- Boost - lists.boost.org

boost::adaptors::keys for mutable ranges of pairs
by Valentin Ziegler 04 Mar '12

04 Mar '12

Hello, map.hpp defines two overloads for adaptors::values, one for StdPairRange& and another for const StdPairRange const& rng. On the other hand, there is just a single adaptors::keys which takes StdPairRange const&. In particular, the code below will not compile: boost::find_if( boost::adaptors::keys(rng), SomePredicate() ).base()->second=NewValue(); Therefore, I would propose an adaptors::keys(StdPairRange& rng) overload similar to adaptors::values(StdPairRange&). Iterators of the returned range shall have a mutable base iterator. Further, iterators of the returned range should have a reference type of range_value<StdPairRange>::type::first_type& , so the elements in keys are mutable where StdPairRange admits this. Regards, Valentin -- Valentin Ziegler | vziegler(a)think-cell.com Software Engineer think-cell Software GmbH | Chausseestr. 8/E | 10115 Berlin | Germany http://www.think-cell.com | phone +49 30 666473-10 | US phone +1 800 891 8091 Amtsgericht Berlin-Charlottenburg, HRB 85229 | European Union VAT Id DE813474306 Directors: Dr. Markus Hannebauer, Dr. Arno Schoedl

2 1

[config] std::unique_ptr, std::ref detection?
by Peter Dimov 04 Mar '12

04 Mar '12

Suppose I want to implement a shared_ptr constructor taking std::unique_ptr. To avoid breaking pre-C++11 compilers, I need to, at least, know whether <memory> contains std::unique_ptr. (I'd also like to know if there's std::ref in <functional> because the specification of the constructor uses it.) But the config library doesn't seem to provide a way. There's no macro for std::unique_ptr, and there's no macro for "generic C++11 compiler and library".

9 23

Possible Bug in Boost Geometry's Huiller Algorithm
by Brian Ahr 03 Mar '12

03 Mar '12

Hi Folks. I was looking at the implementation of the huiller algorithm in boost, and it seems to be partially incorrect. See boost/geometry/strategies/spherical/area_huiller.hpp In the apply() function, we have if(lon2 < lon1){ E =-E;} which I believe should instead be calculation_type lon;if( lon2 < lon1 ) lon = lon1 - lon2;else lon = lon1 - lon2 + two_pi;if(2* half_pi < lon ) E =-E; according to Graphics Gems 5, pg 45 - 46. The comments at the top of the C++ file state that the code was derived from code in Graphics Gems 4, though the code was later corrected in Graphics Gems 5. I noticed this when trying to calculate areas of polygons on a sphere, where one of the points was (0, 0). In that case, it would always return a negative area, even though the polygon was defined counter-clockwise. Using a point other than (0,0) seemed to alleviate the problem, which got me wondering what was going on. Can anybody else confirm that this is an issue? Any suggestions for cleanly overriding boost's implementation until this is fixed? Thanks! - Brian

2 2

[predef] Thanks for all the comments..
by Rene Rivera 01 Mar '12

01 Mar '12

I would like to than everyone who reviewed and/or commented on my library. Instead of trying to respond individually to each point raised I will do a single post about all the points I've gathered, and will gather when I reread all the posts. Hopefully I can write that up by Monday. Needless to say there are a few changes I need to do, regardless of acceptance results :-) Again, thanks to all. And thanks to Joel for managing the review. Rene. -- -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org - grafik/redshift-software.com -- 102708583/icq - grafikrobot/aim - grafikrobot/yahoo

1 0

Adapting Polygon to work with point attributes
by Paul Miller 01 Mar '12

01 Mar '12

Boost::Polygon looks like exactly what I need to do some boolean operations on polygons but my polygons have additional attributes per point that I want to carry through the operation. But I can't figure out how to get polygon to create sensible values for the attributes other than X,Y, and I can't find an example that covers this scenario. It seems from the documentation that this is possible however. Can someone point me to an example, to say, adapt a polygon with this Point type? struct Point { int x, y; float r, g, b, a; }; Ideally it would average new r, g, b, a values together to make new points and polygon join points, but it could just copy one of the existing Points' attributes instead. I'm not picky. Cheers!

