- Boost-users - lists.boost.org

[Boost.Test] Three-way comparison operator
by Leo Carreon 17 May '20

17 May '20

Hi Boosters, Does anyone know if Boost.Test has been updated to provide support for the three-way comparison operator? I’m asking because I’ve encountered a situation where my user-defined class provides the three-way comparison operator but when I perform comparisons using Boost.Test it complains that my class has no ‘operator==’ which is true because it provides ‘operator<=>’. Please note that I’m using Boost 1.73.0 and g++ 10.0.1. Is there any plan to add the three-way comparison operator support? Kind regards, Leo

2 1

warning C4275, C4271when using boost::noncopyable
by Roman Savchenko 14 May '20

14 May '20

Hi All, I'm trying to update from boost 1.62 to boost 1.73. Boost 1.62 was built with VS15 140 toolchain with shared runtime (/MD) and as static lib. Boost 1.73 was built with VS19 140 toolchain with /MD and as static lib. I have a class: #include <boost/noncopyable.h> class __declspec(dllexport) A : private boost::noncopyable when I'm using 1.62 I do not have such warning. when I'm using 1.73 I do have such warning. I'll appreciate if anyone could help me to understand what I'm missing. Why does new version of header generate this warning? Thanks, Roman

2 3

(no subject)
by Ritajit Kundu 11 May '20

11 May '20

I am trying to compute a 2 Dimensional numerical integration with the help of Boost. I've been reading the documentation for a couple of days and my CPP language knowledge is just up to legacy CPP level so I'm a noob. I saw this page <https://www.boost.org/doc/libs/1_73_0/libs/math/doc/html/math_toolkit/gauss…> in documentation which calculates the quadrature using* Gauss-Kronrod *rule. I was trying to build upon it and calculate a 2-d integral using Gauss-Kronrod rule. My understanding is *a 2D integral is basically 2 1D integrals with the Cartesian product.* So I was thinking If I can nest two integration function then I'm done. I've started from the aforementioned example and extended it a bit so the code for a 2d integration looks like, #include <boost/math/quadrature/gauss_kronrod.hpp> #include <iostream> int main(int argc, char *argv[]) { using namespace boost::math::quadrature; auto f1 = [](double t, double s) { return std::exp(-(t*t+s*s+t*s) / 2); }; double error; double Q = gauss_kronrod<double, 15>::integrate(gauss_kronrod<double, 61>::integrate(f1, 0, std::numeric_limits<double>::infinity(), 5), 0, std::numeric_limits<double>::infinity(), 5, 1e-9, &error); std::cout << Q << " " << error<<std::endl; return 0; } . I've just added a lambda function (which is the integrand) of two variables and nested 2 1D Gauss-Kronrod integration functions. I'm compiling this with g++ -Wall -I /path/to/boost_1_73_0 2d-gauss.cpp This gives error and compilation fails. I don't know what wrong I'm doing. How do you suggest to do such an integral? Thanks in advance.

2 1

Google Summer of Code 2020 selected projects
by David Bellot 09 May '20

09 May '20

Dear Boost community, I would like to welcome our Google Summer of Code. 2020 students. This year, 10 students will work very very hard on improving and adding new features to many of our libraries. Once again, the competition has been very hard. A few numbers: - we received 93 proposals in total. 85 of them were complete, then we kept 47 proposals only - finally we selected 10 of them, covering 7 different Boost libraries - and countless hours of work from the most incredible team of mentors! Here are the list of projects for 2020: *Boost.Real* Kishan Shukla will be working on making the Real library ready for revision. His main goal is to reimplement the division algorithm so it can be used with any number base representation and the Karatsuba multiplication algorithm. Also, it will be working on implementing some operators and common irratianal numbers as Pi. Mentors: Laouen Belloli and Damian Vicino. *Boost.Astronomy* Syed Ali Hasan will be working on the astronomical coordinate system, he will implement different coordinate conversion using SOFA library Mentor: Pranam Lashkari Gopi Krishna Menon will be completing the parser for FITS file and also optimise the already developed components. Mentor: Sarthak Singhal *Boost.GIL* Debabrata Mandal will be implementing histogram computation as a built-in feature, algorithms for histogram analysis, image processing algorithms based on histogram and possibly integration with Boost.Histogram library. *Mentor*: Mateusz Loskot, Pranam Lashkari Olzhas Zhumabek will be extending GIL's portfolio of image processing algorithms, implementing for image scaling and sampling, Perona–Malik diffusion, Non-Maximum Suppression for object detection, Hysteresis threshold and some more. *Mentor*: Mateusz Loskot *Boost.Multiprecision* Dimitris Los will work on extending and optimizing parts of Boost.Multiprecision to higher precision of many thousands of bits or more. *Mentor*: Christopher Kormanyos *Boost.uBlas* Ashar Khan will stabilize and improve Boost.uBlas using the C++20-Standard. The main focus of his project will be enhancing code-coverage and documentation, integrating static code analysis and generating tensor concepts. *Mentor*: Cem Bassoy Tom Kwok will improve and finalize the implementation of data frames in uBlas, with the aim of delivering production-ready by the end of the summer. Data frame in uBlas are similar in functionalities to other implementations like in R, Python.Pandas or Julia. *Mentor*: David Bellot *Boost.Geometry* Tinko Bartels will work on robust yet efficient geometric predicates for Boost Geometry. *Mentor*: Vissarion Fisikopoulos *Boost.Real* Vikram Singh Chundawat will work in taking Real to revision ready state. His focus will be in improving the internal representation datatypes and providing functions required to conveniently support Taylor series. *Mentors*: Damian Vicino and Laouen Belloli Best regards, David Bellot

