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

[chrono] mac thread_clock support
by Chris Brown 15 Dec '13

15 Dec '13

I noticed that chrono doesn't seem to support thread_clock on mac (the posix implementation doesn't work for it) so I wrote up an implementation. https://github.com/boostorg/chrono/pull/1 I was pretty surprised this didn't already exist, but haven't been able to find anything in the mailing list or trac regarding it. Curious to hear if there was a reason to avoid this implementation or if it's just something no one got around to. Posting here for visibility since the git / Github workflow is relatively new to a lot of people here.

3 7

[fusion][mpl] transform_view and as_vector compilation error
by Karsten Ahnert 15 Dec '13

15 Dec '13

Hi, we have a problem in our code base. I extracted a minimal example, see below. The code works with g++ version up to 4.8.1 and C++11 but fails to compile with gcc-4.8.1 and clang3.2 and newer. Any ideas what goes wrong here? Should I further dig into the problem? I am using boost 1.54. #include <boost/fusion/include/vector.hpp> #include <boost/fusion/include/as_vector.hpp> #include <boost/fusion/include/transform_view.hpp> struct t1 { struct result_type1 {}; }; struct t2 { struct result_type1 {}; }; struct my_transform { template< class > struct result; template< class F , class T > struct result< F( T ) > { typedef typename T::result_type1 type; }; }; int main( int argc , char *argv[] ) { namespace fusion = boost::fusion; typedef fusion::vector< t1 , t2 > vector_type; typedef fusion::transform_view< vector_type , my_transform > transform_type; typedef fusion::result_of::as_vector< transform_type >::type transformed_vector_type; return 0; }

2 2

Project
by Utsav Tiwary 15 Dec '13

15 Dec '13

I am a third year engineering under graduate from India in interested in Open MPI Programming. Could you please furnish some basic details about the project involving Open MPI done under GSoC 2013,like pre requisites ,basic idea behind the project etc. ?

1 0

Re: [boost] [serialization] new portable binary archive based, on HDF5
by Daniel Koester 15 Dec '13

15 Dec '13

On Fri, 13 Dec 2013 22:39, Mateusz Loskot <mateusz(a)loskot.net> wrote: >> My suggestion would be for Robert Ramey and >> other interested developers to consider the inclusion of my work in the >> official serialization code. I am not sure about the exact procedure in >> this case. > > I *guess*, it may qualify for the Fast Track Review: > http://www.boost.org/community/reviews.html I will check out the procedure. >> I have forked from the Git serialization archive. All supplied tests >> have passed for gcc on Linux and MSVC 10 on Windows. My work can be >> pulled from >> >> https://github.com/dk1978/serialization/tree/development > > Hmm, have you updated the development branch with the upstream lately, > as well as the master in your fork? > > AFAICT it is a few months behind the current upstream, the master too, > it seems impossible to generate a diff usable enough to review your additions. You are right, I have neglected to synchronize with the upstream master, and mostly ignored my own master branch. Please try again now. Best regards, Daniel Koester

2 1

[git] Getting set up to test?
by Beman Dawes 15 Dec '13

15 Dec '13

Before making changes to, say, the develop branch of a library, I want to set the super-project to the latest master so tests run in the most realistic environment, as Robert Ramey has advocated for years. Presumably that means doing something like this: cd modular-boost git checkout master git pull --recurse-submodules cd modular-boost/libs/mylib git checkout develop Is "git pull --recurse-submodules" the correct command to update the submodules? Do I also need to do "git submodule update" or possibly "git submodule update --remote"? Thanks, --Beman

2 1

[conversion] try_lexical_cast and 200$
by Antony Polukhin 15 Dec '13

15 Dec '13

There is a bounty for implementing a `try_lexical_cast` from Facebook: https://www.bountysource.com/issues/1352296-exceptionless-lexical_cast Because of that three questions arise: * Is there some official policy for "feature implementation sponsorship"? * `try_lexical_cast` must be added or not? And if yes, what signature is preffered: bool try_lexical_cast<Out, In>(Out&, const In&) bool try_lexical_cast<Out, In>(const In&, Out&) boost::optional<Out> try_lexical_cast<Out, In>(const In&) * Is there a bank account "for Boost's needs" that is used to do hosting payments, conference room rental payments and other support stuff (to were those 200$ may be transfered)? -- Best regards, Antony Polukhin

