May 2017 - Boost-users - lists.preview.boost.org

[Fiber] Can a fiber promise be set in a different OS thread on which has no fibers running
by lin li 11 May '17

11 May '17

I have tested the following example code with Boost 1.61. And it worked (ignore the potential race conditions when accessing the variable 'p'). But I am not sure the fiber promise can be used in this way or I am just being lucky. #include <cstdlib> #include "boost/fiber/fiber.hpp" #include "boost/fiber/future.hpp" #include "boost/thread.hpp" #include <set> #include <iostream> #include <boost/date_time/posix_time/posix_time_duration.hpp> #include <boost/thread/pthread/mutex.hpp> #include <boost/thread/lock_guard.hpp> #include "boost/thread/pthread/thread_data.hpp" #include "boost/chrono/duration.hpp" using namespace std; boost::fibers::promise<int>* p; void wait() { p = new boost::fibers::promise<int>; auto future = p->get_future(); auto status = future.wait(); std::cout << "Value: " << future.get() << std::endl; delete p; } void f() { boost::fibers::fiber fiberJob(&wait); fiberJob.join(); } void setf() { // Is it safe to do this? boost::this_thread::sleep(boost::posix_time::seconds(5)); p->set_value(20); } void f1() { boost::fibers::fiber fiberJob(&setf); fiberJob.join(); } int main(int argc, char** argv) { boost::thread t(&f); setf(); cout << "Here" << endl; t.join(); cout << "Here 2" << endl; return 0; }

2 1

Cutting headers
by Igor Mironchik 10 May '17

10 May '17

Hello, How can I cut specific library from Boost, if, for example, I don't want to place all Boost but only one library from Boost? Is it possible? --- This email has been checked for viruses by AVG. http://www.avg.com

11 17

boost::asio::deadline_timer cancel and expires_from_now
by jupiter 10 May '17

10 May '17

Hi, Is it necessary to call the timer.cancel() before timer.expires_from_now? I think it is not just unnecessary, it will be a bug to call the timer.cancel() before timer.expires_from_now, I think that will run the cancel if the previous timer is not expired, is it correct? Thank you

2 2

database access library in boost?
by Paolo Bolzoni 09 May '17

09 May '17

Dear list, I was wondering, why there is no database access library in boost? (am I wrong?) It seems an imporant feature to not-have in a such library collection, so there must be a reason. Is there some "official" raccomandation? Soci, cppdb? Your faithfully, Paolo

3 2

Fwd: configure Jamfile boost 1_63_0 tutorial to build from command line with bjam
by Giulio De Luca 09 May '17

09 May '17

I´m totally new to boost, I just downloaded boost_1_63_0 and I´m trying to build the tutorial example located in \boost_1_63_0\libs\python\example\tutorial by following the tutorial documentation <http://www.boost.org/doc/libs/1_62_0/libs/python/doc/html/tutorial/index.ht…> http://www.boost.org/doc/libs/1_62_0/libs/python/doc/html/tutorial/ index.html The example is quite simple: it only contains 4 files: 1. hello.cpp 2. hello.py 3. Jamfile 4. ../Jamroot For simplicity I´m pasting below the content of each file : // hello.cpp #include <boost/python/module.hpp> #include <boost/python/def.hpp> char const* greet() { return "hello, world"; } BOOST_PYTHON_MODULE(hello_ext) { using namespace boost::python; def("greet", greet); } // hello.py import hello_ext print(hello_ext.greet()) And finally the boost configuration file: // Jamfile import python ; project tutorial : requirements <location>. ; python-extension hello_ext : hello.cpp ; run-test hello : hello_ext hello.py ; alias test : hello ; explicit test ; // ../Jamroot import python ; if ! [ python.configured ]{ ECHO "warning: no Python configured in user-config.jam" ; ECHO "warning: will use default configuration" ; using python ;} # Adjust the following if Boost.Python isn't installed in a default location lib boost_python ; project : requirements# <include>/path/to/boost/python <library>boost_python; rule run-test ( test-name : sources + ){ import testing ; testing.make-test run-pyd : $(sources) : : $(test-name) ;} build-project quickstart ; build-project tutorial ;if [ python.numpy ]{ build-project numpy ;} When I execute bjam.exe from my cmd prompt (after properly adding in the environment variables the path where bjam.exe is located) I get the following errors: ...found 12 targets......updating 5 targets... compile-c-c++ hello.obj hello.cpp hello.cpp(7) : fatal error C1083: Cannot open include file: 'boost/python/module.hpp': No such file or directory call "C:\Users\...\AppData\Local\Temp\b2_msvc_11.0_vcvarsall_x86.cmd" >nul cl /Zm800 -nologo @"hello.obj.rsp" ...failed compile-c-c++ hello.obj......skipped <p.>hello_ext.pyd for lack of <p.>hello.obj......skipped <p.>hello_ext.pdb for lack of <p.>hello.obj......skipped <p.>hello for lack of <p.>hello_ext.pyd......failed updating 1 target......skipped 4 targets... When I try and build the project from VS2012 I can build it properly because i know hot to add additional include libraries to the project, but I cannot find the proper way to config the Jamfile to make it build from command line. I tried to config the Jamroot file by changing the path to the boost/python libraries local path but with no luck. I wonder where and how should I add the absoulute path of the boost/python libraries. Any help would be much appreciated Thanks in advance!

