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

[mpl] at<> behaves differently for vector_c<> and vector<>
by ☂Josh Chia (謝任中) 05 Feb '17

05 Feb '17

I asked earlier in a post "[mpl] Working with map, copy and back_inserter". This is a related question. The common question is: What does compile-time equality means in MPL, particularly for map and set key equality? Does it mean std::is_same? The following code works whether USE_VECTOR_C is defined as 0 or 1 -- the last 2 lines are static_asserts that have conditions determined by USE_VECTOR_C so that they succeed, highlighting the difference in behavior between USE_VECTOR_C == 0 and USE_VECTOR_C == 1. USE_VECTOR_C == 0 is the normal case where everything is working as I expect. USE_VECTOR_C == 1 is the unexpected case where the static_asserts in the last 2 lines have conditions altered so that they succeed. Why do vector_c and vector give me different "element values"? In general, what should I be careful about when trying to use the "compile-time values" as keys to avoid the situation elucidated by the last line, where I expect two "compile-time values" to be the same but they are not, resulting in two set/map items instead of one? To clarify, I'm using boost 1.63 with gcc 6.3.1 and -std=c++14 on Arch Linux. Josh #include <type_traits> #include <boost/mpl/at.hpp> #include <boost/mpl/insert.hpp> #include <boost/mpl/set.hpp> #include <boost/mpl/size.hpp> #include <boost/mpl/size_t.hpp> #include <boost/mpl/vector.hpp> #include <boost/mpl/vector_c.hpp> using namespace std; namespace mpl = boost::mpl; #define USE_VECTOR_C 0 using Zero = mpl::size_t<0>; #if USE_VECTOR_C == 1 using Vector = mpl::vector_c<size_t, 0>; #else using Vector = mpl::vector<Zero>; #endif using ZeroElt = typename mpl::at<Vector, Zero>::type; using EmptySet = mpl::set<>; using NonEmptySet = typename mpl::insert<typename mpl::insert<EmptySet, Zero>::type, ZeroElt>::type; static_assert(ZeroElt::value == 0, ""); static_assert(is_same<ZeroElt::value_type, size_t>::value, ""); static_assert(is_same<ZeroElt, Zero>::value == !USE_VECTOR_C, ""); static_assert(mpl::size<NonEmptySet>::value == USE_VECTOR_C + 1, "");

2 1

context library of boost: compiling for windows 32 on ubuntu 64
by Jack Kumba 04 Feb '17

04 Feb '17

Hello, I want to compile libboost_context.a for windows 32 on my ubuntu 64. for that, I specify the following in my user-config.jam file using gcc : : "i686-w64-mingw32-g++" ; and then, I run the following: b2 --with-context toolset=gcc target-os=windows variant=release link=static address-model=32 binary-format=pe architecture=x86 this gives me the following error: libs/context/src/unsupported.cpp:7:2: error: #error "platform not supported" #error "platform not supported" ^ "i686-w64-mingw32-g++" -ftemplate-depth-128 -O3 -finline-functions -Wno-inline -Wall -march=i686 -mthreads -m32 -DBOOST_ALL_NO_LIB=1 -DBOOST_CONTEXT_SOURCE -DBOOST_DISABLE_ASSERTS -DNDEBUG -I"." -c -o "bin.v2/libs/context/build/gcc-mingw-5.3.1/release/binary-format-pe/build-no/link-static/target-os-windows/threading-multi/unsupported.o" "libs/context/src/unsupported.cpp" ...failed gcc.compile.c++ bin.v2/libs/context/build/gcc-mingw-5.3.1/release/binary-format-pe/build-no/link-static/target-os-windows/threading-multi/unsupported.o... full message is posted here: https://thepasteb.in/p/y8hzcR93o3XO7t2 It seems like the context library of boost is not supported to compile for windows on linux? Is it the case? Thanks, Aja

3 4

Can I config process_id to log file name in the .ini setting file using Boost.Log?
by Bruce 04 Feb '17

04 Feb '17

Hi, I am using Boost.Log to output log file and using .ini file to config it. Now I need to start the exe file from the same directory for more than two instance with different start id, something like this: >start my_app.exe 1 >start my_app.exe 2 But currently only the first instance output the log content and the second one can't get the log file because the two exe uses the same log file. So I need to tell the app to use the different log file. So I want to config the .ini file to add process_id to the output file name, but I don't know how? I want to config the ini file like this but it failed:-( FileName="./log/mylog_%ProcessID%.log" Is there a way to add precess id to the output file name? Thanks.