13 25

Splitting/rearranging Boost.Utility
by Peter Dimov 14 Dec '13

14 Dec '13

Currently, Boost.Utility is something of a mixed bag and consequently, a dependency knot. Nearly everything in Boost depends on Utility via (say) boost/assert.hpp, while at the same time Utility itself depends on much of Boost via Iterator (and others). Something Must Be Done. At minimum, assert.hpp, current_function.hpp, ref.hpp, checked_delete.hpp and addressof.hpp need to be moved out. I suspect that call_traits.hpp, compressed_pair.hpp, enable_if.hpp also don't belong. assert.hpp and current_function.hpp should probably go into their own library, Boost.Assert. ref.hpp might go to Boost.Bind. checked_delete.hpp might perhaps go to SmartPtr. No idea where addressof.hpp needs to go. These are just suggestions. I'm open to any resolution, as long as the dependency (and, to a lesser extent, maintenance) problem is resolved to everyone's satisfaction.

8 23

Any interest in a framework for memoization?
by Giacomo Drago 14 Dec '13

14 Dec '13

Hi, I would like to submit a small framework for memoization I am working on (http://projects.giacomodrago.com/c++memo/) At the moment, it is still in a beta stage and does not comply with Boost guidelines. I would like to know if you might be interested. The idea is to have a tool for rapid prototyping of software components implementing dynamic programming algorithms or, anyhow, requiring memoization. I paste here a simple "fibonacci" example: ---------------------- int fibonacci(int i, std::function<int(int)> prereqs) { if (i == 0) return 0; if (i <= 2) return 1; return prereqs(i-1) + prereqs(i-2); } cppmemo::CppMemo<int, int> cppMemo; std::cout << cppMemo.getValue(30, fibonacci) << std::endl; ---------------------- The code will print 832040, without recursive calls to the "fibonacci" function. A stack is managed "under the hood" and appropriate calls are made to the "fibonacci" callback to gather pre-requirements and then call the function again when they are available. This is just a toy example. There is more to know about, including important caveats (if you are interested, please read the documentation carefully): some defaults for template and method parameters are useful for conciseness but are EVIL if you don't know about them. The framework also provides automatic parallelization (when possible). It is important to remark that at this moment the framework relies on fcmm (fast concurrent memoization map), a custom hashmap that is optimized for this very purpose. This can be changed quite easily, as fcmm is not exposed to the client. A C++11 compliant compiler is needed. Thanks for your attention. Giacomo

3 6

Re: [boost] [serialization] new portable binary archive based on HDF5
by Daniel Koester 14 Dec '13

14 Dec '13

In reply to Hartmut Kaiser: >> I would like to share a contribution to the boost::serialization module. >> It provides an answer to the challenge of supporting portable binary >> archives by using the well-known and stable HDF5 library >> (http://www.hdfgroup.org/HDF5) After first signals of interest by other >> users I have decided to fully integrate my work with the current >> boost::serialization module. My suggestion would be for Robert Ramey and >> other interested developers to consider the inclusion of my work in the >> official serialization code. I am not sure about the exact procedure in >> this case. > Two questions: > > 1) What license do you use for your contribution? The source files don't > mention it. I plan to use the Boost Software License, and document this and the archive interface later when the future of this project is decided. > 2) Do you have performance data comparing the new archive with the existing > ones? I'd be very interested to use this archive for HPX > (https://github.com/STEllAR-GROUP/hpx) as an alternative for our homegrown > derivative of the original portable binary archive available as an example > in Boost.Serialization. But for this I'd like to known what performance > advantages/disadvantages this might cause. I do not have performance data. Do you have suggestions for an interesting benchmark, and a link to your own binary archive implementation? I will be happy to get back to you on this point. Best regards, Daniel Koester

2 1

[build] duplicate files in converted history
by Vladimir Prus 14 Dec '13

14 Dec '13

I am getting this: $ git ls-tree 49ba2cf8dd415c100fef66f35e6686b362abe6ee:jam_src <snip> 100644 blob 92965f3a17e1439dc322b555a2c17c5cf03ffa2e jam.c $ git ls-tree 49ba2cf8dd415c100fef66f35e6686b362abe6ee:historic/jam/src <snip> 100644 blob 92965f3a17e1439dc322b555a2c17c5cf03ffa2e jam.c So, we appear to have this file at two locations in the converted history. The tree I am examined is the one where jam sources were imported to svn trunk. Is this the way things are supposed to be. If not, are there other implications of this bug? Now, I already have my toolbox ready, and will probably change rewrite script to fix this, no problem, but still. - Volodya

1 1