July 2012 - Boost - lists.boost.org

[1.51][thread] Compile errors when chrono is disabled
by Andrey Semashev 29 Jul '12

29 Jul '12

Hi, There is a problem with Boost.Thread on windows, when BOOST_THREAD_USES_CHRONO macro is defined. In win32/basic_timed_mutex.hpp there are references to Boost.Chrono not guarded with this macro, which results in compile errors. Could this be fixed before the 1.51 release? I can apply the change, if permission is granted.

2 5

[offtopic] C++11 useful trick
by Eric Niebler 29 Jul '12

29 Jul '12

I recently ran into a limitation in recursive functions that make use of decltype in trailing return types. I've come up with a work-around that I think is a little non-obvious, so I thought I'd share. The problem came up while trying to implementing something like a variadic back() function. The "obvious" implementation doesn't work: template<typename T> T back(T && t) { return t; } template<typename T, typename ...U> auto back(T && t, U &&... u) -> decltype(::back(static_cast<U &&>(u)...)) // ERROR { return ::back(static_cast<U &&>(u)...); } int main() { int i = ::back(1,2,3); } The problem occurs on the line marked "HERE". Trouble is, the variadic "back" function is not yet in scope until after the return type. However, we'd like to use it when specifying the return type. The solution uses a default function template parameter to make the ::back symbol dependent, thereby deferring it's lookup until phase 2, when the symbol is visible. Check it: struct S { template<typename T> static T back(T && t) { return t; } template<typename T, typename ...U, typename Impl = S> static auto back(T && t, U &&... u) -> decltype(Impl::back(static_cast<U &&>(u)...)) { return Impl::back(static_cast<U &&>(u)...); } }; int main() { int i = S::back(1,2,3); } The trick is the "typename Impl = S" followed by a (now dependent) invocation of Impl::back. Voila. Apologies if this is a known technique, and I'm spamming the list with old news. It's news to me. -- Eric Niebler BoostPro Computing http://www.boostpro.com

9 26

[type_erasure]broken link in any.html#any(const U&, const static_binding&)
by Larry Evans 29 Jul '12

29 Jul '12

The html page fragment: http://steven_watanabe.users.sourceforge.net/type_erasure/libs/type_erasure… contains a link, CopyConstructible, which sends my browser to: http://steven_watanabe.users.sourceforge.net/trunk/doc/html/CopyConstructib… which is not there. Maybe it should go to: http://www.boost.org/doc/libs/1_50_0/libs/utility/CopyConstructible.html or something similar? Also, why couldn't there be any number of arguments, something like: template<typename Map, class... U> explicit any(const static_binding< Map > & binding, U &&... arg); I think this would then allow any constructor to be used for the contained type. Of course then the requirement would have to be modified to be something like: The underlying type has a constructor accepting arguments, U&... arg. -regards, Larry

1 0

[fusion] typo in documentation
by Steve Lorimer 29 Jul '12

29 Jul '12

In the fusion documents for algorithm pop-back, the example shows usage of pop_back as: assert(___pop_back__(make_vector <http://www.boost.org/doc/libs/1_50_0/libs/fusion/doc/html/fusion/container/…>(1,2,3)) == make_vector <http://www.boost.org/doc/libs/1_50_0/libs/fusion/doc/html/fusion/container/…>(1,2)); this should read: assert(pop_back(make_vector <http://www.boost.org/doc/libs/1_50_0/libs/fusion/doc/html/fusion/container/…>(1,2,3)) == make_vector <http://www.boost.org/doc/libs/1_50_0/libs/fusion/doc/html/fusion/container/…>(1,2)); link: http://www.boost.org/doc/libs/1_50_0/libs/fusion/doc/html/fusion/algorithm/…

1 0

[type_erasure]any(const binding<Concept>&, ...) compile error
by Larry Evans 28 Jul '12

28 Jul '12

The attached, when compiled with with gcc4.8-20120624, produces compile error: ../../../boost/type_erasure/detail/construct.hpp: In instantiation of 'boost::type_erasure::any<Concept, T>::any(const boost::type_erasure::binding<Concept>&, U0&) [with U0 = value<boost::type_erasure::_a>; Concept = boost::mpl::vector<boost::type_erasure::copy_constructible<boost::type_erasure::_a>, boost::type_erasure::copy_constructible<boost::type_erasure::_b> >; T = boost::type_erasure::_a]': binding_of.cpp:71:32: required from here ../../../boost/type_erasure/detail/construct.hpp:99:9: error: invalid use of void expression ) ^ Am I doing something wrong? TIA. -Larry

2 13

[1.51][ICL] Last minute commit
by Joachim Faulhaber 28 Jul '12

28 Jul '12

Hi, unaware of the fact that commits to the release branch need permission this week I did a commit today (Revision 79767). Sorry for that. So I'd like to ask for permission now. In other words, is it possible to not reverse this commit? Thanks, Joachim -- Interval Container Library [Boost.Icl] http://www.joachim-faulhaber.de

3 4

Re: [boost] [type_erasure] Review
by Thomas Jordan 28 Jul '12

28 Jul '12

Hi, A couple of comments on the documentation (apologies my comments are not in the review template format): 1. The documentation should explain why the library is called 'TypeErasure' (as opposed to something like, e.g., 'Run Time Concepts' (RTC), for example). 2. The documentation needs a section placing the library in the context of 'Related Work', for example Adobe::Poly and related research papers. Some indication of the key differences and similarities would also be useful in this regard. 3. The description of the problem(s) the library addresses - i.e., the motivation for the library - in context of best-practice in 'modern C++', needs to be expanded in the documentation. This could be expedited by pointing to aforementioned outside papers/research on the subject, which already explain things like run-time concepts. 4. The documentation should give some brief glossary and/or include references to technical terms such as 'concept' 5. The documentation could do with some details regarding performance Vs classical OO. 6. The further work mentions support for C++11 rvalue references, what about move emulation for C++03? Cheers, Tom.

