February 2015 - Boost - lists.preview.boost.org

Friend class with namespace gives compiler error for C6000 TI
by Alin Popa 25 Feb '15

25 Feb '15

Hi, I compile my project with boost header only files, the compiler is C6000 Texas Instruments. The compile errors occurs for example in "boost/algorithm/string/find_iterator.hpp" I have extracted the cause of the problem in the next example code: namespace boost { namespace iterators { class iterator_core_access { public: iterator_core_access(){} ~iterator_core_access(){} }; } using iterators::iterator_core_access; } class testFriend { private: // don't work, compiler error: error: namespace "boost" has no actual member "iterator_core_access" friend class ::boost::iterator_core_access; // here is the problem // it works, compiles ok, this was my cure //friend class ::boost::iterators::iterator_core_access; }; int main () { // it works ::boost::iterator_core_access a; return 0; } Do you guys know what is the reason of this ? With msvc and g++ works fine. Thanks and best regards, Alin

2 1

[build] check __aarch64__ Aarch64
by Oliver Kowalke 24 Feb '15

24 Feb '15

Hi, I've added an pull request for jam.h (build/engine) to set os.platform correctly to 'ARM' on Aarch64 architecture. This is required in order to build boost.context correctly on the64bit ARM platform. thx, Oliver

1 0

Any online C++ compiler with Boost version to select?
by Boris Schäling 23 Feb '15

23 Feb '15

Does anyone know of an online C++ compiler which does not only come with Boost but allows you to select the Boost version? While http://coliru.stacked-crooked.com/ comes with Boost 1.57.0, http://ideone.com/ and http://cpp.sh/ with 1.55.0 and http://codepad.org/ with 1.34.1 (really!), I can't even tell whether these websites support Boost explicitly or whether it's more of an accident that I can include Boost headers. :) http://isocpp.org/blog/2013/01/online-c-compilers also mentions http://liveworkspace.org/ and http://gcc.godbolt.org/. But I can't open these websites (not sure whether they are temporarily broken or gone for good?). Boris

3 3

[config] test macro for std::index_sequence<>?
by Oliver Kowalke 23 Feb '15

23 Feb '15

Hi, I can not find an macro (BOOST_NO_CXX11_...) for testing the absence of std::index_sequence<> - may I have overlooked it? so long, Oliver

2 2

Re: [boost] [gsoc15] Attn: We need 2015 mentors and project ideas for Boost!
by Antonio Colino 22 Feb '15

22 Feb '15

Here is a random thought for a project idea. Agner fog created a vectorization lib for the sse and avx instruction sets as part of a project called vector class. Intel created the equivalent opencl kernels as part of a project that is called ispc ( Intel simd program compiler ). Research both and submit a single vector lib that auto vectorizes avx, sse both gpu and cpu overloading the std::vector class operators.Just a random thought. Intel SPMD Program Compiler)

1 0

[mpl] Fix to Ticket #11001
by Bruno Dutra 22 Feb '15

22 Feb '15

I just finished a pull request <https://github.com/boostorg/mpl/pull/14> with a fix for Ticket #11001 <https://svn.boost.org/trac/boost/ticket/11001>, which, by the way, I reported myself, so I'm quite confident it's fixed. Could one of the maintainers kindly review it? Oh and it is worth it to mention that I also included a new test case to ensure the fix is working. Kind regards, *Bruno C. O. Dutra*

1 1

GSoC: Enhanced vector and deque containers
by Benedek Thaler 22 Feb '15

22 Feb '15

Hi, I'm looking at the "Enhanced vector and deque containers"[0] item of the GSoC projects. I have to following interface idea: boost::devector: Like std::vector, but the first call to push_back inserts the element to the middle of the buffer. Likewise, reserve puts the contents to the middle of the new buffer (if any). This allows a cheap push_front, pop_front, push_back, pop_back. boost::deque: Like std:deque, but segment size can be specified in the constructor. To make buffer operations easier, a `segment_end` would be provided: iterator segment_end(iterator begin) Which returns the end of the largest consecutive range starting at `begin`. I'm also thinking about the following twist: Assuming this layout: [a,b,c,d] [e,f,g,_] Brackets denote segments, letters are elements, underscore is empty space. Here, inserting `x` after `d` normally results several moves and this layout: [a,b,c,d] [x,e,f,g] But instead, we could allocate a new segment: [a,b,c,d] [_,x,_,_] [e,f,g,_] To make the insertion cheaper. This behavior is similar to a segmented list, i.e: std::list<boost::devector<T>> Are these similar to the "extra functionality" mentioned in the docs you were thinking about? Thanks, Benedek [0]: https://svn.boost.org/trac/boost/wiki/SoC2015#a5.Enhancedvectoranddequecont…

1 0

Enquiry about gsoc project Enhanced vector and deque containers
by Jasdeep Singh 21 Feb '15

21 Feb '15

I had a few doubts regarding this project and was wondering if someone could help me out.It would be much appreciated.

1 0

[MultiIndex] Memory consumption of ordered indices
by Olaf van der Spek 21 Feb '15

21 Feb '15

Hi, What's the memory consumption of ordered indices? Is it implemented as a linked list, so two pointers per element? In that case for a write-only data structure an external ptr vector would make more sense wouldn't it? -- Olaf

3 4

[filesystem] temp_directory_path() behavior on Windows
by Beman Dawes 20 Feb '15

20 Feb '15

temp_directory_path() on Windows is currently implemented by calling the Windows API GetTempPath() function. Ticket https://svn.boost.org/trac/boost/ticket/5300 points out that GetTempPath() does not work as expected for environmental variables longer than roughly 130 characters. I've added a test to Boost.Filesystem that verifies that boost::filesystem::temp_directory() is affected. The suggested fix is to use GetEnvironmentVariable to in effect implement GetTempPath() the way we would like it to work. That makes sense to me, so I looked up the specs for GetTempPath(): The GetTempPath function checks for the existence of environment variables in the following order and uses the first path found: 1. The path specified by the TMP environment variable. 2. The path specified by the TEMP environment variable. 3. The path specified by the USERPROFILE environment variable. 4. The Windows directory. (4) Makes me queasy. I suspect it will fail anyhow because of permissions errors, but I would prefer to have temp_directory_path() report an error if 1, 2, and 3 are not present. temp_directory_path() already reports an error if a path is found but it is not a directory. OTOH, excluding the Windows directory is a breaking change for anyone currently depending on that behavior, so I thought it best to ask for comments before charging ahead. Thoughts? --Beman

8 39