++14 to -std=c++17 be a breaking change when using Boost? The fact that Boost.Variant and Boost.Optional don't use std::launder - is it an oversight or a conscious decision? Thanks!

2 1

Does boost have something like small_function (with customizable buffer size)?
by Ivan Matek 16 Feb '22

16 Feb '22

I tried googling, but I found nothing. I presume that that means it does not exist, but it is possible it is maybe hidden under some name I could not think of. I am ok with using internal header of some library in case it is code that is covered by tests. regards, Ivan

2 2

MSVC Hint File for Boost.Test macros?
by Dominique Devienne 16 Feb '22

16 Feb '22

Hi. I've just noticed for the first time MSVC (2019 in my case) "squiggles" in my Boost.Test-using unit-test sources, which pointed me to https://docs.microsoft.com/en-us/cpp/build/reference/hint-files?view=msvc-1… and I wonder if anyone has already defined a "Hint File" for Boost.Test, to improve the Navigation Bar experience in Boost.Test-using sources? A quick search didn't turn up anything. Thanks, --DD

1 0

Very intermittent segmentation fault with boost::multiprecision
by tjisana (泰志) 11 Feb '22

11 Feb '22

Team, Trying to get to the root of a very intermittent seg fault in my program. Here's the snipped (not able to reproduce) as the section of code is part of a callback and I'm not sure how to reproduce it. Here's the tickPrice method and the trace from gdb is below that ... any thoughts? TJ void ib_trading_engine::tickPrice(TickerId tickerId, TickType field, double price, const TickAttrib& attribs) { this -> current_time(); ticksnapshot_complete_ = false; last_tick spy_call_trade; last_tick spy_put_trade; last_tick qqq_call_trade; last_tick qqq_put_trade; spy_call_trade.contract(spy_call_contract_); spy_call_trade.datetime(local_time_); spy_call_trade.symbol(spy_call_symbol_); spy_put_trade.contract(spy_put_contract_); spy_put_trade.datetime(local_time_); spy_put_trade.symbol(spy_put_symbol_); qqq_call_trade.contract(qqq_call_contract_); qqq_call_trade.datetime(local_time_); qqq_call_trade.symbol(qqq_call_symbol_); qqq_put_trade.contract(qqq_put_contract_); qqq_put_trade.datetime(local_time_); qqq_put_trade.symbol(qqq_put_symbol_); try { if (tickerId == 1001 && field == TickType::LAST) { spy_strike_ = (unsigned int)std::trunc(price); ticksnapshot_complete_ = true; last_spy_price_ = bmp::cpp_dec_float_50(price); } else if (tickerId == 1002) { switch (field) { case TickType::BID: last_call_bid_ = bmp::cpp_dec_float_50(price); break; case TickType::ASK: last_call_ask_ = bmp::cpp_dec_float_50(price); break; case TickType::LAST: spy_call_trade.last(bmp::cpp_dec_float_50(price)); trade_queue_.push(spy_call_trade); break; default: break; } } else if (tickerId == 1003) { switch (field) { case TickType::BID: last_put_bid_ = bmp::cpp_dec_float_50(price); break; case TickType::ASK: last_put_ask_= bmp::cpp_dec_float_50(price); break; case TickType::LAST: spy_put_trade.last(bmp::cpp_dec_float_50(price)); trade_queue_.push(spy_put_trade); break; default: break; } } else if (field == TickType::LAST && tickerId == 2001) { qqq_strike_ = (unsigned int)std::trunc(price); ticksnapshot_complete_ = true; last_qqq_price_ = bmp::cpp_dec_float_50(price); } else if (tickerId == 2002) { switch (field) { case TickType::BID: last_call_bid_ = bmp::cpp_dec_float_50(price); break; case TickType::ASK: last_call_ask_ = bmp::cpp_dec_float_50(price); break; case TickType::LAST: qqq_call_trade.last(bmp::cpp_dec_float_50(price)); trade_queue_.push(qqq_call_trade); break; default: break; } } else if (tickerId == 2003) { switch (field) { case TickType::BID: last_put_bid_ = bmp::cpp_dec_float_50(price); break; case TickType::ASK: last_put_ask_= bmp::cpp_dec_float_50(price); break; case TickType::LAST: qqq_put_trade.last(bmp::cpp_dec_float_50(price)); trade_queue_.push(qqq_put_trade); break; default: break; } } } catch (const std::exception& exception) { std::cerr << "ib_trading_engine::tickPrice() exception: " << exception.what() << "\n"; } } Core was generated by `./atm_call_paper_2.1'. Program terminated with signal SIGSEGV, Segmentation fault. #0 0x00000000004221fd in boost::multiprecision::backends::cpp_dec_float<50u, int, void>::pow2 (p=30) at /usr/include/boost/multiprecision/cpp_dec_float.hpp:2936 2936 cpp_dec_float<Digits10, ExponentType, Allocator> t; [Current thread is 1 (Thread 0x7fc4418027c0 (LWP 20245))] (gdb) bt #0 0x00000000004221fd in boost::multiprecision::backends::cpp_dec_float<50u, int, void>::pow2 (p=30) at /usr/include/boost/multiprecision/cpp_dec_float.hpp:2936 #1 0x00000000004157ba in boost::multiprecision::backends::cpp_dec_float<50u, int, void>::operator=<double> (this=0x7fc440fe2f40, a=2.9500000000000002) at /usr/include/boost/multiprecision/cpp_dec_float.hpp:2290 #2 0x0000000000411c17 in boost::multiprecision::backends::cpp_dec_float<50u, int, void>::cpp_dec_float<double>(double, std::enable_if<std::is_floating_point<double>::value&&(!std::is_same<double, __float128>::value), void>::type*) (this=0x7fc440fe2f40, val=2.9500000000000002) at /usr/include/boost/multiprecision/cpp_dec_float.hpp:249 #3 0x000000000040fcb6 in boost::multiprecision::number<boost::multiprecision::backends::cpp_dec_float<50u, int, void>, (boost::multiprecision::expression_template_option)1>::number<double>(double const&, std::enable_if<std::is_convertible<boost::multiprecision::detail::canonical<double, boost::multiprecision::backends::cpp_dec_float<50u, int, void> >::type, boost::multiprecision::backends::cpp_dec_float<50u, int, void> >::value&&(!boost::multiprecision::detail::is_restricted_conversion<boost::multiprecision::detail::canonical<double, boost::multiprecision::backends::cpp_dec_float<50u, int, void> >::type, boost::multiprecision::backends::cpp_dec_float<50u, int, void> >::value), void>::type*) (this=0x7fc440fe2f40, v=@0x7fc440fe1d60: 2.9500000000000002) at /usr/include/boost/multiprecision/number.hpp:64 #4 0x0000000000447332 in ib_trading_engine::tickPrice (this=0x1dd0650, tickerId=1002, field=BID, price=2.9500000000000002, attribs=...) at ../src/ib_trading_engine.cpp:700 #5 0x000000000049c5c4 in EDecoder::processTickPriceMsg (this=0x1dd30c0, ptr=0x7fc43c0077a6 "00%", endPtr=0x7fc43c0077a6 "00%") at ../src/client/EDecoder.cpp:65 #6 0x00000000004a7ea6 in EDecoder::parseAndProcessMsg (this=0x1dd30c0, beginPtr=@0x7fc440fe33e8: 0x7fc43c007790 "1", endPtr=0x7fc43c0077a6 "00%") at ../src/client/EDecoder.cpp:2247 #7 0x00000000004bee93 in EReader::processMsgs (this=0x1dd30b0) at ../src/client/EReader.cpp:282 #8 0x0000000000444630 in ib_trading_engine::process_messages (this=0x1dd0650) at ../src/ib_trading_engine.cpp:405 #9 0x000000000044465f in ib_trading_engine::current_time (this=0x1dd0650) at ../src/ib_trading_engine.cpp:410 #10 0x0000000000447086 in ib_trading_engine::tickPrice (this=0x1dd0650, tickerId=1002, field=BID, price=2.96, attribs=...) at ../src/ib_trading_engine.cpp:666