1 0

"Invalid handle" error trying to use boost::process
by tom＠quarendon.net 09 May '17

09 May '17

I'm trying to run a simple program with boost::process, with boost 1.64 I'm using Windows, and visual studio 2015. I'm just trying to run the supplied "async_cio.cpp" example: My program is: boost::asio::io_service ios; boost::asio::streambuf buffer; bp::child c( "cmd /c dir", bp::std_out > buffer, ios ); ios.run(); (I've changed if from launching "test.exe" to launching "cmd /c dir"). If I run this I get: Exception thrown at 0x49409D2A in ConsoleApplication3.exe: 0xC0000008: An invalid handle was specified. Running in the debugger, it is in the destructor of the async_pipe object, in "win_iocp_handle_service::close_for_destruction". I've tried various of the examples with varying degrees of success. But any time I seem to use an async_pipe, I seem to get an "invalid handle" exception when it is calling CloseHandle on the completion port handle. Any ideas what I'm doing wrong? Thanks.

3 3

[Boost.Variant] Is there an r-value version of boost::get for variant?
by dariomt＠gmail.com 08 May '17

08 May '17

As seen in the std docs [1], std::get for variant provides r-value overloads. Does Boost.Variant have them? I'm interested in having this overload in particular: template<typename U, typename T1, typename T2, ..., typename TN> U && get(variant<T1, T2, ..., TN> && operand); Thanks in advance! [1] http://en.cppreference.com/w/cpp/utility/variant/get

2 2

[Units] Units in unit tests
by Stephan Menzel 08 May '17

08 May '17

Hello all, please excuse the weird subject line. I have just introduced boost units as length quantity for vectors into my project here. What previously was: typedef boost::qvm::vec<double, 3> vec3; is now: typedef boost::units::quantity<length, double> qlength; typedef boost::qvm::vec<qlength, 3> vec3; Which seems to work OK so far, I'm not done testing. While converting all my unit tests I came to wonder is how this is best adopted. I have lots of lines like: vec3 v = get_some_value(); and then I test the individual component values like this: BOOST_CHECK_CLOSE(X(v), 42.0, 0.001); // check for length of 42 meters Now this would have to be rewritten to: BOOST_CHECK_CLOSE(X(v).value(), 42.0, 0.001); But doing this, I imply the length is in meters, right? Doesn't that defeat the purpose a little? I feel this would be better: BOOST_CHECK_CLOSE(X(v), 42.0 * meters, 0.001); but the Macro won't accept this as this is not a floating point type anymore. Does anybody have some hint as to how this is best addressed while still have maximum readability? All my best... Stephan