2 2

[TypeTraits] Patch: improving detectability of clang's intrinsics
by Michel Morin 01 Mar '12

01 Mar '12

Attached a patch to improve detectability of `__is_pod` and `__is_empty` intrinsics in clang. This is done by newly introduced `__has_feature(__is_pod__)` and `__has_feature(__is_empty__)` macros. (Clang trunk now accepts feature names with preceding and following double underscores.) Regards, Michel

2 1

Re: [boost] [git] neglected aspects
by Brian Wood 01 Mar '12

01 Mar '12

Dave Abrahams: > I don't see any advantage at all in having a "Boost supertree." If I > need a cluster of five libraries to get a job done, I don't see a reason > to mess around with anything more than those five. And with the > dependencies properly encoded, I won't even have to ask for all of those > explicitly. +1 for keeping things simple. This effort has been discussed for years. Is it close to being ready? Regards, Brian Wood Ebenezer Enterprises http://webEbenezer.net

1 0

[PREDEF] Review
by Ioannis Papadopoulos 29 Feb '12

29 Feb '12

> - What is your evaluation of the design? The design is straightforward and I find it nice, with some exceptions: - BOOST_ARCHITECTURE_* is too long (BOOST_ARCH_* is better) - BOOST_LANGUAGE_* is too long (BOOST_LANG_* is better) - BOOST_CXX would be better if called BOOST_COMPILER or BOOST_COMP - There has to be some macro for compilers that do emulation (for example Intel, PathScale etc appear also as gcc). - I do not like the fact that all macros are defined with a 0 when the compiler/arch/whatever is non-existant. I prefer to do # ifdef BOOST_CXX_GNU ... # endif rather than # if BOOST_CXX_GNU ... # endif - BOOST_VERSION_NUMBER allows only 3 numbers. I don't think that this is the case for all compilers/libraries. - There are some platforms that have unique features (for example Cray machines, IBM BlueGene etc). A BOOST_PLATFORM_* macro would be nice. > - What is your evaluation of the implementation? Including one file includes pretty much everything. It would be more straightforward if everything was in one file (for example merge everything in architecture/ in architecture.h). This also solves the issue of defining macros for things that don't exist. > - What is your evaluation of the documentation? No considerable problems with the documentation. Maybe it should be mentioned what the library supports. > - What is your evaluation of the potential usefulness of the library? I already have something similar in a lot of my codes. I would gladly remove them and use predef - especially if it detects also platforms. > - Did you try to use the library? With what compiler? Did you have any problems? Yes, PathScale is recognized as both gcc and pathscale. > - How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? A quick reading of the documentation and the code. I ran a small test case. > - Are you knowledgeable about the problem domain? Yes. > - Do you think the library should be accepted as a Boost library? I think that it should be ACCEPTED with the requirement that all issues have been resolved; especially the compiler recognition.

1 0

Determining interest: Pure imaginary number library
by Matthieu Schaller 29 Feb '12

29 Feb '12

Dear all, I have recently been playing a lot on numerical analysis problems containing complex numbers and functions. I have been using the SL std::complex<> library and could observe that some of the algorithms weren't optimized in cases where complex numbers have to be multiplied by pure imaginary numbers (i.e. x*sqrt(-1)). I have implemented a class to represent these numbers and overloaded all relevant operators in a more optimized fashion thanks to the knowledge that the real part of these complex numbers is 0. I could observe an important speed-up in some calculations. As an example, this piece of code (1D Schroedinger equation using finite differences): for(size_t i(0); i<nbPoints; ++i) { Psi_new[i] = Psi_old[i] + ii * hbar * deltaT / (12.*m*deltaX*deltaX) * (-Psi_cur[i-2] + 16.*Psi_cur[i-1] - 30.*Psi_cur[i] + 16.*Psi_cur[i+1] - Psi_cur[i+2]) - 2. * ii * V(x[i]) / hbar * deltaT * Psi_cur[i]; } is more than 2 times faster (g++ 4.6.1 core i7 2820QM) when the imaginary unit (constant) variable ii is defined to be Imaginary<double> ii(1.); than when it is defined as std::complex<double> ii(0.,1.); Some other expressions also benefit from this improvement and can run faster. Expressions including square roots or exponentials of pure imaginary numbers are improved a lot. The range of application of such a class is quite narrow but it may be helpful to speed-up some simple numerical problems. The source code as well as a working example is available here: http://code.google.com/p/cpp-imaginary-numbers/ Thanks, Matthieu

8 21

How to build on VC11?
by Olaf van der Spek 29 Feb '12

29 Feb '12

Hi, How does one build Boost 1.49 on VC11? It keeps using VC10. -- Olaf

2 1