- Boost-users - lists.boost.org

boost::unordered_map takes too long when I have only two elements in it
by Ram 14 Jul '17

14 Jul '17

Hi, I am using a boost::unordered_set but I sometimes have just two elements in it. It takes a long time then. Is it wrong if I have too few elements? Thanks, Ram

3 3

rational - 1.64 version shear?
by Nicholas Devenish 14 Jul '17

14 Jul '17

With boost/rational.hpp, the code: boost::rational<int> a{1}; long b{1}; auto c{a * b}; Works fine up to 1.63.0, but fails to compile in 1.64+ on the multiplication with "error: invalid operands to binary expression". This seems to be due to some restructuring that changed the way multiplications are declared, which (as far as I can tell), meant the operation declaration became conditional on number of representable digits. This breaking change isn't listed on the release notes or the ::rational documentation, so I'm wondering if it was accidental and therefore a bug (whose behaviour might change), or a deliberate change to prevent downcasting and just undocumented? Nick (I've come onto a project stuck on 1.56 and trying to update it)

2 1

⁨[process] problem with boost::process::on_exit
by Sandy Martel 14 Jul '17

14 Jul '17

The code below outputs: 1- child about to sleep for 2s 2- child about to sleep for 4s 3- child finish sleeping for 2s 6- sleep 4s child finished 5- child finish sleeping for 4s 4- sleep 2s child finished in a posix environment. I would expect the output to be in-order, looks like I get the wrong on_exit notification for the child processes I launch. Q: is this a bug or am I doing something wrong ? What is the correct way to get notified of specific process termination ? Regards, Sandy. code: #include <boost/process.hpp> #include <boost/asio.hpp> #include <thread> #include <iostream> int main( int argc, char **argv ) { if ( argc > 2 && strcmp( argv[1], "sleep" ) == 0 ) { auto s = atoi(argv[2]); std::cout << (s/2) << "- child about to sleep for " << s << "s" << std::endl; std::this_thread::sleep_for( std::chrono::seconds(s) ); std::cout << (s+1) << "- child finish sleeping for " << s << "s" << std::endl; return 0; } namespace bp = boost::process; boost::asio::io_service ios; // launch a child that will sleep for 2s auto c1 = bp::child( argv[0], "sleep", "2", ios, bp::on_exit( [](int, const std::error_code&) { std::cout << "4- sleep 2s child finished" << std::endl; }) ); // wait a bit, make sure the child startup for my test std::this_thread::sleep_for( std::chrono::milliseconds(10) ); // launch a child that will sleep for 4s auto c2 = bp::child( argv[0], "sleep", "4", ios, bp::on_exit( [](int, const std::error_code&) { std::cout << "6- sleep 4s child finished" << std::endl; }) ); // wait for the notifications ios.run(); }

1 0

[serialization] "Archive" type turns into abstract class during save/load!
by r＠primef.actor 13 Jul '17

13 Jul '17

In a member load function function I call a static method that loads another archive file, i.e. something like this: ```c++ struct Bar; struct Foo { std::shared_ptr<Bar> data; template <class Archive> void load(Archive& archive, const unsigned /*version*/) { std::string key; archive >> key; data = Bar::loadSomewhere<Archive>(std::move(key)); } }; struct Bar { template <class Archive> std::shared_ptr<Bar> loadSomewhere(std::string key) { std::ifstream input(key); Archive ar(input); std::shared_ptr<Bar> bar(new Bar()); ar >> *bar; return bar; } }; ``` The problem is that `Bar::loadSomewhere` is passed a `boost::archive::polymorphic_iarchive`, while I invoke the serialization library with a `boost::archive::xml_iarchive`. This is an abstract class, so I can't instantiate it! So somewhere down the road the `xml_iarchive` is turned into a `polymorphic_iarchive`. I would like to prevent this from happening, because I instantiate an `Archive` object in the `Bar::loadSomewhere` function. How can I get around this problem?

1 0

Re: [Boost-users] Boost 1.65.0 Beta 1 Release Candidate 2
by Matthieu Brucher 11 Jul '17

