March 2019 - Boost - lists.preview.boost.org

[filesystems]
by Florian Lindner 14 Mar '19

14 Mar '19

Hello, I have this innocent piece of code: namespace fs = boost::filesystem; auto path = getFilename(); // returns a string fs::create_directories(fs::path(path).parent_path()); std::ofstream ofs(path + "~"); ofs << info; ofs.close(); fs::rename(path + "~", path); which causes the exception: boost::filesystem::rename: No such file or directory: "../9f/061b4f7a5e529c964659226eedd4e5~", "../9f/061b4f7a5e529c964659226eedd4e5" However, I have no idea how that could happen. I use the rename, so that a reading process never sees an empty file, but only no file or filed with info. Is there any race involved between ofs.close() and fs:rename()? The code was executed on a distributed network filesystem (lustre). Any ideas anyone? Thanks, Florian

5 9

Re: [boost] [Boost-users] Networking TS + Beast, NEW Tutorials, Read this to learn std::net !!!
by Vinnie Falco 14 Mar '19

14 Mar '19

On Thu, Mar 14, 2019 at 10:50 AM james <dirtydroog(a)gmail.com> wrote: > This is excellent news. Timeouts are ASIO's biggest downfall. I'll definitely check it out. > If it's low level, and performant enough, then it paves the way for ASIO-ed connectors for things like Redis and ZeroMQ. Thanks! And we are very happy to answer questions, feel free to open an issue or come to the #beast Slack channel. Regards

1 0

Cannot serialize pointer through base without void_cast_register and register_type.
by Nicolas Guillot 14 Mar '19

14 Mar '19

Hello. *** CONTEXT *** I use your serialization library since several years on linux (gcc). Now, I'm moving to windows (_MSC_VER == 1600) (boost 1.55, but same issues on boost 1.69). For my example, I use a class A in a DLL, another class B inheriting from A in its DLL and an exe, linked to those DLLs, which serializes a B object, through its A base. _______ | A | ------- /_\ | | _______ | B | ------- I make a class serializable like that (I summurize on purpose (no #include, ...)). For A it's the same thing, but it has no base class. ----> B.h ---------------------------------------- #ifdef B_EXPORTS #define B_DLL __declspec(dllexport) #else #define B_DLL __declspec(dllimport) #endif class B_DLL B: public A { public: B(); virtual ~B(); private: friend class boost::serialization::access; template <class Archive> void serialize(Archive & ar, const unsigned int version); int i; }; BOOST_CLASS_EXPORT_KEY(B) ----> B.cpp ---------------------------------------- B::B() : i(123) {} B::~B() {} template <class Archive> void B::serialize(Archive & ar, const unsigned int) { ar & SERIALIZE_BASE(ClassA); ar & SERIALIZE(i); } #define FULL_SCOPED_SERIALIZABLE_CLASS B #include "Common/InstanciateSerializationTemplate.h" ----> InstanciateSerializationTemplate.h ---------------------------------------- This file purpose is to instanciate the template B::serialize methods for each archive type, and to register the class. Here is its contents (once again, no namesapce nor include to make it easier to read): // explicit instantiate code for all archive types expected to be use: // MUST COME BEFORE BOOST_CLASS_EXPORT_IMPLEMENT // binary: template __declspec(dllexport) void FULL_SCOPED_SERIALIZABLE_CLASS::serialize<binary_iarchive>(binary_iarchive & ar, const unsigned int version); template __declspec(dllexport) void FULL_SCOPED_SERIALIZABLE_CLASS::serialize<binary_oarchive>(binary_oarchive & ar, const unsigned int version); // XML: template __declspec(dllexport) void FULL_SCOPED_SERIALIZABLE_CLASS::serialize<xml_iarchive>(xml_iarchive & ar, const unsigned int version); template __declspec(dllexport) void FULL_SCOPED_SERIALIZABLE_CLASS::serialize<xml_oarchive>(xml_oarchive & ar, const unsigned int version); // TXT: template __declspec(dllexport) void FULL_SCOPED_SERIALIZABLE_CLASS::serialize<text_iarchive>(text_iarchive & ar, const unsigned int version); template __declspec(dllexport) void FULL_SCOPED_SERIALIZABLE_CLASS::serialize<text_oarchive>(text_oarchive & ar, const unsigned int version); BOOST_CLASS_EXPORT_IMPLEMENT(FULL_SCOPED_SERIALIZABLE_CLASS) *** MY ISSUE *** All that stuff works in linux, gcc, but not in Windows. If I write in my exe this serialization code: std::ofstream ofs; ofs.open("D:\\TestSerial.xml"); boost::archive::xml_oarchive ar(ofs); shared_ptr<B> b = shared_ptr<B>(new B()); shared_ptr<A> a = b; try { ar << SERIALIZE(a); } catch (boost::archive::archive_exception& ex) { std::cout << "Serialization to stream error, what: " << ex.what() << std::endl; } catch (...) { std::cout << "Catch ... Serialization to stream error" << std::endl; } ofs.close(); With this code, I have an archive_exception: "unregistered class - derived class not registered or exported". I have to write the serialization like that: ar.register_type(static_cast<B*>(0)); boost::serialization::void_cast_register(static_cast<B*>(0), static_cast<A*>(0)); ar << SERIALIZE(a); *** QUESTION *** At least, my question: I cannot do like that: what I would like to achieve: when a new class is made serializable like above, I expect the class to be registered automatically (probabley with BOOST_CLASS_EXPORT_IMPLEMENT in InstanciateSerializationTemplate.h). It's too painfull to write the register_type and void_cast_register each time I want to serialize. What am I missing? Thank you very much for any help or recommandations. Nico