1 0

Boost iostreams for binary data
by Robert Dailey 03 Feb '17

03 Feb '17

I have my own BinaryInputStream and BinaryOutputStream classes that simply take a reference to some container that meets the requirements of random access and contiguous internal memory. It doesn't own this container, but rather serves as an adapter to the container to allow stream operators to be used for streaming out data. For example, you can stream a `std::uint32_t` to the binary stream, and it will insert 4 elements into an internal byte vector which is just `std::vector<std::uint8_t>`. I did not implement these in terms of IO streams, because those seemed a bit overkill and make a lot of assumptions on the data being text-only. Also I can't stream binary data into vectors, lists, and other containers because the data simply can't be parsed. Each stream out of a binary vector must be done so with a finite count of bytes. Would it make sense to implement these objects in terms of the IOstream facilities provided by boost? I know these were designed to make defining stream classes easier, but honestly I don't have a lot of experience with them. Any feedback on this? Thanks in advance. I can provide some example code if needed.

2 2

Mail List Migration - Update
by Michael Caisse 02 Feb '17

02 Feb '17

The Boost mail lists are now migrated to a new set of servers. This change affected all mail lists and web resources (signing up, changing passwords, moderator access, archives...) that were associated with the domain lists.boost.org . If you are having difficulties, please drop me an email. (Oh the irony) Happy coding! michael -- Michael Caisse Ciere Consulting ciere.com

1 0

Mail Lists Migrate at 00:00 UTC Feb 3
by Michael Caisse 02 Feb '17

02 Feb '17

The following mail lists will be switched to new infrastructure at 00:00 UTC on February 3rd. There will be a small interruption while the DNS/MX records propagate. * Boost - Boost developers' mailing list * Boost-announce - Boost announce-only mailing list * Boost-bugs - Bugs reported/updated through Boost's Trac * Boost-build - Boost.Build developer's and user's list * Boost-cmake - Discussion of the CMake-based build system for Boost * Boost-commit - Boost Subversion commit messages * Boost-docs - Discussion of Boost Documentation * Boost-Interest - Moderated announcements of interest to the Boost community. * Boost-maint - Boost community maintenance mailing list * Boost-mpi - Discussion of Boost.MPI development * Boost-test - Test list for the boost mailing lists * Boost-Testing - Running Boost regression tests * Boost-users - Boost Users mailing list * Geometry - Boost.Geometry library mailing list * glas - Generic Linear Algebra Software * proto - Discussions about Boost.Proto and DSEL design * threads-devel - Discussions about the boost.thread library * ublas - ublas mailing list -- Michael Caisse Ciere Consulting ciere.com

1 0

Cross compiling context library of boost for windows 32 on a linux 64-bit machine
by Jack Kumba 01 Feb '17

01 Feb '17

Hello, Did someone succeed cross compiling context library of boost for windows 32 on a linux 64-bit machine? OR Is it even possible? I see the following error: libs/context/src/unsupported.cpp:7:2: error: #error "platform not supported" #error "platform not supported" Also, there is this ticket on this issue https://svn.boost.org/trac/boost/ticket/7262 Please advise me. Thanks, Jack

1 0

Cross compiling boost (context, coroutine, filesystem, and system libraries) on ubuntu 64-bit for windows 32-bit (running on intel atom e3845 that is x86-64 bit)
by Jack Kumba 31 Jan '17

31 Jan '17

Hello, I want to cross compile context, coroutine, filesystem, and system libraries of boost for windows 32-bit (running on Intel Atom E3845) on my ubuntu 64-bit. To do that, I create a user-config.jam file in my homedirecotry and put the following: using gcc : mingw : i686-w64-mingw32-g++ : <compileflags>-m32 <linkflags>-m32 ; then, I would like to run following command: b2 --with-context --with-coroutine --with-system --with-filesystem --prefix=/home/jagdish/COMMELEC/prerequisits/real_boost_installation_dir/ toolset=gcc-mingw target-os=windows variant=release link=static address-model=32 runtime-link=static threading=multi architecture=x86 instruction-set=x86 abi=ms install is it good? Specifically, the options architecture, instruction-set, address-model? thanks in advance for your help. best, jack

1 0

[thread] understanding multiple statically-linked boost::thread in same process
by Simon Haegler 31 Jan '17

31 Jan '17

hi boost users, i am debugging an application where several shared libraries contain statically-linked copies of boost::thread (and other boost components). this means the static part of boost::thread is multiple times present in the process. on macos (apple clang 7.3) we see rare (but reproducible) access violations caused by boost::thread_specific_ptr::get() unexpectedly returning nullptr. if we enable assertions we get an assert in thread.cpp:151 if we dynamically link boost the problems go away. i can superficially understand that linking only one "common" instance of boost::thread is "more safe". what could be the explanation for this observation? is this setup even supported? one theory is that the implementations in pthread/thread_heap_alloc.hpp interfere with each other... the specific line where we get an unexpected nullptr (actually a bad reference) is here: https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_CGAL_cgal_b… the definition of CGAL_STATIC_THREAD_LOCAL_VARIABLE is here?: https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_CGAL_cgal_b… boost version is 1.59 TIA, simon -- Simon Haegler | Software Engineer Esri R&D Center Zurich | Foerrlibuckstrasse 110 | 8005 Zurich | Switzerland T +41 44 204 60 80 | shaegler(a)esri.com

