- Boost - lists.boost.org

[interprocess] Platform-specific flags for mapped_region
by Andrey Semashev 12 Feb '13

12 Feb '13

Hi, I wanted to try using Linux huge pages in my application, which uses Boost.Interprocess currently for interprocess communication. However there seems to be no way to pass MAP_HUGETLB flag to mmap in the interface of mapped_region. I understand that this is a platform-specific option, so probably it would be best to just add an opaque unsigned int flags parameter to the mapped_region constructor with the default value of 0. Or is there a better way of doing this? I'll create a ticket for this feature, but I wanted to know if the general idea of this extension is welcome.

2 3

[Boost.Range] An issue with the documentation of 'remove_erase_if' algorithm
by Adam Romanek 12 Feb '13

12 Feb '13

Hi, I think there's an issue with the documentation of 'remove_erase_if' algorithm from Boost.Range library. It says: "remove_erase_if removes the elements x that satisfy pred(x) from the container. This is in contrast to the erase algorithm which merely rearranges elements." I believe it's not in contrast to the "erase" algorithm but rather the "remove" algorithm. WBR, Adam Romanek

5 8

Future of Boost Windows Installers
by Dave Abrahams 11 Feb '13

11 Feb '13

Hi, As many of you already know, I'm going to work at Apple in late February, and BoostPro is closing its doors. For many years, the Boost Getting Started Guide has suggested that Windows users pop over to BoostPro's website and use the latest installer to get Windows binaries. We'd like to donate to the community the technology used to build those installers. If anyone would like to pick up the job of creating them, we'd be happy to offer guidance. Right now I'm on a plane where an uplink to GitHub is creeping along at 4KB/s, so I've been pushing the stuff up in bits and pieces and the latest stuff isn't up there, but it's going to http://github.com/boostpro/installer and will be complete within 24 hours. Whatever happens with this installer code, someone is going to need to update the Getting Started Guide, either to point at the installer's new home, or to remove references to it. -- Dave Abrahams BoostPro Computing Software Development Training http://www.boostpro.com Clang/LLVM/EDG Compilers C++ Boost

3 7

Release branch now open for merging
by Marshall Clow 11 Feb '13

11 Feb '13

The release branch (https://svn.boost.org/svn/boost/branches/release) is open for merging all stable changes from trunk, including bug fixes and major upgrades to existing libraries. Breaking changes should be coordinated via the mailing list with the developers of the other libraries affected. New libraries may be added, but require permission of a release manager. In other words, merging *does not* require permission from a release manager, except for new libraries. Boost Calendar with the release schedule: http://www.boost.org/development/index.html Release schedule explanation: http://svn.boost.org/trac/boost/wiki/ReleaseSchedule Release practices for developers: http://svn.boost.org/trac/boost/wiki/ImprovingPractices -- The Release Managers, Beman Dawes Daniel James Eric Niebler Marshall Clow Rene Rivera Vladimir Prus

2 1

[iostreams] mapped_file_source fails for zero sized file
by Michael Marcin 11 Feb '13

11 Feb '13

While testing out iostreams memory mapped file source I ran into an issue where it throws when opening a zero sized file (testing on windows 7 vc11 x64). for example something like: namespace io = boost::iostreams; namespace fs = boost::filesystem; fs::path path("C:\\boost\\boost_1_53_0\\libs\\context\\src\\dummy.cpp"); io::stream<io::mapped_file_source> file( path ); Will throw: failed create mapping: The volume for a file has been externally altered so that the opened file is no longer valid. From: http://tinyurl.com/cx2b9ru "When you do not want the size of the file to change (for example, when mapping read-only files), call CreateFileMapping and specify zero for both dwMaximumSizeHigh and dwMaximumSizeLow. Doing this creates a file mapping object that is exactly the same size as the file. Otherwise, you must calculate or estimate the size of the finished file because file mapping objects are static in size; once created, their size cannot be increased or decreased. An attempt to map a file with a length of zero in this manner fails with an error code of ERROR_FILE_INVALID. Programs should test for files with a length of zero and reject such files." I don't see anything in the docs that would lead me to believe this should throw. Rather I would expect: std::istream_iterator<char>(file) == std::istream_iterator<char>()