1 1

Re: [boost] MSVC 2019 and 1.70 beta
by Peter Dimov 14 Mar '19

14 Mar '19

> Tom Kent wrote: > > > Build support for VS 2019 RC (msvc-14.2) is on develop in boost build. > > Do we want to merge that to master for the release? > > I think we do. 1.70 and 2019 will release at about the same time. In fact looking at the calendar, 2019 will release (Apr 2) just before 1.70 is closed for release (Apr 3.)

3 3

boost-install-ing zlib (Was: Re: [Release] Boost 1.70.0 Beta 1 Release Candidate 2)
by Peter Dimov 13 Mar '19

13 Mar '19

> I'll need to try and see what happens with <source> in user-config. > Probably nothing good, because one would expect neither <build-name> nor > <tag> to be specified in such a case. If I remove the check for ZLIB_SOURCE, and just do alias boost_zlib : /zlib//zlib ; libraries-to-install += boost_zlib ; unconditionally in libs/iostreams/build/Jamfile, this is what happens: - when `using zlib : : <source>C:/Projects/zlib-1.2.11 ;` is in user-config, the default `b2 install` fails because `address-model=32,64 variant=debug,release` conflict as all try to build `libz.lib`. When address-model=32 variant=debug is given, `libz.lib` is installed in C:\Boost\lib, and the CMake configuration for boost_zlib correctly points to it. - when using a prebuilt `using zlib : : <include>C:/Projects/include <search>C:/Projects/lib <name>zlib-8 ;`, the CMake configuration file for boost_zlib points to "${_BOOST_LIBDIR}/zlib-8", which is not correct.

2 2

(no subject)
by Marshall Clow 13 Mar '19

13 Mar '19

Boost release 1.70.0 beta 1 is now available at: <https://dl.bintray.com/boostorg/beta/1.70.0.beta1/source/> The SHA256 checksums are as follows: dbf89645e509acb2e53eac3c633834f1dec80bd8622e6e652a9741dcf0807f17 ./boost_1_70_0_b1.7z 957b984b58656a3ec00661ecad7ef82e3a094bad36e37b9e6624843237854679 ./boost_1_70_0_b1.tar.bz2 76d149c1fb9e8a5ceb3bc1e5587edc40fa79a43e8b990d877c2210f0382b9421 ./boost_1_70_0_b1.tar.gz 71e990b2c26880521fd98459b184ee172ce1c46dcf554b6eaab7776b8b6369a6 ./boost_1_70_0_b1.zip For details of what's in the release, see < https://www.boost.org/users/history/version_1_70_0.html>. Please download the beta, give it a try, and report any problems you encounter. Thanks, -- The Boost Release Team

2 3

[rational] permission to merge cmake change?
by James E. King III 13 Mar '19

13 Mar '19

Before 1.70.0 beta started, some cmake changes from Mike Dev were merged into rational. There was a typo in one of the cmake options (BOOST_RATIONA_INCLUDE_TESTS). This does not affect code nor the boost build system. Would it be okay to merge this now or should I wait? https://github.com/boostorg/rational/pull/35 Thanks. - Jim

1 1

[release] Boost 1.70.0 Beta 1 Release Candidate 3
by Marshall Clow 12 Mar '19

12 Mar '19

A new release candidates for the first 1.70.0 beta release are now available at: https://dl.bintray.com/boostorg/beta/1.70.0.beta1.rc3/source/ Changes since RC2: * Relative paths for --prefix, --libdir, etc. are now bound relative to the current working directory, instead of being relative to the Jamfile. Thanks to Tom Kent for the report, and Steven and Peter for the fixes. Changes since RC1: * The Jamfile for Boost.Predef has been updated to allow running the tests from /status. Thanks to mike.dev(a)gmx.de for the report, and to Rene for the fix. The release notes are not yet available. The SHA256 checksums are as follows: dbf89645e509acb2e53eac3c633834f1dec80bd8622e6e652a9741dcf0807f17 ./boost_1_70_0_b1_rc3.7z 957b984b58656a3ec00661ecad7ef82e3a094bad36e37b9e6624843237854679 ./boost_1_70_0_b1_rc3.tar.bz2 76d149c1fb9e8a5ceb3bc1e5587edc40fa79a43e8b990d877c2210f0382b9421 ./boost_1_70_0_b1_rc3.tar.gz 71e990b2c26880521fd98459b184ee172ce1c46dcf554b6eaab7776b8b6369a6 ./boost_1_70_0_b1_rc3.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. -- The Boost Release Managers

