November 2013 - Boost - lists.preview.boost.org

Re: [boost] [TypeIndex] Peer review period for library acceptance begins, ending Thurs 21st Nov
by Antony Polukhin 12 Nov '13

12 Nov '13

2013/11/13 Niall Douglas <s_sourceforge(a)nedprod.com> <...> > On 12 Nov 2013 at 11:59, Robert Ramey wrote: > > > On Tuesday, November 12, 2013 11:34:27 AM UTC-8, Niall Douglas wrote: > > > > > Boost community feedback is requested for the formal peer review of > > > the TypeIndex library by Antony Polukhin. Feedback requested > > > includes ... > > > > > > The above phrasing is a little unclear to me. Is this is a formal review > of > > the library? > Yes, it's a formal review. > > I've just glanced over the documentation but I do have a couple of > > questions. > > > > The serialization library contains a component "extended_typeinfo". Am I > > correct > > in my understanding that this component is meant to address the same > > issues that extended_typeinfo does? If so, how are the two solutions > > different? > In short: * extended_typeinfo is not compatible with std::type_info and is hard to use * TypeIndex is a correctly working std::type_info that does not require RTTI + has other bonuses * detail/sp_typeinfo.hpp implements part of TypeIndex functionality, without workarounds for compiler issues Longer description: Serialization library is created to make a storeable in archive and portable between compilers type identifications. It's interface is not close to the interface of Standard Library, it requires A LOT of code to make typeids work (see example at http://www.boost.org/doc/libs/1_55_0/boost/serialization/extended_type_info… ) TypeIndex is designed to be simple, fast and lightweight. It's a drop-in replacement for std::type_info with big set of workarounds for issues of different compilers and an extension to get human readable type names + it contains a C++11 type_index class and works without RTTI. It does not require user's support code to work. Take a look at this page: http://apolukhin.github.io/type_index/boost_typeindex/getting_started.htmlto see how Standard Library classes and methods are mapped to the TypeIndex ones. TypeIndex is close to detail/sp_typeinfo.hpp (it's just a very very enchanted sp_typeinfo that does not adds macro, embeds workarounds and has interface of std::type_info + other features from paragraph above) -- Best regards, Antony Polukhin

1 0

[1.55] [uuid] Docs header says 'development version' on the website
by Andrey Semashev 12 Nov '13

12 Nov '13

Hi, On the website, Boost.UUID docs have a header that says that these docs are for the 'development version of boost': http://www.boost.org/doc/libs/1_55_0/libs/uuid/uuid.html I looked at a couple other libraries and the header is not displayed, so it's probably Boost.UUID specific. Does anyone know what could be the problem?

2 3

[multi_index] multiple definition errors since 86597
by Jürgen Hunold 12 Nov '13

12 Nov '13

Hi, revision 86597 changed boost/multi_index/detail/bucket_array.hpp Add were: static const std::size_t sizes[]; static const std::size_t sizes_length; member variables, initialised with +template<> +const std::size_t bucket_array_base<true>::sizes[]={ + BOOST_PP_SEQ_ENUM(BOOST_MULTI_INDEX_BA_SIZES) +template<> +const std::size_t bucket_array_base<true>::sizes_length= + sizeof(bucket_array_base<true>::sizes)/ + sizeof(bucket_array_base<true>::sizes[0]); + Now gcc (4.8.1) reports multiple definition errors when using multi_index in different translation units: file1.cpp.o:(.rodata+0x0): multiple definition of `boost::multi_index::detail::bucket_array_base<true>::sizes_length' file2.cpp.o:(.rodata+0x0): first defined here file1.cpp.o:(.rodata+0x20): multiple definition of `boost::multi_index::detail::bucket_array_base<true>::sizes' file2.cpp.o:(.rodata+0x20): first defined here CMakeFiles/graph_test.dir/main.cpp.o:(.rodata+0x0): multiple definition of `boost::multi_index::detail::bucket_array_base<true>::sizes_length' file2.cpp.o:(.rodata+0x0): first defined here file3.o:(.rodata+0x20): multiple definition of `boost::multi_index::detail::bucket_array_base<true>::sizes' file2.cpp.o:(.rodata+0x20): first defined here According to http://stackoverflow.com/questions/1553854/template-static-variable and my current c++ template knowledge, those instantiation belong to a .cpp file to avoid this. This does not work for a header-only library, of course. The current test suite does not catch this because it only contains single file tests. Yours, Jürgen -- * Dipl.-Math. Jürgen Hunold ! * voice: ++49 4257 300 ! Fährstraße 1 * fax : ++49 4257 300 ! 31609 Balge/Sebbenhausen * jhunold(a)gmx.eu ! Germany

2 2

Data Structure and Algorithms yet to be implemented
by Manmohit Rekhi 12 Nov '13

12 Nov '13

Hi, I would really be grateful if could get to know 1)which Data Structures and Algorithms are yet to be implemented in boost? 2)How can i get started on them? Thanking you, Manmohit Rekhi

