- Boost-users - lists.boost.org

[Interval] Help debugging compiler optimization
by Tim van Erven 12 Mar '18

12 Mar '18

Dear all, I am trying to understand why I am getting different numerical results with the interval library depending on the optimization level of the compiler. I am attaching the smallest example I have been able to create: # On my Mac laptop Apple LLVM version 9.0.0 (clang-900.0.39.2) boost 1.66 (installed via homebrew) $ g++ foo.cpp -o foo $ ./foo third1 = 0.3333333333333333148296162562473909929394721984863281250000000000 third2 = 0.3333333333333333148296162562473909929394721984863281250000000000 v1 = (0.9999999999999998889776975374843459576368331909179687500000000000,1.0000000000000000000000000000000000000000000000000000000000000000) v2 = (0.9999999999999998889776975374843459576368331909179687500000000000,1.0000000000000000000000000000000000000000000000000000000000000000) $ g++ -O2 foo.cpp -o foo $ ./foo third1 = 0.3333333333333333148296162562473909929394721984863281250000000000 third2 = 0.3333333333333333148296162562473909929394721984863281250000000000 v1 = (1.0000000000000000000000000000000000000000000000000000000000000000,1.0000000000000000000000000000000000000000000000000000000000000000) v2 = (1.0000000000000000000000000000000000000000000000000000000000000000,1.0000000000000000000000000000000000000000000000000000000000000000) I would expect to get the same output in both cases, but the lower end-points are different in the second case, and seem wrong to me since third2 * 3.0 < 1.0. # On my Linux machine the effect is different: gcc version 5.4.0 20160609 boost 1.58 on Ubuntu 16.04.9 $ g++ foo.cpp -o foo $ ./foo third1 = 0.3333333333333333148296162562473909929394721984863281250000000000 third2 = 0.3333333333333333148296162562473909929394721984863281250000000000 v1 = (0.9999999999999998889776975374843459576368331909179687500000000000,1.0000000000000000000000000000000000000000000000000000000000000000) v2 = (0.9999999999999998889776975374843459576368331909179687500000000000,1.0000000000000000000000000000000000000000000000000000000000000000) $ g++ -O2 foo.cpp -o foo $ ./foo third1 = 0.3333333333333333148296162562473909929394721984863281250000000000 third2 = 0.3333333333333333148296162562473909929394721984863281250000000000 v1 = (0.9999999999999998889776975374843459576368331909179687500000000000,1.0000000000000000000000000000000000000000000000000000000000000000) v2 = (1.0000000000000000000000000000000000000000000000000000000000000000,1.0000000000000000000000000000000000000000000000000000000000000000) Can anyone explain what is going on? Thanks in advance, Tim -- Tim van Erven<tim(a)timvanerven.nl> www.timvanerven.nl

7 14

[Boost.Serialization] (De-)serialization with class versioning
by Lars Ruoff 11 Mar '18

11 Mar '18

Hi, i use the class versioning feature of Boost.Serialization, with an external (non-intrusive) serialization function like this: BOOST_CLASS_VERSION(GAMEINFO, 1) template <typename Archive> void serialize(Archive &ar, GAMEINFO& gameSettings, const unsigned int version) { ar & gameSettings.NPODS; // ... if (version > 0) ar & gameSettings.ARENA_GOAL_EXCLUSION_RADIUS; } Now as soon, as i set BOOST_CLASS_VERSION(GAMEINFO, 1), i get version==1 in this serialize function, even when i *deserialize* (load from archive). I would have expected to get the version of the data in the saved archive in this case. Say i saved with an old version=0. Then my class evolved and i set BOOST_CLASS_VERSION(GAMEINFO, 1). But when reading the old archive, i expect version==0 in the function so i can react on it. Instead, i get always version==1, as soon as i compile with BOOST_CLASS_VERSION(GAMEINFO, 1), even when i'm reading old files! What am i missing? regards, Lars R.

3 7

boost::serialization - Non intrusive serialization of user-defined members?
by Lars Ruoff 05 Mar '18

05 Mar '18

Hi, I have a class structure like class MemberClass; class ContainerClass { MemberClass special_member; int other_members; }; Following the example in http://www.boost.org/doc/libs/1_66_0/libs/serialization/doc/ i want to write a non-intrusive serialization: template<class Archive> void serialize(Archive & ar, ContainerClass & c, const unsigned int version) { ar & c.other_members; // easy - built-in type serialize(ar, c.special_member, version); //use same version??? } Assuming there's is also a non-intrusive void serialize(Archive & ar, MemberClass & m, const unsigned int version); this compiles and works but i feel this is not the way to do it because it reuses the version number of the container class for the member class? So how to handle that? Must the Container class manage version of its members individually? I suggest to add a section "Non-intrusively Serializable Members" to documentation with an example. Cheers, Lars R.

3 7

[fiber] Fiber context switching during stack unwinding...
by Oliver Seiler 01 Mar '18

01 Mar '18

(Hopefully I'm not writing an essay for something that is a gcc/libstdc++/Linux bug) I've been using Boost.Context for a number of years now, to do something similar to what Boost.Fiber does. We recently upgraded our development target environment to Debian 9 (gcc 6.3), and moved up to Boost 1.62.0, so I've been looking at Boost.Fiber to replace how we're using Boost.Context (because it looks like it could greatly simplify our code, among other things). One thing I ran into with Boost.Context was that it interacted in some interesting ways with exception handling, and I haven't really seen any discussion about it. Specifically, switching contexts with an active exception (so during stack unwinding, or in a catch block) would result in weird behaviour in the other context (std::uncaught_exception() returning true, std::current_exception() returning the exception from the other context, try/catches in the new context causing new problems when switching back to the original context which is in the middle of stack unwinding, etc). I chalked that up to limitations of the context switching, though I haven't seen any mention of avoiding context switches with active exceptions; presumably this is some interaction with gcc's exception implementation, and would maybe need support from libunwind to save/restore the exception state (I don't recall if I checked if clang produced similar results, though I'll probably give it a go when I have the chance). So I wasn't surprised when I ran into similar problems with Boost.Fiber. I am wondering why this aspect of usage isn't discussed in any of the documentation; it would be nice if these libraries could work well with language features like exceptions, but I understand that the language spec doesn't really speak to the stack switching happening under-the-hood. It actually wouldn't surprise me that this wasn't even a problem in all environments, and just happens to be an artifact of the Linux implementation, but then we're either talking about unspecified or undefined behaviour, but would be nice to know. Even in the documents going through the C++ standards committee I don't see much mention of how things like coroutines should interact with something like std::current_exception() (perhaps because it is assumed to "just work" and doesn't need additional comment?) Anyone else ever run into issues related to this? Seems like any use of RAII for exception handling would easily run into problems. Mostly I've just found ways to work around it (e.g., not allowing the context switch if std::uncaught_exception() or std::current_exception() indicate it would be unsafe to do so, typically by throwing another exception) but this tends to make libraries harder to use correctly (e.g., can't really allow context switching in a destructor, or in a function called from a destructor, etc). Cheers Oliver

3 4

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