3 3

C++Now 2017 Registration is Open
by Jon Kalb 31 Jan '17

31 Jan '17

Boost Users, Registration for C++Now 2017 is now open: http://bit.ly/CppNow2017IsOpen The conference will be in Aspen, Colorado from May 15th to 20th. The deadline for submissions is this Friday: http://cppnow.org/2017-conference/announcements/2017/01/06/call-for-submiss… Jon

1 0

Re: [Boost-users] [boost::asio] Wait for multiple async_read()'s
by Marat Abrarov 30 Jan '17

30 Jan '17

Hi Adam, > I'd like to write data to these streams then wait for everyone to reply before performing computation on the resulting data. What about Chris's coinvoke (http://blog.think-async.com/2008/10/asynchronous-forkjoin-using-asio.html)? Could it help for you case? Regards, Marat Abrarov.

1 0

[preprocessor]how to compute include with .hpp in name?
by Larry Evans 28 Jan '17

28 Jan '17

With the attached, I get the error: g++ computed_include.hpp.cpp -o a.out In file included from computed_include.hpp.cpp:4:0: computed_include.hpp.cpp:5:22: error: pasting "computed_include" and "." does not give a valid preprocessing token #define FILENAME_INC computed_include ^ computed_include.hpp.cpp:7:58: note: in expansion of macro ‘FILENAME_INC’ #define FILENAME_CAT_STR BOOST_PP_STRINGIZE(BOOST_PP_CAT(FILENAME_INC,HPP_EXT)) ^ computed_include.hpp.cpp:8:10: note: in expansion of macro ‘FILENAME_CAT_STR’ #include FILENAME_CAT_STR ^ Is there some other way this can be done? TIA. -regards, Larry

1 0

[boost::asio] Wait for multiple async_read()'s
by Adam Zegelin 28 Jan '17

28 Jan '17

Hi, I'm writing a boost::asio application that communicates with multiple streams. I'd like to write data to these streams then wait for everyone to reply before performing computation on the resulting data. What is the best way to implement this? The route I've ventured down involves me calling async_write() on each stream, with the callback hander executing async_read(). But this where I'm stuck. What is the best way to execute one final hander once all read operations have completed? One solution I've found, though it feels "un-asio" is to pass the use_future option as the hander for the reads. The result is futures for each read, which I can then place into a list<future>, and have an additional io_service.post()'ed handler call get() on these futures. The only issue with this is that I must call io_service.run() from more than one thread, since the handler calling get() blocks — otherwise I can potentially deadlock the whole application. Another alternative I can think of is that the function that schedules all the async_writes knows ahead-of-time the number of reads that need to be completed. Each async_read() handler would put its results into a shared list and then, if the list's size is equal to the expected number of results, fires the "completion" handler. While easy enough to implement, this seems fraught with issues. If there is more than one thread calling io_service.run(), then access to the shared list has to be synchronised. Also, it doesn't seem very generic — I have a number of places where this write n, read n, execute (or is it map, reduce?) pattern needs to happen, and I'd prefer a solution that was write one, worked everywhere. Any help would be appreciated. - Adam

7 9

[Parameter] Warnings from Preprocessor
by Phillip Seaver 27 Jan '17

27 Jan '17

I recently updated to 1.63 from 1.57, and I'm now getting warnings from usages of BOOST_PARAMETER_MEMBER_FUNCTION on Windows (msvc-12.0) like: warning C4003: not enough actual parameters for macro 'BOOST_PP_SEQ_DETAIL_EMPTY_SIZE' warning C4003: not enough actual parameters for macro 'BOOST_PP_SEQ_DETAIL_IS_NOT_EMPTY' The code appears to be working, but since it's used in headers, I get a lot of warnings. Here's one of the usages: BOOST_PARAMETER_MEMBER_FUNCTION((Revision_t), CreateForegroundWhite, tag, (optional (imgclass, (Class_t), ClassPrint) (fromimage, (bool), true) (blackasgray, (bool), false) )) Should I be concerned about the warnings? Any way to get rid of them? Thanks, Phillip

2 2

Fwd: compiling boost for windows 32 on ubuntu 64 machine
by Jack Kumba 27 Jan '17

27 Jan '17

Hello, I'm trying to compile boost for windows 32-bit on my ubuntu-64 but I'm not able to. what I do is as follows: - I created a file called user-config.jam in my ubuntu home directory and had put the below in it using gcc : : "i686-w64-mingw32-g++" - then, i run ./bootstrap.sh --with-libraries=filesystem --prefix=/home/jagdish/ commelec_api/installation_dirs/boost - and then, b2 toolset=gcc target-os=windows - finally, i see the following error messages: https://thepasteb.in/p/ y8h6q5Qq1yzcO Did someone encounter the same error messages or can tell me why I'm seeing them? Thanks, Kja