3 2

Boost 1.55.0 released!
by Marshall Clow 12 Nov '13

12 Nov '13

Release 1.55.0 of the Boost C++ Libraries is now available. These open-source libraries work well with the C++ Standard Library, and are usable across a broad spectrum of applications. The Boost license encourages both commercial and non-commercial use. This release contains one new library and numerous enhancements and bug fixes for existing libraries. === New Libraries === • Predef: This library defines a set of compiler, architecture, operating system, library, and other version numbers from the information it can gather of C, C++, Objective C, and Objective C++ predefined macros or those defined in generally available headers, from Rene Rivera. === Known bugs with Visual Studio 2013/Visual C++ 12 === Visual Studio 2013 was released quite late in the release process, so there exist several unresolved issues. These include: • Serialization can't compile because of a missing include. • Using has_member_function_callable_with from Boost.Container's allocator_traits causes a compile error (#9332). • For several libraries overloads with intializer lists are incorrectly reported as ambiguous. This is a Visual C++ bug and it isn't clear if there's a good workaround. • In libraries such as Unordered and MultiIndex, calling overloaded functions with initializer lists can result in a compile error, with Visual C++ claiming that the overloads are ambiguous. This is a Visual C++ bug and it isn't clear if there's a good workaround. This won't affect code that doesn't use initializer lists, or uses an initializer list that doesn't require an implicit conversion (i.e. an initializer list of the container's exact value type). • Thread: ex_scoped_thread compile fails (#9333). For details, including download links, see http://www.boost.org/users/news/version_1.55.0 You can also download directly from SourceForge: http://sourceforge.net/projects/boost/files/boost/1.55.0/ To install this release on your system, see http://www.boost.org/doc/libs/release/more/getting_started/index.html Thanks, --The Boost release team Vladimir Prus Rene Rivera Marshall Clow Eric Niebler Daniel James Beman Dawes

2 2

Re: [boost] sqlpp11: SQL for C++
by Brian Wood 12 Nov '13

12 Nov '13

From: Bjorn Reese > On 11/09/2013 11:03 PM, Roland Bock wrote: >> >> * define types representing tables and columns, > > There seems to be room for improvement here. You should consider an > integration with Boost.Fusion. This way you can avoid having to build > a code generator. Hmm. If you want a second opinion, these guys http://springfuse.com are building a code generator. -- Brian Ebenezer Enterprises - So far G-d has helped us. http://webEbenezer.net

2 1

[Functional/Factory] Value factory not accepts rvalues?
by Lin, Yi-Li 12 Nov '13

12 Nov '13

I found that boost::value_factory did not accept rvalues in some case. In my case, temporary value is produced and passed to value factory directly. But boost::value_factory<T>::operator() did not accept rvalue, they take lvaue reference instead. Maybe it'll be better to provide another version of operator takes const lvalue? (Additionally, use std::forward to forward universal reference in C++11?) Sample code which compiler complains: int get_value(); struct var { var( ); explicit var(int); }; void caller() { boost::value_factory<var> f; f(get_value()); }

2 2

Changes to trunk for clang using VC++RTL
by Edward Diener 12 Nov '13

12 Nov '13

I have held off making some changes to trunk in SVN for clang with VC++ RTL until Boost 1.55 was out. I can do so now based on my tests with clang but I am concerned that the changes will not be carried over to the Boost git development branch when Boost switches to git. Should I be waiting until the git changeover is made and everyone learns to deal with git and the setup for using Boost with git, or should I start making changes now and watch the regression tests cycle ? I am sure from my tests that the changes, although involved in low-level libraries like config, mpl, type_traits, will not adversely affect any other compiler and will help in testing clang using VC++ RTL for Boost as clang inches closer to full support in Windows using VC++ headers and libraries.

1 0

Re: [boost] [release][1.55.0] Visual C++ known bugs
by Daniel James 11 Nov '13

11 Nov '13

On 11 November 2013 20:40, Robert Ramey <robertmacleranramey(a)gmail.com> wrote: > > > On Monday, November 11, 2013 12:04:22 PM UTC-8, Daniel James wrote: >> >> On 11 November 2013 12:36, Sergey Cheban <s.ch...(a)drweb.com> wrote: >> > >> > Btw, what do you think about an additional SVN branch for the post-RC2 >> > VC12-related bugfixes? >> >> I just created a branch at: >> >> https://svn.boost.org/svn/boost/branches/maintenance/1_55_0 >> > > Since we've tagged the current state of the release branch with 1_55_0 do we > really have to create another branch? Can't > we just open the release branch for checkins? That would be easier. This branch is just for fixes. The release branch can have other changes, including breaking changes.

2 2

Re: [boost] [release][1.55.0] Visual C++ known bugs
by Daniel James 11 Nov '13

11 Nov '13

On 11 November 2013 21:23, Robert Ramey <robertmacleranramey(a)gmail.com> wrote: > > > On Monday, November 11, 2013 12:42:07 PM UTC-8, Daniel James wrote: >> >> On 11 November 2013 20:40, Robert Ramey <robertmac...(a)gmail.com> wrote: >> > >> > >> > On Monday, November 11, 2013 12:04:22 PM UTC-8, Daniel James wrote: >> >> >> >> On 11 November 2013 12:36, Sergey Cheban <s.ch...(a)drweb.com> wrote: >> >> > >> >> > Btw, what do you think about an additional SVN branch for the >> >> > post-RC2 >> >> > VC12-related bugfixes? >> >> >> >> I just created a branch at: >> >> >> >> https://svn.boost.org/svn/boost/branches/maintenance/1_55_0 >> >> >> > >> > Since we've tagged the current state of the release branch with 1_55_0 >> > do we >> > really have to create another branch? Can't >> > we just open the release branch for checkins? That would be easier. >> >> This branch is just for fixes. The release branch can have other >> changes, including breaking changes. > > > Soo... if we have fixes, we should check them into BOTH branches? If you want to, or just check it into release. It's a way to collect patches, make of that what you will.

1 0