4 7

Crash in async_handshake function
by Ales Mulej 08 Feb '22

08 Feb '22

Hi All, I'm writing TLS client/server and my server is continuously crashing in the async_handshake function. I'm running out of ideas of what I'm doing wrong or how to solve the issue. I'm using VS 2022 and the latest boost from github. Tested with 1.78.0 as well. Using OpenSSL 3.0.1. Code is compiled as x64, /std:c++20. Using a single thread. Simplified function call: on_accept: auto server = boost::make_shared<cserver_session_ws>(std::move(socket), m_ioc, m_ctx); server->run(); run: boost::asio::dispatch(m_ws.get_executor(), boost::beast::bind_front_handler(&cserver_session_ws::on_run, shared_from_this())); on_run: m_ws.next_layer().async_handshake(boost::asio::ssl::stream_base::server, boost::beast::bind_front_handler(&cserver_session_ws::on_handshake, shared_from_this())); on_handshake function never gets called, due to the crash in async_handshake function. m_ws is defined as: boost::beast::websocket::stream<boost::beast::ssl_stream<boost::beast::tcp_stream>> m_ws; Exception call stack: Unhandled exception thrown: read access violation. **this** was 0x1. boost::asio::detail::scheduler::post_immediate_completion(boost::asio::detail::scheduler_operation ...... Line 349 boost::asio::detail::select_reactor::schedule_timer<boost::asio::detail::chrono_time_traits<std::chrono ..... Line 66 boost::asio::detail::deadline_timer_service<boost::asio::detail::chrono_time_traits ..... Line 273 boost::asio::basic_waitable_timer<std::chrono::steady_clock,boost::asio::wait_traits ...... Line 806 boost::asio::detail::completion_handler_async_result<boost::beast::basic_stream ..... Line 485 boost::asio::async_initiate<boost::beast::basic_stream<boost::asio::ip ..... Line 865 ..... Here are only the last few functions from the call stack. I can share the entire call stack if requested. I would appreciate any help on this as I can't figure out what caused the crash. Regards, A