1 0

boost::filesystem::perms on Windows
by Dominique Devienne 07 May '20

07 May '20

Hi. I'm not having any luck with finding out whether a folder is writable or not, on Windows. Is Boost.Filesystem (1.64) supposed to work on Windows for permissions? Especially in "special" folders like those in "C:\Program Files"? C:\Program Files\Vendor\App>echo foo > bar Access is denied. But Boost.Filesystem reports permissions as rw-rw-rw-, when printing them using the BFS-equivalent to the example available on CppReference https://en.cppreference.com/w/cpp/filesystem/perms Despite obviously the folder not being writable, w/o being an administrator. I'm naively using simple code like below: ``` fs::path path(name); fs::file_status stat = fs::status(path); fs::perms p = stat.permissions(); demo_perms(p); ``` Am I missing something obvious? Thanks, --DD PS: I worked around the above, actually trying to create a file, but that has corner cases too, and I'd much prefer a reliable and cross-platform .permissions() from Boost. Smoothing-over platform differences is one "raison d'etre" from Boost after all, no?

2 2

Fwd: [serialization] is the weak_ptr parent kept inside an iarchive
by ypotin 06 May '20

06 May '20

Hello, I was running some tests involving std::weak_ptr using boost serialization, and I realized that when I read a std::weak_ptr from an archive that doesn't contain the "parent" std::share_ptr, the behavior of the read pointer is "hard to understand". Here is an example : #include <cassert> #include <sstream> #include <memory> #include <boost/archive/binary_iarchive.hpp> #include <boost/archive/binary_oarchive.hpp> #include <boost/serialization/export.hpp> #include <boost/serialization/shared_ptr.hpp> #include <boost/serialization/weak_ptr.hpp> class MyInt { public: MyInt() : value(0) { } int value; template<class Archive> void serialize(Archive &ar, unsigned int) { ar & value; } }; int main() { std::stringstream data; { std::shared_ptr<MyInt> shared = std::make_shared<MyInt>(); assert(shared.use_count() == 1); shared->value = 12; /* Write a weak_ptr */ { std::weak_ptr<MyInt> weak(shared); assert(shared.use_count() == 1); boost::archive::binary_oarchive ar(data); ar << weak; assert(shared.use_count() == 1); } assert(shared.use_count() == 1); /* Read a weak_ptr */ std::weak_ptr<MyInt> weak; { boost::archive::binary_iarchive ar(data); ar >> weak; assert(shared.use_count() == 1); assert(weak.use_count() == 1); { std::shared_ptr<MyInt> loaded = weak.lock(); assert(loaded); /* It looks good */ assert(loaded->value == 12); /* Let's see... */ assert(loaded != shared); shared->value = 42; assert(loaded->value == 12); /* So another shared_ptr exists somewhere... */ assert(shared.use_count() == 1); assert(loaded.use_count() == 2); assert(weak.use_count() == 2); } /* But where ? */ assert(shared.use_count() == 1); assert(weak.use_count() == 1); } /* Let's see */ assert(shared.use_count() == 1); assert(weak.use_count() == 0 /* weak.expired() */); /* Hmm, was it inside the archive itself ? */ } /* Let's try again, we need the input stream to be back to the beginning */ data.seekg(0); std::weak_ptr<MyInt> weak; { boost::archive::binary_iarchive ar(data); assert(weak.use_count() == 0); ar >> weak; assert(weak.use_count() == 1); /* It looks fine */ } assert(weak.use_count() == 0); /* aw, on a second thought, it really doesn't */ return 0; } I am building and running this using g++-9.3.0 with boost-1.72.0-1. Clearly, the weak_ptr gets itself a shared_ptr "somewhere" as it can be used after being read. I was wondering "where it is stored and how to access it, if possible, without some `weak.lock()` call ?" It looks like the shared_ptr gets destroyed at the same time as the archive, so I guess it is cleaned up thanks to the delete_created_pointers method when the iarchive is destroyed : does delete_created_pointers actually contains also shared_ptr ? As far as I can see in the code serialization/weak_ptr.hpp, we actually save then load a std::shared_ptr, so I was expecting the weak_ptr to expire as soon as the call returns : since it is not, there *is* another reference. Now, I know I must save the shared_ptr in the archive myself to have my own reference and to get all weak_ptr running properly. But I came across this use case and I can't take it out of my mind, so if you have an explanation, I'm all ears. Thank you