1 0

Re: [Boost-users] NEWBIE: Undefined references: How do I find the library that resolves an undefined reference - Expanded with log example and hope for a general solution
by JENNINGS, LYNN M NH-03 USAF AFMC 517 SMXS/MXDPB 26 Jan '17

26 Jan '17

Thanks again Andy! I am sending this note to thank Andy for his gracious help and to document the resolution for future newbies. Andy's suggestion to "nm -Cg" the boost.log directories was just what I needed! (Sorry, I should have remembered the -C option when I got nothing but garbage.) What I found was that the header files I use, generate a reference to ...::v2s_mt_nt5::..., while the libraries only define the namespace with ...::v2_mt_nt5::... For example, my code has the following: #include <boost\log\core.hpp> #include <boost\log\trivial.hpp> #include <boost\log\expressions.hpp> #include <boost\log\utility\setup\file.hpp> #include <boost\log\utility\setup\common_attributes.hpp> namespace logging = boost::log ; namespace src = boost::log::sources ; namespace keywords= logging::keywords ; namespace sinks = boost::log::sinks ; void init_logger() { logging::add_file_log( "sample.log", keywords::rotation_size = 10 * 1024 * 1024, keywords::time_based_rotation = sinks::file::rotation_at_time_point(0,0,0), keywords::format = "[%TimeStamp%]: %Message%", keywords::auto_flush=true ) ; logging::core::get()->set_filter (logging::trivial::severity >= logging::trivial::info ) ; logging::core::get()->add_global_attribute("TimeStamp", boost::log::attributes::local_clock()); } ... I compiled with the following command from my makefile: cc -g -DCPU=PENTIUM -DMC68000 -DDoing2VxWorks -IC:\boost -I../source -I../include -MMD -MP -c -o obj\logLib.o logLib.cpp Which I built with the following command from my makefile: cc -g -DCPU=PENTIUM -DMC68000 -DDoing2VxWorks -IC:\boost -I../source -I../include -MMD -MP -c -o obj\logLib.o logLib.cpp gfortran -o unit01test.exe obj\unit01test.o ..\mcpmain\obj\rtinit.o ..\mcpmain\obj\bite2.o ..\mcpmain\obj\bmain.o ..\mcpmain\obj\keypm2.o ..\mcpmain\obj\ksvc.o ..\mcpmain\obj\scen1.o ..\mcpmain\obj\disup.o ..\mcpmain\obj\idisup.o ..\mcpmain\obj\keyin.o ..\mcpmain\obj\disdr.o ..\mcpmain\obj\led_task.o ..\mcpmain\obj\spip_task.o ..\mcpmain\obj\iff_task.o ..\mcpmain\obj\inmsgmgr.o ..\mcpmain\obj\cppmainUnit01Test.o ..\runhistory\obj\rhlog_api.o ..\utilities\obj\main2aux.o ../../vxWorksLib/vxWorksLib/vxWorks/source/obj\logLib.o ../../vxWorksLib/vxWorksLib/vxWorks/source/obj\msgQlib.o ../../vxWorksLib/vxWorksLib/vxWorks/source/obj\pciConfigLib.o ../../vxWorksLib/vxWorksLib/vxWorks/source/obj\semLib.o ../../vxWorksLib/vxWorksLib/vxWorks/source/obj\sharedMemLib.o ../../vxWorksLib/vxWorksLib/vxWorks/source/obj\sockLib.o ../../vxWorksLib/vxWorksLib/vxWorks/source/obj\sysLib.o ../../vxWorksLib/vxWorksLib/vxWorks/source/obj\taskLib.o ../../vxWorksLib/vxWorksLib/vxWorks/source/obj\wdLib.o -LC:\boost\stage\lib -lboost_log_setup-mgw48-mt-d-1_61 -lboost_log-mgw48-mt-d-1_61 -lboost_thread-mgw48-mt-d-1_61 -lboost_system-mgw48-mt-d-1_61 -LC:\Windows\SysWOW\en-US -lws2_32 -lstdc++ -lboost_log_setup-mgw48-mt-d-1_61 -lboost_log-mgw48-mt-d-1_61 which gives me the error: ../../vxWorksLib/vxWorksLib/vxWorks/source/obj\logLib.o: In function `Z11init_loggerv': c:\1PortMUTES2MinGW\ProjectRepository-Build\vxWorksLib\vxWorksLib\vxWorks\source/logLib.cpp:16: undefined reference to `boost::log::v2s_mt_nt5::sinks::file::rotation_at_time_point::rotation_at_time_point(unsigned char, unsigned char, unsigned char)' But when I check the nm output, I find the following ::v2_mt_nt5:: namespace defined in all of the various forms of the Boost log libraries: d000608.o: U _head_libboost_log_mgw48_mt_d_1_61_dll 00000000 I _imp___ZNK5boost3log9v2_mt_nt55sinks4file25rotation_at_time_intervalclEv 00000000 T boost::log::v2_mt_nt5::sinks::file::rotation_at_time_interval::operator()() const I built my boost libraries using the directions for a MinGW build. I built my code using the same MinGW framework. However, I found the reason for the two conflicting name spaces, v2_mt_nt5 and v2s_mt_nt5 at: http://www.boost.org/doc/libs/1_56_0/libs/log/doc/html/log/rationale/namesp… The magic key is to recognize the <linkage> name space. I resolved most of my Boost log problem by adding -DBOOST_ALL_DYN_LINK to the compile operation. The correct compile command is: cc -g -DCPU=PENTIUM -DMC68000 -DDoing2VxWorks - DBOOST_ALL_DYN_LINK -IC:\boost -I../source -I../include -MMD -MP -c -o obj\logLib.o logLib.cpp Note that I was able to resolve my remaining errors where gcc just listed the mangled symbol by grep'ing the library file.

