- Boost - lists.preview.boost.org

[Graph] Efficiently finding cycles in a (multi)graph
by Louis Dionne 11 Apr '13

11 Apr '13

Dear Boost, I have implemented the `hawick_circuits` algorithm described in [1] for finding all the elementary circuits of a directed (multi)graph. I implemented it in a generic fashion, planning to propose it for inclusion in Boost. Boost.Graph currently contains the `tiernan_all_cycles` algorithm for finding all the cycles in a graph. However, it is undocumented and it has a time bound (for which no precise analysis is provided) that makes it unsuitable even for moderately sized graphs. In fact, I implemented `hawick_circuits` because `tiernan_all_cycles` did not cut it for my application. My implementation is available at [2]. It is licensed under the Boost license and was tested using the test suite at [3]. I am posting this message for two reasons: 1. To know whether there is interest for such an algorithm in Boost.Graph. 2. To spread the word that this implementation exists for C++, to save people the hassle of reinventing the wheel before Boost.Graph officially provides it. I will create a Trac ticket if there is interest, but note that I won't submit a patch immediately because I am in deep juice with school right now. Regards, Louis Dionne P.S.: More information on the test suite can be found at [4]. [1] http://bit.ly/12LMwIr [2] http://bit.ly/10MSAhW [3] http://bit.ly/YdLHVn [4] http://bit.ly/16SvlW6

1 1

GSoC 2013
by İslam Yaşar 11 Apr '13

11 Apr '13

Hi, I'm a graduate student and work since 2 years as a C++ developer on Unix-like systems (FreeBSD and Linux), i want to implement a library for Boost. It is a thread-safe queue that is implemented on disk. You can use it as a IPC. Is anyone interested to mentor this project ? Thanks Islam Yasar

3 2

Boost.uBLAS for GSoC 2013
by Sathyam Vellal 11 Apr '13

11 Apr '13

Hello. I'm Sathyam Vellal, a 3rd year undergraduate student pursuing Computer Science. I'm interested in contributing to Boost.uBLAS in unifying the vector and matrix representations. I've looked at the docs ( http://www.boost.org/doc/libs/1_53_0/libs/numeric/ublas/doc/index.htm) to get an idea of the library. Does anyone have any suggestions on how I could proceed more and connected to this project? Thank you -- Sathyam M Vellal Computer Science and Engineering, P.E.S. Institute of Technology, Bangalore, India.

2 1

GSoC 2013 : Boost.uBLAS
by Jignesh Jain 11 Apr '13

11 Apr '13

I am interested in this project for Gsoc, which involves unification of vector and matrix libraries. I have been browsing boost ublas code, but have no idea about how should I begin as I am totally new to Boost. Can Someone please guide me some approach for getting started in this project. Thank you.

1 0

Boost.Multiprecision GSoC 2013
by Nikhar Agrawal 10 Apr '13

10 Apr '13

Hi, My name is Nikhar Agrawal. I'm a 2nd year Computer Science student. I'm interested to work for Boost in GSoC and I found the project Boost.Multiprecision really really interesting. Any pointers on how should I go about it? Meanwhile, I'll go through this book:- http://www.loria.fr/~zimmerma/mca/mca-cup-0.5.1.pdf Thanks

3 3

[thread] using Boost.Thread with move-only arguments
by Fernando Pelliccioni 10 Apr '13