3 6

Boost SSL handshake(http_server_sync_ssl) : still reachable memory keeps increasing
by Sandeep Bhardwaj 03 Feb '22

03 Feb '22

Hi all, I am using boost 1.76.0 version. I have slightly modified the example program http_server_sync_ssl.cpp to just dump the payload it receives and i have used certificates with ssl context. When I run the program with valgrind, the "still reachable" memory keeps increasing with time. This is one example: ==17719== 38,880 bytes in 60 blocks are still reachable in loss record 740 of 740 ==17719== at 0x4C312EF: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==17719== by 0x567F438: CRYPTO_zalloc (in /usr/lib64/libcrypto.so.1.1) ==17719== by 0x52AA5CA: SSL_SESSION_new (in /usr/lib64/libssl.so.1.1) ==17719== by 0x52AAE22: ??? (in /usr/lib64/libssl.so.1.1) ==17719== by 0x52C8A06: ??? (in /usr/lib64/libssl.so.1.1) ==17719== by 0x52B7FF6: ??? (in /usr/lib64/libssl.so.1.1) ==17719== by 0x52A3F43: SSL_do_handshake (in /usr/lib64/libssl.so.1.1) ==17719== by 0x419630: boost::asio::ssl::detail::engine::do_accept(void*, unsigned long) (engine.ipp:316) ==17719== by 0x4193E3: boost::asio::ssl::detail::engine::perform(int (boost::asio::ssl::detail::engine::*)(void*, unsigned long), void*, unsigned long, boost::system::error_code&, unsigned long*) (engine.ipp:248) ==17719== by 0x418F6D: boost::asio::ssl::detail::engine::handshake(boost::asio::ssl::stream_base::handshake_type, boost::system::error_code&) (engine.ipp:148) ==17719== by 0x419778: boost::asio::ssl::detail::handshake_op::operator()(boost::asio::ssl::detail::engine&, boost::system::error_code&, unsigned long&) const (handshake_op.hpp:47) ==17719== by 0x424DB8: unsigned long boost::asio::ssl::detail::io<boost::asio::basic_stream_socket<boost::asio::ip::tcp, boost::asio::any_io_executor>, boost::asio::ssl::detail::handshake_op>(boost::asio::basic_stream_socket<boost::asio::ip::tcp, boost::asio::any_io_executor>&, boost::asio::ssl::detail::stream_core&, boost::asio::ssl::detail::handshake_op const&, boost::system::error_code&) (io.hpp:38) ==17719== by 0x421332: boost::asio::ssl::stream<boost::asio::basic_stream_socket<boost::asio::ip::tcp, boost::asio::any_io_executor>&>::handshake(boost::asio::ssl::stream_base::handshake_type, boost::system::error_code&) (stream.hpp:389) ==17719== by 0x41D5FC: boost::beast::ssl_stream<boost::asio::basic_stream_socket<boost::asio::ip::tcp, boost::asio::any_io_executor>&>::handshake(boost::asio::ssl::stream_base::handshake_type, boost::system::error_code&) (ssl_stream.hpp:344) ==17719== by 0x406971: do_session(boost::asio::basic_stream_socket<boost::asio::ip::tcp, boost::asio::any_io_executor>&, boost::asio::ssl::context&, std::shared_ptr<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const> const&) (boost_server.cpp:219) I have attached the test program. Can anyone point out what needs to be done here? Thanks & Regards, Sandeep

3 4

Boost DLL shared library handling
by Nicholas Yue 02 Feb '22

02 Feb '22

Hi, I am learning about <boost/dll/shared_library.hpp> So far I have manage to follow along with the tutorial I believe https://www.boost.org/doc/libs/1_77_0/doc/html/boost_dll/tutorial.html Using tutorial 5 as an illustration https://github.com/apolukhin/Boost.DLL/blob/develop/example/tutorial5/load_… On this line, I will have the ability to reference the object of my_api_sum I believe https://github.com/apolukhin/Boost.DLL/blob/develop/example/tutorial5/load_… I have the following questions: 1. If I have a class that is e.g. a shape, say rectangle (of different sizes), what approach may I use to create different instances of the class from the plugin? 2. If my class requires parameters for it's construction, how do I pass it via the shared library loading mechanism ? Or have I been looking to use boost dll shared_library wrongly and that I might look at the plugin being some sort of factory design approach ? Thank you in advance. Cheers -- Nicholas Yue https://www.linkedin.com/in/nicholasyue/

1 0