1 0

boostcpp.jam: No such file or directory
by JIA Pei 26 Jan '17

26 Jan '17

Hi, all: It seems boost has changed a lot in building? I strictly followed boost *Getting Started on Windows* but when I tried to install boost using command *b2 *. I got the following error messages: D:\boost_1_63_0>C:\boost\bin\b2.exe --build-dir=C:\boost toolset=msvc > --build-type=complete stage > *boostcpp.jam: No such file or directory* > Jamroot:137: in modules.load > ERROR: rule "boostcpp.set-version" unknown in module > "Jamfile<D:\boost_1_63_0>". > D:/boost_1_63_0/tools/build/src/build\project.jam:325: in load-jamfile > D:/boost_1_63_0/tools/build/src/build\project.jam:64: in load > D:/boost_1_63_0/tools/build/src/build\project.jam:145: in project.find > D:/boost_1_63_0/tools/build/src\build-system.jam:535: in load > D:\boost_1_63_0\tools\build\src/kernel\modules.jam:295: in import > D:\boost_1_63_0\tools\build\src/kernel/bootstrap.jam:139: in boost-build > D:\boost_1_63_0\boost-build.jam:17: in module scope > D:\boost_1_63_0> Did anybody meet this issue as well? Most of the time, I work under Linux, which seems to be more convenient.... Cheers -- Pei JIA, Ph.D. Email: jp4work(a)gmail.com cell in Canada: +1 778-863-5816 cell in China: +86 186-8244-3503 Welcome to Vision Open http://www.visionopen.com

1 0

Re: [Boost-users] NEWBIE: Undefined references: How do I find the library that resolves an undefined reference - Expanded with log example and hope for a general solution
by JENNINGS, LYNN M NH-03 USAF AFMC 517 SMXS/MXDPB 26 Jan '17

26 Jan '17

Thanks Richard and Andy for the timely response! Andy's answers provides the perfect example of my dilemma. When Andy said to check the MSDN documentation for ms742213, I quickly found that I needed the Ws2_32.lib library, which I was able to solve my issue. However, I also have the following: 1. boost::log::add_log_file(...) with a sinks::file::rotation_at_time_point(...) 2. boost::random::random_device range; I made a guess with the log and added boost_log_setup_mgw48-mt-d-1_61 and boost_log_ mgw48-mt-d-1_61, but these libraries didn't resolve my issue. When I check the boost documentation (I realize that many of the boost additions just require header files, which the examples are careful to specify. I figure this must require a library because I am getting an undefined error.), there were many mentions of a library, but I couldn't find where the name of the library was specified! So I hope there is a general way to determine the name of the boost library so I do not need to keeping asking about particular undefined errors. In the past, I have been able to google the undefined error and find the library reference, but Thanks again for all of your help! -----Original Message----- On 1/24/2017 3:44 PM, Richard Hodges wrote: > That would be the winsock library. > > https://msdn.microsoft.com/en-us/library/ms740673.aspx > > > > On 24 January 2017 at 20:33, JENNINGS, LYNN M NH-03 USAF AFMC 517 > SMXS/MXDPB <lynn.jennings_at_[hidden]> wrote: > >> NEWBIE question: Undefined references: How do I find the library that >> resolves an undefined reference. For example, the output for the gcc >> linker running in MinGW shows: >> >> C:/boost/boost/asio/detail/impl/winsock_init.ipp:39: Undefined reference >> to `WSAStartup_at_8' >> >> I have googled the documentation a number of ways, but I cannot find the >> library that will resolve this error. I am sure that this is a common >> problem. Please point me to the documentation so I can fix this and the >> undefined references. >> >> Thanks! In the future, you can check Microsoft's documentation [1] for info on their functions. I searched for "WSAStartup", and found [2]. At the bottom of that page is a "Requirements" box, which tells you the library it's in. Andy 1. http://msdn.microsoft.com 2. https://msdn.microsoft.com/en-us/library/windows/desktop/ms742213(v=vs.85).…

2 1

circular_buffer: rotate() vs erase_begin() for scalars
by Robert Dailey 25 Jan '17

25 Jan '17

I'm using circular_buffer to represent a FIFO queue of bytes: boost::circular_buffer<std::uint8_t> m_queue; When I read bytes from the "front", I simply do: m_queue.erase_begin(num_read_bytes); I do so because the documentation for erase_begin() specifically mentions constant time performance for scalar types since no destructor needs to be called. However, I feel that rotate() better expresses what I'm trying to do. Is there a reason to prefer one over the other for scalar types?