11 Jul '17

This happens a lot when people develop on gcc or visual studio and then try building with clang. Having a function and a class with the same name is bound to generate errors. So properly should be tagged as a bug. Le 11 juil. 2017 11:14 AM, "Raffi Enficiaud via Boost-users" < boost-users(a)lists.boost.org> a écrit : Le 11.07.17 à 02:14, Marshall Clow via Boost-users a écrit : The second release candidates for the 1.65.0 beta release are now > available at: > > > https://dl.bintray.com/boostorg/beta/1.65.0.beta.1/source/ > boost_1_65_0_beta1_rc2.7z > > https://dl.bintray.com/boostorg/beta/1.65.0.beta.1/source/ > boost_1_65_0_beta1_rc2.tar.bz2 > > https://dl.bintray.com/boostorg/beta/1.65.0.beta.1/source/ > boost_1_65_0_beta1_rc2.tar.gz > > https://dl.bintray.com/boostorg/beta/1.65.0.beta.1/source/ > boost_1_65_0_beta1_rc2.zip > > The SHA256 checksums are as follows: > > d54d29568c4e4933cbfe40a292b6cb84a7f68121e34f6a83f7d0f7c652b6b2d8 > boost_1_65_0_beta1_rc2.7z > 092676557513e992347d30dd0cd058fb8af2a401e5da4f4413f5b577852d726b > boost_1_65_0_beta1_rc2.tar.bz2 > 31fa3c807e2205e18f7d5c74f77f540677f7b0937af1e37f595c137e1e87bf84 > boost_1_65_0_beta1_rc2.tar.gz > 83c8d8b76fa9591572226a1a49fd191d76db3011b52b907cbbc8d37d46213bd2 > boost_1_65_0_beta1_rc2.zip > > > This RC contains a single fix to the stacktrace library. > > As always, the release managers would appreciate it if you download > the candidate of your choice and give building it a try. Please report > both success and failure, and anything else that is noteworthy. > > Thanks! > > -- The release managers > Hi, I have this error during the build on OSX. Built with: ./b2 cxxflags="-std=c++11 -stdlib=libc++" linkflags="-stdlib=libc++" --without-log --layout=versioned install ...on 14200th target... common.mkdir bin.v2/libs/fiber/build/darwin-4.2.1/release/threading-multi common.mkdir bin.v2/libs/fiber/build/darwin-4.2.1/release/threading- multi/numa darwin.compile.c++ bin.v2/libs/fiber/build/darwin-4.2.1/release/threading- multi/numa/pin_thread.o libs/fiber/src/numa/pin_thread.cpp:23:10: warning: pin_thread() not supported [-W#pragma-messages] # pragma message "pin_thread() not supported" ^ 1 warning generated. darwin.compile.c++ bin.v2/libs/fiber/build/darwin-4.2.1/release/threading- multi/numa/topology.o libs/fiber/src/numa/topology.cpp:23:10: warning: topology() not supported [-W#pragma-messages] # pragma message "topology() not supported" ^ libs/fiber/src/numa/topology.cpp:27:14: error: use of undeclared identifier 'topology'; did you mean 'fibers::numa::topology'? std::vector< topology > topology() { ^~~~~~~~ fibers::numa::topology ./boost/fiber/numa/topology.hpp:38:21: note: 'fibers::numa::topology' declared here std::vector< node > topology(); ^ libs/fiber/src/numa/topology.cpp:27:14: error: template argument for template type parameter must be a type std::vector< topology > topology() { ^~~~~~~~ /Applications/Xcode8.1.app/Contents/Developer/Toolchains/Xco deDefault.xctoolchain/usr/bin/../include/c++/v1/vector:456:17: note: template parameter is declared here template <class _Tp, class _Allocator = allocator<_Tp> > ^ 1 warning and 2 errors generated. "g++" -ftemplate-depth-128 -O3 -Wall -dynamic -gdwarf-2 -fexceptions -Wno-inline -fPIC -arch x86_64 -std=c++11 -stdlib=libc++ -DBOOST_ALL_NO_LIB=1 -DBOOST_CONTEXT_DYN_LINK=1 -DBOOST_DISABLE_ASSERTS -DBOOST_FIBERS_DYN_LINK=1 -DBOOST_FIBERS_SOURCE -DBOOST_FILESYSTEM_DYN_LINK=1 -DBOOST_SYSTEM_DYN_LINK=1 -DNDEBUG -I"." -c -o "bin.v2/libs/fiber/build/darwin-4.2.1/release/threading-multi/numa/topology.o" "libs/fiber/src/numa/topology.cpp" ...failed darwin.compile.c++ bin.v2/libs/fiber/build/darwin -4.2.1/release/threading-multi/numa/topology.o... _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org https://lists.boost.org/mailman/listinfo.cgi/boost-users

2 1

Boost 1.65.0 Beta 1 Release Candidate 1
by Marshall Clow 09 Jul '17

09 Jul '17

The release candidates for the 1.61.0 release are now available at: https://dl.bintray.com/boostorg/beta/boost_1_65_0_beta1_rc1.7z https://dl.bintray.com/boostorg/beta/boost_1_65_0_beta1_rc1.tar.bz2 https://dl.bintray.com/boostorg/beta/boost_1_65_0_beta1_rc1.tar.gz https://dl.bintray.com/boostorg/beta/boost_1_65_0_beta1_rc1.zip The SHA256 checksums are as follows: cd7429340bb4aae38e0ad2e9f59c39e66335302c41577956123e55fc1e2a28ee boost_1_65_0_beta1_rc1.7z fda076f7b3c994524d9cd41926c72f4a9b7c017bef310a055956fdd0bed7c358 boost_1_65_0_beta1_rc1.tar.bz2 0d1473041f7a9fa9ce1e256c870cc52af6f98abf723cf4f1ac61c2b45f45d38a boost_1_65_0_beta1_rc1.tar.gz f45bbb75b84cdc0c66f8a571d77c518cfbac076d7a6fc7937c4812f0f242a792 boost_1_65_0_beta1_rc1.zip As always, the release managers would appreciate it if you download the candidate of your choice and give building it a try. Please report both success and failure, and anything else that is noteworthy. Thanks! -- The release managers