5 8

[Release] Boost 1.70.0 Beta 1 Release Candidate 2
by Marshall Clow 12 Mar '19

12 Mar '19

A new release candidates for the first 1.70.0 beta release are now available at: https://dl.bintray.com/boostorg/beta/1.70.0.beta1.rc2/source/ Changes since the last RC: * The Jamfile for Boost.Predef has been updated to allow running the tests from /status. Thanks to mike.dev(a)gmx.de for the report, and to Rene for the fix. The release notes are not yet available. The SHA256 checksums are as follows: 26dba8156ad08041c6da52c84a2654aefda292180afe3546a9a74dda0dfb3781 ./boost_1_70_0_b1_rc2.7z ee794d98429bc2bfae054f2fab6fde53ea95f7c5967d8b267e57c894a4d46e57 ./boost_1_70_0_b1_rc2.tar.bz2 bca5f189d87b26ab903838dee2fd4c54e9043bd35d27b8923cdef42651ae7261 ./boost_1_70_0_b1_rc2.tar.gz 6c8f3af4e64d3a04a3d3a5e98499d2f6bb460ad82c1f33ff7da800b5ab182b0b ./boost_1_70_0_b1_rc2.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. -- The Boost Release Managers

4 25

Re: [boost] [Boost-users] [release] Boost 1.70.0 Beta 1 Release Candidate 1
by Tom Kent 11 Mar '19

11 Mar '19

On Fri, Mar 8, 2019 at 9:44 PM Tom Kent <lists(a)teeks99.com> wrote: > > > On Fri, Mar 8, 2019 at 8:57 PM Tom Kent <lists(a)teeks99.com> wrote: > >> >> >> On Fri, Mar 8, 2019 at 6:42 PM Marshall Clow via Boost-users < >> boost-users(a)lists.boost.org> wrote: >> >>> On Fri, Mar 8, 2019 at 10:21 AM Marshall Clow <mclow.lists(a)gmail.com> >>> wrote: >>> >>>> The release candidates for the first 1.70.0 beta release are now >>>> available at: >>>> >>>> https://dl.bintray.com/boostorg/beta/1.70.0.beta.1.rc1/source/ >>>> >>>> The release notes are not yet available. >>>> >>>> The SHA256 checksums are as follows: >>>> >>>> a7c570d6307d05dc8d45a3b4c69a0afdcba2b22eac0c4da4d496559f53db7b89 >>>> ./boost_1_70_0_b1_rc1.7z >>>> 4b91b9a203cf12f6001cf39de83c7cc005fbc5390e2b4b2fc29ce7b19c0537a7 >>>> ./boost_1_70_0_b1_rc1.tar.bz2 >>>> 460dd6499bba9df04c6bfa219b26097bad803b8758304ba45f2fc291aa42c96e >>>> ./boost_1_70_0_b1_rc1.tar.gz >>>> f209c9a086ed2c7882a0bc8755792f436eb141eaeb8c823ed57295ca9ce25b93 >>>> ./boost_1_70_0_b1_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. >>>> >>> >>> >> I'm having a problem with B2 in this release. >> >> For many releases now, I have run a command such as: >> b2 -j16 --build-type=complete toolset=msvc-10.0 address-model=64 >> architecture=x86 --prefix="./" --libdir=lib32-msvc-10.0 >> --includedir=temp_inc install >> >> to get the windows builds of each toolset. For the install, I set the >> prefix as the current directory, so that I'll get a lib{arch}-msvc-{ver} >> directory for each toolset in there. However, with this 1.70 release, it >> seems to ignore my prefix and put the headers in >> libs/headers/build/temp_inc/boost-1_70/boost, instead of >> temp_inc/boost-1_70/boost. For the libraries, it scattered them throughout >> the build in directories such as libs/coroutine/build/lib32-msvc-10.0/. >> >> Any thoughts on what might be causing this? Am I using the --prefix >> --libdir and --inlcudedir correctly? Or at least not badly enough to >> warrant this? >> > > Also, even though I have BZIP2_SOURCE and ZLIB_SOURCE set, I'm not seeing > the appropriate bzip2 and zlib libraries being generated. > > After updating my scripts to us "staging" instead of "install" to avoid the above issue, I was able to get everything else to build successfully for Visual Studio (msvc-10.0 through msvc-14.1, note this is the first release I haven't built against msvc-8.0 and 9.0). The only exception being the lack of bzip2 and zlib support mentioned above. toolset arch compile Link Execute msvc-10.0 32 X X X msvc-10.0 64 X X X msvc-11.0 32 X X X msvc-11.0 64 X X X msvc-12.0 32 X X X msvc-12.0 64 X X X msvc-14.0 32 X X X msvc-14.0 64 X X X msvc-14.1 32 X X X msvc-14.1 64 X X X Compile means that the b2 command completed without errors Link means that visual studio was able to link a sample executable to a library (libboost_thread-vcXXX-mt[-gd]-1_XX.lib) generated Execute means that the linked program executed without errors. Full build logs can be found at https://gist.github.com/teeks99/0aeed77c9251ba02090d9ab7372e2c73 Tom

4 4