2 3

C++Now / BoostCon 2017 -- Call For Submissions
by Michael Caisse 23 Jan '17

23 Jan '17

This is a quick reminder that the C++Now 2017 Call For Submissions is *open*. All submissions must be made by February 3rd. For more information, see http://cppnow.org/

1 0

How to build Boost with Clang/libc++ and NOT link against libstdc++?
by Deniz Bahadir 20 Jan '17

20 Jan '17

Hi guys, I am struggling with building the Boost (1.58.0) libraries with Clang (3.8) and libc++ (on Ubuntu 16.04). I am able to build the Boost libraries using Clang and flags "-std=c++14 -stdlib=libc++", however, I noticed that several libraries are still linked against GCC's standard-library "libstdc++". I then tried adding the flag "-lc++abi" (later I tried "-Wl,-lc++abi") to Boost.Build options "cxxflags" and "linkflags" to explicitly link against "libc++abi". With no luck. Several boost-libraries are still linked against "libstdc++". Does anyone know how and why this link-dependency gets added and how to get rid of them? Does it come in through some other link-dependencies? At least none of them seem to link to "libstdc++"? The Boost shared-libraries that link to "libstdc++" are: libboost_chrono-clang38-mt-1_58_0.so.1.58.0 libboost_context-clang38-mt-1_58_0.so.1.58.0 libboost_coroutine-clang38-mt-1_58_0.so.1.58.0 libboost_date_time-clang38-mt-1_58_0.so.1.58.0 libboost_filesystem-clang38-mt-1_58_0.so.1.58.0 libboost_graph-clang38-mt-1_58_1.so.0.58.0 libboost_iostreams-clang38-mt-1_58_0.so.1.58.0 libboost_locale-clang38-mt-1_58_0.so.1.58.0 libboost_log-clang38-mt-1_58_0.so.1.58.0 libboost_log_setup-clang38-mt-1_58_0.so.1.58.0 libboost_math_tr1-clang38-mt-1_58_0.so.1.58.0 libboost_math_tr1f-clang38-mt-1_58_0.so.1.58.0 libboost_math_tr1l-clang38-mt-1_58_0.so.1.58.0 libboost_program_options-clang38-mt-1_58_0.so.1.58.0 libboost_python-clang38-mt-1_58_0.so.1.58.0 libboost_random-clang38-mt-1_58_0.so.1.58.0 libboost_regex-clang38-mt-1_58_0.so.1.58.0 libboost_signals-clang38-mt-1_58_0.so.1.58.0 libboost_thread-clang38-mt-1_58_0.so.1.58.0 libboost_timer-clang38-mt-1_58_0.so.1.58.0 libboost_unit_test_framework-clang38-mt-1_58_0.so.1.58.0 libboost_wave-clang38-mt-1_58_0.so.1.58.0 The once that are fine and do not link against "libstdc++" are: libboost_atomic-clang38-mt-1_58_0.so.1.58.0 libboost_container-clang38-mt-1_58_0.so.1.58.0 libboost_math_c99-clang38-mt-1_58_0.so.1.58.0 libboost_math_c99f-clang38-mt-1_58_0.so.1.58.0 libboost_math_c99l-clang38-mt-1_58_0.so.1.58.0 libboost_serialization-clang38-mt-1_58_0.so.1.58.0 libboost_system-clang38-mt-1_58_0.so.1.58.0 libboost_wserialization-clang38-mt-1_58_0.so.1.58.0 Interestingly, all libraries are also linked against "libc++abi" except for: libboost_atomic-clang38-mt-1_58_0.so.1.58.0 libboost_container-clang38-mt-1_58_0.so.1.5 For completeness, my build-command is this: cd /home/jenkins/workspace/Pipeline_Boost/sources/boost_1_58_0 LINK_LIBCXXABI="-Wl,-lc++abi" COMMON_COMPILER_FLAGS="-m64 -fpic -fdiagnostics-color=always -O3" CFLAGS="$COMMON_COMPILER_FLAGS" CXXFLAGS="$COMMON_COMPILER_FLAGS -std=c++14 -stdlib=libc++ -isystem/usr/include/libcxxabi $LINK_LIBCXXABI" CC=/usr/bin/clang-3.8 CXX=/usr/bin/clang++-3.8 /home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost-build/b2 -j1 -q \ --build-dir=/home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost \ --build-type=minimal --layout=versioned address-model=64 install \ --prefix=/home/jenkins/workspace/Pipeline_Boost@3/clang38/installed/opt/BENOCS/clang38 \ --libdir=/home/jenkins/workspace/Pipeline_Boost@3/clang38/installed/opt/BENOCS/clang38/lib/x86_64-linux-gnu \ dll-path==/opt/BENOCS/clang38/lib/x86_64-linux-gnu \ toolset=clang-3.8 \ cflags="$CFLAGS" CXXFLAGS="$CXXFLAGS" \ linkflags="-Wl,--as-needed $LINK_LIBCXXABI" For the environment-variable LINK_LIBCXXABI I tried "-lc++abi" as well and also tried providing it only in the cxxflags or the linkflags or both (as above). I would be happy, if anyone could explain that behavior or could even provide a workaround/fix for me, so that "libstdc++" is no longer linked. Thanks, Deniz PS: BTW, does anyone know if it is correct, that shared libraries (in general) that are linked against "libc++abi" do always also link to "libgcc_s"? (What is it for?) -- BENOCS GmbH Dipl.-Inform. Deniz Bahadir Winterfeldtstr. 21 10781 Berlin Germany Phone: +49 - 30 / 577 0004-22 Email: deniz.bahadir(a)benocs.com www.benocs.com Board of Management: Michael Wolz, Dr.-Ing. Oliver Holschke, Dr.-Ing. Ingmar Poese Commercial Register: Amtsgericht Bonn HRB 19378