2 1

building for msvc-14.1
by Isaac Lascasas 09 Jul '17

09 Jul '17

Hello I have serious problems building boost for vs2017. My win10 install has just visual studio community 2017 installed right now, with windows SDK 10.0.15063.0 and 8.1. I run the vs2017 x64 native prompt and then I do something like this: b2 -j12 toolset=msvc-14.1 address-model=64 architecture=x86 link=static threading=multi runtime-link=shared --build-type=minimal stage --stagedir=stage/x64 some headers give weird stl container errors: c:\libs\boost_1_64_0__64\boost/archive/detail/basic_serializer_map.hpp(47): error C2039: 'set': is not a member of 'std' Thanks, Isaac.

1 0

[review] Mp11 review starts in one week
by Bjorn Reese 08 Jul '17

08 Jul '17

The formal review of Peter Dimov's Mp11 library is scheduled for July 15 - July 24, 2017 [1]. Mp11 is a C++11 metaprogramming library for compile-time manipulation of data structures that contain types. It’s based on template aliases and variadic templates and implements the approach outlined in the article "Simple C++ metaprogramming" [2] and its sequel [3]. These articles are useful background information for the review. * Mp11 aims to make simple usage simple, and to support complex usage without complicating the simple use cases. * Mp11 works on any type-list, whether its own type-list mp_list, or standard type-lists such as std::tuple and std::variant, or user-defined type-lists. * Mp11 works with any meta-function, such as C++11 or Boost type-traits, or user-defined type-traits. Mp11 can be found here: * Documentation: https://rawgit.com/pdimov/mp11/master/doc/html/mp11.html * Source code: https://github.com/pdimov/mp11/tree/master [1] http://www.boost.org/community/review_schedule.html [2] http://pdimov.com/cpp2/simple_cxx11_metaprogramming.html [3] http://pdimov.com/cpp2/simple_cxx11_metaprogramming_2.html

1 0

Serialization support for small_vector
by Brian Wood 07 Jul '17

07 Jul '17

Shalom I've added support for small_vector to the C++ Middleware Writer. While testing the support, I got a warning about dereferencing a type-punned pointer breaking strict-aliasing rules. That was on line 340 of small_vector.hpp. I was using g++ 7.1 on FreeBSD/ TrueOS. I used the following "Middle" code for testing: message-lengths variable example_messages -out (message_id_8, ::boost::container::small_vector<int32_t,30>) -in (::boost::container::small_vector<int32_t,30>) } >From what I could tell the serialization/deserialization worked fine. If small_vector were broken out into a separate library that would be helpful from my perspective. It would make it easier for me to add small_vector to my example code. Brian Ebenezer Enterprises - In G-d we trust. http://webEbenezer.net

2 1

[serialization] unable to link on OS X when MACOSX_DEPLOYMENT_TARGET=10.7
by Pavol Juhas 06 Jul '17

06 Jul '17

Hello, I'd like to build the boost_serialization library with a compatibility setup for Mac OS X version 10.7, because I need to use it with Anaconda Python binaries, which are built for OS X 10.7. The OS X compatibility version can be set either by passing `-mmacosx-version-min=10.7` as a compile and link option or by setting environment variable MACOSX_DEPLOYMENT_TARGET=10.7. If I do so, the boost_serialization library builds without complaints, however I am not able to link it with my test program. Here are the steps I take to compile boost_serialization and the test program starting from a fresh boost 1.64.0 tree: # -------------------------------------------------------------- $ export MACOSX_DEPLOYMENT_TARGET=10.7 $ ./bootstrap.sh $ ./b2 -j4 variant=release link=shared --with-serialization stage # The stage/lib directory now contains libboost_serialization.dylib. # Use it to build a test program from tarchive.cpp: $ c++ tarchive.cpp -I. -Lstage/lib -lboost_serialization Undefined symbols for architecture x86_64: "boost::archive::basic_binary_oprimitive<boost::archive::binary_oarchive, char, std::char_traits<char> >::init()", referenced from: boost::archive::binary_oarchive_impl<boost::archive::binary_oarchive, char, std::char_traits<char> >::init(unsigned int) in tarchive-5e24a9.o "boost::archive::basic_binary_oprimitive<boost::archive::binary_oarchive, char, std::char_traits<char> >::save(std::string const&)", referenced from: void boost::archive::save_access::save_primitive<boost::archive::binary_oarchive, std::string>(boost::archive::binary_oarchive&, std::string const&) in tarchive-5e24a9.o ... # -------------------------------------------------------------- The only thing the test program (attached) does is to instantiate the boost::archive::binary_oarchive class. The procedure above works fine when MACOSX_DEPLOYMENT_TARGET is set to 10.9 or higher. Am I missing some option that would fix the failed linking when building for OS X 10.7? I suppose this could be related to object visibility options as the serialization library is compiled with "-fvisibility=hidden" and "-fvisibility-inlines-hidden" options. Thank you, Pavol

2 3