1 0

Boost 1.73, beast and -fconcepts
by David P. Riedel 04 May '20

04 May '20

Hi It seems that beast still does not work with gcc 9.3 and -fconcepts. If I try to build the beast test programs using -fconcepts, I still get a lot of errors such as those below. I thought this was fixed in 1.73. ???? Thanks --------------------- In file included from ../../../boost/beast/core/basic_stream.hpp:1457, from ../../../boost/beast/core.hpp:16, from doc/core_snippets.cpp:19: ../../../boost/beast/core/impl/basic_stream.hpp:781:1: error: no declaration matches ‘auto boost::beast::basic_stream< <template-parameter-1-1>, <template-parameter-1-2>, <template-parameter-1-3> >::async_connect(const endpoint_type&, ConnectHandler&&)’ 781 | basic_stream<Protocol, Executor, RatePolicy>:: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ In file included from ../../../boost/beast/core.hpp:16, from doc/core_snippets.cpp:19: ../../../boost/beast/core/basic_stream.hpp:1202:5: note: candidates are: ‘template<class Protocol, class Executor, class RatePolicy> template<class Iterator, class ConnectCondition, class IteratorConnectHandler> requires completion_token_for<IteratorConnectHandler, void()> typename boost::asio::async_result<typename std::decay<RangeConnectHandler>::type, void(boost::system::error_code, Iterator)>::return_type boost::beast::basic_stream< <template-parameter-1-1>, <template-parameter-1-2>, <template-parameter-1-3> >::async_connect(Iterator, Iterator, ConnectCondition, IteratorConnectHandler&&)’ 1202 | async_connect( | ^~~~~~~~~~~~~ ../../../boost/beast/core/basic_stream.hpp:1141:5: note: ‘template<class Protocol, class Executor, class RatePolicy> template<class Iterator, class IteratorConnectHandler> requires completion_token_for<IteratorConnectHandler, void()> typename boost::asio::async_result<typename std::decay<RangeConnectHandler>::type, void(boost::system::error_code, Iterator)>::return_type boost::beast::basic_stream< <template-parameter-1-1>, <template-parameter-1-2>, <template-parameter-1-3> >::async_connect(Iterator, Iterator, IteratorCon

2 2

[review] [LEAF] LEAF Review Starts May 22
by Michael Caisse 01 May '20

01 May '20

The Boost formal review for LEAF (Lightweight Error Augmentation Framework) will take place May 22, 2020 through May 31st. LEAF is brought to us by Emil Dotchevski, the author of Boost.Exception. Similar to Boost.Exception, this library allows arbitrary error objects to be returned; however, unlike Boost.Exception it does not require dynamic memory. The library can be used with or without exception handling. Error handling is such an import part of every software system. LEAF provides a unique solution to the problem. The excellent documentation can be found here: <https://zajo.github.io/leaf/> and includes a tutorial, examples, reference, design rationale, and comparisons to other error handling approaches/libraries. The source is available at github. The review branch will be used during the review period. <https://github.com/zajo/leaf/tree/review> The purpose of this announcement is to give people some time to start looking at the documentation and source code for the library. Feel free to begin asking questions. Most of all, get involved! Everybody has some experience with handling errors. Tell the Boost community what you think of LEAF. Thank you for your participation! michael -- Michael Caisse Ciere Consulting ciere.com

1 0

Three-way comparison operator
by Leo Carreon 30 Apr '20

30 Apr '20

Hi Boosters, Does anyone know if Boost.Test has been updated to provide support for the three-way comparison operator? I’m asking because I’ve encountered a situation where my user-defined class provides the three-way comparison operator but when I perform comparisons using Boost.Test it complains that my class has no ‘operator==’ which is true because it provides ‘operator<=>’. Please note that I’m using Boost 1.73.0 and g++ 10.0.1. Is there any plan to add the three-way comparison operator support? Kind regards, Leo

4 4

Windows boost::filesystem canonical on NTFS network drives mounted in a directory
by gga 29 Apr '20

29 Apr '20

When on Windows, if you use boost::filesystem::canonical() to resolve a NTFS network drive mounted in a directory, the result is what I see as a useless result. For example: C:\ntfs_mount\other\dirs\file boost::filesystem returns: C:\?VolumeXXX-XXX-XXX-XXX\ other\dirs\file This volume notation is useless as it cannot be passed to any file function. Also, the path is divided by a space to other\dirs. It would be better if canonical does not mess with the network drive at all and returns the directory as is, like boost::filesystem::absolute() does. -- Gonzalo Garramuño

1 0