2 2

Fwd: Boost coroutine/context libraries
by Gregory Laird 20 Jan '17

20 Jan '17

--------------- Forwarded message (begin) Subject: Boost coroutine/context libraries From: Gregory Laird <glaird(a)northgatebowl.com> Date: Tue, 17 Jan 2017 17:10:57 -0800 Newsgroup: gmane.comp.lib.boost.user I am trying to write a cooperative task scheduler in c++ and I have discovered the boost context and coroutine libraries. I have written these sorts of schedulers many times in assembly language so I appreciate the issues in the methodology. I have written lots of c code but am less versed in c++. I have been trying to find some straightforward examples of the libraries use without a lot of extraneous c++ language elements (e.g. binds, lambdas, etc.). I did find one example http://stackoverflow.com/questions/11716291/boost-context-class that is very nice, but it will not compile now with the current boost library. I get an error saying that boost::context::fcontext_t has been removed from the public api. Could someone direct me to some examples that demonstrate the context or coroutine functionality that would be easier to understand. I want to write a scheduler that is very similar to the example listed above where coroutines yeild back to a main caller which then chooses the next coroutine to continue its processing. Thanks. --------------- Forwarded message (end)

5 6

Program abort at canonical()
by Andy 20 Jan '17

20 Jan '17

I have: #include <boost/filesystem/path.hpp> #include <boost/filesystem/operations.hpp> boostPath = canonical(boostPath); While I debug it is OK, but when I run debug or release x64 program standalone from cmd - it abort at canonical without any info.

4 6

Multiprecision -- cubing a number gives unexpected result
by Craig Henderson 19 Jan '17

19 Jan '17

Hello, I am new to Boost.Multiprecision and have come across a result that I cannot understand. I have the following code to calculate the sum of three cubes, using bigint_t = boost::multiprecision::mpz_int; bigint_t a = 2220422932; bigint_t b = 2218888517; b = -b; bigint_t c = 283059965; c = -c; std::cout << '(' << a << ")^3+(" << b << ")^3+(" << c << ")^3 = "; bigint_t a1 = a * a * a; bigint_t b1 = b * b * b; bigint_t c1 = c * c * c; std::cout << '(' << a1 << ")+(" << b1 << ")+(" << c1 << ") = "; std::cout << bigint_t(a1+b1+c1) << '\n'; and it produces the correct result (2220422932)^3+(-2218888517)^3+(-283059965)^3 = (10947302325566084787191541568)+(-10924622727902378924946084413)+(-22679597663705862245457125) = 30 However, if I use this implementation, with a,b and c assigned the result of the multiplication, a = a * a * a; b = b * b * b; c = c * c * c; std::cout << '(' << a << ")+(" << b << ")+(" << c << ") = "; std::cout << bigint_t(a+b+c) << '\n'; ... I get a very strange result: (24307641127223864542936438774018437376)+(24240519923499804093545671548118385521)+(6419686120902663137493915211500625) = 48554580736844571299619604237348323522 What is going on? Is this an invalid use of the multiplication that I have missed? I get the same result with cpp_int and gmp; using bigint_t = boost::multiprecision::cpp_int; using bigint_t = boost::multiprecision::mpz_int; I'm using g++ (Ubuntu 5.4.0-6ubuntu1~16.04.4) 5.4.0 20160609, Boost 1.58 Many thanks Craig