3 5

Re: [Boost-users] Help with Boost Serialization library!
by label 03 May '17

03 May '17

Hi Tina, When you build the boost libraries, is there a way to generate .lib files for them? Below, I have documented the steps I took to build the library. Please take a glance at it and let me know if I missed something that will allow me to obtain the .lib file. I cannot see your pictures (A non-text attachment was scrubbed...), so I can only guess what goes wrong. My way building boost is (on Windows with VC++ 2015) cd c:\boost_1_64_0 .\bootstrap .\b2 --without-mpi --without-python --build-type=complete --reconfigure msvc stage This generates library files, for example libboost_serialization-vc140-mt-1_64.lib in the stage subdirectory of my boost root directory, here c:\boost_1_64_0\stage __________________________________ Freundliche Grüße * Best regards * Met vriendelijke groeten * Meilleures salutations Holger Gerth _________________________________________________________________________________ EB-Soft Gesellschaft fuer elektronische Beschriftungs-Software mbH Neckargartacher Str. 114, 74080 Heilbronn, Germany Tel: +49 (0)7131 / 9466-0 Email: info(a)ebsoft.de Fax: +49 (0)7131 / 1099-0 Homepage: www.ebsoft.de Amtsgericht-Registergericht-Stuttgart HRB 102882 Geschaeftsfuehrer: Michael Ehring, Helmut Sigler _________________________________________________________________________________

1 0

externing template instantiations of Boost.Multiprecision numbers
by Dani Brake 02 May '17

02 May '17

Hi Boost, I happily use Boost.Multiprecision inside a largely templated codebase, which involves the entirely-templated Eigen as well. My compile times are high. To gain insight into why compile time is so ridiculous, I used Templight <https://github.com/mikael-s-persson/templight> with Clang to query how many times my templates are instantiated, and how long they take. The answer was overwhelmingly that Boost.Multiprecision's number class is to blame, with 8264 instantiations of bmp::number<bmp::backends::mpft_float_backend<0,allocate_dynamic>, bmp::expression_template_option::et_on> 5052 of the same with the gmp_int backend, and 2689 for mpz_rational. I think the cure might be an extern template declaration on these three types, but where is the correct place to do this? In a header *I* have written which provides a few extensions / additions I needed? Or do I need to patch my install of BMP and put it there? As a follow-up, I have a cycle of instantiations I believe, but do not yet know, are expression templates. Can I extern a few common instantiations, and let the compiler do the rest? Any help or techniques are greatly appreciated. Especially as I (slowly) move toward applying expression templates to complex multiprecision arithmetic, too. Thanks a ton, Dani Brake University of Notre Dame Applied and Computational Mathematics and Statistics

1 0

Help with Boost Serialization library!
by Alam, Tina 02 May '17

02 May '17

Hi Boost Users, I need the Boost Serialization Library in order to use a MATLAB make file. The make file is looking for a static library file named either boost_serialization.lib or libboost_serialization.lib. However, when I built the serialization library, I got a lot of object files and rsp files rather than a .lib file. Do you know how to create the .lib file? I am not sure if it matters, but I am using a Windows computer. When you build the boost libraries, is there a way to generate .lib files for them? Below, I have documented the steps I took to build the library. Please take a glance at it and let me know if I missed something that will allow me to obtain the .lib file. 1. Download and unzip the boost folder http://www.boost.org/users/history/version_1_63_0.html 1. Look inside folder and make sure you find bootstrap.bat 1. Open the Visual Studio Developer Command Prompt and navigate to the boost folder [cid:image001.jpg@01D2BF3C.88D418B0] 1. Type bootstrap into the command line and hit enter [cid:image002.jpg@01D2BF3C.88D418B0] 1. After it finishes bootstrapping, type in .\b2 --with-serialization and hit enter. The serialization library will be built. I still have not figured out how to generate the .lib file. [cid:image003.jpg@01D2BF3C.88D418B0] Tina Alam IT Flex - Software Engineer

2 1

[BGL] null edge descriptor
by Ireneusz Szcześniak 01 May '17

01 May '17

Hi, I'm using Boost Graph Library (BGL), and I need a null edge. For years I was using G::edge_descriptor(), i.e., a default-constructed edge descriptor, and the code was working. However, I wonder if it is correct. Are all default-constructed edges equal? If I get a default-constructed edge descriptor e1 at some point in code, and another default-constructed edge descriptor e2 at some other point in code, will there descriptors always equal, i.e., will e1 == e2 always hold? Is it guaranteed that a default-constructed edge does not refer to an existing edge in a graph? Best, Irek

1 0