1 0

Picking overload with the most specific concept
by Louis Dionne 11 Feb '13

11 Feb '13

Hi, I was playing around today and I achieved to do something I find quite cool. I would like to know whether the technique/general principle is of any interest. Basically, it allows to overload functions based on the modeling of a concept by a type without having to write complicated enable_ifs. It does come with some problems and it is not very general for the moment, but it's a first shot. Here we go: #include <duck.hpp> // the library where I implemented that #include <forward_list> #include <utility> #include <vector> // Generates some boilerplate required to enable concept overloading for // a function named `my_distance`. Ignore this for now. DUCK_ENABLE_CONCEPT_OVERLOADING( my_distance, my_distance(std::declval<Iterator>(), std::declval<Iterator>(), State()), typename Iterator ) struct random_access_version_picked { }; struct forward_version_picked { }; // The State template parameter will be used to carry information during // the resolution. tag::my_distance was defined by the above macro. template <typename Iterator, typename State = tag::my_distance> // This can be seen like an enable_if with some more functionality. typename duck::requires< // This overload is picked iff Iterator models RandomAccessIterator // and no other overload for a more specific concept exist. // Concepts are MPL metafunction classes, so there's no magic here. duck::model_of<duck::RandomAccessIterator, Iterator> // Could add more duck::model_of clauses here , duck::overload_resolution_state<State> // required for the mechanics , duck::template_parameters<Iterator> // ditto , random_access_version_picked>::type // return type my_distance(Iterator, Iterator, State = State()) { return random_access_version_picked(); } // Same thing for the ForwardIterator implementation. template <typename Iterator, typename State = tag::my_distance> typename duck::requires< duck::model_of<duck::ForwardIterator, Iterator> , duck::overload_resolution_state<State> , duck::template_parameters<Iterator> , forward_version_picked>::type my_distance(Iterator, Iterator, State = State()) { return forward_version_picked(); } int main() { std::forward_list<int> list; std::vector<int> vec; forward_version_picked forward = my_distance(begin(list), end(list)); random_access_version_picked random = my_distance(begin(vec),end(vec)); } This example works with Clang 3.3 and GCC 4.8. For those interested in seeing more, the full library is available at: http://github.com/ldionne/duck Look for the `next` branch. The file involved with the above mechanics is include/duck/requires.hpp. Also see the test in test/requires.cpp. Regards, Louis Dionne

1 0

[Git] Status update
by Dave Abrahams 10 Feb '13

10 Feb '13

Please see https://github.com/ryppl/boost-svn2git/wiki if you're interested in the status of modularized history conversion. -- Dave Abrahams

4 3

[smart_ptr] !defined(BOOST_NO_CXX11_SMART_PTR) does not imply !defined(BOOST_NO_CXX11_RVALUE_REFERENCES)
by Michel Morin 10 Feb '13

10 Feb '13

`!defined(BOOST_NO_CXX11_SMART_PTR)` does not necessarily mean that the compiler support of rvalue references is available. For example, when I use shared_ptr (on trunk) with clang + libc++ in a C++03 mode, I get warnings something like boost/smart_ptr/shared_ptr.hpp:556:51: warning: rvalue references are a C++11 extension [-Wc++11-extensions] shared_ptr & operator=( std::unique_ptr<Y, D> && r ) Regards, Michel

3 4

[preprocessor] enabling variadics support
by Gennadiy Rozental 09 Feb '13

09 Feb '13

Hi, It appears PP variadics support is disabled on wide range of compilers where it is in fact working (at least It seems to be). This includes Clang and PGI. Can we enable it now? Gennadiy

3 2

[variant] Please vote for behavior (Was: Basic rvalue and C++11 features seupport)
by Antony Polukhin 08 Feb '13

08 Feb '13

Current implementation of recursive_wrapper move constructor is not optimal: recursive_wrapper<T>::recursive_wrapper(recursive_wrapper&& operand) : p_(new T(std::move(operand.get()) )) { } During descussion were made following proposals: I: Leave it as is - bad performance - can not provide noexcept guarantee II: Set operand.p_ to NULL, add BOOST_ASSERT in get operations + good performance + provides noexcept guarantee for move constructor + optimization will be used even if varinat has no type with trivial default constructor + easy to implement - triggers an assert when user tries to reuse moved object - adds an empty state to the recursive_wrapper III: Make recursive_wrapper and variant cooperate, enable move for varinat in the presence of recursive_wrappers only when there is at least one type that is nothrow-default-constructible, regardless of whether it's current. It is easyer to understand by example: typedef variant<int, recursive_wrapper<foo>> V; V v1( std::move(v2) ); This move-constructs v1 from v2 and leaves int() into v2. + good performance + provides noexcept guarantee for rcursive_wrapper + does not adds an empty state - optimization won't trigger if varinat has no type with trivial default constructor - hard to implement - user may be obscured by the fact, that v2 from the example now contains int IV: After move away, delay construction of type held by recursive_wrapper till it will be be required. +/- good performance? but more checks + provides noexcept guarantee for rcursive_wrapper + does not adds an explicit empty state + optimization will be used even if varinat has no type with trivial default constructor +/- not hard to implement --- additional requirement for type T of recursive_wrapper: it must be default constructible - less obvious behavior for user (for example get() function would construct values, allocate memory and can possibly throw) Please, vote for solution or propose a better one. P.S.: As I know Joel de Guzman votes for solution II (and totally agains solution I, because it is too slow). I wanted to leave as is (solution I) or use solution IV, but Joel talked me to solution II. P.P.S.: More info can be obtained from ticket #7718 https://svn.boost.org/trac/boost/ticket/7718 and previous series of letters. -- Best regards, Antony Polukhin

23 126