July 2012 - Boost - lists.boost.org

My take on an allocate_shared_array and make_shared_array implementation
by Glen Fernandes 31 Jul '12

31 Jul '12

Hi all, I've desired an allocate_shared_array for a while (and, looking at some existing threads about the topic, it seems others have also). Recently, while working with a code base that accepted both shared_ptr<T> and shared_array<T> for a sequence of T elements, I decided to write both (for shared_ptr<T> since shared_array<T> does not support allocators). The code for both my make_array<T> and allocate_array<T> initial implementation can be found at: https://github.com/glenfe/glenfern/tree/master/glenfern I was wondering if there is any desire for these (after necessary tweaking, refinement, and testing) to become part of Boost. I'm not certain how useful they are in their current form for people using shared_array<T> that can't switch to shared_ptr<T>. Is there any desire to update shared_array<T> so that it supports custom allocators? In that event, my implementation can then easily be applied to shared_array<T> as well. Alternatively, are there any plans to provide a separate smart pointer type for an array with allocator support? Cheers, Glen

1 0

[block_ptr] Update
by Phil Bouchard 31 Jul '12

31 Jul '12

Hi, It's been more than a year since a completed the block_ptr and it seems as if more recent library are being reviewed first. This would suggest the FIFO of the schedule is not followed. I was just wondering since at the same time the reviewing of block_ptr is cost / beneficial. Regards, -Phil Author of: http://www.fornux.com/ http://www.finitetheory.com/

4 7

Fwd: deep-thought reboot
by DongInn Kim 31 Jul '12

31 Jul '12

Dear all, Can we reboot our NFS server sometime tomorrow? If nobody objects this reboot, I will go ahead to reboot deep-thought at 9am (E.T.) on July 31st, 2012. Date: Tuesday, July 31 2012 Time: - 6:00am-6:10am Pacific US time - 7:00am-7:10am Mountain US time - 8:00am-8:10am Central US time - 9:00am-9:10am Eastern US time - 1:00pm-1:10pm GMT The following services would not be available during the reboot. - NFS $HOME - web sites (e.g., www.open-mpi.org, svn.mpi-forum.org, www.osl.iu.edu/publications, ...) Please let me know if you have any issues or question about the outage. Regards, -- - DongInn --------------------------- CREST System administrator dikim(a)cs.indiana.edu Indiana University Bloomington, IN Begin forwarded message: > From: Rob Henderson <robh(a)indiana.edu> > Subject: deep-thought reboot > Date: July 30, 2012 8:36:41 AM EDT > To: DongInn Kim <dikim(a)osl.iu.edu> > > DongInn, we need to reboot deep-thought in order to get updated raid controller firmware in place and boot to the new kernel. This should be very quick with only a few minutes of downtime. Would there be a good time for us to do this? > > Thanks, > > --Rob

1 1

[type_erasure] Performance
by Julien Nitard 31 Jul '12

31 Jul '12

Hi all, I took time to write a simple benchmark for boost::any. Please find the code attached. It requires C++11. I could only test with g++ because clang 3.0/Ubuntu segfaults on the code. I'd be more than happy to get feedback on whether the tests are relevant. I think the compiler can't impact the result by optimizing away the payloads because there's the virtual call barrier preventing it from doing so. The results are quite good for a library that is not optimized yet. Here are some typical results on my machine (though results may vary slightly on each run): Times are in seconds for 100 million iterations, using -O3 -std=c++0x and no other special flags. * Classic inheritance replacement: Inherited: 0.261134 Erased: 0.260092 Erased + ctor: 0.343596 Erased + ctor 3 concepts: 0.295584 * boost::function replacement: boost::bind: 0.260356 boost::function: 0.469181 std::function: 0.515477 te::any + boost::bind: 0.494241 te::any + lambda: 0.379945 Basically that means that TypeErasure has a cost that's zero or at worse neglectible when compared to normal inheritance. It probably means that those simple use cases can't be optimized since the theoretical performance is achieved. If you create the any object each time, then there's something like a 30% loss. That's against an empty virtual function which means that for non trivial payloads the cost is still low. Good job ! The only thing I can't explain is why creating an any with 3 concepts in an mpl::vector is faster than a single primitive concepts. Any ideas ? Nice as well is that callable<> matches or bests boost::function in the simple cases I have tested (likely to be common thanks to C++ lambdas). Actually, I guess that the results would be even better in "normal" circumstances because the way I wrote the test any has 2 virtual call to perform. I guess that in that case again, optimization is not required before acceptance and release. Regards, Julien

1 0

[multiprecision] Huge testcase datafile in commit r79802
by Lars Viklund 31 Jul '12

