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

why boost does not have skip list implementation?
by David Frank 01 Mar '18

01 Mar '18

like java concurrentSkipListSet for example. Thanks, David

1 0

PPA for boost
by legalize+jeeves＠mail.xmission.com 28 Feb '18

28 Feb '18

[Please do not mail me a copy of your followup] Hello fellow boosters, I am one of the maintainers of the C++ language track on exercism.io. When I first implemented this language track, we didn't have good regex support in g++'s standard library yet, so we needed boost.regex to fill that gap. Since we were already going to use boost.regex, I used boost.test for the unit test support. At the time, I found a PPA that was providing prebuilt binaries for ubuntu. Since that time, the need for boost.regex has gone away because the standard library implementation has been completed for g++. Along the way, the biggest source of problems in getting people going with C++ has been boost. For Windows users, it's annoying that CMake won't "just find" boost when you install it because there's no uniform mechanism between boost, packaging and CMake to say where boost got installed. For Mac users, they end up having difficulty with homebrew and fighting with Xcode. For Linux users, the PPA has gone stale and they end up being forced to build from scratch and get sent off into a ditch. In the meantime, we've got conan at stable 1.0 version right now, but that introduces a python dependency and then a conan dependency and for beginners I don't know if that's a path less fraught with peril. We've done some work on a branch to ditch boost.test and therefore boost entirely and switch the unit test framework to Catch. We have yet to "drop the hammer" on that and bring it onto master because reasons. (I am not the developer of that branch, but I suspect it hasn't happened yet because of the other integration issues that are needed; namely a bunch of documentation has to be updated and there will be teething pains with anyone who has already downloaded existing exercises and written them against boost.test.) Now, I find it hard to believe that I am the only one that has suffered from these issues all this time. Sure, if you only use header-only libraries then you don't need to build anything, but there are more than a few libraries in boost that either require or benefit significantly from being compiled. In all this time, have we not managed to come up with a stable, reliable PPA archive with ongoing support for boost for linux users? All I seem able to find are half-heartedly maintained PPAs with stale versions of boost on them. I have even tried to navigate going about setting up my own PPA in order to solve the problem myself, but I found launchpad.net and the whole PPA process so poorly documented and explained that I couldn't make head-nor-tails of it and gave up. (And this is coming from someone who has done more than my fair share of packaging and installer work on multiple platforms.) -- "The Direct3D Graphics Pipeline" free book <http://tinyurl.com/d3d-pipeline> The Terminals Wiki <http://terminals-wiki.org> The Computer Graphics Museum <http://computergraphicsmuseum.org> Legalize Adulthood! (my blog) <http://legalizeadulthood.wordpress.com>

2 1

Missing libs
by Tim Burgess 28 Feb '18

28 Feb '18

Hi, I downloaded the modular Boost master from Git Hub, as per instructions and built it on Windows as follows: .\Bootstrap .\b2 headers .\b2 This generated (amongst others) the following libs for boost::filesystem: libboost_filesystem-vc141-mt-gd-x32-1_67.lib libboost_filesystem-vc141-mt-gd-x64-1_67.lib libboost_filesystem-vc141-mt-x32-1_67.lib libboost_filesystem-vc141-mt-x64-1_67.lib I then tried to build a static release X64 version of my application using VS2017 and got: 1>LINK : fatal error LNK1104: cannot open file 'libboost_filesystem-vc141-mt-s-x64-1_67.lib' Obviously, I don't have this file, though I had thought that .\b2 was supposed to build all targets. So, I tried: .\b2 link=static This still does not generate the required file - where am I going wrong, please? BTW, I have checked that my project's library path points to the appropriate boost\stage\lib folder. Best wishes. Tim Burgess

2 2

[Thread] Version 1.66.0 incompatible with ISOLATION_AWARE_ENABLED
by Jim Wilburn 28 Feb '18

28 Feb '18

I work on a Windows code base that I have been updating from Boost 1.64.0 to Boost 1.66.0. One of the projects sets ISOLATION_AWARE_ENABLED to 1 (refer to https://msdn.microsoft.com/en-us/library/windows/desktop/bb773175(v=vs.85).…) #define ISOLATION_AWARE_ENABLED 1 The project also includes <boost/thread.hpp>. The Windows SDK version is 10.0.16299.0. The include chain is: <boost/thread.hpp> <boost/thread/thread.hpp> <boost/thread/thread_only.hpp> <boost/thread/win32/thread_data.hpp> <boost/thread/win32/thread_primitives.hpp> <boost/detail/winapi/dll.hpp> <boost/winapi/dll.hpp> The last include contains constructions such as: BOOST_SYMBOL_IMPORT boost::winapi::HMODULE_ WINAPI LoadLibraryA(boost::winapi::LPCSTR_ lpFileName); With a nonzero ISOLATION_AWARE_ENABLED, this results in redefinition errors, e.g.: .../boost/winapi/dll.hpp(38): error C2375: 'IsolationAwareLoadLibraryA': redefinition; different linkage ...\include\10.0.16299.0\um\winbase.inl(287): note: see declaration of 'IsolationAwareLoadLibraryA' Jim Wilburn