1 0

Interest in a tweener library
by Julien Jorge 27 Jul '12

27 Jul '12

Hello, I wrote a C++ library to create tweeners and I'm thinking about porting it to Boost. The purpose of the tweeners is to interpolate the intermediate values between two given values during a time interval. Various easing functions can be applied to control the computation of the intermediate values. The library provides the following functionalities: - Interpolation is done on the value of a given variable or using an user-provided callback function, - The interpolated values are computed using a predefined easing function or any user-provided function respecting the contract, - Several tweeners can easily be executed simultaneously or in a row, The documentation of the library is available at http://libclaw.sourceforge.net/tweeners.html and the source code is available at http://sourceforge.net/projects/libclaw/files/libclaw/1.7.1/ as part of a larger library named Claw (check the claw/tween and claw/code/tween folders) Is there an interest to add this tweener library to Boost and what would be the adaptations to do? Regards, -- Julien Jorge

8 24

Last minute commit
by Joachim Faulhaber 27 Jul '12

27 Jul '12

Hi, unaware of the fact that commits to the release branch need permission this week I did a commit today (Revision 79767). Sorry for that. So I'd like to ask for permission now. In other words, is it possible to not reverse this commit? Thanks, Joachim -- Interval Container Library [Boost.Icl] http://www.joachim-faulhaber.de

1 0

[type_erasure] Review
by Karsten Ahnert 27 Jul '12

27 Jul '12

Hi, here is my review for Steven Watanabe's TypeErase library. First of all, I want to thank Steven for developing this library. My overall evaluation is that TypeErasue should be accepted as boost library. I did not looked very deeply into all details or the implementation but I tried to solve some problems with the library. So, I only looked at the library from the users perspective. = 1. What is your evaluation of the design? = The design looks clear and concise. I have only two small point: * Why do I need to specify relaxed_match to get a default constructable any? Maybe a empty() function would be nice to check if any holds a value (in case it is default constructable). I think a hint about a default constructable any should be in the tutorial. It is important for users. * Sometimes it is not clear, why some constructors work and others not, at least for me. Details are shown below. = 2. What is your evaluation of the implementation? = I did not looked into the implementation. = 3. What is your evaluation of the documentation? = * The documentation looks good. But I think more real-world examples would be very useful. I also encountered some problems where I could not find a hint how to solve it. For example, consider the case where you need an any (for example in a recursive data structure) to be a member of some other class: struct my_class { private: any_type a; }; What kind of constructor do I need to set up this any properly? I came up to define two constructors: struct my_class { template< class T > my_class( T t ) : a( std::move( t ) ) { } my_class( any_type a_ ) : a( a_ ) { } private: any_type a; }; = 4. What is your evaluation of the potential usefulness of the library? = The library is very useful and can be applied to a large number of problems. The library provides an alternative to classical OOP and can be used in many situations. = 5. Did you try to use the library? With what compiler? Did you have any problems? = I tried with gcc-4.7 and gcc-4.6 without problems. I wrote a test application with some kind of an expression system which can be build at runtime. (Maybe this example would be useful for the docs?) My main any definition is template< class Return , class Context > struct requirements : public mpl::vector< te::copy_constructible<> , te::typeid_<> , te::callable< Return( Context ) > , te::relaxed_match > { }; // ... template< class Return , class Context > using expr = te::any< requirements< Return , Context > > ; template< class Context > using boolean_expr = expr< bool , Context >; Then you can build base classes for unary and binary expressions which model the expr concept and contain the subexpression, for example: template< class Return , class Context > class unary_expr { public: typedef expr< Return , Context > expr_type; template< class T > unary_expr( T child ) : m_child( std::move( child ) ) { } unary_expr( expr_type child ) : m_child( child ) { } expr_type& child( void ) { return m_child; } const expr_type& child( void ) const { return m_child; } protected: expr_type m_child; }; At this point I needed some time to figure out that you need at least these two overloads of the constructor. Actually I am not sure if these two overloads are sufficient for all use cases. Some clarification in the docs or any hints would be very useful here. Having the base classes it is very easy to implement concrete expressions, for example: // Boolean expressions struct true_ { template< class Context > bool operator()( Context ) const { return true; } }; struct false_ { template< class Context > bool operator()( Context ) const { return false; } }; template< class Context > class not_ : struct unary_expr< bool , Context > { template< class T > not_( T child ) : unary_expr< bool , Context >( child ) { } bool operator()( Context c ) { return ! this->m_child( c ); } }; This works quite nice, but I always had problems to figure out how the any type are constructed, for example using expr_type = boolean_expr< void >; using not_ = not_< void >; using and_ = and_< void >; using or_ = or_< void >; true_ t; false_ f; // why does this not work? // expr_type e3( not_( f ) ); // cout << e3() << endl; // why does this work? expr_type e4( and_( t , or_( f , t ) ) ); cout << e4() << endl; The first example expr_type e3( not_( f ) ) does not compile. It fails because not_( f ) is interpreted as a function pointer: rule_engine.cpp:35:16: error: too few arguments to function ‘expr_type e3(not_)’ rule_engine.cpp:34:15: note: declared here I have no idea why the second one compiles. = 6. How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? = I spend about 10 hours with studying the library. = 7. Are you knowledgeable about the problem domain? = Some experience in boost.any, boost.function and a self written type erasure application. Not too much. Thank you again! Karsten

2 2