31 Jul '12

Hi list. My boost-commit@ mbox made my mail agent go into a nice crawl today when displaying the commit details for r79802. It turns out that it's about 6.3 MiB large, most of which is a massive 100k line test case in sandbox/big_number/libs/multiprecision/performance/delaunay_data.txt Will this behemoth file end up in the real repository in case of library acceptance, and even worse, will it end up in the deployed tarballs? Are there any guidelines on keeping tests neat, or possibly generating performance test data if it's excessively huge? While in the discussion, has there been any thoughts of separating documentation from the tree, and/or having codegen of the source files that are clearly generated (vector200.hpp, etc.) in the build tree? It's very straining on disk space to have multiple Boost trees lying around, particularly if you're working in some kind of homedir environment. Combating-bloat'ly-yours, -- Lars Viklund | zao(a)acc.umu.se

2 2

[type_erasure]suggest tuple doc contain requires clause specifying T... are unique
by Larry Evans 31 Jul '12

31 Jul '12

http://steven_watanabe.users.sourceforge.net/type_erasure/libs/type_erasure… Specifies that the T... are placeholders; however, it doesn't specify that they be unique; hence, tuple<concept,_a,_a> x(int(0),std::vector<int>(0)); would satisfy the current description but wouldn't be right. Of course no one would intentionally do that; however, if the Tags are calculated in some meta-program, it might happened accidentally. Hence the requires class should also say, "If T... are not unique, no diagnostic occurs", or something similar. The reason I think "no diagnostic occurs" is that I've looked briefly in tuple.hpp, at the: ::boost::type_erasure::cons<Concept, T...> impl; in the class tuple definition and the cons definition, and I don't see any check for duplicates. -regards, Larry

2 3

[preprocessor metaprogramming] tools
by Nathan Ridge 31 Jul '12

31 Jul '12

Hello, I've noticed preprocessor metaprogramming has become more popular of late, both as a means to provide a cleaner alternative interface to a library (e.g. the DEFINE and ADAPT macros in Boost.Fusion or the proposed macro wrappers for Boost.TypeErasure) and as a means to emulate new language features (e.g. Boost.LocalFunction, Boost.Generic, and Boost.Contract). Sadly, IDE support for writing macro metaprograms is fairly poor in comparison to IDE support for writing regular C++ code. I wanted to share with you something I've found very useful for tracing/debugging macro metaprograms: the Eclipse CDT IDE has a "macro stepper" that allows you to step through the expansion of a macro step-by-step from the initial invocation to the final expansion. (This feature can be accessed by using the "Explore Macro Expansion" context menu item on a macro invocation.) I would be curious to hear what tools you have found useful for writing or debugging macro metaprograms. Regards, Nate

7 11

Re: [boost] BlueGo Beta 0.0.1
by Julian Amann 30 Jul '12

30 Jul '12

> How can I get the sources? the sources can be found at https://bitbucket.org/Vertexwahn/bluego Julian

1 0

[date_time] Who is in charge?
by Philippe Vaucher 30 Jul '12

30 Jul '12

According to the tracker, it looks like the last ticket with activity is https://svn.boost.org/trac/boost/ticket/5257 which is 16 months ago. Also, according to the mailing list I see a lot of discussion/patches dieing out of inactivity ("[boost] [date_time] warning removal", "[boost] [Date Time] bug in documentation example (Localization Demonstration)"). I'm not here to point fingers, we all have busy times where we just cannot handle open-source responsabilities anymore, but maybe this lib should be switched to "no active maintainers" to avoid any confusion? Also, and probably I should have begin with this, is there a page that lists who is the maintainer of what lib beside the Trac? I couldn't find any, but I didn't look very hard either. Philippe

11 18

[type_erasure]any(U,static_binding) stores wrong type
by Larry Evans 30 Jul '12

30 Jul '12

When the attached is run, it produces output: ./static_binding.exe std::is_copy_constructible<ph_erasee_type>=0 53:ph_ctor_arg<boost::type_erasure::_a>::CTOR.default:my_instance=0 creating a_fst from a_ctor_arg 64:ph_ctor_arg<boost::type_erasure::_a>::CTOR.copy:my_instance=1:(void*)this=0x66ab50 :*fst_cast=ph_erasee<boost::type_erasure::_a>.my_instance=0:(void*)fst_cast=0x66ab50 dynamic_cast<ph_ctor_arg_type*>(value_base_ptr)=0x66ab50 dynamic_cast<ph_erasee_type*>(value_base_ptr)=0 Indicating that a ph_ctor_arg_type is being stored in a_fst instead of the expected ph_erasee_type. -regards, Larry

2 6