2 1

[Boost.Math] Weird compiler-error with Clang 3.8
by Deniz Bahadir 19 Jan '17

19 Jan '17

Hi guys, Our build-server (Jenkins) tried to build Boost 1.58 with Clang 3.8 and libc++ and tripped over an error that I do not understand. Maybe someone can shed some light on it? The specific error is this: clang-linux.compile.c++.pch /home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost/boost/bin.v2/libs/math/build/clang-linux-3.8/release/threading-multi/../src/tr1/pch.hpp.pth clang: error: cannot specify -o when generating multiple output files Does this look familiar to anyone? My build-environment is Ubuntu 16.04 x86_64 and I invoked Boost.Build like this: cd /home/jenkins/workspace/Pipeline_Boost/sources/boost_1_58_0 COMMON_COMPILER_FLAGS="-m64 -fpic -fdiagnostics-color=always -O3" CFLAGS="$COMMON_COMPILER_FLAGS" CXXFLAGS="$COMMON_COMPILER_FLAGS -std=c++14 -stdlib=libc++ -isystem/usr/include/libcxxabi -lc++abi" CC=/usr/bin/clang-3.8 CXX=/usr/bin/clang++-3.8 /home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost-build/b2 -j1 -q \ --build-dir=/home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost \ --build-type=minimal --layout=versioned address-model=64 install \ --prefix=/home/jenkins/workspace/Pipeline_Boost@3/clang38/installed/opt/BENOCS/clang38 \ --libdir=/home/jenkins/workspace/Pipeline_Boost@3/clang38/installed/opt/BENOCS/clang38/lib/x86_64-linux-gnu \ dll-path==/opt/BENOCS/clang38/lib/x86_64-linux-gnu \ toolset=clang-3.8 \ cflags="$CFLAGS" CXXFLAGS="$CXXFLAGS" linkflags="-Wl,--as-needed" The more detailed error-message is the following: common.mkdir /home/jenkins/workspace/Pipeline_Boost(a)3/clang38/build/boost/boost/bin.v2/libs/math/build common.mkdir /home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost/boost/bin.v2/libs/math/build/clang-linux-3.8 common.mkdir /home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost/boost/bin.v2/libs/math/build/clang-linux-3.8/release common.mkdir /home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost/boost/bin.v2/libs/math/build/clang-linux-3.8/release/threading-multi common.mkdir /home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost/boost/bin.v2/libs/math/build/clang-linux-3.8/release/threading-multi/.. common.mkdir /home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost/boost/bin.v2/libs/math/build/clang-linux-3.8/release/threading-multi/../src common.mkdir /home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost/boost/bin.v2/libs/math/build/clang-linux-3.8/release/threading-multi/../src/tr1 clang-linux.compile.c++.pch /home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost/boost/bin.v2/libs/math/build/clang-linux-3.8/release/threading-multi/../src/tr1/pch.hpp.pth clang: error: cannot specify -o when generating multiple output files rm -f "/home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost/boost/bin.v2/libs/math/build/clang-linux-3.8/release/threading-multi/../src/tr1/pch.hpp.pth" && "clang++" -x c++-header -m64 -fpic -fdiagnostics-color=always -O3 -std=c++14 -stdlib=libc++ -isystem/usr/include/libcxxabi -lc++abi -m64 -fpic -fdiagnostics-color=always -O3 -O3 -Wno-inline -Wall -pthread -fPIC -m64 -m64 -fpic -fdiagnostics-color=always -O3 -std=c++14 -stdlib=libc++ -isystem/usr/include/libcxxabi -lc++abi -m64 -fpic -fdiagnostics-color=always -O3 -DBOOST_ALL_NO_LIB=1 -DBOOST_BUILD_PCH_ENABLED -DBOOST_MATH_TR1_DYN_LINK=1 -DNDEBUG -I"." -I"libs/math/src/tr1" -Xclang -emit-pth -o "/home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost/boost/bin.v2/libs/math/build/clang-linux-3.8/release/threading-multi/../src/tr1/pch.hpp.pth" "libs/math/build/../src/tr1/pch.hpp" ...failed clang-linux.compile.c++.pch /home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost/boost/bin.v2/libs/math/build/clang-linux-3.8/release/threading-multi/../src/tr1/pch.hpp.pth... ...skipped <p/home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost/boost/bin.v2/libs/math/build/clang-linux-3.8/release/threading-multi>assoc_laguerre.o for lack of <p/home/jenkins/workspace/Pipeline_Boost@3/clang38/build/boost/boost/bin.v2/libs/math/build/clang-linux-3.8/release/threading-multi>../src/tr1/pch.hpp.pth... ...failed updating 1 target... ...skipped 1 target... ...updated 11871 targets... Can someone tell me what is going on and how to fix or work around this problem? Thanks, Deniz -- BENOCS GmbH Dipl.-Inform. Deniz Bahadir Winterfeldtstr. 21 10781 Berlin Germany Phone: +49 - 30 / 577 0004-22 Email: deniz.bahadir(a)benocs.com www.benocs.com Board of Management: Michael Wolz, Dr.-Ing. Oliver Holschke, Dr.-Ing. Ingmar Poese Commercial Register: Amtsgericht Bonn HRB 19378

3 4