2 1

all roots of a polynomial
by mdebonis 22 Feb '18

22 Feb '18

Hello, I found Boost while searching for a C++ library which can do accurate polynomial root finding. The problem is that the documentation does not make it very clear how to implement this tool. Does anyone by any chance have some sample code on implementing this feature or steer me in the direction of an example in the documentation? I would like to use the best decimal place accuracy (it appears that you can get 50-place here?) Thanks! -- Sent from: http://boost.2283326.n4.nabble.com/Boost-Users-f2553780.html

4 3

Boost improperly including NTVERP.H
by Brian Catlin 22 Feb '18

22 Feb '18

Boost 1.66.0 is causing the Windows file NTVERP.H to be included (defining macros such as VER_DEBUG, VER_PRODUCTVERSION_STR, etc), preventing the definition of those macros by the application. This behavior didn't exist in 1.65.1, and breaks a standard Windows paradigm (it is very common for a Windows program to have a Version.h file containing the version resource macros (that are also defined in NTVERP.H), and then include Version.h in the main CPP file and the Resources.RC file). The NTVERP.H file contains the version information for the Windows SDK, and should NOT be included in an application! NTVERPH.H is included in: boost.1.66.0.0\lib\native\include\boost\predef\platform\windows_uwp.h boost.1.66.0.0\lib\native\include\boost\winapi\config.hpp

1 0

compilation error
by Dk Jack 21 Feb '18

21 Feb '18

Hi, I am getting a compilation error when trying to compile my app. I've detailed my issue on stack overflow. https://stackoverflow.com/questions/48897103/compile- error-in-boost-header-file-fails-on-centos-7-compiles-on-ubuntu-16-04 Would appreciate if someone here can provide some guidance on how to resolve it. Thanks. D.

1 0

Serialization of optional<std::string>
by Chardrazle 21 Feb '18

21 Feb '18

As part of upgrading a product to a later version of boost the compilation has tripped up on serialization of a boost::optional<std::string>. The /optional.hpp/ header now has the comment: ...and this fires for std::string. What, if any, is the recommended way forward for this? An example code snippet to illustrate: Previous version of boost: 1.50; new: 1.6x Compiling with g++ 4.6.4 (although same problem with 4.8.5): -W -Wall -std=c++0x... (The real code is actually using the portable_binary_archive, but let's avoid muddying the waters at this stage :) -- Sent from: http://boost.2283326.n4.nabble.com/Boost-Users-f2553780.html

2 6

Math and numerics: Odeint - problem with adaptive step size
by William Heitler 20 Feb '18

20 Feb '18

Hello, I'm a moderately experienced C++ programmer using Visual Studio 2017, but I'm new to using the boost odeint library. I'm trying to integrate the Hodgkin-Huxley equations, which are a set of ODEs that describe the generation of nerve impulse. The output should be a time-series waveform which is mainly flat but with occasional spikes in it. I have a functor HH which works fine with a standard fixed-step Runge-Kutta routine runge_kutta4< state_type > stepper; integrate_const(stepper, HH, x, 0.0, 20.0, 0.04,trace_observer()); This code produces exactly the right waveform However, I would like to use an adaptive step size because the waveform is pretty flat between the spikes. I tried the following: integrate_adaptive(make_dense_output(1.0e-6, 1.0e-6, runge_kutta_dopri5< state_type >()), hh, x, 0.0, 10.0, 0.04, trace_observer()); This works up to about the peak of the first spike, but then crashes with the error message: Debug Assertion Failed! ... Expression: vector iterator not incrementable It dies in the Visual Studio vector library code _Vector_const_iterator& operator++() in a call from boost::norm_inf where iterator first1 reports as first1 {2.105352686232e-314#DEN} std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<double> > > I certainly don't understand how the boost code works under the hood, but I wondered if anyone could suggest a change in my calling code that might make it work? I have hand-crafted some adaptive step size code which works OK (and does run faster than the fixed step size), but it is a bit of a kludge and I would rather use code developed by people who actually understand the maths. For reasons of brevity I haven't put the HH code in this message, but could do if it would help. Thanks for any help Dr W J Heitler School of Biology University of St Andrews 01334 463490 www.st-andrews.ac.uk/~wjh<http://www.st-andrews.ac.uk/~wjh>

1 0

[Iostreams] Known issue with 32-bit and iostreams::file_descriptor_sink?
by Tim Landscheidt 19 Feb '18

19 Feb '18

Hi, I'm trying to debug several issues in ledger (https://www.ledger-cli.org/) under Fedora 26 (boost 1.63; https://github.com/ledger/ledger/issues/534) and Debian Stretch (boost 1.62; https://github.com/ledger/ledger/issues/539) These issues occur only on 32-bit architectures (i386/i686), and on Fedora 26 (boost 1.63) some segmentation faults can be /fixed/ by executing: | std::ostream *dummy = new iostreams::stream<iostreams::file_descriptor_sink>(2, iostreams::never_close_handle); but this does not "work" on Debian Stretch (boost 1.62). Does that ring a bell for anyone? (NB: I don't want to im- ply that there is a bug in Boost, but maybe someone recog- nizes the scenario and can suggest where the real issues probably lie.) TIA, Tim

1 1