10 Apr '13

Hi Vicente, I have an issue with Boost.Thread contructor using move-only type arguments. The "Thread Constructor with arguments" section of the documentation says ... template <class F,class A1,class A2,...> thread(F f,A1 a1,A2 a2,...); Preconditions: F and each An must by copyable or movable. http://www.boost.org/doc/libs/1_53_0/doc/html/thread/thread_management.html… "An" could be copyable or movable, but the following example fails to compile... error: use of deleted function 'A::A(const A&)' //Begin code ------------------------------------ #include <boost/thread.hpp> struct A { A() = default; A( const A& other) = delete; A& operator=( const A& other) = delete; A( A&& other) = default; A& operator=( A&& other) = default; void run() {} }; void f( A a ) { a.run(); } void exec_boost() { A a; boost::thread t( f, std::move(a) ); t.join(); } //End code -------------------------------------- I think we should modify ... template <class F,class A1> thread(F f,A1 a1,typename disable_if<boost::thread_detail::is_convertible<F&,thread_attributes >, dummy* >::type=0): thread_info(make_thread_info(boost::bind(boost::type<void>(),f,a1))) { start_thread(); } ... by ... template <class F,class A1> thread(F&& f,A1&& a1,typename disable_if<boost::thread_detail::is_convertible<F&,thread_attributes >, dummy* >::type=0): thread_info(make_thread_info(boost::bind(boost::type<void>(),f,boost::forward<A1>(a1)))) { start_thread(); } ( In the all family of related ctors in "boost/thread/detail/thread.hpp" ) But, I don't know if boost::bind is forwarding its arguments I think we should modify Boost.Bind to be similar to std::bind. Is there any reason why Bind has not changed so far? Have you found any limitations? If Bind can not be changed, I think Boost.Thread should use some other tool similar to Bind to make forwarding parameters. GCC Libstdc++ use a Bind-lite tool for std::thread. I can work on this if you want. What do you think? Environment: Windows 7 64bits MinGW - GCC 4.7.2 MSVC 2010 MSVC 2012 Boost 1.53.0 Non-Variadic template code branch Regards, Fernando Pelliccioni

2 10

Re: [boost] GSoC 2013 : Implementing Dynamic Programming Algorithms
by Aman Gupta 09 Apr '13

09 Apr '13

Hi, Thanks for the reply. The first thing I want to implement is a Integer-Knapsack problem which takes in a precision error(i.e. an epsilon which is a indicator of the allowed error in the input) The input parameters are Weights, Values and the error. The first thing to do is massage the values according to the error (e.g. if the value is ~10^4 divide it down to ~10^3 according to the error) then run the a dynamic programming algorithm, which has a running time of O(n^3/e) (n : number of items, e :error) Hence, an arbitrarily close approximation is possible with a running time trade-off. Also, a vanilla variety of the knapsack problems(with small instances) which gives the optimal solution can also be implemented. I also want to implement the assignment problem, which also arises in a lot of situations in real life ( http://en.wikipedia.org/wiki/Assignment_problem) using the Hungarian Algorithm. I realize this requires effort but I'm really interested in adding a feature to the boost library. Regards, Aman Gupta

1 0

GSoC 2013 - Boost registration?
by Paul A. Bristow 09 Apr '13

09 Apr '13

http://www.google-melange.com/gsoc/events/google/gsoc2013 March 18: 19:00 UTC Mentoring organizations can begin submitting applications to Google. March 29: 19:00 UTC Mentoring organization application deadline. Just checking that Boost has/is registering? Paul --- Paul A. Bristow, Prizet Farmhouse, Kendal LA8 8AB UK +44 1539 561830 07714330204 pbristow(a)hetp.u-net.com

3 3

GSoC 2013 : Implementing Dynamic Programming Algorithms
by Aman Gupta 09 Apr '13

09 Apr '13

Hi, I am Aman Gupta from BITS-Pilani India, I am a 4th year Undergraduate Student majoring in Computer Science. I've noticed that boost does not contain dynamic programming algorithms like Integer-Knapsack, Longest Common Sub-sequence. I was thinking that implementing this would be a good project. I realize most of these problems are non-polynomial in complexity but I was thinking implementing them to a certain size will not be problem. Also, since some of the dynamic programming algorithms have alternate greedy algorithms which compromise on correctness but are faster, implementing those can be useful. The knapsack problem(which comes in a lot of situations) for which there exists an algorithm which gives an arbitrarily close approximation to the optimal value can be implemented. ( e.g. f(weights,values,error) could be the function, the user has a trade-off between error and computation time). I think it could be useful in situations where the user just decides the error he wants or perhaps how fast he wants his computation. Please let me know if this seems like a valid project idea to you! Thanks, Aman Gupta

2 1

[move] move's is_convertible does not compile with aligned data
by Tobias Loew 09 Apr '13

09 Apr '13

The implementation of move's is_convertible-type has the same problems on MSVC with aligned types as reported for type_trait's is_convertibale in Ticket #5271 https://svn.boost.org/trac/boost/ticket/5271 As type_trait's implementation has been fixed, a simple solution would be letting the user (and not the code in move.hpp:41) [not] define BOOST_MOVE_AVOID_BOOST_DEPENDENCIES regards Tobias -- View this message in context: http://boost.2283326.n4.nabble.com/move-move-s-is-convertible-does-not-comp… Sent from the Boost - Dev mailing list archive at Nabble.com.

2 2